takeyama gordon towse developments in the economics of copyright
DESCRIPTION
asdasdsaTRANSCRIPT
Richard Leon GoveBook Fund
Developments in the Economics of Copyright
To my mom, Margery Takeyama, with love, and in loving
memory ofmy dad, Wallace Takeyama.LNT
To the memory of Sam Postbrief, and to thefuture of his
grandson, Theo Sam Rubin-Petrovic.
WJG
I Developments in the
j
Economics of i
Copyright ;
Research and Analysis
Edited by
Lisa N. TakeyamaAmherst College, and San Francisco State University, USA
Wendy J. Gordon
Boston University, USA
Ruth Towse
Erasmus University, The Netherlands
Edward ElgarCheltenham, UK Northampton, MA, USA
Lisa N. Takeyama, Wendy J. Gordon, Ruth Towse, 2005
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system or transmitted in any form or by any means, electronic,
mechanical or photocopying, recording, or otherwise without the prior
permission of the publisher.
Published byEdward Elgar Publishing Limited
Glensanda House
Montpellier Parade
CheltenhamGlosGLSOlUAUK
Edward Elgar Publishing, Inc.
136 West Street
Suite 202
NorthamptonMassachusetts 01060
USA
A catalogue record for this bookis available from the British Library
ISBN 1 84376 930 1
Printed and bound in Great Britain by MPG Books Ltd, Bodmin, Cornwall
Contents
List offigures vi
List of contributors vii
Preface Lisa N. Takeyama xiv
General outline of the book xvi
1 . Should economics play a role in copyright law and policy? 1
Pamela Samuelson
2. Risk sharing and the distribution of copyright collective
income 23
Arthur Snow and Richard Watt
3. MP3s and copyright collectives: a cure worse than the disease? 37
Stan Liebowitz
4. Peer-to-peer, piracy and the copyright law: implications for
consumers and artists 60
Anne Duchene and Patrick Waelbroeck
5. 'Fair use' as policy instrument 80
Timothy J. Brennan
6. Towards a differentiated products theory of copyright 103
Christopher S. Yoo
7. Private appropriability and sharing of knowledge: convergenceor contradiction? The opposite tragedy of the creative
commons 120
Giovanni B. Ramello
8. IMS Health or the question whether copyright still deserves a
specific approach in a market economy? 142
Alessandra Narciso and Paul L. C. Torremans
9 The basics matter: at the periphery of intellectual property 1 59
F. Scott Kiejfand Troy A. Paredes
Index 191
Figures
3.1 LP albums sold per capita 38
3.2 Album revenues 49
4. 1 Consumer demand with information-push technology 69
4.2 Consumer demand with information-pull technology 72
4.3 Application with 9 = 0.15 75
4.4 Application with cp= 0.30 75
5. 1 Consumer surplus, profits, and costs for work n: 'equal
diversion' 85
5.2 Consumer surplus, profits, and costs for work n: profit-
maximizing price> fair use price 90
5.3 Consumer surplus, profits, and costs for work n: profit-
maximizing price< fair use price 9 1
vi
Contributors
Timothy J. Brennan is Professor of Public Policy and Economics at the
University of Maryland, Baltimore County and Senior Fellow at Resources
for the Future. Prior to coming to UMBC, he was an associate professor of
telecommunications policy at George Washington University and was a
staff economist with the Antitrust Division of the US Department of
Justice. From 1996-97, he served as the senior economist for industrial
organization and regulation on the staff of the White House Council of
Economic Advisers. His primary research areas are in antitrust, regulated
industries, intellectual property, and communications policy, with over 75
articles and chapters in journals and books covering economics, law, energy
policy, telecommunications, philosophy and politics. He serves on the
editorial boards of the Journal of Regulatory Economics, Information
Economics and Policy, and Communications Law and Policy. He is co-
author of Alternating Currents: Electricity Markets and Public Policy. Hereceived his PhD in economics from the University of Wisconsin, Madison
in 1978.
Anne Duchene completed her PhD in economics at CERAS - Ecole
Nationale des Ponts et Chaussees, after graduate studies at Universite de
Paris 1 (La Sorbonne). Her thesis focuses on intellectual property rights,
and more specifically on Internet piracy, agency relationships between
patent lawyers and innovators, and the (dis)functioning of patent offices.
Wendy J. Gordon is Professor of Law and Paul J. Liacos Scholar in Law at
the Boston University School of Law. With Richard Watt she co-edited The
Economics of Copyright (Edward Elgar 2003), and has published on four
continents scholarship using economics and philosophy to analyze the
merits of the legal institutions that regulate information and culture. She
currently serves as a Vice President for the Society for Economic Research
on Copyright Issues; the Chair Elect for the Intellectual Property Section
of the Association of American Law Schools; and as a Visiting Scholar in
Comparative Media Studies at the Massachusetts Institute of Technology.
Among Professor Gordon's honors are: being named a Fulbright Scholar;
being thrice cited by the US Supreme Court; being elected to the Visiting
Senior Research Fellowship at St. John's College, Oxford University; and
vii
viii Developments in the economics of copyright
receiving a residency at the Rockefeller Foundation retreat in Bellagio,
Italy. Her editorial board service includes The Encyclopedia of Law and
Economics and the Review of Economic Research on Copyright Issues,
and she served as the 2002-03 Area Organizer for Intellectual
Property/Electronic Commerce for the American Law and Economics
Association. Professor Gordon's many articles include An inquiry into the
merits of copyright' (Standard Law Review), 'Fair use as market failure'
(Columbia Law Review), A property right in self-expression' (Yale Law
Journal), 'On owning information' (Virginia Law Review), 'Copyright
norms and private censorship' (forthcoming in the Oxford University Press
volume, Copyright and Free Speech), and the Oxford Handbook on LegalStudies chapter on 'Intellectual Property'. Her writings also address areas
such as rhetoric, restitution and property theory.
F. Scott Kieff is an associate Professor in Technology Law and Business in
Washington University School of Law and 2003-05 W. Glenn Campbelland Rita Ricardo-Campbell National Fellow at the Hoover Institution of
Stanford University. He previously served as Visiting Assistant Professor at
the University of Chicago Law School and the Northwestern University
School of Law and as John M. Olin Senior Research Fellow in Law,
Economics, and Business at Harvard Law School. He is also a member of
the founding faculties of the Munich Intellectual Property Law Center and
the Canadian Centre for Intellectual Property Policy at McGill University
Faculty of Law. Scott graduated with a degree in molecular biology and
applied microeconomics from MIT. He was admitted to the New York Bar
after graduation from the University of Pennsylvania Law School and
served as Law Clerk to the Honorable Giles S. Rich on the US Court of
Appeals for the Federal Circuit. Scott has delivered numerous articles and
speeches about obtaining and enforcing intellectual property rights. Heedited the book Perspectives on the Properties of the Human Genome Project
(Elsevier), and co-authored the popular treatise and casebook Principles ofPatent Law (Foundation Press), now in its third edition. His research inter-
ests generally involve the interface among law, economics, ethics, and crea-
tive endeavors such as science, engineering, medicine, and art, with a focus
on technology law and business, intellectual property, contracts, unfair
competition, antitrust, complex litigation, and the allocation of decision-
making ability and authority in disputes involving technological facts.
Having practiced as an associate with the firm of Pennie & Edmonds in
New York, and as an associate and counsel with the firm of Jenner & Block
in Chicago, he has been called to testify as a legal expert before federal
courts and agencies and maintains his connection to the law firm and busi-
ness communities through his ongoing consulting practice, which focuses
Contributors ix
on the strategy and management of intellectual property, competition, and
litigation.
Stan Liebowitz was trained at Johns Hopkins (BA) and UCLA (PhD). Heis currently a Professor of Economics in the School of Management at the
University of Texas at Dallas after having been on the faculty at the
University of Western Ontario, University of Rochester, and a FacultyFellow at the University of Chicago. In addition to five books, he has pub-lished over 60 articles in journals including the American Economic Review
and the Journal of Political Economy, as well as more popular outlets such
as the Wall Street Journal and CIO Magazine. He is director of the Center
for the Analysis of Property Rights and Innovation (CAPRI) and is Vice
President of the Society for Economic Research on Copyright Issues
(SERCI). He also serves on the Editorial Board of the Review ofEconomic
Research on Copyright Issues (RERCI) and is an adjunct scholar at the Cato
Institute. Professor Liebowitz's research interests include the economic
impact of new technologies on copyright owners (most recently, the impactof filesharing), the economics of networks, pricing issues, and antitrust. His
work (with Steve Margolis) on network effects and lock-in, culminated in
two books, Winners, Losers and Microsoft (Independent Institute, 1999)
and The Economics of Qwerty (NYU Press/Palgrave, 2001, edited by Peter
Lewin). His most recent book titled Rethinking the Networked Economywas published by the American Management Association in the fall of
2002. His research has been the focus of articles in The Economist, the Wall
Street Journal, the New York Times, the Financial Times, and a program on
the BBC. He has consulted and testified in the United States and Canadaon issues related to technology and intellectual property.
Alessandra Narciso is a doctoral researcher in the Department of Law of the
University of Leeds (UK). She holds a first degree in law from La Sapienza
University in Rome and an LLM degree from the University of Leicester
(UK). Her area of specialization is the legal protection of folklore.
Troy A. Paredes is an Associate Professor of Law at Washington University
School of Law in St. Louis. His work primarily focuses on corporate
governance and securities regulation, as well as on property theory and law
and economics. He has written numerous articles on a wide range of topics,
such as corporate governance and venture capital in developing countries,
hostile takeovers, executive compensation, CEO overconfidence, informa-
tion overload and decision theory, asset securitization, intellectual property
and bankruptcy, and hedge funds. Before joining Washington University,
Paredes was a corporate and regulatory attorney.
x Developments in the economics of copyright
Giovanni B. Ramello is Assistant Professor of Economics at the Universita
Carlo Cattaneo - LIUC, Italy. He received an education both in econom-
ics (PhD) and in computer sciences in Italy (Universita degli Studi di
Torino, Universita Bocconi di Milano) and in France (Universite Jean
Moulin - Lyon 3 and CNSM, Lyon). He has been research fellow at the
Fondazione Giovanni Agnelli, Torino (Italy), at the Banff Centre (Alberta,
Canada), and has served as an advisor for the EU Phare-Twinning Project
on the Harmonisation of Antitrust in the European Countries (topic: inter-
play between antitrust and intellectual property), for the Forum for the
Information Society of the Italian Presidency of the Ministers' Council, for
the Federation of the Italian Universities and in some Italian antitrust
cases. He is currently a member of the editorial board of the EuropeanJournal of Comparative Economics. His main research interests concern
industrial economics, antitrust, economic analysis of law and institutions,
intellectual property, information goods and knowledge production. Hehas published in these fields a number of papers in national and interna-
tional journals and books.
Pamela Samuelson is a Chancellor's Professor of Law and Information
Management at the University of California at Berkeley. She is a Director
of the Berkeley Center for Law and Technology, as well as an advisor to the
Samuelson High Technology Law and Public Policy Clinic at Boalt Hall.
She teaches courses on intellectual property, cyberlaw and information
policy. She has written and spoken extensively about the challenges that
new information technologies pose for traditional legal regimes, especially
for intellectual property law. She is co-author of the Software and Internet
Law casebook and of The law and economics of reverse engineering'
(jointly with Suzanne Scotchmer), published in 2002 in the Yale LawJournal and of 'Intellectual property rights in data?' (jointly with J.H.
Reichman) in 1997 in Vanderbilt Law Review, as well as sole author of a
large number of law review and computing journal articles. She is a Fellow
of the Association for Computing Machinery (ACM) and a ContributingEditor of Communications of the ACM. A 1971 graduate of the University
of Hawaii and a 1976 graduate of Yale Law School, Samuelson practiced
law as a litigation associate with the New York law firm Willkie Farr &Gallagher before turning to more academic pursuits. From 1981 throughJune 1996 she was a member of the faculty at the University of Pittsburgh
Law School, from which she visited at Columbia, Cornell, and Emory LawSchools. She has been a member of the Berkeley faculty since 1996.
Since 2002, she has been an Honorary Professor of the University of
Amsterdam.
Contributors xi
Arthur Snow is a Professor in the Department of Economics in the Terry
College of Business at the University of Georgia. He earned a PhD in eco-
nomics from the University of Wisconsin-Madison in 1979, and was
employed as Assistant Professor of economics at the University of Virginia
and at Georgetown University before joining the faculty at the University
of Georgia. His research on the economics of asymmetric information,
decision making under uncertainty, and on issues in public economics has
appeared in the Journal of Political Economy, the Journal of Public
Economics, the Journal ofRisk and Uncertainty, the International Economic
Review, the Economic Inquiry, and Economics Letters. He presided as pres-
ident of the Risk Theory Society at the 2004 Seminar, and currently serves
on the editorial board of the Geneva Papers on Risk and Insurance Theory.
Lisa N. Takeyama received her PhD in economics from Stanford University
under a National Science Foundation Fellowship. She currently holds a
Research Associate position in the department of economics at Amherst
College. Prior to coming to Amherst, she was on the faculty in the depart-
ment of economics at the University of Oregon. Included among her most
well-known publications on piracy and intellectual property are TheWelfare Implications of Unauthorized Reproduction of Intellectual
Property in the Presence of Demand Network Externalities' (Journal
of Industrial Economics) and 'The Intertemporal Consequences of
Unauthorized Reproduction of Intellectual Property' (Journal of Law and
Economics). She has also published other important work on a wide variety
of topics including durable goods monopoly, auctions and economic
growth in the Journal of Industrial Economics, the European Economic
Review, Economics of Innovation and New Technology, and Economics
Letters. Her article (with Eric B. Budish), 'Buy Prices in Online Auctions:
Irrationality on the Internet?' (Economics Letters) was listed by the pub-lisher as being among the top ten most requested articles for 2001 . In 1997,
she was named among the top 15 economics scholars at the Assistant
Professor rank at liberal arts colleges. She currently serves on the editorial
board of the Review ofEconomic Research on Copyright Issues. In 2003, she
was the local organizer for the second annual Congress of the Society for
Economic Research on Copyright Issues, held in Northampton, MA,USA.
Paul L.C. Torremans is the City Solicitors' Educational Trust Professor of
Intellectual Property Law at the School of Law of the University of
Nottingham (UK). He also holds an appointment as Professor of Private
International Law at the Faculty of Law of the University of Gent in
Belgium. He was awarded his PhD and LLM degrees by the University of
xii Developments in the economics of copyright
Leicester (UK) and he also holds law, notary law and law teaching degrees
from the University of Leuven (Belgium). He specializes in intellectual
property law and private international law and more specifically, in the
interaction between intellectual property law and private international law
and between intellectual property law and competition law. He has pub-lished various articles and books, including Holyoak and Torremans
Intellectual Property Law, (3rd edition, Butterworth, 2001; 4th edition,
Oxford University Press, 2005), and Intellectual Property and Private
International Law, with J.J. Fawcett (Clarendon Press, 1998). His latest
book is entitled Copyright and Human Rights, P. Torremans (ed.), (KluwerLaw International, 2004).
Ruth Towse is Reader in Cultural Industries at Erasmus University
Rotterdam. Her main area of expertise is in cultural economics with special
reference to the economics artists' labor markets and copyright in the cul-
tural industries. She has published widely in cultural economics and on the
economics of copyright, editing major collections in both fields, as well as
being the author of a number of books and articles in journals and in other
publications. She was Joint Editor of the Journal of Cultural Economics
from 1993-2002. She is President of the Society for Economic Research on
Copyright Issues (SERCI) (2004-6).
Patrick Waelbroeck is a research fellow from the National Science
Foundation (FNRS) in Belgium. He earned his PhD in economics from
Universite de Paris 1 (La Sorbonne) in December 2000 with highest honors.
He also holds an MA from Yale University and is a Fulbright alumnus. His
current research proposes both empirical and theoretical perspectives on
Internet piracy and technological protection in the music industry.
Richard Watt is Professor of Economic Theory at the Universidad
Autonoma de Madrid. After completing a Bachelor's and a Master's
degree in economics at Canterbury University in New Zealand, he earned
his doctorate in economic theory from the Universidad Autonoma de
Madrid in 1 990. His interest in the economics of copyright is manifested
in his book Copyright and Economic Theory: Friends or Foes! (Edward
Elgar, 2000). He is also actively researching the economics of risk bearingand risk sharing. He has published several papers in international journals
on this topic, as well as on the theory of oligopoly. He is currently the
General Secretary of the Society for Economic Research on CopyrightIssues (SERCI) and the managing editor of the Review of Economics on
Copyright Issues (RERCI).
Contributors xiii
Christopher S. Yoo is Associate Professor of Law at the Vanderbilt
University Law School. A graduate of Harvard College, the Anderson
School at UCLA, and the Northwestern University School of Law, his
research focuses primarily on how technological innovation and cutting-
edge economic theories are transforming the regulation of telecommunica-
tions and the electronic media. He is also pursuing research on the
application of imperfect competition models to copyright. His principal
work in these areas has been published in the New York University Law
Review, the Northwestern University Law Review, the Cornell Law Review,
the Georgetown Law Review, the Vanderbilt Law Review, the Southern
California Law Review, the Emory Law Journal, and the Yale Journal on
Regulation. Before joining the Vanderbilt law faculty, he clerked for Justice
Anthony M. Kennedy of the Supreme Court of the United States and was
associated with the DC law firm of Hogan & Hartson.
Preface
Lisa N. Takeyama
The Society for Economic Research on Copyright Issues (SERCI) was
established in 2001 as a response to the surging interest in copyright and the
expanding body of research applying economic analysis and theory to
copyright issues. As stated on the society's website (www.serci.org),
SERCI's objective is to actively promote and internationally distribute
quality academic research that relates economic theory to all aspects of
copyright and intellectual property of a cultural nature.
Since its inception, SERCI has held annual international congresses in
Spain, the USA, and Italy. From these congresses, two collections of papershave been published, including the present book. Additionally, SERCItakes great pride in the launch of its new journal entitled, Review ofEconomic Research on Copyright Issues (RERCI). Its inaugural issue has
already appeared, and RERCI will now appear twice yearly.
The first book-length collection of papers, The Economics of Copyright:
Developments in Research and Analysis (2003), edited by Wendy J. Gordonand Richard Watt and published by Edward Elgar, contains papers selected
from the inaugural SERCI congress held in Madrid in 2002. The present
book presents further developments in the economics of copyright. It con-
tains selections from the second annual SERCI congress held in
Northampton, Massachusetts, in 2003. Both congresses were richly fueled
by the productive interactions among the many economists, legal scholars,
and other copyright professionals present.
The chapters in this book, each of which has undergone peer-review and
revision, consider a wide variety of topics and issues in the economics of
copyright, including the role of economics in copyright law and policy,
peer-to-peer music file sharing, optimal fair use standards, the benefits of
copyright collectives, copyright and market entry, alternatives to copyright,
the impact of copyright on knowledge production, the proper balance
between copyright and competition law, and the application of systematic
principle to issues that arise at the periphery of intellectual property law.
The chapters also contain a diversity of methodologies employed - from
formal theoretical economic modeling to institutional and case study anal-
yses. Given the diversity of topics and analyses employed, this book should
xiv
Preface xv
be of interest not only to economists, but also to legal scholars and other
professionals interested in copyright issues and management.We would like to thank all of those involved in making this book pos-
sible, including, of course, our contributors, peer-reviewers, and congress
participants, all of whom contributed insightful ideas and suggestions that
helped make this a better book. We would also like to thank Richard Watt,General Secretary of SERCI, without whose vision and efforts SERCIwould not exist. We are also grateful to the Spanish Author's Society
(SGAE) for their continued support of, and belief in, SERCI. Many thanks
are also owed to Edward Elgar Publishing for their assistance in bringingthis volume to press.
Finally, I would like to thank my wonderful co-editors, Wendy Gordonand Ruth Towse, who provided an abundance of invaluable input on this
project and with whom I greatly enjoyed working.
General outline of the book
The book begins with a chapter by Pamela Samuelson, a legal scholar
famous for her pioneering work on the copyright issues raised by new infor-
mation technologies. In Chapter 1,she addresses what may be perhaps the
most fundamental question regarding economics and copyright: just what
role does economics have in copyright law and policy? Samuelson suggeststhat economics has not thus far been as influential in intellectual propertylaw and policymaking as it has been in other areas of economic regulation,
such as antitrust. She offers several possible explanations for this, includ-
ing for example, lack of economic expertise on the part of the relevant pol-
icymakers, and 'the tight nexus between the copyright industry and the
policymaking community'. Samuelson explains how economics does have
an important and useful role to play in copyright law and policy- touch-
ing also on ways economics can be misused - and offers several examplesand case studies.
The chapter by Samuelson, in addressing some of the institutional and
methodological difficulties that contribute to a law/economics divide,
serves to introduce the book as a whole. The primary focus of the book's
succeeding sections is the application of economic tools and analysis to
examine particular issues within copyright. They begin with consideringthe economics of copyright collectives.
In Chapter 2, Arthur Snow and Richard Watt offer a novel view of the
benefits of copyright collectives in their capacity to pool the risks of their
members. While it is well-known that collective administration of copyrightcan reduce transaction costs, the potential risk-sharing benefit of copyrightcollectives has not been previously considered. Using a formal analytical
model, Snow and Watt show that when the final market value of each
creator's work is uncertain, a copyright collective can improve the welfare
of its risk-averse members by alleviating individual risks. Snow and Watt
also point out that despite this potential risk-sharing benefit, the actual
allocation rules employed by copyright collectives in fact preclude this
benefit for their members.
In Chapter 3, Stan Liebowitz carefully examines the benefits and costs of
certain proposals that have been put forth as alternative solutions to the
problem of peer-to-peer music filesharing. Under these proposals, a copy-
right collective would be formed to administer a tax-and-subsidy system
xvi
General outline of the book xvii
whereby the record industry's traditional revenue sources would be
replaced (or supplemented) by revenue generated from taxation of ancil-
lary products (e.g., CDs, ISP services, etc.). Such a system could either
entirely replace the current copyright system or it could serve as a restitu-
tion mechanism to compensate copyright owners for losses induced byunauthorized copying. Some of the benefits of such a system mightinclude, for example, decriminalization of widespread illegal behavior,
enhanced incentives for the production of creative works, and increased
consumption. Liebowitz argues, however, that the costs of such a system,
including for example, those induced by tax distortions and a mispriced
royalty rate, as well as the potential for 'gaming' the system, make such pro-
posals (as his title indicates) 'a cure worse than the disease'.
Peer-to-peer music filesharing is also addressed by Anne Duchene andPatrick Waelbroeck in Chapter 4. However, in contrast to Chapter 3's
emphasis on proposals to resolve the problems induced by peer-to-peermusic filesharing, Chapter 4 is premised on the idea that firms may benefit
from peer-to-peer networks in their capacity to convey information to con-
sumers about new music cheaply, as compared to the more costly tradi-
tional mechanisms (e.g., advertising and promotions). Duchene andWaelbroeck's main thesis is that increasing copyright protection may have
significantly disparate welfare consequences, depending upon whether
information is conveyed to consumers via peer-to-peer networks ('demand-
pull') or via traditional mechanisms ('demand-push'). They develop a
model from which they derive welfare implications of increasing copy-
right's criminal penalties under both demand-pull and demand-push mech-anisms when firms can also employ their own additional technological
protection against piracy (e.g., digital rights management). They find,
among other things, that increasing criminal penalties benefits firms
employing demand-push mechanisms and induces them to increase their
level of technological protection, while the reverse is true for firms employ-
ing peer-to-peer networks. The effect of potential audience size on the
firm's preferred information transmission mechanism is considered as well.
In Chapter 5, Timothy Brennan addresses the issue of fair use, the
flexible doctrine employed by US law that allows behavior to occur - such
as some copying in educational, private, and journalistic contexts - that
would otherwise constitute prima facie copyright infringement. Brennan
examines the fair use doctrine in light of the diversity of policy goals that
have been proposed for its application. To this end, he develops a modelfrom which he derives optimal fair use standards under several different
policy objectives. Specifically, in addition to economic efficiency, he consid-
ers four possible alternate policy objectives: maximization of number of
works supplied, aggregate consumer surplus, gross benefit, and numbers of
xviii Developments in the economics of copyright
uses. By deriving ordered rankings for the optimal levels of fair use under
each of the above standards, Brennan shows how fair use might be adjusted'to address the breadth of policy goals that might be considered'.
In Chapter 6, Christopher Yoo criticizes the standard economic analysis
of copyright that typically assumes copyright turns creators into monopo-lists. Yoo argues that the well-known 'access-incentives' tradeoff is largely
driven by the assumed monopolization of the market. If one instead con-
siders the impact of copyright within a framework that allows for productdifferentiation and entry, the access-incentives tradeoff may be less
significant. By increasing producer appropriability and profit, increased
copyright protection can stimulate entry of competitors producing similar
works, resulting in lower prices, increased product variety, and increased
access. Yoo suggests that this approach also makes a broader set of policyinstruments potentially available for use, but disentangling the effects of
one from another can be quite complicated.In Chapter 7, Giovanni Ramello argues that the 'standard' economic
analysis of copyright is too narrowly focused on individual creation. In par-
ticular, economists have failed to consider the inherently social nature of
knowledge production, which takes place in what Ramello calls the 'social
creative commons'. While economists have recognized that copyrightrestricts access to output (i.e., to the copyrighted good), they have been
insufficiently sensitive to the ways in which copyright also restricts access
to inputs (i.e., to the creative commons and the production of knowledge).Ramello argues that restricting access to the creative commons and hence,
to knowledge sharing, reduces the future productive capacity of the crea-
tive commons, a consequence which Ramello calls an 'opposite tragedy of
the creative commons'. For these reasons, Ramello argues for a 'minimal-
ist' rather than a 'maximalist' approach to copyright.In Chapter 8, Alessandra Narciso and Paul Torremans examine the
interaction and proper balance between copyright and competition (anti-
trust) law. Their analysis largely focuses on the German IMS Health case,
a case that well illustrates the tension that exists between copyright and
competition law. Narciso and Torremans' main thesis is that the applica-tion of competition law to cases in which copyright exists should factor in
and avoid undermining the structural system originally put in place by
copyright. So, for example, in antitrust cases involving copyright, it may be
inappropriate to employ as precedent court decisions in non-copyright
cases, at least without explicit recognition of the special structure intention-
ally put into place by copyright. Narciso and Torremans apply this notion
to several other policy issues raised by the IMS Health case, with a focus
on the essential facilities doctrine.
In Chapter 9, Scott Kieff and Troy Paredes provide a strong argument
General outline of the book xix
for applying, both at the judicial and legislative levels, 'a basics approach'to issues in dispute that arise at 'the periphery of intellectual property law'.
Too often, they argue, such disputed matters are resolved in a piecemealfashion by creating specialized approaches and unique doctrines. This
approach, they argue, 'selectively emphasizes or alternatively ignores par-
ticular features of various legal disciplines in crafting specialized doctrines
for IP'. In contrast, under their competing 'basics approach,' Kieff and
Paredes propose that such matters would be best resolved by applying the
basics of the relevant substantive law, such as intellectual property law,
antitrust law, and property law. Although Kieffand Paredes focus on patent
law, their analysis applies equally to copyright. For example, they arguethat the anti-circumvention provisions of the Digital Millennium
Copyright Act are inconsistent with the 'basics approach' since the trans-
actions it regulates are better governed under existing contract law or the
IP law of indirect infringement.
As this brief review of the contents of this volume shows, both econo-
mists and legal scholars are applying their analytical skills to a broad rangeof theoretical and applied areas in the economics of copyright. We hopethat the volume will stimulate further research and debate in this importantarea.
1 . Should economics play a role in
copyright law and policy?
Pamela Samuelson
1.1 INTRODUCTION
The principal justification for intellectual property (IP) laws in the Anglo-American tradition is economic. 1 Without a grant of exclusive rights,
authors and inventors would have too little incentive to invest in socially
beneficial innovations for this simple reason: developing the first embod-iment of an innovation generally requires very costly investments; subse-
quent copies are generally far less costly to make, and are often trivially
cheap and easy.2 If creators cannot prevent others from selling products
embodying their innovations - which copyists can offer at a lower price if
able to free-ride on the innovator's first copy costs - innovators will not
be able to recoup their research and development (R&D) costs and justify
further investments in innovations. 3Although the non-excludable, non-
rival character of intellectual creations resembles public goods,4
it is
socially desirable for private actors, rather than the government, to makethe investments to overcome this public goods problem.
5 Intellectual
property rights are an ingenious device for addressing this problem: the
government grants rights to qualifying innovators, and the rights provideincentives to individuals to invest in innovation because the rights
granted provide assurance that innovators can recoup their R&D costs if
their intellectual creations prove to be valued by the public.6Original
works of authorship are particularly vulnerable to market-destructive
copying, which is why few question the economic wisdom of a copyright
regime.7
Intellectual property would, for this reason, seem to be 'a natural field
for economic analysis of law.' 8It would, hence, be logical for economic
analysis to have considerable influence in shaping copyright and other intel-
lectual property laws.9 Yet, as compared with other fields of economic regu-
lation, particularly antitrust law, economics has had very little influence
thus far in the intellectual property law and policymaking process. This
chapter considers why this might be so in the field of copyright law and
2 Developments in the economics of copyright
whether economics might become more influential in this field of law and
policy in the future.
The lack of influence is not attributable to inattention from the scholarly
community. Law and economics scholars have produced a substantial and
growing literature that analyzes many aspects of copyright law. 10 Amongthe copyright rules that are widely viewed as economically sound are: the
idea/expression distinction (that is, copyright protects only an author's
expression, not her ideas),11 the independent creation defense,
12 the deriva-
tive work right (that is, the right to control adaptations of the work, such
as a motion picture version of a novel),13 the separation of the ownership
of a copy of a work from ownership of the copyright in it,14 the ability of
authors to assign 'slices' of their exclusive rights,15 limits on the duration of
copyright,16 and the work made for hire rule (which treats employers as the
'author' of works by employees performed within the scope of employ-
ment).17 The economics of copyright collectives and of compulsory licens-
ing has also garnered scholarly attention. 18 Cultural economics, a subfield
of economics, has studied the market for art and cultural products, such as
opera and theatre, and offered insights about policies needed to induce
investment in these cultural forms. 19 One bold work considers whether eco-
nomics is friend or foe of copyright,20 but rare is the work that attempts to
analyze the economics of copyright as a whole.
Among the most studied subjects has been the economics of the fair use
doctrine of US copyright law. 21 Economists and economically minded legal
scholars often view uses of copyrighted works as 'fair' if high transactions
costs or other factors have impeded the effective establishment of a market
for clearing rights.22
Particularly influential has been Wendy Gordon's
analysis of the fair use issues in the Sony Betamax case. Gordon argued that
use of Betamax machines to tape broadcast television programs for later
viewing should be fair because the costs of negotiating rights clearances
would greatly exceed fees collectable for this kind of socially beneficial
use. 23 Because the economic effects of a challenged use are almost alwaysof central importance in judging whether the use will be deemed non-
infringing as a fair use, economics is especially useful in providing guidanceon this copyright issue.
24
Economists and economically minded legal scholars have questioned or
criticized certain copyright rules as economically unsound. For example, it
is unnecessary for personal letters and diaries to be automatically qualified
for copyright protection, given that these works would be created without
the economic incentives of copyright.25
Special exemptions from copyright
liability for certain events organized by agricultural cooperatives and vet-
erans' groups seem to be the product of special interest lobbying, rather
than rigorous economic analysis.26 Extension of the duration of copyrights
Should economics play a role in copyright law andpolicy? 3
in existing works27 and so-called 'restoration' of foreign copyrights that had
long been in the public domain for failure to comply with US formalities28
(e.g., putting copyright notices on published copies of the work) are also
economically questionable.29
As interesting and provocative as the economics of copyright literature is,
even its most ardent fans would have to admit that economics has rarely
played a significant role in the copyright law and policymaking process.
Section 1 .2 will consider various reasons why economics has not been more
influential in copyright up until now. Section 1 .3 will discuss the economics
of extending the terms of existing copyrights and the Supreme Court's deci-
sion in the Eldred vs. Ashcroft case. 30 Although an initial reading of the
Supreme Court's decision might suggest that economics will have little influ-
ence in future court decisions interpreting copyright law, Section 1 .3 explains
why this interpretation of the Eldred decision is incorrect. Section 1 .4 will
discuss particular venues in which economics would be most useful in the
copyright law and policymaking process going forward. Legislators should
(and occasionally do) seek independent economic expertise when consider-
ing copyright and related intellectual property proposals. Courts mayfind economic analysis a useful input to sound decision-making, especially
in cases involving new technology issues unanticipated when legislatures
adopted copyright laws. Section 1.5 will consider some reasons why copy-
right professionals might resist the incursion of economics into the copy-
right policy process, and why economics should nonetheless play a more
important role in copyright policymaking in the future.
1 .2 WHY HAS ECONOMICS HAD SO LITTLEIMPACT ON COPYRIGHT SO FAR?
One salient factor explaining the low impact of economic analysis on copy-
right law and policy is a lack of economic expertise in the relevant policy-
making community. Many copyright professionals have backgrounds in
fields other than economics and are more inclined to embrace a romantic
conception about art and literature, and the people and firms who create and
commercialize copyrighted works that would disincline them to look to eco-
nomics for guidance about how the law should be crafted. 31 Unlike the US
Department of Justice Antitrust Division, the Federal Trade Commission,and the Federal Communications Commission (FCC), for example, all of
whom regularly participate in policymaking as to other economic regula-
tions, the US Copyright Office (as well as its patent and trademark coun-
terpart) has no in-house economic experts who might contribute insights
about economic effects of various policy proposals.32
4 Developments in the economics of copyright
For path-dependent historical reasons, copyright and other intellectual
property policy matters in the US have generally been considered by sub-
committees of the House and Senate Judiciary Committees, rather than byCommerce Committees, from which economic regulations often emerge.
These subcommittees have relied heavily on industry witnesses when
considering copyright and other intellectual property-related legislation.
Copyright industry groups, in fact, often write the laws that the legislature
enacts. 33 A revolving door enables copyright experts to move from the
legislature or legislative committee staff positions to law firms represent-
ing major copyright firms to industry associations to government agencies
responsible for framing copyright law and policy.34 This contributes to
insularity in copyright policy analysis.
The US is not alone in its systematic, if perhaps unconscious, exclusion
of or inattention to economic analysis in the copyright policymaking
process. Nations whose legal rules are grounded in the natural right of
authors in their works,35
especially those that affirm that authors have
moral rights in their works, have less reason to perceive economics as a
useful input in the policymaking process than the US which at least histor-
ically conceptualized copyright in utilitarian terms.
Although the tight nexus between the copyright industry and the policy-
making community is the main reason economics has thus far had so little
influence on copyright law and policy, it is not the only factor at work.
An impediment to influence is that economists who study copyright issues
often speak in a language inaccessible to policymakers and copyright pro-
fessionals. Few copyright professionals are able, for example, to penetrate
the mathematical expressions that often pervade the academic economic
literature. If economists wish for their work to inform the copyright policy
process, they will have to learn to translate their insights into a more ver-
nacular form. They may also need to publish policy-relevant work in venues
where policymakers would be likely to read it.
Cultural differences are also a factor. Lawyers and policymakers view
some problems that deeply engage economists (e.g., the optimal duration
of a patent or of copyright protection for computer software36) as a waste
of time. Other issues of interest to economists (e.g., network externalities)
may have policy implications that are not readily apparent to copyright
professionals.37 A legal and policy audience may find incomplete an eco-
nomic analysis whose policy implications are not spelled out. If economists
want their work to inform the copyright policymaking process, they will
have to invest the extra time necessary to articulate the policy implications
that flow from their analyses.
The standard economic practice of articulating simplifying assumptionsbefore constructing an economic analysis does not mesh well with the
Should economics play a role in copyright law andpolicy? 5
mindset of lawyers and policymakers either. If the whole edifice of an
analysis depends on assumptions A, B, C, D and E being true and at least
one of them is questionable, lawyers and other policy analysts may be
inclined to discredit the analysis as a house of cards.
Influencing copyright law and policy will also require economists to give
this field somewhat greater attention. Economic research on intellectual
property issues has tended to focus on subjects for which funding from
government or foundation grants is available. Until quite recently, grant-
makers rarely supported research on copyright issues. Far more economic
research has been done on patent issues, mainly because of their significance
for the predominant industrial economy of the nineteenth and twentieth
centuries, but also because more funding has historically been available to
study the economics of patents. The emergence of an information-based
economy in the late twentieth century has attracted the attention of econo-
mists and economically-informed lawyers to exploring the economics of
copyright.38
Also an impediment to influence is the plain fact that economics will not
always yield an unimpeachable policy prescription for copyright issues.
This is partly due to the complex intricacies of the many industries that rely
upon or intersect with copyright law. Since copyright law has many provi-
sions of general applicability that affect a wide range of industries - from
book publishing to architecture to computer software to theatrical
performances, just to name a few - it may be very difficult, if not impossi-
ble, to assess with any precision the economic effects likely to flow from
adoption of particular copyright rules, let alone the economics of the copy-
right regime as a whole. 39
It is to be expected, moreover, that on many issues, economists will differ
in their analyses of the same issue. In A&M Records, Inc. vs. Napster, Inc.,
for example, economic experts offered widely varying analyses of the effects
of peer-to-peer file sharing on the market for digital music.40 Some econo-
mists considered file sharing to be benign or beneficial, while others asserted
it was harmful to the market for sound recordings.41
Policymakers will
understandably be reluctant to put much weight on economic expertise
if equally qualified experts offer irreconcilably conflicting analyses of the
same phenomenon.Another barrier to influence is the understandable reluctance of copy-
right lawyers and policy professionals to acknowledge the public choice
problems presented by the current copyright law and policy-making
process.42 These problems are well-documented in the scholarly literature.43
Major copyright firms are well-organized and well-funded. They typically
have a common interest in getting stronger legal rules from the legislature.
Hence, it may be a sound investment for them to lobby to achieve the
6 Developments in the economics of copyright
concentrated benefit a legislature can grant them. The public may ultim-
ately have to pay higher costs if copyright lobbyists are successful, but these
costs are diffuse and distributed over a broad base of people and firms.
Collective action problems make it difficult for parties that will be nega-
tively affected by higher protection rules to organize effective resistance to
copyright industry lobbying. This mix of concentrated benefits and dis-
tributed costs is likely to yield the best laws money can buy. The CopyrightTerm Extension Act (CTEA),
44 the Digital Millennium Copyright Act
(DMCA) anti-circumvention provisions,45 and state 'super-DMCA laws46
are three widely recognized examples from the US experience.
1 .3 WHAT EFFECT WILL ELDRED VS. ASHCROFTHAVE ON THE FUTURE OF ECONOMICS INCOPYRIGHT?
A stark example of the battle between economic and non-economic rea-
soning in the copyright law and policy process is evident in the Eldred vs.
Ashcroft case decided in 2003 by the US Supreme Court.47 Eric Eldred, an
online publisher of public domain works, challenged the CTEA, which
extended copyright terms an additional twenty years, based on an eco-
nomic theory of the Intellectual Property Clause of the US Constitution.48
Article I, Sec. 8, Cl. 8 gives Congress power 'to promote the progress of
Science and useful Arts, by securing to Authors and Inventors for limited
times the exclusive Right in their respective Writings and Discoveries.'
Boiled down to its essence, Eldred's theory was that this clause prohibited
Congress from granting exclusive rights to authors without the quid pro
quo of a 'progress-promoting' act (i.e., authorship of a newly created orig-
inal work) in return.49 Stated more colloquially, Congress violated the
Constitution by extending the term of existing copyrights for an additional
twenty years because the public got nothing in return for that grant. The
CTEA was purely a windfall to existing copyright owners. Such a quidpro
nihilo, Eldred argued, failed to satisfy constitutional standards. 50
Seventeen economists -including five Nobel Prize winners -
signed a
friend-of-the-court brief to the US Supreme Court in support of Eldred's
challenge to the CTEA. 51 The brief made the obvious point that the CTEAcould not provide incentives for the creation of works already in existence. 52
It went on to consider the deadweight loss that a longer term of above-cost
pricing would bring, as well as the substantial transactions costs the CTEAwould impose on subsequent users. 53
Tracking down copyright owners and
negotiating with them is costly, even if the owners do not insist on paymentsfor reproducing or distributing copies of their works. CTEA impedes
Should economics play a role in copyright law andpolicy? 7
public access to many works that are no longer commercially exploited, but
that may still be of interest, as well as preventing the creation of many newderivative works. 54 The economists' brief also pointed out that life of the
author plus 70 years was virtually equivalent to the perpetual term of copy-
right that the Constitution forbids. 55
The Supreme Court rejected Eldred's challenge to the CTEA. Congresshad a rational basis, the Court decided, for believing that CTEA would lead
to more investments in distributing and preserving existing works. 56 The
Constitution gave to Congress the responsibility of crafting copyright legis-
lation, and even if it exercised this power unwisely, the Court was reluctant
to second-guess legislative decisions. 57Only Justice Breyer would have
struck down theCTEA on an economic interpretation of the Constitution.58
The Eldred opinion is starkly different from many of the Court's rulings
on economic regulations. Economic analysis is generally quite influential
when the Supreme Court assesses economic regulations, especially amongthe more conservative members of the Court. 59 Given how often the Court
has emphasized the economic incentive rationale for the existence of copy-
right law,60
it is striking how little attention was paid to economics in Eldred.
Justice Ginsburg, who authored the majority opinion, is not, of course, the
most economically minded of the Justices. But the de-emphasis on eco-
nomics in Eldred has other roots.
There is a deep rift within the Court about constitutional powers of
Congress and about how much deference to give to legislation emanatingfrom it.
61 Eldred's lawyers drew upon several Supreme Court precedents
interpreting the constitutional powers of Congress in a restrictive manner
and establishing rigorous standards for judging the constitutionality of
certain kinds of legislation.62 Further limitations on Congress' power under
Article I, sec. 8, cl. 8 were, Eldred's lawyers argued, consistent with these
recent precedents,63 as well as with the Court's prior rulings under the
Intellectual Property clause.64 Eldred's lawyers hoped the analogy to these
other restrictive rulings would woo the more conservative members of the
Court to further restrictions on Congress' constitutional powers in Eldred.
Eldred's lawyers used a different strategy to persuade Justices Breyer and
Stevens (who had dissented in the earlier challenges to Congress' power) to
support Eldred's cause. 65
There was, in short, a lot going on in the Eldred case that had nothingwhatever to do with copyright law, but a good deal to do with the Court's
internal debate on the scope of Congress' powers under the Constitution.
Other factors were also at play. Most significant was the indisputable fact
that Congress had extended the term of existing copyrights numerous times
before: in 1790, 1831, 1909, and 1976, as well as several times leading up to
enactment of the Copyright Act of 1976.66 Although Eldred's lawyers
8 Developments in the economics of copyright
developed an elaborate chart to distinguish the CTEA from other exten-
sions,67 this chart may have backfired because it underscored the frequency
of these enactments. To comprehend the scope of Congress' power [to
extend terms of existing copyrights] under the Copyright Clause,' said the
Court,'
"a page of history is worth a volume of logic."'68
The Court was also understandably concerned about the implications of
a ruling in Eldred's favor. If the CTEA was invalid, the retroactive term
extensions in the Copyright Act of 1976 would almost certainly be chal-
lenged, and it was not immediately evident how the Court could find a prin-
cipled basis for distinguishing this extension from the CTEA. 69 Had the
Court acceded to the heightened scrutiny for which Eldred argued,70
it
might have opened the floodgates of constitutional challenges to manyforms of intellectual property legislation, not just to copyright rules. 71
Finally, Eldred was willing to accept that life of the author plus 70 years
was a 'limited time' under the Constitution applied prospectively.72 In view
of this, the Court thought that life plus 70 years must also be sufficiently
limited applied retroactively.
Taking into account the range of explanatory factors, the Eldred decision
may be less of a setback for a significant role for economic analysis in copy-
right law and policy than a cursory glance might suggest. The Court was
not willing to interpret the Constitution as mandating economically sound
results, but this does not mean that the Court believes that copyright law
and policy should be uninformed by economic analysis. The next section
will consider several contexts in which economic analysis has been and is
likely to be a useful input into the policymaking process for copyright and
related forms of intellectual property law.
1 .4 WHEN MIGHT ECONOMICS BE USEFUL?
When legislatures consider proposals to add a new subject matter to copy-
right or to adopt new copyright-like forms of legal protection, they should
(and occasionally do) seek input from impartial economic experts about
whether such legislation is necessary to address a market failure that is
deterring socially optimal investments in particular classes of innovations
or information resources. Economic expertise, for example, had some influ-
ence in the US when the National Commission on New Technological Uses
of Copyrighted Works (CONTU) made economic as well as doctrinal
arguments urging that copyright law be used to protect computer pro-
grams.73 The European Union too relied on economic analysis in crafting
appropriate copyright rules for computer programs.74 Owing in substantial
part to the decisions of the US and EU to protect programs by copyright,
Should economics play a role in copyright law andpolicy? 9
this rule became an international standard in 1 994 after member states of
the World Trade Organization (WTO) acceded to the Agreement on Trade-
Related Aspects of Intellectual Property Rights (TRIPS), which required
copyright protection for programs.75
The US Congress has a more mixed record in relying on economic
analysis when considering sui generis (of its own kind) forms of legal pro-
tection for other innovations. It relied heavily on economic argumentswhen adopting in 1984 a new law to protect original layouts of circuits in
semiconductor chips.76However, it adopted a copyright-like form of legal
protection for vessel hulls and other boat parts in 1998 without serious
economic evidence or expertise to support it.77
Yet, economic analysis has
had some cautionary influence on Congress in the debate about whether
to adopt an equivalent sui generis right in the data in databases such as the
EU adopted in 1996. 78
The European Commission asserted a market failure rationale to justify
the issuance of a Directive to member states of the EU requiring them to
adopt a sui generis form of intellectual property protection for controlling
extraction and reuse of data in databases. 79 Without such a law, the theory
went, there would be too few incentives to invest in database develop-
ment. 80 The new law confers on those who invest substantial resources in
database development with 15 years of exclusive rights to control extrac-
tion and reuse of all or substantial parts of the databases (renewable for
additional 15-year terms so long as substantial investments continue).81
Although the Commission invoked an economic rationale for this law, it
neglected to provide any empirical evidence of market failure or rigorous
economic analysis of the regime proposed to cure the perceived failure. Hadthe Commission been more rigorous in its use of economics, the EuropeanParliament might have realized there was less of a need for this sui generis
law than the Commission posited, and might have adopted a somewhat
narrower form of legal protection for data in databases. 82
Economic input should also be sought when copyright laws are
amended. Had the US Congress, for example, sought impartial advice from
economists about the effects of copyright term extension before enacting
the CTEA, it is conceivable that economic arguments against its adoptionwould have persuaded some in Congress to oppose it.
83 Economic analysis
might also have been useful in assessing the anti-circumvention provisions
of the Digital Millennium Copyright Act (DMCA), which have been
heavily criticized as overbroad and anti-competitive.84
Copyright rules are also sometimes promulgated in the context of agency
rule-making, and economic analysis should be a useful input to decision-
making in this venue as well. For example, the Copyright Office is required
to make determinations about claims for allocation of compulsory license
10 Developments in the economics of copyright
revenues to particular copyright owners. 85 Economic analysis would be
helpful in making such assessments. Proposals for new compulsory licenses
or revisions to existing compulsory licenses should likewise be scrutinized
for economic soundness. 86 The US Copyright Office has been given powerto conduct rule-makings every three years to consider whether to exemptcertain classes of works or uses of works from anti-circumvention rules. 87
This should be a forum in which the Office might be educated about the eco-
nomic effects of broad anti-circumvention rules.
Once rules have been legislated or issued by an appropriate agency, eco-
nomics may sometimes have a useful role to play, and sometimes not, in
how the law is interpreted thereafter. If a legislature has adopted an open-ended rule (e.g., that copyright protects an author's expression, not his or
her ideas or methods), economic analysis may be a useful tool in interpret-
ing the law in an economically sound way. However, if the legislature has
precisely crafted certain rules that prove to be economically harmful, it maybe more appropriate to invoke economic analysis to support proposedamendments to the law than to interpret the rules in litigation.
An instructive example of the pros and cons of economic analysis in
the interpretation of open-ended copyright rules can be found in the USjudicial experience interpreting copyright law as applied to computer
programs.88 Whelan Associates vs. Jaslow Dental Labs., Inc. was the first
American appellate decision to consider whether the 'structure, sequence,
and organization' ('SSO') of computer programs could be protected by
copyright law. 89 Whelan relied upon both doctrinal and economic groundsin ruling that SSO was copyright-protectable.
The doctrinal analysis relied on this syllogism: computer programs are
considered to be literary works under US copyright law, and since the struc-
ture, sequence, and organization of literary works are generally protectable
by copyright law, the structure, sequence, and organization of programsshould be protected by copyright law as well.90
Complementing this doctrinal analysis was an economic argument that
focused on the need for software developers to have sufficient protec-
tion to recoup development costs and on the locus of value in computer
programs:
By far the larger portion of the expense and difficulty in creating computer pro-
grams is attributable to the development of the structure and logic of the
program, and to debugging, documentation and maintenance, rather than to the
coding. See Frank, Critical Issues in Software 22 (1983) (only 20% of the cost of
program development goes into coding); Zelkowitz, Perspective on Software
Engineering, 10 Computing Surveys 197-216 (June, 1978). See also InfoWorld,Nov. 11, 1985 at 13 ('the "look and feel" of a computer software product often
involves much more creativity and often is of greater commercial value than the
Should economics play a role in copyright law andpolicy? 1 1
program code which implements the product . . .'). The evidence in this case is
that Ms. Whelan spent a tremendous amount of time studying Jaslow Labs,
organizing the modules and subroutines for the Dentalab program, and workingout the data arrangements, and a comparatively small amount of time actually
coding the Dentalab program.91
The Third Circuit took seriously a friend-of-the-court brief submitted bya software industry organization which argued that if copyright protection
did not extend to the SSO of a program, the software industry would be
jeopardized because the law would provide too little protection to induce
an optimal level of investment in development of computer programs.92
The Third Circuit directly responded to this plea: 'The rule proposed here,
which allows copyright protection beyond the literal computer code, would
provide the proper incentive for programmers by protecting their most
valuable efforts, while not giving them a stranglehold over the developmentof new computer devices that accomplish the same end.'93
Between 1986 and 1992, the Whelan decision was influential in subse-
quent US cases, both in its doctrinal and economic reasoning.94 The first
decision that challenged Whelarfs hegemony was Computer Associates vs.
Altai, Inc. in 1992.95 On one important point, the Second Circuit Court
of Appeals agreed with Whelan: the structure, sequence and organization
of a program could, in an appropriate case, be protected by copyright
law.96 Unlike Whelan, which regarded program structure as exempt from
infringement only when there was essentially no other way to structure the
program,97 the court in Altai reasoned that similarities in the structure of
two programs might, however, be due to functional constraints such as the
need to develop a program that would interoperate with another program,
efficiency considerations, or use of the same standard programming tech-
niques, none of which was protected by copyright law.98
The court in Altai directed that these and other unprotectable elements
of programs be 'filtered out' before infringement analysis began99 to ensure
compliance with section 102(b) of the US copyright statute which states
that '[i]n no case shall copyright protection for an original work of author-
ship extend to any idea, procedure, process, system, method of operation,
concept, principle, or discovery, regardless of the form in which it is
described, explained, illustrated, or embodied in such work.' 100 The court
ultimately ruled that some structural similarities in Altai were attribut-
able to the fact that both Computer Associates (CA) and Altai were
developing programs to interoperate with the same three IBM operating
system programs, and other structural similarities were to be expected in
programs of that kind. 101Hence, it affirmed the lower court's finding of
noninfringement.
12 Developments in the economics of copyright
As in Whelan, CA and its amid predicted dire economic consequencesfor the software industry if courts did not provide strong protection to
program structure. However, the Second Circuit took a different view:
[The Supreme Court's decision in] Feist teaches that substantial effort alone
cannot confer copyright status on an otherwise uncopyrightable work. . . .
[Djespite the fact that significant labor and expense often goes into computerprogram flow-charting and debugging, that process does not always result in
inherently protectable expression. Thus, Feist implicitly undercuts the Whelan
rationale, 'which allow[ed] copyright protection beyond the literal computercode ... [in order to] provide the proper incentive for programmers by protect-
ing their most valuable efforts. . . .' [citation omitted]. We note that Whelan wasdecided prior to Feist when the 'sweat of the brow' doctrine still had vitality. In
view of the Supreme Court's recent holding, however, we must reject the legal
basis of CA's disincentive argument.102
The Second Circuit went on to say:
[W]e are unpersuaded that the test we approve today will lead to the dire conse-
quences for the computer program industry that plaintiff and some amici
predict. To the contrary, serious students of the industry have been highly crit-
ical of the sweeping scope of copyright protection engendered by the Whelan
rule, in that it 'enables first comers to "lock up" basic programming techniquesas implemented in programs to perform particular tasks.' 103
The Second Circuit warned that if courts heeded purely economic argu-
ments for broadening the scope of copyright protection for computer pro-
grams, this would impair the integrity of copyright law:
While incentive based arguments in favor of broad copyright protection are
perhaps attractive from a pure policy perspective, . . . ultimately, they have a cor-
rosive effect on certain fundamental tenets of copyright doctrine. If the test wehave outlined results in narrowing the scope of protection, as we expect it will,
that result flows from applying, in accordance with Congressional intent, long-
standing principles of copyright law to computer programs.104
The Second Circuit pointed out that patent protection might be a moresuitable way to protecting some program innovations than copyright.
105 If
copyright proved to be too 'thin' to provide proper incentives to program
developers, this was a matter for Congress, not the courts, to consider. 106
Altai holds that copyright protection is unavailable to elements of pro-
grams necessary for interoperation with other programs (that is, program
interfaces).107
Judged purely in terms of the creativity and judgment
required to design them, program interfaces might initially seem to be 'ori-
ginal' enough to be protectable expression as a matter of copyright law. 108
Should economics play a role in copyright law andpolicy? 1 3
However, once developed, program interfaces unquestionably constrain the
design choices of subsequent programmers seeking to develop software
capable of successfully interacting with an existing program.109
The competition policy significance of copyright protection for inter-
faces was recognized in Europe in part because the Competition Policy
Directorate of the European Commission had taken action against IBMarising from its practice of changing interfaces in a manner that had exclu-
sionary impacts on European developers of computer peripherals.110 If
European software developers were going to be able to compete with USsoftware in the world market, these developers would need to be able to
use interface information from American programs. The EC Competition
Policy Directorate intervened in negotiations surrounding the drafting of
a directive on the legal protection for computer software that resulted in
these economic considerations being brought to bear on the scope of
copyright protection for software in Europe.111 The US and EU consen-
sus on limiting copyright to enable interoperability influenced other
nations as well. 112
Sega Enterprises, Ltd. vs. Accolade, Inc. 113 addressed a related interoper-
ability issue as to whether programmers could lawfully decompile or disas-
semble other firms' programs in order to get access to information needed
to develop an interoperable program. This was a dicey issue for copyrightlaw because the decompilation and disassembly process inevitably requires
the making of a number of copies of the target program.114 Such copies
arguably run afoul of the exclusive right that copyright law confers onauthors to control the reproduction of their works in copies.
115 Yet unless
copyright law recognized at least a limited right to decompile or disassem-
ble programs, the decision not to protect interface information would be
significantly undermined.116
Accolade relied on the fair use defense of US copyright law in respond-
ing to Sega's charge of infringement.117 Courts consider four factors in
determining whether a use is fair: the purpose of the defendant's activ-
ities, the nature of the copyrighted work, the amount and substantiality
of the use, and the effects of the use on the market for the copyrightedwork. 118 Accolade persuaded the court that decompilation was the only
way Accolade could obtain access to the unprotectable interface informa-
tion in Sega's game programs (apart from entering into an unacceptable
licensing agreement) and that accessing and reusing this information
had made it possible for Accolade to develop new non-infringing works
to compete in the market with Sega's works. The court consequentlydecided that intermediate copying of programs for the purpose of achiev-
ing interoperability was fair use. 119 The court relied on economic reason-
ing in support of the ruling, although its analysis is not as systematically
14 Developments in the economics of copyright
economic as some of the law review commentary reaching the same legal
conclusion. 120
Both Altai and Sega vs. Accolade have been widely accepted as sound
rules in the US and abroad,121 and the software industry has continued to
flourish in the decade after these decisions were rendered. 122 Two lessons
from the software copyright caselaw are, first, that courts should be waryof unsubstantiated economic claims that without broad protection, the
affected industry will be severely harmed, and second, that considering evi-
dence of economic impacts of the court's ruling in particular contexts can
assist judges in making sound decisions in copyright cases. While software
law has now achieved a stability it lacked in the 1980s and early 1990s, chal-
lenges posed by new technologies for the application of copyright law to
digital works continue to be troublesome. Perhaps the most potent examplehas been the entertainment industry lawsuits against the developers of
peer-to-peer file sharing technologies.123 While economic analysis has been
of some assistance to judges forced to grapple with the thorny legal issues
these cases present,124
it may be necessary for legislatures to address what
standards ought to apply in regulating technologies with infringing and
non-infringing uses. 125
It is worth noting that the Canadian Supreme Court recently considered
another new technology issue that was certainly not anticipated when the
Canadian legislature enacted its copyright law. The Court concluded that
the economic rights of an artist were not infringed when a purchaser of
copies of his work transferred the copyrighted image from paper to canvas
and then resold the copies.126
Although not invoking economic analysis as
such, the care with which the Canadian Court considered the economic
impacts of the challenged use was impressive and bodes well for further
uses of economic arguments in copyright cases in Canada.
1.5 CONCLUSION
Copyright is an important form of economic regulation. This chapter has
considered several reasons why economic analysis has had relatively little
effect on copyright law and policy thus far. It has also argued that eco-
nomics has much to offer in assisting policymakers and judges in crafting
sound copyright law and policy.
There are several reasons why courts and policymakers may resist the
incursion of economics into the copyright law and policymaking process
and in the interpretation of copyright law. As noted in Section 1 .2, manyin the copyright field, including policymakers, have little or no economic
expertise and little inclination to seek it out. Second, copyright industry
Should economics play a role in copyright law andpolicy? 1 5
professionals have been successful in getting much if not all of what theywant without the aid of economics, so they may perceive little need for
this analytical tool. Third, copyright industry professionals may be waryof embracing economics because although it may sometimes provide
support for stronger protection rules, it may not always cut in their favor.
They may resist embracing a tool that might provide grounds for policy
outcomes they disfavor. Fourth, until some major participant in the copy-
right policy-making process decides that economic analysis would be valu-
able as input to the policy process, the inertia of the current situation is
likely to persist.
Even so, I predict that economic analysis will have greater impact on
copyright in the future. A new generation of scholars is building up a rich
corpus of policy-relevant copyright research and is articulating the policy-
relevant implications of this research. 127 There is now a Society for
Economic Research on Copyright Issues (SERCI), which hosts an annual
international conference and publishes edited volumes of research on the
economics of copyright and a journal, the Review ofEconomic Research on
Copyright Issues (RERCI).128 The increasing importance of copyright
industries to the US and world economy should make policymakers more
receptive to insights that economics can provide.129
Moreover, the inclusion
of copyright and other intellectual property rights within the framework
for promoting world trade by virtue of adoption of the TRIPS Agreementas part of the agreements establishing the World Trade Organizationshould lead to greater use of economic learning to shape global intellectual
property norms.13 This should lead to greater reliance on economic expert-
ise at the national legislative and policymaking process.
Perhaps it is only a matter of time before the US Copyright Office andsister agencies in other countries will hire not just a resident economist, but
a group of them, one of which might be designated a Chief Economist,and give serious consideration to economic effects before endorsing par-ticular legislative initiatives. Economic principles can contribute to evolv-
ing our understanding of the access/incentives tradeoff so that copyrightwill continue - or perhaps return to - its long-standing tradition of pro-
moting the progress of science and the useful arts to the benefit of
humankind.
ACKNOWLEDGMENT
I wish to thank Wendy Gordon and Richard Watt for inviting me to give a keynote address at
the annual conference of the Society for Economic Research on Copyright Issues in
Northampton, MA in June 2003.
1 6 Developments in the economics of copyright
NOTES
1 . See, e.g., Paul Goldstein, Copyright's Highway: From Gutenberg to the Celestial Jukebox,168-69 (1994); Mark A. Lemley, The Economics of Improvement in Intellectual PropertyLaw, 75 Tex. L. Rev. 989, 1074-76 (1997). Although 'original' works of authorship are
not necessarily 'novel' in a patent sense, I am using the terms 'innovators' and 'innova-
tion' to encompass the works of both authors and inventors.
2. See, e.g., Wendy J. Gordon, Asymmetric Market Failure and Prisoner's Dilemma in
Intellectual Property, 17 U. Dayton L. Rev. 853 (1992).
3. See, e.g., Pamela Samuelson, Randall Davis, Mitchell D. Kapor, and J.H. Reichman,A Manifesto Concerning the Legal Protection of Computer Programs, 94 Colum. L. Rev.
2308 (1 994) (making this sort of economic argument for legal protection against cloningof computer program innovations).
4. See, e.g., Wendy J. Gordon, Fair Use as Market Failure: A Structural and Economic
Analysis of the Betamax Case and its Predecessors, 82 Colum. L. Rev. 1600, 1610-11
(1982).
5. Id. at 161 1-12.
6. See, e.g., J.H. Reichman, Electronic Information Tools - The Outer Edge of WorldIntellectual Property Law, 24 Int'l Rev. Indus. Prop. & Copyright L. 446 (1993).
7. See, e.g., William M. Landes and Richard A. Posner, An Economic Analysis ofCopyright Law, 18 J. Legal Stud. 325, 333-44 (1989) (presenting a formal modelof copyright). But see Stephen Breyer, The Uneasy Case for Copyright: A Study ofCopyright in Books, Photocopies, and Computer Programs, 84 Harv. L. Rev. 281 (1970)
(questioning the economic need for copyright). See also Barry W. Tyerman, TheEconomic Rationalefor Copyright Protection for Published Books: A Reply to Professor
Breyer, 18 UCLA L. Rev. 1100 (1971); Stephen Breyer, Copyright: A Rejoinder, 20UCLA L. Rev. 75 (1972).
8. Landes and Posner, supra note 7, at 325.
9. See, e.g., Richard A. Posner, Economic Analysis of Law (5th Ed. 1998) (surveying eco-
nomic analysis of many bodies of law).
10. Among the classic works are: Breyer, supra note 7; Gordon, supra note 4; Gillian K.
Hadfield, The Economics of Copyright: A Historical Perspective, 38 Copyright L.
Symposium (1992); Robert M. Hurt and Robert M. Schumann, The Economic Rationale
of Copyright, 56 Am. Econ. Rev. Papers & Proceedings 421 (1966); Landes and Posner,
supra note 7; Arnold Plant, The Economic Aspects of Copyright in Books, 1 Econometrica167 (1934). These and other important articles on copyright can be found in Wendy J.
Gordon and Richard Watt, Economics of Copyright: Developments in Research and
Analysis (2003); Ruth Towse and Rudi Holzhauer, The Economics ofIntellectual Property(2002 Vol. I, Parts II-V). Other important, but more recent, work is cited throughout this
article.
11. 17 U.S.C. sec. 102(b). See, e.g., Landes and Posner, supra note 7, at 347-53.
12. Id. at 345-^7.
13. 17 U.S.C. sec. 106(2). See, e.g., Landes and Posner, supra note 7, at 354-55.
14. 17U.S.C.sec.202.
15. 17 U.S.C. sec. 201(d).16. 17 U.S.C. sec. 302-304. See infra notes 51-55 and accompanying text for economic argu-
ments in favor of limits on the duration of copyrights.17. 17 U.S.C. sec. 201 (b). See, e.g., I.T. Hardy, An Economic Understanding of Copyright
Law's Work-Made-For-Hire Doctrine, 12 Colum.-VLA J.L. & Arts 181 (1988).18. See, e.g., Stanley M. Besen, Willard G. Manning, and Bridger M. Mitchell, Copyright
Liability for Cable Television: Compulsory Licensing and the Coase Theorem, 21 J. Law& Econ. 67 (1978); Stanley M. Besen, Sheila Kirby and Steven Salop, An Economic
Analysis of Copyright Collectives, 78 Va. L. Rev. 383 (1992); Robert P. Merges,
Contracting Into Liability Rules: Intellectual Property Rights and Collective Rights
Organizations, 84 Calif. L. Rev. 1293 (1996).
Should economics play a role in copyright law andpolicy? 1 7
19. See, e.g., Joni M. Cherbo and Margaret Jane Wyszomirski, eds. The Public Life
of the Arts in America (2000); Ruth Towse, ed., A Handbook on Cultural Economics
(2003).
20. See Richard Watt, Copyright and Economic Theory: Friends or Foesi (2000).
21 . See, e.g., Timothy J. Brennan, Harper & Row vs. The Nation, Inc.: Copyrightability andFair Use, 33 J. Cop. Soc'y 368 (1986); William T. Fisher III, Reconstructing the Fair Use
Doctrine, 101 Harv. L. Rev. 1659 (1988); Gordon, supra note 4; Stanley J. Liebowitz,
Copying and Indirect Appropriability: Photocopying of Journals, 93 J. Pol. Econ. 945
(1985); Landes and Posner, supra note 7, at 357-61; Robert P. Merges, Are You MakingFun of Me? Notes on Market Failure and the Parody Defense of Copyright, 21 AIPLAL.Q. 305 (1993); Alfred C. Yen, When Authors Won't Sell: Parody, Fair Use, and
Efficiency in Copyright Law, 62 U. Colo. L. Rev. 79 (1991). See also Lisa N. Takeyama,
Welfare Implications of Unauthorized Reproduction of Intellectual Property in the
Presence of Demand Network Externalities, 42 J. Ind. Econ. 155 (1994).
22. See, e.g., Gordon, supra note 4, at 1652-57.
23. Id. at 1602-22. Gordon argues that fair use should be awarded in a wide range of other
cases as well. See id. at 1627-36 (discussing, e.g., externalities, antidissemination motives,
and nonmonetizable interests); Wendy J. Gordon, Excuse and Justification in the Law ofFair Use: Transaction Costs Have Always Been Only Part of the Story, 13 J. Copyright
Society 149(2003).24. The effect of a challenged use on the market, or potential market, for a copyrighted work
is one of the principal criteria for judging whether a use is fair. See 17 U.S.C. sec. 107.
25. See, e.g., William M. Landes, Copyright Protection of Letters, Diaries, and Other
Unpublished Works: An Economic Approach, 21 J. Legal Stud. 79 (1992).
26. See 17 U.S.C. sec. 1 10(6), 1 10(10).
27. See Sonny Bono Copyright Term Extension Act, Pub. L. No. 105-298, 112 Stat. 2827
(1998), codified at 17 U.S.C. sees. 302(a), 302(c), discussed infra notes 47-72 and accom-
panying text.
28. 17 U.S.C. sec. 104A. See, e.g., Edward Lee, The Public's Domain: Evolution of LegalRestraints on the Government's Power to Control Public Access Through Secrecy or
Intellectual Property, 55 Hastings L.J. 91 (2003) (discussing economic and constitutional
problems with 'restoration' of these copyrights).
29. See, e.g., Richard A. Epstein, The Dubious Constitutionality of the Copyright Term
Extension Act, 36 Loy. L.A.L. Rev. 123 (2002); Dennis S. Karjala, Judicial Review of
Copyright Term Extension Legislation, 36 Loy. L.A.L. Rev. 199 (2002).
30. Eldred vs. Ashcroft, 123 S.Ct. 769 (2003).
3 1 . See, e.g., Peter A. Jaszi, Toward a Theory of Copyright: The Metamorphoses ofAuthorship,1991 Duke L.J. 455 (discussing the romantic conception of copyright). See also Breyer,
supra note 7, at 284-91 (discussing non-economic justifications for copyright).
32. A decade ago, I recommended that government offices charged with responsibilities for
intellectual property policy should hire staffeconomists. See Pamela Samuelson, Will the
Copyright Office Be Obsolete in the Twenty-First Century?, 13 Cardozo Arts & Ent. L.J.
55 (1994).
33. See, e.g., Jessica Litman, Digital Copyright (2001); William F. Patry, Copyright and the
Legislative Process: A Personal Perspective, 14 Cardozo Arts & Ent. L.J. 139, 141 (1996).
34. The President of the Association of American Publishers, for example, is Pat Schroeder,
formerly a Congresswoman. The recently retired head of the Business Software Alliance,
Emery Simon, was with the US Trade Representative Office where he worked on intel-
lectual property issues before joining BSA. Bruce Lehman who served as head of the
Patent & Trademark Office during the Clinton Administration had been a legislative
staffer in Congress before becoming a lawyer who represented copyright industry groups.Shira Perlmutter, formerly with the US Copyright Office, is now a Vice President and
Associate General Counsel for Intellectual Property at AOL Time Warner.
35. France and Germany are among the countries that protect authors' rights on this basis.
See, e.g., Goldstein, supra note 1, at 168-69; Alain Strowel, Droit d'auteur and
1 8 Developments in the economics of copyright
Copyright: Between History and Nature, in Of Authors and Origins: Essays on Copyright
Law, 235 (Brad Sherman and Alain Strowel, eds, 1994).
36. See, e.g., Richard Gilbert and Carl Shapiro, Optimal Patent Length and Breadth, 21
RAND J. Econ. 106 (1990) (arguing for a thin scope and long-term of protection for
patents); Peter S. Menell, Tailoring Legal Protectionfor Computer Software, 39 Stan. L.
Rev. 1045 (1989) (arguing for short duration for legal protection for software);
Samuelson et al., supra note 3, at 2423 (arguing for a short duration of legal protection
against cloning of computer program behavior and other industrial design features of
programs).37. See, e.g., Michael L. Katz and Carl J. Shapiro, Network Externalities, Competition and
Compatibility, 75 Am. Econ. Rev. 424 (1985) (discussing economics of network externali-
ties). An important contribution to the legal literature on network effects is Mark A. Lemleyand David F. McGowan, Legal Implications of Network Economic Effects, 86 Cal. L. Rev.
479 (1998) (drawing out several possible legal implications of network externalities).
38. See, e.g., Joseph Farrell, Standardization and Intellectual Property, 30 Jurim. J. 35 (1 989);
Wendy J. Gordon, On Owning Information, 78 Va. L. Rev. 149 (1992); Lemley &McGowan, supra note 37; Menell, supra note 36; J.H. Reichman, Legal Hybrids Between
the Patent and Copyright Paradigms, 94 Colum. L. Rev. 2432 (1994); Samuelson et al.,
supra note 3. See also sources infra note 127.
39. See, e.g., David McGowan, Copyright Nonconsequentialism 69 Missouri L. Rev. 1 (2004)
(emphasizing indeterminate relationship between copyright rules and desired outcomes).Cf. George Priest, What Economists Can Tell Lawyers About Intellectual Property, in Res.
Law & Econ. 19 (J. Palmer and R. Zerbe, eds. 1986) (questioning the usefulness of eco-
nomic analysis in respect of intellectual property law because of inadequacies in empiri-cal understanding of innovation).
40. 239 F.3d 1004, 1016-17 (9th Cir. 2002).
41 . Plaintiffs' economic expert David Teece focused on the harmful effects on the plaintiffs'
ability to offer competing digital music services. Id. at 1017. Defendants' expert arguedthat use of Napster for sampling purposes enhanced the willingness of users to buyrecorded music. Id. Plaintiffs commissioned two empirical studies of the effects of peer-
to-peer filesharing on sales of recorded music which concluded that among college stu-
dents, purchases had declined because of Napster filesharing. Id. at 1016.
42. On public choice generally, see, e.g., William N. Eskridge, Jr., Implications of Public
Choice Theoryfor Statutory Interpretation, 74 Va. L. Rev. 275 (1988); Daniel A. Farber
and Philip P. Frickey, Law and Public Choice: A Critical Introduction (1991).
43. See, e.g., Glynn S. Lunney, Jr., The Death of Copyright: Digital Technology, Private
Copying and the Digital Millennium Copyright Act, 87 Va. L. Rev. 813 (2001); Neil W.
Netanel, Locating Copyright Within the First Amendment Skein, 54 Stan. L. Rev. 1
(2001); Patry, supra note 33; Stewart E. Sterk, Rhetoric and Reality in Copyright Law, 94
Mich. L. Rev. 1197(1996).44. Pub. L. No. 105-298, sec. 102(b) and (d), 112 Stat. 2827-2828 (amending 17 U.S.C. sees.
302, 304). A Canadian legislative proposal to extend the term of copyrights in unpub-lished works has been controversial. See James Adams, Estates' Rights in Canadian
Copyright Re-examined, Globe and Mail, Sept. 23, 2003, available at http://www.the-
globeandmail.com/servlet/ArticleNews/TPStory/LAC/20030923/COPY23/TPEntertain
ment/TopStories.45. Pub. L. No. 105-304, 112 Stat. 2860 (1998), anti-circumvention provisions codified at 17
U.S.C. sec. 1201.
46. See, e.g., Alex Breeding, What Are the Implications of the New Super-DMCA Laws?,
TechRepublic, 16 July, 2003, available at http://techrepublic.com/5100-6314-5054407.html. A status report on enactment of state super-DMCA laws can be found at
http://www.freedom-to-tinker.com/superdmca.html.47. Eldred vs. Ashcroft, 123 S.Ct. 769 (2003). Justices Stevens and Breyer dissented in sepa-
rate opinions, the former offering an alternative historical and constitutional analysis,
and the latter mainly focusing on the economic effects of the CTEA. See id. 790-801
(Stevens dissent) and 801-15 (Breyer dissent).
Should economics play a role in copyright law and policy? 1 9
48. See Brief for Petitioners in Eldred vs. Ashcroft, at 15-28, available at http://eldred.cc/
legal/supremecourt.html. Eldred also argued that the CTEA violated the First
Amendment. See id. 34-48.
49. Id. at 23.
50. Id.
51. See Brief of George A. Akerloff et al. in Eldred vs. Ashcroft, available at http://eldred.cc/
legal/supremecourt.html. Akerloff, Kenneth Arrow, James Buchanan, Ronald Coase,
and Milton Friedman were the Nobel Prize winning signatories of this brief. See id.,
Appendix A.
52. Id. at 8.
53. Id. at 11-13.
54. Id. at 12- 14.
55. Id. at 8.
56. Eldred, 123 S.Ct. at 781-82.
57. Id. at 783-84.
58. Id. at 801 (Breyer, J., dissenting) ('The economic effect of this 20-year extension - the
longest blanket extension since the Nation's founding- is to make the copyright term
not limited, but virtually perpetual. . . . And more importantly, its practical effect is not
to promote, but to inhibit the progress of "Science" - by which word the Framers meant
learning or knowledge. . . .').
59. See, e.g., Eastman Kodak Co. vs. Image Technical Services, Inc., 504 U.S. 451 (1992);
Jefferson Parish Hosp. Dist. No. 2 vs. Hyde, 466 U.S. 2 (1984).
60. See, e.g., Mazer vs. Stein, 347 U.S. 201, 219 (1954); Sony Corp. of Am. vs. Universal City
Studios, Inc., 464 U.S. 417, 429 (1984).
61 . See, e.g., United States vs. Lopez, 514 U.S. 549 (1995); United States vs. Morrison, 529
U.S. 598 (2000). In both decisions, Justices Breyer, Ginsburg, Souter and Stevens dis-
sented from rulings that Congress lacked power under the Commerce Clause to enact
legislation to regulate guns near schools and gender-based crimes of violence. See, e.g.,
Paul J. Heald and Suzanna Sherry, Implied Limits on the Legislative Power The
Intellectual Property Clause as an Absolute Constraint on Congress, 2000 U. 111. L. Rev.
1119 (2000) (discussing numerous Supreme Court decisions restrictively interpreting
Congress' power and considering their implications for various intellectual property
provisions).
62. See Brief for Petitioners, supra note 48, at 1 1-12 (invoking Lopez and Morrison).
63. Id.
64. Id. at 12-13, 20-21.
65. In view of Justice Breyer's earlier economic skepticism about copyright, see Breyer, supranote 7, Eldred's lawyers expected him to be skeptical of the CTEA on economic groundsand receptive to arguments based on the constitutional history of the Intellectual
Property Clause indicating that the Framers embedded limits in it in order to guard
against rent-seeking. See Brief for Petitioners, supra note 48, at 23-26. Eldred's lawyers
thought Justice Stevens would be receptive to their arguments because his opinion in
Sony Corp. of Am. vs. Universal City Studios, Inc., 464 U.S. 417, 429 (1984) emphasized
copyright as a limited monopoly, the primary purpose of which was to promote public
access to knowledge. (The monopoly privileges that Congress may authorize are neither
unlimited nor primarily designed to provide a special private benefit. Rather, the limited
grant is a means by which an important public purpose may be achieved. It is intended
to motivate the creative activity of authors and inventors by the provision of a special
reward, and to allow the public access to the products of their genius after the limited
period of exclusive control has expired.')
66. Eldred, 1 23 S.Ct. at 775-76.
67. See Addendum to Reply Brief for the Petitioners in Eldred vs. Ashcroft (chart
distinguishing CTEA from other extensions), available at http://eldred.cc/legal/
supremecourt.html.68. Eldred, 123 S.Ct. at 778, quoting New York Trust Co. vs. Eisner, 256 U.S. 345, 349 (1921).
69. Eldred, 123 S.Ct. at 790.
20 Developments in the economics of copyright
70. Brief for Petitioners, supra note 48, at 39-47; Eldred, 123 S.Ct. at 788-90.
71. See, e.g., Pamela Samuelson, The Constitutional Law of Intellectual Property AfterEldred vs. Ashcroft, 50 J. Cop. Soc'y 547, 556 (2003).
72. Eldred, 123 S.Ct. at 775.
73. See Final Report of the National Commission on New Technological Uses of CopyrightedWorks 14-26 (1979). TheCONTU analysis focused on the economics of developing indi-
vidual programs, but did not consider the appropriate scope of copyright protection for
computer programs or vexing issues such as the interoperability and decompilationissues discussed infra notes 98-120 and accompanying text.
74. See Council Directive 91/250 on the Legal Protection of Computer Programs, 1991 O.J.
(LI 22) 42 (cited hereinafter as 'Software Directive'). Article 6 of this Directive permits
decompilation of computer program code for purposes of achieving interoperability
among programs. See, e.g., Pamela Samuelson, Comparing U.S. and E.C. CopyrightProtection For Computer Programs: Are They More Different Than They Seem?, 13
J. Law & Comm. 279, 287-88 (1994) (discussing the concerns of the EuropeanCommission's competition directorate about the software directive that influenced the
rule in favor of decompilation for interoperability purposes).
75. See Agreement on Trade-Related Aspects of Intellectual Property Rights, in The LegalTexts: The Results of the Uruguay Round of Multilateral Trade Negotiations 321, art.
10(1) at 325 (Cambridge 1994) (cited hereinafter as TRIPS').76. See Semiconductor Chip Protection Act, Pub. L. No. 98-620, 98 Stat. 3347 (1 984), now
codified at 17 U.S.C. sec. 901 et seq. See, e.g., Prepared Testimony of F. Thomas Dunlap,
Jr., Corporate Counsel and Secretary of Intel Corp., Hearings Before the Subcommittee
on Courts, Civil Liberties, and the Administration of Justice of the House Committeeon the Judiciary, on H.R. 1028, 98th Cong., 1st Sess. (8/3/83) explaining the industry's
need for this legislation; Prepared Statement of Sen. Charles McC. Mathias, Jr., House
Hearings, supra, at 3: '[Chip] innovators are being ripped off by onshore and offshore
"chip pirates" who, for a fraction of the developer's cost, can now legally appropriateand use these chip designs as their own.' Of particular concern was the loss to Japanese
industry of a substantial share of the market for random access memory chips to
Japanese competitors whose superior quality control made their chips very competitive.See Stephen P. Kasch, The Semiconductor Chip Protection Act: Past, Present, and Future,
7 High Tech. L.J. 71,79(1993).77. See Vessel Hull Design Protection Act, Pub. L. No. 105-304, tit. V, 1 12 Stat. 2905 (1998),
codified at 17 U.S.C. sec. 1301-1332. Had this law been subjected to impartial economic
analysis, it is difficult to believe Congress would have passed it.
78. See, e.g., J.H. Reichman and Paul F. Uhlir, Database Protection at the Crossroads: Recent
Developments and Their Impact on Science and Technology, 14 Berkeley Tech. L.J. 793
(1999) (discussing the evolution of legislative proposals).79. See European Parliament and Council Directive 96/9/EC of 1 1 March 1 996 on the Legal
Protection of Databases, 1996 O.J. (L. 77) 20, available at http://europa.eu.int/comm/
internal_market/en/intprop/docs/. For a history of the justification for and evolution of
this Directive, see, e.g., J.H. Reichman and Pamela Samuelson, Intellectual Property
Rights in Data?, 50 Vand. L. Rev. 51, 80-97 (1997).
80. EU Database Directive, supra note 79, recitals 1-12.
81. Id., arts. 7, 10.
82. See, e.g., Reichman and Samuelson, supra note 79, at 113-36 (criticizing the EU sui
generis regime because of its potential harmful effects on competition and innovation as
well as for science).
83. See, e.g., Brief of AkerlofT, supra note 51; Karjala, supra note 29.
84. See, e.g., Pamela Samuelson and Suzanne Scotchmer, The Law and Economics ofReverse Engineering, lllYale L.J. 1575, 1630-49 (2002) (critical of anti-circumvention
regulations).
85. See, e.g., 17 U.S.C. sees. 1003-1007.
86. See, e.g., Neil W. Netanel, Impose a Noncommercial Use Levy to Allow Free Peer-to-Peer
File Sharing, 17 Harv. J.L. & Tech.l (2003).
Should economics play a role in copyright law andpolicy? 2 1
87. 17 U.S.C. sec. 1201(a)(l)(C). See Memorandum to James Billington from Mary Beth
Peters regarding Recommendations of RM 2002-4, Rulemaking on Exemptions from
Prohibition on Circumvention of Copyright Protection Systems For Access Control
Technologies, dated 27 Oct., 2003, available at http://www.copyright.gov/1201/docs/
registers-recommendation.pdf.88. The subsequent paragraphs on Whelan, Altai and Sega vs. Accolade derive from Pamela
Samuelson, Economic and Constitutional Influences on Copyright Law in the United
States, 23 Eur. Intell. Prop. Rev. 409 (Sept. 2001).
89. 797 F.2d 1 222 (3rd Cir. 1 986).
90. Id. at 1239-40.
91. Id. at 1231.
92. See Brief Amicus Curiae of ADAPSO, The Computer Software and Services Industry
Association, Inc., Whelan Associates, Inc. vs. Jaslow Dental Laboratory, Inc. (Civ. A.
No. 85-1358) (1986).
93. Whelan, 797 F.2d at 1237.
94. See, e.g., Johnson Controls, Inc. vs. Phoenix Control Systems, Inc., 886 F.2d 1 173 (9th
Cir. 1989); Lotus Development Corp. vs. Paperback Software Int'l, 740 F. Supp. 37
(D. Mass 1990).
95. 982 F.2d 693 (2nd Cir. 1992).
96. Id. at 702-3.
97. Whelan, 797 F.2d at 1 240.
98. Altai, 982 F.2d at 708-11.
99. Id. at 707.
100. 17 U.S.C. sec. 102(b).
101. Altai, 982 F.2d at 710.
102. Id. at 71 1-12.
103. Id. at 712 (citation omitted).
104. Id.
105. Id.
106. See id.
107. Id. at 707-11.
108. See, e.g., Arthur R. Miller, Copyright Protectionfor Computer Programs, Databases, and
Computer-Generated Works: Is Anything New Since CONTU?, 106 Harv. L. Rev. 977
(1993) (arguing that interfaces should be protectable by copyright).
109. /4/to/,982F.2dat710.110. See, e.g., Alan K. Palmer and Thomas C. Vinje, The EC Directive On The Legal
Protection Of Computer Software: New Law Governing Software Development, 2 Duke
J. Comp. & Int'l L. 65 (1992).
111. Id. at 71-78.
112. See, e.g., Jonathan Band and Masanobu Katoh, Interfaces on Trial: Intellectual
Property and Interoperability in the Global Software Industry 271-97 (1995).
113. 977 F.2d 1510 (9th Cir. 1992). The Ninth Circuit had endorsed Whelan in Johnson
Controls, Inc. vs. Phoenix Control Systems, Inc., 886 F.2d 1173 (9th Cir. 1989), but
repudiated it in Sega, 977 F.2d at 1524-25.
1 14. See, e.g., Andrew Johnson-Laird, Software Reverse Engineering In The Real World, 19
U. Dayton L. Rev. 843 (1994).
115. 17 U.S.C. sec. 106(1).
116. Although some software developers openly publish APIs (application programming
interfaces) for their programs, many do not. The willingness of program developers to
license APIs to other firms varies considerably. Some firms only license APIs on terms
that some developers regard as unacceptable (e.g., giving up the right to make a version
of the same program for other platforms). See Sega, 977 F.2d at 1514.
117. 17 U.S.C. sec. 107.
118. The fair use defense is discussed in Sega, 975 F.2d at 1 520-28.
1 19. See, e.g., Sega, 977 F.2d 1510, 1527-8.
120. For scholarly commentary making economic arguments in favor of fair use for
22 Developments in the economics of copyright
decompilation for purposes of interoperability, see, e.g., Lawrence Graham andRichard O. Zerbe, Jr., Economically Efficient Treatment of Computer Software: Reverse
Engineering, Protection and Disclosure, 22 Rutg. Comp. & Tech. L.J. 61, 132-34 (1996);
Lemley and McGowan, supra note 37, at 525; Samuelson and Scotchmer, supra note
84, at 1621-26.
121 . See, e.g., Band and Katoh, supra note 1 12, at 271-82, 294-97; Jonathan Band, SoftwareReverse Engineering Amendments in Singapore and Australia, J. Internet L. 17, 20
(Jan. 2000).
1 22. See, e.g. ,PriceWaterhouseCoopers, Contributions of the Packaged Software Industry to
the Global Economy (1999), http://www.bsa.org/usa/globalip/econ/pwcl999.pdf.123. See, e.g., A&M Records, Inc. vs. Napster, Inc., 239 F.3d 1004 (9th Cir. 2001) (prelim-
inary injunction against peer-to-peer software developer as a contributory and vicari-
ous infringer of digital music copyrights because of centralized directory and indexing
functionality); In re Aimster Litigation, 334 F.3d 643 (7th Cir. 2003) (summary judg-ment against peer-to-peer developer as contributory infringer because of direct assist-
ance to infringements). But see Metro-Goldwyn-Mayer Studios, Inc. vs. Grokster, Ltd.,
259 F. Supp.2d 1029 (C.D. Cal. 2003)(summary judgment in favor of peer-to-peer soft-
ware developer because of capability for substantial noninfringing uses).
1 24. See supra note 4 1 .
125. See, e.g., Brief Amicus Curiae of Forty Intellectual Property Professors in MGM vs.
Grokster, available at http://www.sims.berkeley.edu/~pam/papers (arguing that the
legislature, not the courts, should regulate technologies with non-infringing uses).
1 26. See, e.g., Theberge vs. Galierie d'Art du Petit Champlain, Inc., 2002 S.C.C. 24.
127. See, e.g., Yochai Benkler, Intellectual Property and the Organization of Information
Production, 22 Int'l Rev. of L. & EC. 81 (2002); supra note 1 Lemley; Glynn S. Lunney,
Jr., Reexamining Copyright's Incentives-Access Paradigm, 49 Vand. L. Rev. 483 (1996);
Michael Meurer, Copyright Law and Price Discrimination, 23 Cardozo L. Rev. 55
(2001); Christopher S. Yoo, Towards a Differentiated Products Theory of Copyright,
Chapter 6 of this volume (for an expanded version see Christopher S. Yoo, Copyrightand Product Differentiation, 79 N.Y.U.L. Rev. 212 (2004)).
128. See, e.g., Gordon and Watt, supra note 10. For information about SERCI and its con-
ferences and journal, see http://www.serci.org/.
129. See, e.g., International Intellectual Property Alliance, Copyright Industries in the U.S.
Economy: The 1998 Report (Exports of copyright-related products and services are
increasing at an impressive rate - in 1996 their value surpassed that of every other
export sector in the United States.)
130. See, e.g., Pamela Samuelson, Challenges For the World Intellectual Property
Organization and the Trade-Related Aspects of Intellectual Property Rights Council In
Regulating Intellectual Property Rights In The Information Age, 21 Eur. Intell. Prop.Rev. 578 (Nov. 1999).
2. Risk sharing and the distribution of
copyright collective income
Arthur Snow and Richard Watt
2.1 INTRODUCTION
Hollander (1984, p. 200) sums up the basic functions of copyright collect-
ives as follows:
Copyright collectives are associations to whom authors transfer copyrights for
purposes of exploitation. In general, collectives are concerned with the follow-
ing functions: (1) they grant licenses for the use of works in their repertory,
(2) they negotiate and collect royalties, and distribute them to their members,
(3) they take legal action against those who infringe the copyrights to which theyhold title.
Each of these three aspects of collective administration of copyrights (as
well as the monopoly position that such collectives enjoy) has been subject
to the analysis of economic theory. In particular, the literature asserts that
the underlying reason for copyright collectives to form naturally is that by
joining together and marketing a blanket license to the entire repertory,
each individual member can achieve a better utility position than by acting
alone. In this chapter, we shall consider the possible ways in which collect-
ive administration improves the welfare of its members.
In the above quote from Hollander, it is clear that the principal idea is to
share certain transactions costs over collective members (costs of negotiat-
ing, receiving payment for, and protecting license agreements). Clearly,
then, the most obvious reason for collective administration is the welfare
gains that the members can enjoy as transactions costs are diminished.
The Coase theorem implies that final resource allocations may, in general,
be socially inefficient whenever transactions costs exist. When it comes to
intellectual property markets, in which access to intellectual products
between individuals is traded, transactions costs not only exist, but they mayeven be more significant than the final value of the object of the transaction.
The reason for this is the fact that intellectual products are typically of a
public good nature, and so there is a very low cost of free-riding associated
23
24 Developments in the economics of copyright
with them. Once access has been granted to one user, it becomes very costly
to then exclude access being taken by the entire set of potential users.
This problem is often overcome, at least partially, by collective adminis-
tration. The example of copyright is perhaps the most prevalent case. Whena copyrightable product is created, along with it is created an entire set of
access rights, some of which can be transacted with low transactions costs,
but others of which can only be transacted at a very high transactions cost.
For example, consider the case of a musical composition. Once the compos-ition has been fixed onto a physical support, say a CD Rom, it can easily be
listened to by whoever has that CD Rom. When a consumer purchases the
CD Rom from a recognized retailer, the right to a limited set of access priv-
ileges is automatically purchased as well. These include the right to listen to
the music at any time and the right to sell the CD Rom to a third party. In
most countries, the consumer has the right to make a copy of the CD Romfor personal use as well. However, with the purchase price of the CD Rom,the consumer will not acquire the right to make copies of the music for
resale, or the right to play the music to an audience, especially for economic
gain. The problem for the owner of the copyright in the music is how to
monitor the use that the consumer makes of the CD Rom at a reasonable
cost. One partial method lies in collective administration of the copyright.
Whenever the set of users of a particular type of intellectual product is
common to many (say, n) copyright holders, in principle they can, by
joining together and selling a license to the entire repertory, save on the
transactions costs implied by n - 1 individual rounds of negotiation, and
likewise they can also save on both the monitoring costs and the costs of
litigation whenever such recourse is needed. 1 In this way, the option of col-
lective administration can make feasible a set of transactions that is
unprofitable when carried out individually. In economic terms, copyright
collectives enjoy the benefits of a natural monopoly.
However, saving on transactions costs is not the only manner in which
collective administration can be of value to individual copyright holders. If
we assume, as is indeed very natural, that the final market value of each
copyright is a random variable (that is, there is uncertainty as to its final
value), then a copyright collective is able to alleviate the risk-bearing situ-
ation of its members. If the members are risk-averse (again, a very natural
assumption), then an improved risk situation provides for expected utility
gains which are just as important as the gains that the transactions costs
savings provide for. It is curious to note that this aspect of collective admin-
istration seems to have been entirely ignored, both in theoretical contribu-
tions and in real-life cases.
Given the diversity in the market value of each composition in a copy-
right collective's repertory, it is quite common for copyright collectives to
Risk sharing and the distribution of copyright collective income 25
take great pains in analysing the markets in which their repertories are used
in order to determine how much use is made of each particular compos-ition in the repertory. This is puzzling since it goes quite against the idea
that the songs are bundled in the first place, in a blanket license, in order to
save on such transactions costs. 2 The reason why this is done is simply to
be able to base the final distribution of net income on the use made of each
individual copyright. In this way, the copyright holders of the songs that
are most played can be given a proportionately greater share of the collec-
tive's surplus. Abstracting from the fact that monitoring the market use of
individual songs increases the transactions costs of running the collective,
there are important reasons strongly rooted in the economic theory of risk
bearing that suggest that doing this is inconsistent with the best interests of
the pool of members, thereby making current copyright collective activity
inefficient.
In this chapter, we think of a copyright collective as a simple coalition, in
the sense of traditional game theory. Each potential individual member has
the opportunity to join the coalition, or to act alone. In such a setting, it is
very clear that the decision on whether or not to join the coalition is affected
both by how much of the coalition's income is distributed among the
members, and by exactly how that income is distributed. Therefore, the dis-
tribution rule will affect the number of members that the coalition has
(in equilibrium) as well as the welfare of each individual. We shall imposetwo basic requirements of the distribution mechanism of a coalition's net
income. First, we will require that the distribution method be such that indi-
vidual copyright holders prefer to become members of the coalition rather
than acting alone (we refer to this as the 'participation condition'). Second,
we will require that existing members of the coalition prefer new members to
join (we refer to this as the 'incentive compatibility condition'). Given these
two requisites, we first ask if a distribution method that consists of learning
the true market value of each composition and then paying each individual
member accordingly satisfies these requirements. We then propose alterna-
tive distribution mechanisms that satisfy the two requirements and are
preferred by all individuals to the 'learn-then-distribute' option.
2.2 A SIMPLE MODEL OF A COPYRIGHTCOLLECTIVE
Let us assume that there is a set of n individuals, each of whom can create a
composition that has strictly non-negative market value, but where the
market value of the composition is a random variable. Here, we make the
simplest possible assumption, namely that all composition lotteries are
26 Developments in the economics of copyright
identical. Indeed, the assumption that all compositions are identical, albeit
deterministic, is common in the literature (see, for example, Besen et al.
1992).
Moreover, we will assume that the composition lotteries are not only
identically distributed, but that they are also statistically independent with
a probability of success that is common knowledge.3Concretely, we assume
that each composition can be either a 'success', in which case its market
value is x = 1,or a 'flop', in which case its market value is x = 0. We assume
that the probability of success is/?, which is assumed strictly between and
1,and we denote the true outcome of composition i by xr
Now, assume that the (monetary equivalent) costs of creation of each
composition are constant and equal to c, and that if the composition is
indeed marketed by the copyright holder (acting alone), then there exist
transactions costs (costs of negotiation, monitoring, and litigation),
denoted by t.
All copyright holders are assumed to be strictly wealth-loving and
strictly risk-averse. In order to avoid the need for formally considering
utility, instead we take the approach of mean-variance analysis. Concretely,we shall assume that all copyright holders have a strict preference for
greater expected wealth, and a strict preference for lower variance of
wealth. If copyright holder i goes ahead with the creation and marketingof the composition on an individual basis, then the expected value of his
final wealth (assuming that he has no wealth outside of what the compos-ition generates) is given by
(2.1)
On the other hand, the variance of his wealth is given by
E(E(x -c-t}-(x-c- t)Y = E(p- x)2=p(l -p), (2.2)
which depends only on the probability of success.
Now, consider the situation of the coalition. Assuming that the coali-
tion has m < n members, the final market value of the blanket license
follows a binomial distribution. It can take on any of the discrete values
0, 1, 2, 3, . . ., m, depending on exactly how many of the m member com-
positions are successes. Using the well-known formula for the binomial
distribution, the probability that exactly h of the m compositions will be
successful is given by
m\
Risk sharing and the distribution of copyright collective income 27
Given this, the expected final market value of the coalition's repertory is
True to the public good assumption of the compositions in question, and
to the fact that each composition has the same user set, we assume that the
total cost of marketing the blanket license to all the compositions in its
repertory is the same as for any particular individual; that is, the entire
repertory can be marketed at a (transactions) cost of t.4 Given this, the
coalition (with m members) earns a net income of X(m) t for distribution
to the members.
Now assume that the collective can learn the true outcome of each indi-
vidual composition lottery at some cost, which we fix at A:, and then give
individual member / a payment of
t + k
Under this 'learn-then-distribute' system, a member's expected wealth is
'
x - c-i^LJi- c-iAU(i-Jo-c- ) (2.3)
=p-c m
with a variance of
x-c- \-L-c- \\=E(p-xY=p(\-p\ (2.4)mI \
m//
Clearly, under this distribution rule, all members get a final wealth lottery
with the same variance as acting alone. It is also clear from (2.3) that the
expected wealth of each collective member is strictly increasing in m, so that
all collective members will welcome new members to the coalition, imply-
ing that the incentive compatibility condition is satisfied. However, it is not
clear that the participation condition is satisfied. Comparing the expected
wealth in (2.3) with that in (2.1), participation (that is, all individuals prefer
to become members of the coalition) can be guaranteed if and only if
t + k
m
28 Developments in the economics of copyright
Clearly, this requirement is more easily satisfied when k is small, when t is
large, or when m is large, each of which is intuitively understandable.
In short, for any given values of / and A:, we have:
Proposition 2.1 A distribution rule under which a collective learns (at a
cost of k) the final result of each independent composition in its reper-
tory, and distributes final income accordingly, is incentive compatible.
There is a minimum membership, given by m* = (t + k)/t, beyond which
the distribution rule also satisfies the participation condition. 5
Assuming that the coalition has sufficient members, the fact that the
learn-then-distribute system satisfies both incentive compatibility and par-
ticipation is due entirely to the transactions costs savings that the coalition
offers. However, the risk-bearing situation of each coalition member has
not changed, since the coalition offers the same variance of the final wealth
lottery as acting alone. Hence, while the coalition increases the expected
value of income (by saving on transactions costs) it does nothing at all to
alleviate the risk situation of the members. Given that the creators are risk-
averse, it would be welfare enhancing if the coalition could offer a better
risk situation as well as transactions costs savings. We now go on to con-
sider if this is indeed possible.
2.3 RISK SHARING DISTRIBUTION MECHANISMS
Consider first the opportunity that the copyright collective has of offering
full and fair insurance to all members. Since no learning is done as far as
the final result of each individual lottery, the only costs of administering
the set of copyrights is /. Now, assume that the collective pays each
member the expected value of his composition less his share of the trans-
actions costs with certainty. In this case, each creator receives an expectedwealth of
p~ c ~~ <2 - 5>
with zero variance. Clearly, this distribution rule satisfies both participation
and incentive compatibility. Moreover, it also Pareto dominates the learn-
then-distribute rule, since all members get a lower variance of wealth and
a higher expected final wealth. Thus, we have:
Proposition 2.2 A sharing rule stipulating that each individual copy-
right holder receives the expected value of his composition with
Risk sharing and the distribution of copyright collective income 29
certainty satisfies both participation by all individuals and incentive
compatibility. It also strongly Pareto dominates the learn-then-distribute
rule.
The fact that the welfare benefits of the full insurance mechanism are so
obvious begs the question of why this system is not typically used. Perhapsthe answer lies in the fact that, if the collective pays each member his
expected value, then there is a strictly positive probability that the final
market value of the repertory will either exceed or fall short of the total
outlay paid to the members as royalties. Unless the collective has access to
a perfect reinsurance market, the members of the collective would still bear
some risk with respect to their final wealth positions, although this risk
becomes smaller as the membership increases. Clearly though, the full and
fair insurance option may not be feasible. However, there are intermediate
options that have full insurance, although not fair insurance, that satisfy
both participation and incentive compatibility, and that still Pareto dom-inate the learn-then-distribute rule.
Consider the following simple option. The collective distributes a fixed
sum to each individual member such that the expected value of each
member's income is the same as if the distribution rule had been learn-then-
distribute. That is, each member is given a fixed payment of
t + k t kp =p .
m mmOnce again, this rule clearly satisfies incentive compatibility, and satisfies
participation under the same proviso as in Proposition 2. 1 .
6Moreover, this
system is strictly preferred by all copyright holders to the learn-then-
distribute rule since it offers the same expected wealth with a lower variance.
Since no learning is done, this rule implies that total royalty payments are
mp t k, while the total net income of the collective is X(m) t. Under
this system, the probability that the collective runs into* a bankruptcy
problem can be significantly reduced (although, naturally, the reduction
depends on the size of k). For example, if we assume that m =100, and if we
assume that k = 10, then the probability of bankruptcy is under 3 per cent. 7
Naturally, the accumulated but not distributed net income of the collective
would, at certain times, be distributed to the members as profit (as opposedto strict royalty payments). Whether or not such a system is indeed feasible
depends on the true values of the parameters m, p and k, and on the access
that the collective has to other financial products (short-term loans, third
party insurance, etc.), but at least in principle it would seem to be feasible in
most modern societies.
30 Developments in the economics of copyright
Proposition 2.3 A sharing rule stipulating that each member of the col-
lective receives with certainty a payment equal to the expected value of
the payment under the learn-then-distribute rule, satisfies incentive com-
patibility always, and satisfies participation under the same restriction
on total collective membership as in Proposition 2.1. Moreover, this rule
Pareto dominates the learn-then-distribute rule, since it sets the variance
of each copyright owner's final wealth to 0. The probability of bank-
ruptcy under this distribution rule becomes negligible as k increases.
Even if the full but not fair insurance option were not feasible, there are
other options that also Pareto dominate the learn-then-distribute rule and
that never run into bankruptcy problems. To see some of these options,
note that the expected value of the repertory is8
EX(m) = mp (2.6)
and the variance of the market value of the repertory is
E(mp-X(m))
2 = mp(\-p). (2.7)
Now assume that the collective's net income is distributed according to
pre-agreed shares, such that member / receives a proportion \iof the final
net income of the collective, where
in order to avoid all bankruptcy issues (that is, the collective will distribute
exactly all of its net income, whatever that turns out to be). Using this rule,
the expected final wealth of member / is
E\\\X(m)--\-c\ = E\.X(m)-c--mj I
m
= \iEX(m)-c--m
= \mp -c-- (2.8)m
with variance
\mp - c - - -I \
iX(m)- c - -
} }
=E(\.(mp
-X(m)))
2
m \ tn
= \2E(mp-X(m))2.
Risk sharing and the distribution of copyright collective income 3 1
Using (2.7), we find that the variance of final wealth under this distribution
rule is
\}mp(\-p\ (2.9)
Now, since each composition is, ex ante, identical, let us simply assume
that \f= \/m V/. In this case the expected value of final wealth for each
member of the collective is
mwhich is the same as under the full and fair insurance mechanism, and
greater than under the learn-then-distribute rule. On the other hand, the
variance of the final wealth of each of the collective members is
~r-which is greater than any full insurance mechanism (under which variance
is 0), but less than what the learn-then-distribute rule offers (which from
(2.4) is/>(l -/>))
Finally, note that this sharing rule satisfies both incentive compatibility
(expected value is increasing in m and variance is decreasing in m) and par-
ticipation (expected value of being a collective member is strictly greater
than acting alone because of the transactions costs savings, and the vari-
ance that being a member offers is strictly less than that corresponding to
acting alone), and that it Pareto dominates the learn-then-distribute rule
(since it offers both a greater expected value, as the learning costs are saved,
and a lower variance). Thus, we have:
Proposition 2.4 A sharing rule stipulating that each collective memberis paid an equal share of the final net income of the collective is both
incentive compatible and satisfies participation. It also strictly Pareto
dominates the learn-then-distribute rule.
Although we have found a manner in which the net income of a copy-
right collective can be distributed that strictly Pareto dominates the cur-
rently used method, and that will not imply any bankruptcy issues, one
wonders if this method can be improved upon. This type of question has
been analysed extensively in the literature on efficient risk sharing. Twomain results of that literature are worth mentioning here. First, the
so-called 'mutuality principle' of Karl Borch (see Borch 1960) implies that
if a group of risk-averse agents each contribute a random variable to a
common pool (known as a mutual), then a Pareto-efficient risk-sharing
32 Developments in the economics of copyright
rule must make the final wealth of each individual member independentof the particular contribution that he made. In other words, the final
wealth of each member of a copyright collective should depend only on
the final value of X(m), and not on the individual composition lotteries.
This has been achieved in our proposal, where each individual receives an
equal share of X(m), but is clearly not achieved in the learn-then-distribute
rule.
Second, the question naturally arises: can the equal-share rule also be
improved upon. In particular, since it satisfies the mutuality principle, the
equal-share rule fully diversifies all diversifiable risk, but it leaves the final
wealth of each individual dependent upon the aggregate, or social, risk as
determined by the random variable X(m). The question is exactly how this
aggregate risk should be shared. This question was solved by Wilson
(1968), who proves that full risk sharing efficiency will require that each
member of a mutual should receive a fraction of the total pooled wealth
(here X(mj) that is equal to his individual risk tolerance divided by the sumof the risk tolerances of all members.
The sharing rule based on risk tolerances would, almost certainly, implythat members who supply identically valued compositions to the repertoryreceive different shares of the final pooled wealth. Besen et al. (1992,
pp. 395-7) note that discriminatory practices in the distribution of collect-
ive income are often frowned upon by regulatory authorities. 9 Given this,
it remains to be seen if the fully efficient sharing rule offered by Wilson is
indeed feasible in the regulatory scenario of most copyright collectives.
2.4 HETEROGENEOUS AND CORRELATEDCOMPOSITION LOTTERIES
Our simple model of a copyright collective assumes that the compositionlotteries are identical and independent. In this section we briefly examine
the implications of relaxing these assumptions.When different compositions have different probabilities of success, an
additional condition comes into play that is satisfied vacuously when prob-abilities are identical, namely, the distribution rule must satisfy the require-
ment of 'group rationality.' That is, it must not be the case that a subset of
the members prefers to defect to an alternative collective. As an example,
suppose that half of the compositions in the repertory have a high prob-
ability of success while the others have a low probability. Clearly, the equal
sharing rule introduced in the preceding section will not be viable if the
difference in probabilities is so large that those more likely to succeed prefer
defecting to a collective of their own, even though transactions costs are
Risk sharing and the distribution of copyright collective income 33
then spread over half as many members as in the original collective.
However, group rationality can be restored by appropriately increasing the
share allotted to those with a high probability of success. As long as the col-
lective can categorize copyright holders according to their probabilities, the
benefits of risk pooling described in the previous section can be realized byviable copyright collectives.
Although we have focused on the case of independent composition lot-
teries, correlated lotteries would yield the same qualitative results, except in
the unrealistic case of perfect positive correlation where all compositionseither succeed or fail together and there are no insurance possibilities.
Negative correlation introduces hedging opportunities that improve the
collective's own risk situation. Indeed, with perfect negative correlation, the
collective can attain a riskless position, just as it can whenever it has access
to a perfect reinsurance market. In these cases the collective can offer its
members full and fair insurance for their composition lotteries. As correl-
ation among the lotteries and reinsurance opportunities become imperfect,
the insurance benefits a collective can offer become less valuable, but
nonetheless dominate the learn-then-distribute rule.
2.5 ASYMMETRIC INFORMATION
Thus far we have abstracted from the informational asymmetries associ-
ated with moral hazard and adverse selection. The learn-then-distribute
rule provides the strongest incentive to create a profitable copyrightable
work, since the author is rewarded for the effort to produce a successful
composition precisely in proportion to the success actually realized. In this
manner the rule addresses the problem of moral hazard, whereby a com-
poser's costly effort to produce a successful piece cannot be monitored andrewarded directly. However, the rule is an inefficient response to the
problem, since it leaves the risk-averse composer bearing all the risk of
failure when, as we have seen, insurance opportunities are available. If
moral hazard poses a significant problem for the membership of a collect-
ive, then it would be inefficient to take full advantage of the available insur-
ance opportunities, since the members would then have the least incentive
to be productive. As a general matter, an efficient response to moral hazard
entails balancing the benefits of enhanced incentives to put forth produc-tive effort against the risk-bearing costs that this necessarily entails. Anoptimal balance would call for a distribution rule that ties royalty payments
partially to a composition's realized success, offering some incentive for
productive effort, and partially to the realized success of the entire reper-
tory, offering some insurance against failure. Thus, moral hazard would
34 Developments in the economics of copyright
mitigate, but not eliminate the advantage we have identified for distribution
rules that incorporate risk pooling.
Adverse selection, whereby individual composers and the collective
have different perceptions of the likelihood of success, poses a different
problem. In a heterogeneous population, when individual composers are
better informed about the probability of success than the collective, risk
classification through categorization and screening, as commonly practiced
in insurance markets, would allow the collective to provide insurance to its
members, although not as effectively as when information is symmetric.10
For those more likely to be successful, screening mechanisms would pay a
higher proportion of royalties based on actual performance and a smaller
proportion through insurance. Thus, while all of the benefits of risk
pooling could not be realized, it would still be possible for members of the
collective to enjoy some of those benefits through a distribution mechanism
that does not rely exclusively on the learn-then-distribute rule.
It is also possible for the collective to be better informed about the like-
lihood of success than overly optimistic copyright holders. 11Indeed, it is
possible for the divergence in beliefs to be so great that the insurance terms
offered by the collective are unacceptable to the members. In this case the
learn-then-distribute rule would be the only viable system available to a
copyright collective.
2.6 CONCLUSIONS
In this chapter we have proposed that copyright collectives may not be
offering their members as great a welfare benefit as they could. By focusing
all of their attention on the transactions costs savings of their services, they
have ignored the opportunity to offer their members important risk-sharing
benefits. We have described several methods under which the net income of
copyright collectives is distributed to members in a manner that takes
advantage of risk-sharing opportunities and that Pareto dominates exclu-
sive reliance on a distribution rule that calls for first learning the outcome
of each composition lottery and then paying out royalties accordingly.
Two explanations may be offered for the failure of copyright collectives
to depart from complete reliance on the learn-then-distribute rule. First,
a partial explication can be found in the existence of asymmetric informa-
tion - adverse selection and/or moral hazard. In these cases, it is well knownthat the agent (the informed party, here, the creator) must retain risk in order
to satisfy incentive compatibility. Secondly, copyright holders may be unre-
alistically optimistic about their chances of success and therefore unwilling
to accept financially viable alternatives to the learn-then-distribute rule. 12
Risk sharing and the distribution of copyright collective income 35
Both of these explanations imply that the full potential of copyright collec-
tives to provide benefits for their members could be realized if their
members were better informed about the value of the risk pooling services
available through the collectives.
NOTES
1 . We note that in some instances, some of the monitoring costs are borne by the purchas-
ing entity which is required to report usage rates to the collective. In any case, the col-
lective still needs to suffer some costs in processing and administering this information.
2. Blanket licensing may also serve to eliminate competition among composers and allow
them to act as a cartel. However, in many jurisdictions, tariffs are regulated, limiting the
opportunity to exercise monopoly power.3. In sections 2.4 and 2.5 we consider the implications of relaxing these assumptions.4. Really we only require that the collective has a transactions cost of t' <mt\ that is, the
collective is more efficient than the sum of its members. Setting t' =
qt, the requirementis that q< m, and in the chapter we have assumed q
=1 . In fact, the more the underlying
product (here compositions) resembles a pure public good, the lower is the value of q,
and q goes to 1 in the extreme case of a pure public good. It seems likely that for the case
at hand, we will typically have a very low value of q, and so our assumption of q= 1 is
only meant to be a rough approximation that facilitates the argument significantly.
5. Clearly the condition can be expressed equivalently as k<mt-t\ that is, the cost of
learning is less than the total transactions cost savings.
6. Indeed, participation can be achieved with a lower membership than under the learn-
then-distribute rule, since each member gets a better risk-bearing situation. Exactly howthe limit value of m is affected is impossible to determine unless we introduce an analy-sis based on utility functions.
7. Concretely, the probability is 1 .76 per cent if/?= 0.5. This compares to the figure of 46.02
per cent when fair insurance is offered.
8 . These results are not so easy to prove. See, for example, Chapter 4 of Canavos ( 1 989) for
proofs.
9. In spite of this, the American Society of Composers, Authors and Publishers (ASCAP)in particular does discriminate. They distribute income in a way that discriminates
according to how long an individual has been a member (see Besen and Kirby, 1 989 p. 2 1 ).
Also, these authors (p. 10), note that, when comparing copyright collectives over the
world 'a very wide range of methods are employed to distribute the proceeds of blanket
licences'.
1 0. See Crocker and Snow (2000) for a theoretical analysis of risk classification in the insur-
ance industry.
1 1 . The possibility that authors may be overly optimistic about the chances of success oftheir work was discussed by Watt (2000, p. 89), and also by Besen (1987).
12. We note that in some instances the creator is paid a fixed fee by a third party who then
holds the copyright. Risk pooling offers no benefits to these third parties when they are
sufficiently diversified to be risk neutral.
REFERENCES
Besen, S. (1987), 'New technology and intellectual property: An economic analy-
sis', RAND Report N N-2601-NSF.
36 Developments in the economics of copyright
Besen, S. and S. Kirby (1989), 'Compensating creators of intellectual property:Collectives that collect', RAND Report N R-3751-MF.
Besen, S., S. Kirby and S. Salop (1992), 'An economic analysis of copyright collect-
ives', Virginia Law Review, 78, 383-41 1.
Borch, K. (1960), The safety loading of reinsurance premiums', Skandinavisk
Aktuarietskrift, 153-84.
Canavos, G. (1989), Applied Probability and Statistical Methods, New York:
McGraw-Hill.
Crocker, K.J. and A. Snow (2000), The theory of risk classification,' in G. Dionne
(ed.), Handbook of Insurance, Boston, MA: Kluwer Academic Publishers,
pp. 245-76.
Hollander, A. (1984), 'Market structure and performance in intellectual property:The case of copyright collectives', International Journal of Industrial
Organization, 2, 199-216.
Watt, R. (2000), Copyright and Economic Theory: Friends or Foes?, Cheltenham,UK and Northampton MA, USA: Edward Elgar.
Wilson, R. (1968), The theory of syndicates', Econometrica, 36, 1 13-32.
3. MP3s and copyright collectives:
a cure worse than the disease?
Stan Liebowitz
Only one thing is impossible for God: to find any sense in any copyright law onthe planet. Whenever a copyright law is to be made or altered, then the idiots
assemble.
Mark Twain: Mark Twain's Notebook, 23 May, 1903
3.1 INTRODUCTION
The advent of massive unauthorized copying by individuals using
peer-to-peer systems and MP3 files, besides generating an enormousamount of press, has brought the affected copyright industries to the
brink of declaring war against many of their customers. This has taken
the form of a zealous attempt by copyright owners to shut down and limit
peer-to-peer filesharing, with the record industry at the forefront but the
movie industry waiting in the wings. First Napster was shut down. Then
Napster's progeny were brought to court, although the Recording
Industry Association of America (RIAA) has for the moment been frus-
trated in its attempt to have the courts shut down these latter firms. 1
Simultaneously, the record industry has been covertly distributing on
peer-to-peer networks 'spoof files that look like popular MP3 files but
contain no music, in order to hinder downloaders. The industry has also
threatened (obliquely, to be sure) to start engaging in activities that would
prevent computers from being able to download MP3 files, although it is
unclear exactly how this would be done.
Finally, the industry is in the process of the most draconian action of all:
suing users for copyright damages. It already has sued and settled with hun-
dreds of users, some of whom have set up web sites to help them pay for their
settlements. The record industry hopes that fear of prosecution will reduce
the usage of peer-to-peer networks sufficiently to help return the industry to
a better financial state. Indeed, initial statistics from comScore/Media
Metrix, the PEW Internet Project, and Nielsen NetRatings indicate a rather
37
38 Developments in the economics of copyright
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
Figure 3. 1 LP albums soldper capita
sharp decline of about 30 per cent in the number of subscribers to major
filesharing networks after announcement of these law suites. 2
The industry has legitimate grounds for concern. As Figure 3.1 dem-
onstrates, sales since 1999 have been on a steady decline. The decline is
considerably larger than any other decline in the last 30 years, although
the current decline is still smaller than the declines brought about from the
introduction of radio in 1921 and the Great Depression. If real revenues per
capita had been the basis of analysis, the decline from 1978-82 would be of
a similar magnitude to the current decline, although the 1978-82 decline
was, in part, cost-based, due to the transition to less expensive cassettes
from vinyl records.
For many analysts, the behavior of the copyright industries appears mis-
guided. In their view, the corporate entities that are trying to enforce current
copyright laws are seen as antiquated behemoths unable to move with the
times. Critics have suggested that the current set of institutions, including
copyright law and the firms that largely administer the business associated
with copyright, do not effectively address the interests of creators or users. 3
Partly as a backlash to the behavior of the copyright owners, many acade-
mics interested in topics associated with copyright or the Internet have argued
that an alternative to the current copyright regime is in order. The proposals
that have been offered are often subsumed under the rubric of a 'compulsory
license', although most are not technically compulsory licenses; they are in
fact more closely related to performing rights tariffs administered by the
American Society of Composers, Authors and Publishers (ASCAP) and
Broadcast Music, Inc. (BMI). A more appropriate term to describe these pro-
posals would probably be to call them variations of a 'copyright collective'.
Proponents of these proposals generally emphasize several possible
MP3s and copyright collectives 39
positive characteristics of these alternative systems-they appear to have
relatively low administrative costs, they decriminalize behavior that has
become widespread, and they might lead to increased incentive for artistic
creation. Economists writing on the subject would also mention that such
a system would allow consumption to reach a level that is more efficient
than traditional copyright. Reading some of these articles, one can get the
impression that artistic nirvana is right around the corner, if only we have
the courage to scrap the current system.
While I will discuss these proposals in more detail below, the basic idea
is that a pool of money would be generated in a secondary market (pre-
sumably related to MP3s) and transferred to copyright owners. Specifically,
the money would be generated by taxes on ancillary products, such as blank
CDs, CD writers, ISPs, stereo equipment, and so forth. Although somecommentators see a copyright collective as a supplement to the current
copyright system, it is also viewed, particularly by its more passionate
advocates, as a complete replacement of traditional copyright, at least for
recorded music.
While such a system does offer some theoretical advantages, I arguebelow that the defects of such a system have not been sufficiently examined.
Although the current system is obviously imperfect, as any system must be,
it is unlikely that a copyright collective would meet even the modest goalsof a net positive impact, to say nothing of the claims of virtual perfectionthat have been attributed to it.
Just because a system could exist and survive does not mean that it is
efficient from an economic perspective. It is the purpose of this chapter to
examine the probable efficiency characteristics of such a system so that a
better understanding of the alternatives can be gained.
3.2 HOW SUCH SYSTEMS WORK - THE CONCEPT
It is important to note that the compulsion aspect of the 'compulsorylicense' refers to an action compelling the behavior of the copyright owner,not the user of the work. A compulsory licensing scheme is one where the
government requires that copyright owners make their works available to
users, usually at a fixed or capped price. In contrast, a copyright collective
may in principle be voluntary, but often it is a foregone conclusion that
membership is essential to receive copyright payments. In that case, it is a
distinction without much of a difference.
One analogy that is often brought up is the blanket license offered by
performing rights societies (ASCAP and BMI) to radio and television broad-
casters. The blanket license grants broadcasters copyright clearance for the
40 Developments in the economics of copyright
music they broadcast. Mildly simplified, radio and television stations pay a
percentage of their advertising revenues to performing rights societies,
which then distribute the proceeds among their members after deducting
for operating expense. The rates paid to these organizations are set by
semi-government organizations (the courts or a government appointed
board).
This is to be contrasted with a compulsory license favorably put forward
by Lawrence Lessig which, for example, compels the copyright owners of
distant programs to allow cable operators to retransmit these signals to their
viewers without having to negotiate with the copyright owner. Instead, cable
operators pay a small percentage of their revenues, specified by statute, into
a pool (several pools, actually) and this money is then disbursed among
copyright owners with oversight by the Copyright Arbitration Royalty
Panel (CARP), a government entity.
Another compulsory license is for the use of songs on (phono) records.
Once a song has been included on a record that has been made available to
the public, then anyone else wishing to record that song can do so as long
as they pay the composer of the song the deemed compulsory license fee.
Let me illustrate how this works with an historical example. Suppose
Jimmy Webb writes a song with incomprehensible lyrics called 'MacArthur
Park'. He can sell or give the song to the highest bidder (the song was actu-
ally sung by deceased actor Richard Harris). If a song appears likely to be a
hit ('MacArthur Park' went to number one in the charts), the negotiated price
for this first recording might be very high. Once the Richard Harris recording
of the song is distributed to the public, however, the compulsory license kicks
in. Any other performers who wish to sing about cakes melting in the rain can
now also record 'MacArthur Park', if they pay the compulsory license, which
is set by fiat to approximately one half of one cent per minute, per record dis-
tributed.4 Once the compulsory license kicks in, the composer is compelled to
allow additional users to record his song at the regulated price.
Since the blanket and compulsory licenses seem to 'work', to use a phrase
often found in writings on this topic, they are often held out favorably as a
demonstration that such a system can work to replace the record industry's
traditional revenue sources. However, as I explain below, there are important
differences between these current blanket/compulsory licenses and the sug-
gested new licensing systems. Further, these new systems are more likely to
fail a cost/benefits test than is the case for, say, the traditional blanket license.
Restitution vs Replacement
There are at least two possible types of copyright collective systems that can
be imagined. In one system, the copyright collective would entirely replace
MP3s and copyright collectives 4 1
the current market-based copyright system, at least for digital products. For
those industries for which it would be used, such a system would eliminate
copyright as we currently know it. Adherents of copyright replacement
argue that the current copyright system provides few financial incentives for
artists to create products anyway, and that alternative forms of payment(such as concerts) are the true current drivers of incentives to create. In their
view, the digital world provides vast new opportunities to improve on a
poorly functioning copyright system.
The choice doesn't need to be this stark, however. It is possible to allow
the current market-based copyright regime to operate as best it can while
using a copyright collective as a form of restitution for losses in the market
due to unauthorized copying activity.
Take the case of the compulsory license of distant broadcast signals in
cable retransmission. The television broadcast market is one that generatesmost of its revenues from local viewers in a regular marketplace. By almost
anyone's calculations distant signals carriage on cable systems comprise a
minor component of this market. One can argue that in order to limit the
costs of having thousands of cable operators negotiate with hundreds of
broadcasters for relatively minor rights in distant signals, a compulsorylicense makes sense.
Similarly for the compulsory license in records, the transactions
surrounding the first recording allows the market to determine and gener-ate the value of the primary use; again, it is only the later, presumably less
valuable recordings, that can take advantage of the compulsory license.
If a copyright collective system for copying records were to work as an
add-on to traditional copyright, the goal of the copyright collection systemwould be to compensate copyright owners for the damage brought about byMP3 downloads. The system as usually envisioned would provide immun-
ity from copyright violations to private individuals engaging in MP3 down-loads and it would also provide revenues to the copyright owners through a
tax on ancillary products.
Further, the current copyright system would continue to exist in its
present form, providing copyright protection against the many forms of
copying other than file sharing. Such a proposal has recently been putforward by Neil Netanel. I refer to this as a quid pro quo system, where
copyright owners forgo revenues from peer-to-peer file sharers in return for
payments based upon taxes on ancillary products.A somewhat different alternative, and one that hasn't to my knowledge
been proposed, is to put in place a tax-and-subsidy system to recompense
copyright owners without any change in the rights or obligations of the users
with regard to the legality or illegality of their MP3 downloading activities.
The 1992 Audio Home Recording Act's provision for a tax on digital audio
42 Developments in the economics of copyright
tape (DAT) can be thought of as this type of system, as can the current
Canadian tax on blank CDs, which is accumulated in a pool and used to pay
copyright owners even though MP3 uploading would still remain a copy-
right violation. 5 This can be thought of as a status quo with restitution.
The Use of MP3s is Not Exogenous to the Copyright Collective
In both the status quo and quid pro quo systems, a tax and distribution
scheme is set up to compensate copyright owners for losses brought about
by MP3 downloads. There is an important difference between them,
however. The status quo system does not provide users with any new rights
in terms of making copies (or put another way, it doesn't abrogate the tra-
ditional rights of the copyright owners). The quid pro quo system, in con-
trast, essentially creates new rights to the traders of MP3 files who are no
longer in violation of copyright. The proponents of copyright collectives
clearly favor the quid pro quo system. Yet, in this latter world, we might
expect a change in behavior on the part of MP3 downloaders that weakens
the case for a quid pro quo system.
Under the current copyright regime, there is a potential cost to the MP3traders if they are discovered violating copyright. The recent crackdown by
the RIAA on downloaders is an example of increasing this cost. Not only
is there the moral conundrum involved with knowingly violating the law,
but MP3 downloaders must bear the risk of being charged with copyright
infringement.
One can predict that if these costs to MP3 users disappear there will be
a larger usage of MP3 files and a greater impact on the sale of CDs. It is
difficult to predict the magnitudes, however, since it is unclear how muchof a restraint these costs are currently imposing on the behavior of users.
But it is important not to predicate the amount of revenues necessary to be
generated by a quid pro quo system based on the amount of MP3 down-
loading that occurs before the system is in place, although that is exactly
what these proposals do.
3.3 THEORETICAL BENEFITS OF A COPYRIGHTCOLLECTIVE SYSTEM
The current evidence supports a claim that file trading is damaging the
record industry.6 Yet most MP3 traders are normally law-abiding citizens.
Many are students. The typical MP3 trader appears to consider MP3downloads to be a perfectly acceptable form of behavior. The number of
downloaders is very large, constituting an important percentage of the
MP3s and copyright collectives 43
population.7
It is so large that copyright supporters risk labeling a large
portion of their fellow citizens as thieves. Although a bit far-fetched, one
might imagine the downloading issue rending the fabric of the country just
as the drug enforcements efforts have. That, at least, is the more nightmar-ish scenario that has been suggested.
Under these circumstances, society could well decide that the cost of
trying to reduce unauthorized MP3 downloading under the current copy-
right regime is too great- in particular, greater than the costs of switching
to some form of copyright collective. This is particularly true given the well-
known imperfections in the copyright system, which I summarize below.
The Imperfection of the Copyright System
Intellectual products, such as stories and music, are not in and of them-
selves physical commodities. Instead, these products generally must be putinto some physical representation (records, tapes, books, etc.) before theycan be enjoyed. Without some form of intellectual property right, such as
copyright, the creator of an intellectual product would not have control
over the reproduction and sale of his creation. Anyone could reproduce andsell copies of the creator's work, and any producer of reproductions whodid not pay the creator could undercut the price that was charged by pro-ducers who did pay the creator. This state of affairs would be expected to
eliminate from the market any producers who paid the creator, leaving norevenue for the creator of the intellectual property.
Since it is unreasonable to expect creators, or artists, to expend effort
(full-time or professional effort) producing artistic works without the
expectation of concomitant payment (anymore than we would expecthome builders to build homes, or janitors to clean floors), economic analy-sis predicts that without copyright (and patent) laws, the production of
intellectual products would be far below levels that would be considered
socially desirable. 8
Even with the existence of intellectual property protection, the quantityof intellectual products created and consumed will be less than the ideal
level. This is the unavoidable underproduction and underconsumption of
intellectual properties that results from intellectual products being 'non-
rival' goods (sometimes called public goods).9Non-rivalry means that once
a work is produced it does not get used up. A physical manifestation of the
work can be used up, but not the work itself.
To start, there is a problem in the consumption of an intellectual product.There is no cost to society (other than the cost of a reproduction) in letting
an individual consume a unit of a non-rivalrous good, since there is nodiminution in the possible consumption choices of anyone else. Therefore,
44 Developments in the economics of copyright
an economic requirement for efficient consumption of a non-rivalrous goodis that any consumers who would like to consume it (e.g., a particular song)be allowed to consume the good as long as they are willing to pay the repro-
duction cost. Note that the only single markup over the reproduction cost
that could achieve this requirement would be a markup of zero. 10 A markupof zero, of course, would then provide no revenues to the creators (as
opposed to the producers of reproductions) of intellectual products. With
no revenues, many creators would abandon the creation of non-rivalrous
goods, which almost everyone agrees is not an efficient solution.
This is the economic paradox of non-rivalrous good pricing. There is no
practical mechanism that will produce the ideal quantity of titles and the
ideal number of reproductions for any given title.11
Charging a positive
markup will generate revenues, allowing for the creation of non-rivalrous
works, but the consumption would then be inefficient since some individu-
als with a value for the product greater than the reproduction cost will not
purchase it.
There is a further inefficiency involved with non-rivalrous goods. Non-rivalrous goods would be expected to be produced at less than the ideal
level. In an ideal world, every product that has a value to consumers that
surpasses its cost of production will be produced. For any single intellec-
tual work, even a monopoly seller of that work will not be capable of
appropriating the full potential value of that title since some consumers will
pay less than their reservation price while other consumers will be priced
out of the market. Thus, some intellectual products will not be produced,even where the potential value is greater than the cost of production. Onlya perfect price discriminator would be able to appropriate sufficient rev-
enues to guarantee the efficient production of works. 12 This inefficiency is
not restricted to non-rivalrous goods, however. Every potential product has
the possibility of not being produced due to a lack of appropriation- in
other words, there are some markets that do not exist, but would exist in an
ideal world.
These potential production-based imperfections have greater resonance
in the case of non-rivalrous goods because the markets for any particular
type of non-rivalrous good - books, movies, and so forth - consist of
individual, non-homogeneous titles, each of which is essentially a separate
market. Thus the generic market, for books say, exists, and it is easier to
imagine the harm from not having all the titles written that we might want
than it is to imagine the harm from an undefined market that does not
exist at all.
These market inefficiencies are sometimes taken as the intellectual
stepping-stones on the path to suggesting that some mechanism other than
traditional copyright be considered.
MP3s and copyright collectives 45
Copyright Collectives can Lead to Efficient Consumption, But Only at
a Zero Price
Since the cost of the creation of the intellectual product, the music in the
case of records, is a fixed cost, and unlimited copies can be made from this
one original, the opportunity cost to society in allowing any individual to
consume the product is merely the cost of making (and delivering) the copy.This means that society gains each time a potential consumer with a
willingness-to-pay for the product that is greater than the cost of copyingis allowed to consume the product.
As explained above, under a copyright regime, successful projects musthave prices that deter some consumers from generating surplus because the
creator must receive some payment. Consumption, therefore, will be below
the ideal level.
This logic also argues for a copyright term, if copyright is the chosen para-
digm, that is only long enough to allow the creator to receive sufficient
payment to voluntarily provide the work so as to minimize the time duringwhich consumption remains below the ideal level.
These consumption inefficiencies in the copyright system can provideammunition to those arguing for changes. If only we were willing to forgo
providing any system of payment to the creators of works, they suggest,
we could then be guaranteed that consumption of these works wouldbe efficient.
Yet few appear willing to argue that copyright owners should not be com-
pensated at all for their efforts. It has sometimes been suggested, therefore,
that alternatives to copyright, such as government grants or prizes, be used.
The current proposals for copyright collectives suggest that a tax be placedon some ancillary product, such as blank CDs, or CD burners, or perhaps
high fidelity sound equipment. The proceeds from this tax would then be
used to compensate copyright owners for their efforts.
This, of course, introduces its own inefficiency by reducing consumptionin the ancillary market that has the tax. Even ignoring the fact that the tax
may fall on the wrong individuals (not all blank CDs, for example, are used
to copy MP3 files; not all CD burners are used to burn MP3 files), by reduc-
ing the uses of blank CDs or burners (through the higher prices induced bythe tax) a new market distortion is created in these markets. If we lower the
price of the copyrighted work to its marginal cost, but make up the rev-
enues by imposing taxes in another market, we are merely shifting the
inefficiency from the market for the copyrighted work to a related market.
Even if the performers making CDs were to be paid through general
taxes, as opposed to taxes on ancillary products, there would still be the
general inefficiencies involved with the tax system. Therefore, even if a
46 Developments in the economics of copyright
copyright collective system eliminates an inefficiency in its home market,that is hardly grounds for its use since it creates inefficiencies in other
markets.
3.4 REAL-WORLD DIFFICULTIES
It is impossible to determine the proper size and direction of markets in any
way other than by the examination of freely functioning markets. Althoughthe difficulty of emulating markets has historically been grossly underesti-
mated, whether we are talking about centralized economic planners or justthe more prosaic regulatory bodies that abound even in market-based
economies, history has revealed that outside observers cannot divine the
characteristics of a market from alternative sources of information.
Markets arrive at results that are often very difficult to predict in advance.
Often, with enough effort, economists can come to understand why market
outcomes take the particular form that they do. A few simple examples can
be used to illustrate cases where economic theory would even have a
difficult time predicting whether the price is positive or negative, to say
nothing of how much should be paid.
Begin with a case that is often misunderstood -payola. Payola is the
pejorative term used when record companies pay radio stations to play par-
ticular songs. How would an outside analyst determine the proper paymentfor the use of the records? Typically, in the economy, firms pay for their use
of inputs. Radio uses music as its main input. Yet radio stations do not payfor the CDs that they play. Quite the opposite
- there is a history of record
producers paying radio stations in order to play their records. Further,
radio is a substitute for listening to recorded music, and the recorded music
industry would likely benefit if radio could be restricted or eliminated. 13
Although payola is a practice that governments have tried to ban, it is in
fact indicative of competition between record companies.
Similarly, many companies pay television and movie studios for product
placement, as BMW has recently done to have its automobiles and motor-
cycles featured in James Bond movies. M&M Mars is thought to have madeone of the great marketing gaffes of all time when it refused to allow the
makers of ET to use its product; instead, Reese's Peanut Butter Cupsenjoyed the benefits of global publicity from having an immensely popularextraterrestrial enjoying its products.
How would any third-party entity know whether the producers of James
Bond movies should pay for their use of cars, or whether they should get
paid by the automobile companies for having the vehicles appear in the
film? There is no simple economic test other than a market test.
MP3s and copyright collectives 47
Automobiles are costly to produce. Automobiles are an input in James
Bond movies. The producers pay for actors, cameras, microphones, and
most other inputs. Cars turn out to be different only because automobile
manufacturers believe that the publicity associated with the movie will gen-erate sufficient additional sales that they would benefit from giving the
automobiles to the studio. Should this belief change, then once again movie
studios would need to pay for automobiles. In a movie where the automo-
bile was represented as being unreliable, movie studios will need to pay for
the automobiles they use and are likely to use fictitious vehicles to escapethe wrath of the automobile company so characterized.
The point here is to demonstrate the difficulty of determining the most
rudimentary aspect of pricing- whether a product receives a positive or
negative price. Obviously, if the sign of the price is difficult for an analyst
to determine, the magnitude of the price is even more difficult still.
Since copyright is the subject of this chapter, it is useful to examine the
workings of some copyright markets. For example, should authors be paidfor their work, or should they pay publishers to publish their work?
Academic authors often pay journals to have their work published (the
submission fee is nominally to pay, in part, the outside 'referees' for their
time and effort but authors would pay even if referees worked for free).
Academics are willing to pay to be published because their reputationsand incomes are expected to increase when their articles are published.The author's financial gain from publication is greater than the gain of
the journal.
Similarly, many individuals would like to be published authors. It is
gratifying to have a book to show to friends and neighbors. There has
always been a 'vanity' press for individuals who pay publishers to print
their books.
If one restricted one's attention to only these vanity and academic markets
one might think that it was the norm for authors to pay publishers, that fame
was the main purpose in writing. Yet we know that the bulk of the bookmarket (measured in revenues) consists of works where a small number of
authors are paid some very large sums (both up-front and in royalties) to
produce books even though the authors often become quite famous.
If we did not have the market to point the way, how would an observer
or agency know to charge negative prices to one set of authors and positive
prices to another?
Similarly, it might be (and has been) argued that CD sales serve mainlyas a useful device for artists to increase concert endorsement revenues. It
could be the case that this was the market solution. How would a govern-ment entity know the sign or size of the correct price for payments to
recording artists except for the fact that we have the market to tell us?
48 Developments in the economics of copyright
Admittedly, the direction of payment is often not as difficult to predict
as in the examples I have chosen. Nevertheless, even in markets where the
direction of the payments is not difficult to determine, the amounts clearly
are difficult. I cannot over-emphasize how difficult it is to emulate the
results that come from market transactions.
3.5 HOW MIGHT A COPYRIGHTCOLLECTIVE WORK?
The overriding determination that must be made in a copyright collective
system is determining the amount of revenue (the royalty) to be raised. Thetwo secondary decisions that must be made are: (1) how the money is goingto be raised - in other words, who is going to be taxed, and (2) how the
money is going to be distributed. For expositional convenience, we can
assume that there is a 'Copyright License Board' (CLB) whose task is to set
rates and determine payouts.
Boards like the CLB usually determine amounts by having competing
experts (often economists) testify about the conditions in the marketplace.
These experts usually work for competing parties (e.g., performing rights
societies and broadcasters) who are the payers and the payees of the various
copyright royalty rates. The competing parties in this instance are likely to
be the artists and record industry versus the retailers and producers of
ancillary equipment who will bear the burden of the tax.
Determining the Amount of Revenue to Raise
The most difficult issue, in my opinion, is determining the amount of
revenue to be raised, which we can refer to as the 'quantum'. Although there
are many arguments that can be made about how one might make such a
determination - arguments based on ideas such as fairness, cultural value,
and national identity, a primary goal is often thought to be to mimic markets
since markets are thought to provide efficient solutions that maximize
economic values. Boards such as a CLB usually state that they try to emulate
markets, although this is not necessarily the exclusive criterion.
If we charge the CLB with the goal of mimicking the market, it will
become increasingly more difficult for the CLB to choose the quantum as
time goes on since the initial market values will recede into the past.
Initially, the market values will be well known since we have been in a periodwhere the market was the mechanism generating values. If, for example, the
current system were to be scrapped today, the CLB would have a fairly goodidea of what the quantum should be if it wanted to match the revenues of
9000-
8000-
g 7000-
1 6000 -
c'
5000-
4000-
3000-
2000
MP3s and copyright collectives 49
Real album revenue
O> O^ ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON ON
Figure 3.2 Album revenues
m vo r- ooON ON ON ONON ON ON ON
record companies or performers. Both of these statistics are available from
the current market participants. If it is thought that current numbers are
below what they would have been without the untoward influence of MP3downloading, the numbers could be adjusted to take this into account.
Determining the quantum in this case is relatively easy.
As we move further away from the period when values were generated in
markets, it will become increasingly difficult to divine what the market move-
ments would have been. And markets do move. Figure 3.2, for example, gives
inflation adjusted revenues in the market for full-length audio recordings
since 197 3. Note the 50 per cent increase from 1975 to 1978, followed by a 45
per cent drop from 1 978 to 1 982 and a doubling from 1 982 to 1 994. Based on
a history of CLB-like entities, it is very unlikely that such dramatic changeswould be countenanced, approved, or understood. Demographics were
largely unchanged. Income changes did not follow this pattern. Prices of
substitute products did not suddenly rise or fall. Thus, the usual explanatoryvariables that experts look at would not have predicted these changes well.
I suggested in my 2004 article (see note 6) that this increase in revenues
beginning in the early 1980s was caused by the emergence of portable
devices to play recorded music. Even if a CLB had grasped the importanceof portable devices, how would it have known to double the revenues and
not increase them by, say, 40 per cent, or 250 per cent? These very prosaic
difficulties are at the heart of the mispricing that is rampant in non-market
economies.
It is unreasonable to expect even a highly skilled CLB dominated byunbiased market experts to correctly mimic the workings of a market. The
50 Developments in the economics of copyright
reality is that the CLB will have to wade through reams of testimony that
is often far from unbiased and is often outside the realm of the expertise of
the members of the CLB (who are often lawyers or industry professionals).
Although I will not discuss in detail the actual performance of such boards
in this chapter, there are numerous instances of decisions indicating that
these boards, for whatever reason, have reached decisions that are incon-
sistent with basic economic principles. For example, I think it would be
agreed by virtually all knowledgeable commentators that it would be an
error in economic logic for the CLB to set the quantum to grow with the
overall inflation rate, holding the real quantum constant. This would be a
particularly egregious error if the inflation-adjusted market for related
copyrighted goods was rapidly growing at the time. Yet there have been
such decisions in the case of blanket licenses. In a similar vein, the com-
pulsory license for mechanical rights remained constant from 1909 until
1976. Is it even remotely possible that the value of such a right would have
remained constant in nominal terms for 77 years if it were market deter-
mined? Of course not.
If the copyright collective system was one of restitution instead of
replacement, one might argue that these problems would not be as severe
as implied in the preceding paragraphs. If part of the market continued to
function as a market, might not the market signals be used to provide the
type of detail I have argued that a CLB would have difficulty inferring?
There is some truth to this claim, but as conditions changed in the CDmarket, for example, it would be difficult to know whether the then current
CD purchasers represented the typical market anymore. For example, some
audiophiles claim that MP3 compression lowers the quality of the sound
relative to CDs, making audiophiles less likely to participate in MP3 down-
loads. If the market for CDs became dominated by audiophiles, who are
well-known for spending large sums on esoteric audio equipment, the
market price for CDs might rise dramatically. This would hardly be indica-
tive of an overall increase in demand, however. Thus, the problems dis-
cussed in this section would still exist, even though they might be mitigated
to some extent.
Using MP3 Downloads as a Proxy for the Market
It might be (and has been) suggested that the number of MP3 downloads
should be used to determine changes in the quantum set by the CLB. After
all, wouldn't MP3 downloads bear the same relationship to revenues as did
the quantity of records sold? Although there might be such a relationship,
there also might not be. Even if there were, we do not know whether we can
get accurate statistics on total MP3 downloads.
MP3s and copyright collectives 5 1
Not every download is actually a download. Many downloads are unsuc-
cessful or aborted for one reason or another. Not all downloaders are
equal. Some downloaders have the capability to purchase CDs and others
do not. If a large number of five-year-olds are convinced to download a
song from a television cartoon while their parents sleep-in, should that get
equal weight to downloads from adults?
Obviously, not every download represents a lost sale.
Finally, and most important, the measured number of downloads can be
manipulated at relatively low cost. What is to keep record companies, or
other organizations that represent large numbers of recording artists, from
using banks of computers to download songs? What is to keep them from
running contests to convince private individuals to download more songsthan they otherwise would? It would be relatively inexpensive to manipu-late MP3 download figures. Additionally, as we shall see below, individual
creators have incentives to inflate their downloads, also skewing measured
downloads.
Who Pays the Tax?
There are several issues with the placement of the tax (tariff). First, one
needs to determine which ancillary market to tax. Second, one needs to
determine the proper rate to charge.
As a general statement, it is correct to say that taxes distort markets. As a
matter of simple fairness, however, it is appropriate that the inevitable losses
from a tax be borne as much as possible by those who benefit from MP3downloading, and not by individuals having no association with MP3download activity. It is difficult to make a serious case that MP3 download-
ing is of sufficiently important national concern that we all should be willing
to chip in to support it. This would rule out using general tax revenues.
With this as a backdrop, the choices available are not terribly appealing.
Blank CDs or CD-writers, although likely to be used by MP3 download-
ers, are also heavily used by individuals interested in storing computer data.
There is no compelling logic to having other computer users subsidize MP3downloaders.
The purchase of audio equipment, on the other hand, is more directly
related to listening to recorded music than is the case for CDs and CDwriters. For this reason, one might argue that a tax on audio equipmentwould more precisely target music listeners, the group that seems to benefit
from MP3 downloads and thus the group that should pay the tax. There
are still variations within this group, however. Some individuals are more
interested in watching DVDs than listening to music, yet both require
similar equipment.
52 Developments in the economics of copyright
There is a more serious problem in the case where the CD market con-
tinues to exist. A tax on audio equipment would be paid not only by MP3downloaders but also by those who purchase CDs and do not download
MP3s. Although the continued existence of part of the CD market was a
mitigating factor in terms of determining the quantum of the tariff, here it
works to increase the distortion caused by a tax.
Finally, the proposal that is currently enjoying the most support is one
to tax ISPs. The claim is often made that if all American Internet users were
to pay $5/month, they could be allowed to download as much as they want.
Who wouldn't be willing to pay $5 a month for this privilege?
Although this is an interesting rhetorical device, there are two problemswith the logic. First, the $5 per month figure is based on the current esti-
mate of losses suffered by the record industry, and those losses are likely to
increase when everyone gets the green light to stop paying for CDs. A more
complete payment would be in the order of $20 per month.
Second, plenty of individuals would balk at paying these amounts. Manyindividuals using the Internet do not buy CDs and see no reason theyshould subsidize CD consumers.
Who Gets the Revenue?
Once a quantum of money is generated, how would a CLB determine howthe money is to be divided among the copyright owners?
One possibility is that a single copyright collective would be formed, cov-
ering all copyright owners, and that the collective would be charged with
distributing the revenues. Even if this were to be the case, however, the col-
lective would still have to decide how the revenues were distributed amongits members, and it would face the same decisions that I am going to ascribe
to the CLB in the following paragraphs. Further, when there are different
parties involved in the creation of a product (musicians, composers, record-
ing companies), it is likely that there will be multiple groups striving to get
larger shares of the royalty payments, and the Board would have to set or
approve any division that occurs. In the case of retransmission compulsory
license, for example, the Copyright Arbitration Review Panel (CARP)determines the relative amounts received by the sports claimants, the
creators of local programming (news and information), Hollywood pro-
ducers and others. In other words, the Board determines the revenues goingto different genres and it is not impossible that something like that could
happen if the recording industry were put under the auspices of such a
board.
Regardless of how the various parties are aligned, how would revenues
likely be distributed? The suggestion that has most often been made is to
MP3s and copyright collectives 53
use data on MP3 downloads as a basis for rewarding creators. If method-
ologies could be created to measure downloads, this view holds, and we
could award the dollars based upon these measurements.
The comparison is often made to the performing rights societies
(ASCAP and BMI) who sell blanket licenses to broadcasters and use the
proceeds to pay their members based upon estimates of how frequently a
composer's music is played on radio or television. In fact, the distribution
systems are more complex than this, often taking the length of the music
and the type of music into account when determining their distributions.
Nevertheless, the frequency of play clearly is a factor. There are difficulties
in emulating this system in the case of MP3s, however.
One fly in this ointment is that MP3 downloads are amenable to manipu-
lation, as I have mentioned when discussing the determination of the
quantum. It is possible for an individual artist to pay someone to download
massive numbers of files in order to increase that artist's share of the dis-
tributed revenues.
One might suggest that such manipulation can be prevented by some
form of user authentication that could discover instances of the same user
repeatedly downloading the same file. But how would the CLB discover
cases of massive fraudulent downloading? Perhaps there could be some
system that only allows one downloaded instance of a song for each IP
address, but I presume creative coders could find ways to fool such a system.
Might the Board be forced to engage in the same type of behavior that the
RIAA is now engaged in - trying to prevent 'unauthorized' downloading?This would be ironic, since it is largely that behavior which has created the
pretext for creating a copyright collective system to begin with.
Gaming the system in this manner is not a problem under current com-
pulsory or blanket license systems. The products upon which these licenses
are placed are purchased in the market (for example, in the case of per-
forming rights, broadcasters buy television programs, and in the case of
mechanical rights, individuals buy records) and it would be a money-losing
proposition for copyright owners to try to inflate the measured use of their
creations by purchasing large numbers of records or paying television sta-
tions to play programs containing their music. 14
A second problem in using download statistics is that downloads maynot match up well with purchases. Bands could convince their fans to
download songs for the sake of downloading. It is likely easier to get youngfans with much time but little money, to spend some time downloading
songs to help out their favorite band than it might be to get them to buymore CDs. Such results are not likely to mimic the market very well.
Finally, it is not clear that we can measure downloads very well.
Although I have seen estimates on some web sites, I am not sure how
54 Developments in the economics of copyright
accurate they are. Those who purport to measure such numbers often keeptheir techniques shrouded in secrecy.
3.6 COMPARISON TO BLANKET LICENSES
When the argument for copyright collectives is put forward, analogies are
frequently made to the blanket license sold by the performing rights asso-
ciations ASCAP and BMI. The purchase of a blanket license allows the
purchaser to use the entire repertoire of the association for a yearly fee that
is usually related to the revenues of the firm purchasing the license. The
prices of the blanket license (royalty rates) are not market determined but
instead are controlled by Rate Courts and the CARP. The thinking appears
to be that the performing rights markets seem to be functioning well, so
there is no reason to believe that a similar regime for the recording indus-
try might not also work well.
There are three problems with the suggestion that the performing rights
markets are an exemplar to be followed by the record industry. First, it isn't
clear how close the royalty rates that have been chosen by the Rate Courts
and CARP are to the 'correct' rate. Second, the performing rights markets
tend to be small offshoots of larger markets, so any inefficiency would tend
to be relatively small relative to the market as a whole. Third, there are often
front-end market arrangements that would serve to limit the damage done
by inefficient performing rights royalties which are not available in the case
of a replacement for the record industry. I will take each of these points
in turn.
Are Performing Rights Efficiently Priced?
Can we take the history of blanket license rates as a harbinger that a copy-
right collective will work well as a replacement for the record industry?
Are performing rights efficiently priced? The answer is that we have no
earthly idea. 15
There are several markets where blanket licenses are used. In the case of
television and radio, the two largest markets, the blanket license purchased
by the broadcaster confers only the right by the composer to broadcast the
music. Ironically, the performing right does not confer any permissions bythe performer of the music.
How would we know if the tariff rates were set too high or too low?
Perhaps we would know that the rate was too high if composers were
willing to pay broadcasters to include their music in a television or radio
broadcast. Yet, the rate could be higher than the optimal rate without
MP3s and copyright collectives 55
being so high that it entirely eliminated the payment normally going to the
composer of music used in broadcasts. And, of course, the rate could be
too low with no obvious way to make such a determination. If one peruses
the economic arguments that have been put before various copyright tri-
bunals, one finds no clear method to tell whether rates are too high or too
low.
A major saving grace of most performing rights tariffs is that they are
linked to the overall size of the market. The performing rights tariff rate for
radio, for example, is a percentage of advertising revenues. Therefore, the
royalty payments will change as the industry grows or declines. This assures
some modest linkage between them and is likely to keep the royalty pay-
ments from getting too far out of line.16
If the current record industry paradigm is replaced with a copyright col-
lective, however, it will be impossible to tie royalty payments to some indus-
try measure since there will be no market measures to be found. There will
be no restriction on the possibility that the royalty payments might become
much too high or much too low.
Performing Rights are an Ancillary Market
With compulsory and blanket licenses, the tariffs make up a very a small
component of the total market. In the case of television broadcasting, the
performing rights tariff is only in the range of 1-2 per cent of the total
revenue of broadcasters. In the case of radio, it is in the vicinity of
3 per cent. For sound recordings, the compulsory license only covers addi-
tional performances of the song after the initial performance, and the
original performance is likely to be the far more important market.
Similarly, for the case of a compulsory license for cable retransmission of
distant signals, the major market is the broadcast television market and
the compulsory license is a very small component. If the tariff rates were
set incorrectly in any of these markets, it would not have too large an
impact on the overall market due to the very small impact that the tariff
has on the overall market.
Why is this important? In these cases, the potential harm from replacing
a market price with a regulated price is small because these tariffs are
small. In fact, it is informative to examine the reasons why the royalty rates
were not allowed to be set in markets in the first place. In many of these
cases, it was thought that transactions costs would be so large relative to
the value of the rights as to tend to cause market breakdown. Imagine the
difficulties that individual composers would encounter trying to track
down each use of their song in a television or radio broadcast relative to
the value of the right in the broadcast of a single program. Imagine the
56 Developments in the economics of copyright
costs of trying to negotiate rates in each instance, or of broadcasters trying
to track down all the composers to secure the required rights. In this case,
it appears to be far better for everyone involved to have something like a
blanket license.
In principle, you could have market negotiations between ASCAP/BMIand broadcasters to set the blanket tariff rate, but a concern that the per-
forming rights societies would wield monopoly power over broadcasters
(perhaps somewhat unfounded since the broadcasters have been able to
organize into a single bargaining unit of their own) led to the requirementthat either party could appeal to the appropriate regulatory body to have
the rates set.
It is also the case that the rate is set between the seller and the buyers of
the license, with some regulatory authority as an intermediary. Comparethis to the suggested copyright collective for MP3s. The parties involved in
the setting of the rate will be the sellers of the right (record companies,
artists, composers), but the other parties are likely to be the sellers of the
products bearing the tax (ISPs, sellers of blank CDs, and so forth). Their
negotiations are unlikely to be able to lead to anything approaching a
market value since ISPs, for example, are unlikely to care how much music
is created since they are not the consumer of the music. 17
Further, MP3s substitute for the purchase of a record, and the sugges-
tion has been for a compulsory licensing system to replace either a very
large portion of the market or perhaps the entire market, not 1 per cent or
3 per cent. In this case, the copyright collective handles the primary product
being sold. In this case, any mistake will be far more serious.
Unlike Performing Rights, There is No Safety Valve
There is one other important difference that mitigates the damage from
mispricing the royalty rate in a blanket or compulsory license. These
markets often have a separate but related market that can compensate, to
some extent, for errors in the royalty rate.
The television blanket license, for example, is a back-end payment that is
made in addition to front-end payments that are most often paid to com-
posers when their compositions are commissioned for a television program.If the royalty rate is too high, the front-end payment might very well be
reduced since higher royalty rates would make composers more willing to
accept lower front-end payments. Similarly, if the royalty rate were too low,
broadcasters would need to pay higher front-end payments if they wished
to achieve the same quality of music embedded in their programs. The
front-end allows gross mistakes to be at least somewhat ameliorated. There
are other complexities having to do with risk sharing that would require a
MP3s and copyright collectives 57
particular front-end/back-end balance for true efficiency, but I will largely
ignore that issue at this time.
A copyright collective system that provided the revenues for the entire
recording industry enterprise has no 'front-end' that might be used to com-
pensate for an error in the royalty rate. If a mistake is made in this market,
which as we have seen is very likely, there is no safety valve to ameliorate
the problem.
3.7 CONCLUSIONS
The current difficulty with enforcing copyright in the face of MP3 down-
loads is roiling the economic covenant between music creators and music
users, threatening to overturn current institutional arrangements. Academic
writers have proposed various arrangements to assuage these difficulties.
The most popular of these suggestions has been to replace the current system
with a form of copyright collective.
From a distance, a copyright collective has some very attractive features.
One needs to examine such a system in detail, however, to truly gauge the
likelihood that its adoption might lead to an improvement.A copyright collective throws out the markers, the lighthouses if
you will, that can help guide the prices in these markets. It requires that
prices and revenues be set in some arbitrary manner. Setting prices and
revenues are the very questions that any economic system answers by its
choice of rules. The evidence of the last century has led almost all com-
mentators to agree that markets are superior at allowing consumers to
determine which goods producers produce, how much is produced, and at
providing incentives for quality improvements, compared to commandand control methods.
There are good reasons to believe that mispricing by the copyright col-
lective will be a more serious problem for the record industry than it mighthave been for other markets that have adopted some form of copyright col-
lective. And there would be no safety value in the record industry copyright
collective to limit damage.Before we throw out the baby with the bathwater, we need to investigate
more carefully the arguments that are used to support a movement awayfrom the unfettered market and toward some alternative, such as a copy-
right collective. We also need to consider other proposals, such as enhanced
copy protection, known as digital rights management. Finally, we should
not be so quick to abandon the current market. It is not yet clear howonerous enhanced enforcement of current copyright laws will turn out to
be, or whether such enforcement can feasibly protect the industry. Only as
58 Developments in the economics of copyright
a very last resort should we replace the current market system with a
copyright collective.
NOTES
An earlier version of this article was first published by the Progress and Freedom Foundationon IPcentral http://www.ipcentral.org/review/ vln2intro.html.
1 . After its victory in the Napster case, the industry hit a roadblock in its attempt to shut
down Napster progeny Grokster, Streamcast and Kazaa. A judge has ruled that those
non-centralized filesharing systems were little different than VCRs and thus were not
liable for the infringing behavior of their customers.
2. See for example, http://www.pewinternet.org/PPF/r/124/report_display.asp.3. There are numerous academic authors who believe the current system should be
changed. See Fisher (2003), Ku (2002), Lessig (2001), Nadel (2002), and Netanel (2002).An overview of the parties and the ideas involved in this debate can be found at the web
siteGrammy.com: http://www.grammy.com/features/2003/0725_complicenses.html.4. There are some other notification requirements that I ignore here. Also, the price is actu-
ally the higher of eight cents per song of five minutes or one-and-a-half cents per minute.
5. The 1992 Act also required that manufacturers of DAT machines restrict the machinefrom being able to make copies of copies. The DAT technology never was very success-
ful in the US whereas CD burners have become extremely common.6. For evidence on the impact of MP3 downloads on the sales revenues in the recording
industry through 2002 see Liebowitz (2004a).
7. Estimates suggest that 40 million Americans have downloaded at least one copyrightedwork. See 'Downloads to save music biz' by Jane Weaver, MSNBC, 12 August, 2002.
8. It has sometimes been argued by economists and others that copyright is not needed
at all. This argument often boils down to nothing more than a claim that the optimalduration of the copyright is less than the time advantage that one gets from being first.
See Plant (1934).
9. I used to use the term 'public goods' but there are at least two definitions in the profession,one that contains only the non-rivalrous consumption assumption and one that also
includes 'non-excludability' (the inability to prevent users from consuming the good). The
difficulty with the latter definition is that non-excludability is a function of the law and the
amount of resources devoted to excluding nonpayers, the subject of this chapter. The twoattributes are not necessarily related and I have always thought it was counterproductiveto treat non-excludability as an inherent attribute of intellectual products, since it is not.
10. I am assuming that only one price can prevail in the market and that consumers have
different valuations for the product, with these valuations arrayed down to zero. Anypositive price then excludes some potential users. If multiple prices are allowed, however,then charging each consumer just slightly below what he is willing to pay can achieve
optimal results. This is known as perfect price discrimination.
1 1 . The only way to imagine these products being produced at this ideal level is with the
imaginary cases of 'perfect price discrimination' or a perfectly omniscient governmentintervening in the market.
12. Do not confuse the claim that markets do not produce the ideal output with the claim
that markets do not produce the efficient output. The efficient output is not the ideal
output if the ideal output cannot be achieved, as Harold Demsetz reminded us several
decades ago in his classic article (1969).
13. In a recent paper, I examined two natural experiments regarding the impact of radio onrecord sales - the introduction of radio in the US and the introduction of commercialradio in England. In neither instance is there any evidence of a positive effect of radio
play on record sales. If anything, the results tend to go the opposite way. Imagine, for
MP3s and copyright collectives 59
example, that there was no radio. The only way to listen to music in automobiles would
be to listen to prerecorded music, which would certainly increase record sales. See
Liebowitz (2004b). For a detailed history of payola that discusses attempts by the record
industry to outlaw it, see Coase (1979).
14. Payola exists, on the other hand, to try and shape which songs are 'hits' and to increase
record sales, not to increase airtime per se to generate additional performing rights pay-ments.
1 5. For a fuller treatment of this issue see Liebowitz (2002).
16. Although the same linkage had existed in the television performing right tariff, in most
countries and most time periods, that linkage was broken in the US when the Southern
District Court of New York instituted a rule that tied royalty payments to a combin-
ation of changes in the inflation rate and the number of broadcasters, a decision that
threatens to break any linkage between economic value and royalty payments.17. Some percentage of bandwidth will be devoted to music downloads, and ISPs wouldn't
want to lose customers due to having too little music. But the losses to ISPs from lost
customers are likely to be small relative to the payments made in the tariff since most cus-
tomers have multiple uses of the Internet and the number likely to base decisions on the
availability of music is likely to be small.
REFERENCES
Coase, Ronald (1979), 'Payola in radio and television broadcasting', Journal ofLawand Economics, 22, 269-328.
Demsetz, Harold (1969), 'Information and efficiency: Another viewpoint', Journal
ofLaw and Economics, 12, 1-22.
Fisher, W.W. (2003), 'Promises to Keep: Technology, Law, and the Future of
Entertainment', draft.
Ku, Raymond (2002), 'The creative destruction of copyright: Napster and the neweconomics of digital technology', University of Chicago Law Review, 69,
263-324.
Lessig, Lawrence (2001), The Future of Ideas, New York: Vintage Books.
Liebowitz, Stan J. (2002), 'Mission impossible: Determining the value of copyright'in Ysolde Gendreau (ed.), Copyright: Administrative Institutions, CopyrightAdministrative Institutions: Conference Organized by the Centre de recherche
en droit public (CROP) of the Faculty of Law of the Universite de Montreal,11 and 12 October 2001, (Cowansville, Quebec: Les Editions Yvon Blais) 2002,
77-100.
Liebowitz, Stan J. (2004a), 'Will MP3 downloads annihilate the record industry?The evidence so far', Advances in the Study of Entrepreneurship, Innovation, andEconomic Growth, 15, 229-60.
Liebowitz, Stan J. (2004b), 'The elusive symbiosis: The impact of radio on the
record industry', The Review of Economic Research on Copyright Issues, 1,
93-118.
Nadel, Mark S. (2002), 'Questioning the economic justification for (and thus
constitutionality of) copyright law's prohibition against unauthorized copying:
106,'mimeo.
Netanel, Neil W. (2002), 'Impose a noncommercial use levy to allow free P2P file-
swapping and remixing', working paper.
Plant, Arnold (1934), 'The economic aspects of copyright in books', Economica,
May, 167-95.
4. Peer-to-peer, piracy and the
copyright law: implications for
consumers and artists
Anne Duchene and Patrick Waelbroeck
4.1 INTRODUCTION
Many voices in the music industry claim that Internet piracy has reduced
sales of original CDs and that illegal MP3 files have become a substitute to
legal CD purchases. Different from for-profit piracy, end-user piracy seems
to be much more difficult to control. The industry and policymakers have
addressed the issue by reinforcing the copyright law, by implementing tech-
nological protection and by actively enforcing legal protection through
lawsuits targeted at developers and users of peer-to-peer (P2P) networks.
On the other side of the field, advocates of online distribution technologies
have argued that MP3 downloads offer a new way for consumers to try out
new music, a typical experience good, in order to make more informed pur-
chases, which could eventually boost CD sales. The goal of this chapter is
to analyse the effect of increasing copyright protection on the pricing and
protection strategies of the firms and its consequence on consumer surplus.
We consider two distribution technologies that differ as to how consumers
acquire information on new music. Before the formal presentation of the
model, we start by an overview of legal and technological aspects of digital
music distribution and a non-technical summary of our main findings.
4.2 TECHNOLOGICAL AND LEGAL ASPECTS OFINTERNET PIRACY
Digital Music Distribution
Two technologies are currently used to download music files and to listen
to music online without the authorization of copyright owners: file-sharing
technologies and non-authorized audio-streaming technologies.
60
Peer-to-peer, piracy and the copyright law 6 1
The principle of filesharing technologies is very simple. Users run the
search engine of the software, looking for specific files. Typically, a user
types the name of an artist or the title of a particular album or song. In the
second step, the software returns 'file results' found on computers con-
nected to the file-swapping network at the time of the search. In the last
step, the user proceeds to download files directly from other users sharing
the relevant files. Note that filesharing consists not only in downloading but
also in uploading files. While downloading is always active, uploading can
be passive. Indeed, downloaded files are by default on the sharing list and
can thus be automatically uploaded (unless specified otherwise by the
user).1 In addition, users can transfer songs from an original CD into
digital format and upload them in a similar way.
Most P2P technologies include a priority rating system that provides
information about the material shared by users. For instance, the Kazaa
priority rating is a measure of how many megabytes have been uploaded
compared to how many megabytes have been downloaded over a given
period of time. It is clear that such a system benefits users who share large
popular files such as recent theatrical and pornographic movies.
Along with P2P, the Internet also gave birth to audio streaming. Audio
streaming can take two forms. On the one hand, Internet radio stations
owned by sites independent from major technological distribution com-
panies broadcast licensed music. On the other hand, specific streaming
technologies owned by large software and content providers (such as
Microsoft, Apple, AOL) have also obtained license to broadcast music
from copyright owners. Indeed, the Digital Millennium Copyright Act
requires web casters and commercial broadcasters to pay licensee fees:
these fees are set to 0.07 cents per performance with a minimum of $500 a
year; fees are collected by the Royalty Panel (CARP).2
Music as a Digital Product
Music files can be compressed without losing much information or quality,
so digital copies have a technical quality similar to the original. Digital
music files can be assimilated to public goods as they are non-rival and
non-excludable. However, the original digital product is often bundled
with other non-digital components such as a printed booklet (with lyrics,
pictures, song and artist information, etc.) and a CD case for music CDs.
Recent digital technological protection tools that we will discuss later on
can also make digital music files close to private consumption goods.
Finally, music is a typical experience good. When facing new music, con-
sumers like to spend time getting recommendations from music magazines,
listening to music on the radio or in a record store before making their
62 Developments in the economics of copyright
purchase decision. The traditional way for record companies to provideinformation to consumers on the existence and style of new CDs and
artists is to spend large costs on advertising and promotion. However,after the Napster experience, it has become clear that there is a cheaper
way for consumers to obtain this information: by searching, downloadingand testing digital music files made available through P2P or other file-
swapping technologies. Acquiring information is costly and time consum-
ing: downloaded files could differ from the one users expected, mainlybecause the filename has changed; the downloaded file can be badly
encrypted or there can be a download error, and this can be found out
only once the file is downloaded, which makes users waste time. As a
result, copying and use of P2P is limited to consumers with a low oppor-
tunity cost of spending time online, i.e. mainly teenagers and college stu-
dents. To sum up, we consider traditional marketing and promotions as
an information-push technology, and new filesharing networks (and to a
lesser extent audio-streaming), as information-pull technologies, where
consumers, not firms, spend time and resources. 3
The Law
The Copyright Act aims at protecting authors of 'original works of
authorship', including literary, dramatic, musical and artistic works. The
protection is available for published and unpublished works.
The copyright law includes several exemptions to copyright infringe-
ment. The most ambiguous is 'fair use' in the digital era. Four elements
have to be balanced to determine whether an activity lies within fair use:
the purpose of the use; the nature of the work being used; the amountof the work used; the effect of the use on the market for or value of the
original work.
Unlike traditional means of copying, file-sharing technologies provide a
large-scale diffusion channel that is virtually impossible to monitor, as a
single copy can be downloaded by any user across the world. So far, public
policies have addressed this new form of end-user piracy by strengthening
existing copyright laws, as illustrated by the 1 998 US Digital Millennium
Copyright Act (DMCA).4
Along with the strengthening of copyright laws has come a series of
legal actions aimed at shutting down filesharing technologies. None of
those technologies has come under more scrutiny than Napster, an online
operator which allowed registered users to exchange indexed music files in a
compressed format (MP3), and to do so freely and anonymously. Followinga lawsuit filed by the Record Industry Association of America (RIAA),
Napster was found guilty of copyright infringement. The court found that
Peer-to-peer, piracy and the copyright law 63
the use of Napster harms the music industry on two grounds: reduced sales
of CDs, and heightened barriers to entry by the music industry in the online
distribution. Napster's argument according to which sampling leads to
higher CD purchases was supported by an empirical study, but ruled out
as flawed and non-objective by the court. After monitoring activities over
new filesharing networks, the Recording Industry Association of AmericaRIAA filed 261 lawsuits charging music uploaders with copyright infringe-
ment in September 2003. These legal actions are also based on the argu-ment that music downloads are causing substantial damage to the music
industry.5 Under the current legislation, any user of P2P risks legal attacks.
The current reinforcement of legal protection against Internet piracyincreases the perception of that risk. Hence, P2P users have to take into
account the cost of being caught uploading music files. This legal cost can
be represented by the expected penalty of being caught copying.
Technological Protection
The DMCA also enabled 'digital rights management' (DRM), a small
piece of software that can detect, monitor and block (unauthorized) use of
copyrighted material. There are two types of existing DRM technologies:
watermarking (digital identification inserted in digital files) and finger-
printing (converts the files content into a unique identification number).
Despite a promising future, companies that create such technological pro-tection tools struggle to survive. 6
There are several reasons behind this failure. First, some legally pur-chased copy-protected CDs failed to play on some computers and audio
devices. Second, it is almost impossible to find protection devices for such
mass-consumption products that cannot be hacked. Third, there is the issue
of DRM standards. Finally, DRM software threatens the privacy of users.
Samuelson (2003) argues the DRM goes beyond the Copyright Act.
Indeed, DRM can protect any digital content even if it is not protected bythe copyright law such as documents in the public domain. It reduces the
value of fair use and can force consumers to view content that they do not
wish to (such as ads and FBI warnings). Because of such restrictions, DRMsometimes stands for 'digital restrictions management'. Moreover, it can
potentially protect over an infinite amount of time, which is contrary to
the spirit of the Copyright Act. In a way, DRM creates the basis for a
perpetual payment system.
Another controversial aspect of DMCA is the anti-circumvention rule
that prohibits users bypassing technological measures of protection. Manypeople involved in the copyright law have argued that DMCA and its
corollary DRM fail to strike a balance between the benefits of developers
64 Developments in the economics of copyright
and users and would, in addition, impede artistic and scientific progress.
Liebowitz (2002) challenges that view. Because DRM allows some form of
first-price discrimination, he argues, the use will not be reduced (however,
his analysis fails to notice that transformative and creative uses that are
authorized under fair use could be greatly reduced by DRM). Moreover,
first-price discrimination leaves no surplus to consumers and is question-
able from a social point of view if the 'intensive' music listener is a low-
income individual who would be charged more than, say, a wealthy
individual who only listens to music occasionally. Liebowitz (2002) also
challenges the view that DRM will slow 'sequential innovations' that rely
on previous work (as argued by Landes and Posner 1989). Here he argues
that DRM will not hinder such innovation because copyright only protects
the expression of ideas and not the ideas themselves.
In addition to these arguments, we would like to discuss three other
important economic dimensions of DRM. First, because DRM can be
implemented in the hardware, at the operating software level and also at
the player level, each of which are all provided by different firms, the issue
of setting standards and making sure that all platforms are compatiblecannot be neglected. For that reason, some industry observers say that
DRM stands for 'down-right messy'. Second, technological protection
software and its inventors create an additional layer between artists and
consumers, introducing a new source of vertical inefficiency. Third, if
DRM is legally mandated in players and software, this could raise the
barrier of entry in the distribution and the software protection markets
and increase market power.
4.3 A SUMMARY OF THE LITERATURE
Despite the technological breakthrough brought by filesharing technolo-
gies, the debate on the welfare implications of piracy goes a long way back
in the economic literature on unauthorized copies of copyrighted material,
reviewed by Peitz and Waelbroeck (2003 a). Instead of duplicating the
review here, we would like to discuss which arguments fit or do not fit the
characteristics of the music industry. When a copyright owner can monitor
the quantity of copies likely to result from the purchase of original mater-
ial, he or she can indirectly appropriate revenues by charging a higher price
for the original (see Liebowitz 1985; Besen and Kirby 1989; and Bakos et al.
1999). The first argument is related to the pricing of a club membershipand the nature of the cost to copy. The second argument is related to the
literature on bundling and how club formations can reduce the variance of
the demand of the club as a whole compared to individual demands.
Peer-to-peer, piracy and the copyright law 65
Potentially, indirect appropriation could arise if users of filesharing tech-
nologies were be ready to pay a premium to purchase the original version
of a popular hit song in order to improve their priority rating when they
upload the files on a P2P network (for a discussion, see Liebowitz 2002).
However, both arguments are unlikely to play a key role in the case of P2P
technologies since it is extremely difficult to monitor file exchanges and
only a minority of P2P users share files. Fair use policy can involve a
tradeoff between the number of works supplied and the number of uses of
each work. Specifically, expanding fair use could reduce the number of
markets served by copyrighted works. In principle, expansion could be
more costly on both efficiency-based and non-economic policy standards.
For example, increasing the fraction of uses regarded as fair could reduce
the number of works produced with little effect on number of uses per
work, reducing total uses of all works. Such a change would move the
outcomes farther away from both overall economic efficiency and non-
economic objectives relating to volume of use of copyrighted works.
Most of the time, the copy is of lower quality than the original and
product differentiation in many models imply that the increases in con-
sumers' surplus more than compensate the static losses of producers. This
argument is easily understood since the ex-post welfare-maximizing price
is equal to the marginal cost, which is assumed to be zero. However, in
a long-term perspective, such profit loss will result in less incentive to
provide quality on the market (an important contribution to this idea
should be credited to Novos and Waldman, 1984). For the reasons that wehave already mentioned, MP3 files arguably have a lower expected value
than an original CD so that some elements of product differentiation
should be part of the debate on Internet music piracy.
In some cases, positive network externalities generated by copiescan benefit copyright owners as shown by Conner and Rumelt (1991),
Takeyama (1994) and Shy and Thisse (1999). There is a case for weaknetwork effects in music consumption if users value the number of people
listening to the same music. These social network effects can result from the
fact that consumers want to belong to a community or be able to talk about
music in social gatherings. In principle, network effects could depend onboth the number of originals and copies. There is another rationale for the
existence of network effects among copiers using filesharing technologies:
the speed of downloading music files grows along with the size of the
network. Such network effects are endogenous, since users of filesharing
technologies can decide whether or not to share their files. This decision
typically depends on the expected gain in the user's rating (making it easier
for him to download new songs in the future) and the cost of sharing related
to the probability of being caught by legal authorities or receiving a virus
66 Developments in the economics of copyright
and the loss of computer resources. However, if there is only a small
number of users sharing a large number of files, the extent of network
effects will be limited (for a discussion see Peitz and Waelbroeck, 2004b).
Finally, digital copies can provide information on the value of original
products that can be assimilated to an experience good. Takeyama (2003)
shows how copies with information on the characteristics of a durable goodcan solve adverse selection problems. Peitz and Waelbroeck (2003b) show
how a multi-product firm can benefit from better matching consumers to
their ideal products due to better sampling on P2P networks, despite the
negative competition effect due to the availability of digital copies. Webelieve that arguments based on the informational role of copies make a lot
of sense for music, which is a typical experience good, and we will use this
argument in our model.
4.4 NON-TECHNICAL OVERVIEW OF THE MODEL
We propose a simple model of the music industry to analyse welfare impli-
cations of an increase in copyright protection against end-user piracy when
firms can implement technological protection. The model includes three
original and relevant features.
First, we consider two means of intellectual property protection: legal
protection (set by public authorities) and technological protection (imple-
mented by copyright owners). Copyright owners can implement costly
technological protection (DRM) on original material in order to prevent
their duplication (as MP3 files).7However, technological protection alone
cannot eliminate online copying since it is always possible for professional
hackers to crack software/hardware protection, plus anti-protection devices
would be quickly available on P2P networks. Therefore, active enforcement
of copyright laws is also required. Legal protection consists of legal pro-
ceedings against copiers who must internalize the risk of being caught and
the resulting penalty.8
Second, music is an experience good. In addition to the traditional wayof targeting consumers with promotions and ads, we assume that P2P
technologies make it possible for some firms to enter the market at a low
distribution cost, as consumers bear the cost of acquiring information in
lieu of record companies.9 In short, we view traditional distribution as
an information-push technology in which the firm pays to provide infor-
mation to consumers and P2P as an information-pull technology where
consumers spend resources to acquire information on products they have a
potential interest in. Although this is a very simplistic representation of
music distribution, such a conceptual distinction allows us to analyze the
Peer-to-peer, piracy and the copyright law 67
welfare implications of an increase in copyright protection according to
the information transmission technology.
Finally, we assume that a digital copy is less valuable than equivalent
original material, but at a cost that depends on the users' disutility of using
computers and the Internet to download and burn files. On the one hand,
we assume that a digital music file available on a P2P network has a lower
expected value than original material. Some users view digital copies as
poor substitutes to original CDs, mainly because they greatly value lyrics,
pictures and other information such as songwriters and album title. On the
other hand, because the amount of entertainment time is inelastic, we also
assume that consumers have different opportunity costs of spending time
looking for and downloading music files over P2P networks.
Our model contributes to the literature on welfare implications of
illegal copies on the following points. First, we endogenously assess the
level of technological protection chosen by firms, and we emphasize its
connection with legal protection. We find that increasing copyright pro-
tection gives record companies using the information-push technologyincentives to raise the level of technological protection, which increases
their profits. It is detrimental to firms using information-pull technologies.
Second, we show that stronger legal actions against copiers lead to a
smaller consumers' surplus due to the direct negative effect on copiers
and the indirect negative effect on buyers of original products through the
increase in technological protection and prices. Finally, we illustrate the
fact that an increase in copyright protection may decrease profits at
the industry level when there is a large proportion of firms producingsmall-audience music.
4.5 A SIMPLE MODEL OF INTERNET PIRACYWITH AN INCREASE IN COPYRIGHTPROTECTION
We analyze the role of an increase in copyright protection when onlytwo different and extreme forms of distribution technologies are available.
These technologies differ in the ways consumers acquire information.
Information transmission technologies. On the one hand, the information-
push technology allows firms to sell original versions of their records (with
their full set of features: CD box, lyrics, pictures . . .) at a large fixed cost
of marketing and promotions noted by K, which, for simplicity, we assume
to give information about the existence and the characteristics of the
product to all consumers. On the other hand, the information-pull (P2P)
technology allows consumers to search and test digital copies that we
68 Developments in the economics of copyright
assimilate to downgraded versions of originals (with a limited set of fea-
tures). Thus the P2P technology allows firms to freely distribute their prod-
ucts online but only to informed consumers, i.e. to those who have
downloaded digital copies. In other words, only consumers with a low
opportunity cost of spending time online searching, downloading and
testing digital files can become potential buyers. With both distribution
technologies, the fixed cost of production and the marginal distribution
cost of the original are normalized to zero. 10
Technological and legal protection. Regardless of the distribution tech-
nology, there is online end-user piracy, the amount of which depends on
the extent of technological protection and legal enforcement of copyright
protection. The producer can spend resources on technological protection
of originals in order to make digital copies of his or her product more
difficult to create and to find online. 11Technological protection, a, can be
implemented by the firm at an increasing and convex cost function c(a).
In addition to technological protection, original products can be legally
protected. This protection is modeled by legal fines, cp, that are determined
by the enforcement of laws such as the DMCA. Thus, 9 represents the
additional disutility of copying. In other words, <p is the expected cost of
being caught downloading unauthorized digital copies.
Music experience. Consumers incur an opportunity cost s of spendingtime searching and downloading digital music files using the P2P technol-
ogy. We assume that s is uniformly distributed on the segment [0, s\ of
mass 1 . That opportunity cost not only reflects a direct cost of spendingtime on the Internet (related to the user's level of Internet sophistication),
but it also reflects an indirect cost related to the financial value of what
the user gives up to spend time online (his or her wage or other leisure
activities, for instance). When listening to music (through promotions or
through copying), consumers can either enjoy it or not. These states of
nature occur with probability p (the consumer enjoys the music) and 1-
p
(the consumer does not like the music) respectively (with O^p^l).Copiers who like what they have listened to obtain a utility represented bythe function v(l a), whereas buyers of the original product obtain a
utility v>v(l a) for all 0<a<l. The differential between copies and
originals, av, represents the value of the original over the copy, such as
lyrics, booklet, pictures, song information, and so on. 12
Consumers who do not like the music get utility, regardless of whether
they learned the existence and the characteristics of the music through
advertising or through a digital copy, and regardless of the quality differ-
ential between copies and originals. The implicit assumption behind the
latter requirement is that consumers do not care about additional features
of music they do not like.
Peer-to-peer, piracy and the copyright law 69
To sum up, the total expected cost of copying is s + 9. It is worth notingthat this cost could have taken the form of 5(1 + <p) (as in Novos and
Waldman, 1984). However, we do not consider that copiers with a higher
downloading cost s should suffer more from legal protection than others.
Thus, we assume that legal protection creates an additional burden that is
identical for all copiers.
Consumers' expected utility is defined by u(x, y), where xE {0, 1} is the
decision to purchase the original (x=
1) or not (x=
0), and yE. {0, 1 } is the
decision to download the digital copy (y=
1) or not (y=
0). We assume that
11(0,0) =0.
Information-push Technology
In this section, the firm uses the offline distribution technology with the
fixed marketing and advertising cost K. 13 We assume that advertisement
reaches all consumers who are thus aware of the existence and the char-
acteristics of the product. Hence, consumers take their purchasing deci-
sions in a perfect information environment. Consumers who like the
music (in proportion p) can purchase the original or copy the down-
graded version.
The utility of purchasing the original version is w(l,0)=
v-/?, and
the utility of downloading a digital copy is w(0,l)=
v(l -a) -s- 9. Let
s = p-q>-av denote the opportunity cost of the indifferent consumer
between the original and the copy. Consumers purchase the original if it
provides a higher utility than copying or not purchasing at all:
If the first inequality holds, consumers who like the product (in propor-tion p) purchase it if their opportunity cost of using the P2P technology is
such that s > s. This is illustrated in Figure 4. 1 .
When the price is very low, all consumers who like the product (in pro-
portion p) purchase it. This occurs when s < 0. Similarly, when the price
s s
Figure 4. 1 Consumer demand with information-push technology
70 Developments in the economics of copyright
is very high, no consumer purchases (s> s or/?
>v). Therefore, the firm
faces the following demand:
d(p, a)= p^^
S
ifp> v (and d(p,oi)=
otherwise), and maximizes its profit function to set
the optimal level of copyright protection a and the optimal price/?:
s sMax IT = pp ---
c(a) K<*./> s
s.t.p< v.
The optimal price/?* and technological protection a* are such that
l>
. andz* z,s
As a result, a* is increasing in <p:
da*- > O.14
A reinforcement of legal protection (as a consequence of the DMCA for
instance) thus makes firms increase technological protection. The intuition
is the following. Increasing copyright protection decreases the number of
copiers because buying becomes more advantageous. Firms can increase
prices somewhat to capture more surplus from buyers, but they may want
to also increase technological protection to prevent consumers at the
margin switching to the copy because of a higher price.
Moreover, the demand for originals and the optimal price are increasing
functions of copyright protection cp.15Proposition 4.1 describes the impact
of a higher legal protection on a firm's profit and consumers' surplus:
Proposition 4.1 Increasing legal protection cp:
(1) Increases the firm's optimal profit IT*;
(2) Decreases consumers' surplus.
Proof
(1) Applying the envelope theorem shows that
Peer-to-peer, piracy and the copyright law 71
(2) First, consumers who still copy the product after an increase
in 9 have a lower utility. Indeed, (0, 1)=
v(l -a) -5 -9. Since
da*/d(p>
0, v(l -a) decreases as 9 increases, so that w(0, 1) also
decreases. Second, the utility of consumers who still purchase the
product after an increase in 9 decreases because of the higher price.
Third, consumers who switch from the copy to the original have a
lower utility. Previously they had a higher utility of copying. After
9 increases, the utility of copying and buying decreases, and theyswitch to buying. Thus, they are worse off.
Proposition 4. 1 states that a reinforcement of legal protection (as a con-
sequence of the Digital Millennium Copyright Act, for instance) makesfirms increase technological protection. Firms who distribute their prod-ucts using the information-push technology benefit from a stronger legal
protection. However, the welfare gains are lowered by the reduction in
consumers' surplus.
Information-pull Technology
In this section, the firm distributes its product using a file-sharing tech-
nology. We assume that consumers are not initially informed about the
characteristics of the product, but they can acquire information by down-
loading digital copies. Consumers who have copied the product and
enjoyed the music have the possibility to directly purchase the original
from the firm. The expected utility of downloading the digital copy is
w(0, l)=
pv(l -a) -s- 9 (the probability of liking the music being p).16
The consumer who is indifferent between downloading a copy or not is
such that H(0,l)=
w(0,0)<=>,s=
pv(l -a) -9. Hence, only consumers
with a low search cost and who like the music can become potential
buyers. On the one hand, copiers who did not like the music do not pur-chase the original version, as their utility of purchasing the original
product would be-/?-s-<p<0. On the other hand, consumers whohave copied the product (of mass pv(l
-a)-9 /s) and who liked it (in pro-
portion p) have an expected utility w(l, 1)=
p(v(l-a) + va -p)
- s - 9.17
It
is clear that copiers who like the music will purchase the original providedthat /?< av. Indeed, copiers who like what they listened to have an extra
utility av-p of purchasing the original instead of just copying. This is
illustrated in Figure 4.2.
When legal protection is very strong, there are no copiers, and therefore
no potential demand for originals. Even with a strictly positive potential
demand, no consumer purchases when the price is very high. Therefore, the
firm faces the following demand:
72 Developments in the economics of copyright
s s
Figure 4.2 Consumer demand with information-pull technology
if /?<av (and d(p,a)=
otherwise). This leads to the following optimiza-
tion program:
f .Max TT = --c(a)
a,P 5
s.t.p<av and <p<
pv(l a).
Here the level of technological protection, a, is a decreasing function of
copyright protection cp.18
Increasing copyright protection reduces the
potential demand for original products distributed online, as potential
buyers need to download a digital copy first and this activity is more costly
when copyright protection increases. In order to compensate this loss, it is
optimal to reduce the level of technological protection. Overall, profits also
decrease and so does buyers' surplus.
Proposition 4.2 An increase in copyright protection:
( 1 ) Decreases the profits of firms distributing their products using P2P;
(2) Decreases the social loss due to the downloading cost;
(3) Decreases buyers' surplus.
Proof
(1) Applying the envelope theorem shows that
dix* ova
~T~= ~--'
fl(p 5
(2) We first show that
S^ *a
acp
Peer-to-peer, piracy and the copyright law 73
Straightforward calculations yield:
ds da v/5
Substituting the expressions found in note 19 and rearranging
shows that ds/dy < 0. A reinforcement of the copyright law from
cpjto cp2 (with <p2 >(pj) makes a proportion sfo^ s(<p2) stop copy-
ing. There is a reduction of the number of copiers, which means that
there is a smaller proportion of copiers who waste resources on
copying without enjoying the product (1-
pX-stoj)-
.?(<P2)].
(3) The surplus of copiers who become purchasers is given by
/6= <;(pv[l-a'
o
After rearranging, we obtain that. ;-
-<x*(p)]
-J -pv
-
The expression in between brackets is negative. Thus Ubis decreas-
ing in <p.
The intuition behind Proposition 4.2 is that an increase in legal pro-
tection reduces the demand for copies and thus the potential demand for
originals. Since the firm extracts all buyers' surplus, consumers only obtain
the surplus from the copy.
The main difference with the previous distribution technology is that
there is always both a benefit and a cost of copyright protection on con-
sumers' surplus. The reason is that increasing 9 reduces the number of
copiers and thus both the number of copiers who waste resources on down-
loading music they do not like and the number of copiers who become
buyers of the original product.
74 Developments in the economics of copyright
Application: Effect of an Increase in Copyright Protection on the
Profits of the Industry
In this section we assume that C(OL)= ca2
/2.
When the firm distributes its product with the information-push technol-
ogy, it sets a price
+ _ cs(y + s)
2cs v2p
and a level of technological protection
2cs v2p
(the denominator is positive by the second order condition). This yields a
profit of
cp(9 + s)2
1 4cs 2v2p
using the information-push technology.
When the firm uses the information-pull technology, it sets a price
p* = a*v and a level of technological protection
^ = vp(vp-
9)
It is easy to check that the optimal profit is
~<P)
2
22s(cs + 2v2p2)<
In Figures 4.3 and 4.4, we compare the profits using both distribution tech-
nologies for two different legal environments (<p= 0.15 and <p
=0.30). The
other parameters are set to s = 1,v = 1
,K= 0.45, c = 1 .
19 The flattest curve
corresponds to the profit obtained by using information-pull technology,
and the steepest curve to the profits of a firm using the information-push
technology. The trade-off between the two distribution technologies lies in
the potential audience of the music, p.
For any given values of <p, music with a low potential audience p is
not distributed, and firms with an average potential audience choose the
information-pull technology, although they could have used the other.
0.2-
0.1-
Peer-to-peer, piracy and the copyright law 75
Information-push
Information-pull
0.2 0.4S 0.6 0.8 1
-0.1 -
-0.2-
Potential audience
Figure 4. 3 Application with cp= 0. 75
0.6i
0.5 -
0.4-
0.2-
0.1 -
Information-push
Information-pull
-0.10.6 0.8 1
Potential audience
Figure 4.4 Application with q>= 0.30
For instance, in Figure 4.3, firms with 0.8<p<0.9 use the information-
pull technology although the traditional distribution technology was prof-
itable. Finally, firms producing music with a large audience prefer the
information-push technology.
Increasing <p from 0.15 to 0.3 has several noticeable effects (see Figure
4.4). First, the range of p that supports the information-push distribution
technology increases. Second, more firms prefer the information-push tech-
nology. Third, the minimum audience which makes it profitable to enter
the market with the information-pull technology increases too. Thus there
are fewer firms in the industry and a more important proportion chooses
76 Developments in the economics of copyright
the information-push technology. The aggregate effect of an increase in
copyright protection at the industry level depends on the distribution of
audience p. When there is a high proportion of superstars, an increase in
copyright protection leads to an increase of industry profits. However,when there are proportionally more small-audience artists, industry profits
might fall.
4.6 CONTRIBUTION, AGENDA AND CONCLUSION
We have argued that information-pull technologies offer a possibility of
entry through cheap distribution channels, and open the market to newartists and small distribution companies. We have shown that a stronger
legal protection can shift too much the balance between consumers and
producers in favor of producers of large audience music. In a way, these pro-
ducers could use legal protection as a foreclosure tool against new products.
However it is foreseeable that as a growing number of new productsbecomes available, consumers will ignore the existence of one particular
product in which they could have a potential interest. We believe that newdistribution technologies could open the market for informational inter-
mediaries. Indeed, in our model, both distribution technologies are
inefficient at transmitting relevant information. On the one hand, produc-ers who use information-push technologies waste the fixed cost of adver-
tisement and promotion on reaching consumers with no interest in their
music. On the other hand, some P2P users waste resources on download-
ing music files they finally do not like. Informational intermediaries could
reduce these inefficiencies, by promoting and recommending new products
(a system already available on Amazon.com and similar Internet sites), tar-
geting only consumers who have the highest likelihood of purchasing the
original. Some of the new emerging business models are based on cross-
platforms, like Kazaa for instance, who work together with record compa-nies as a new vehicle to reach mass markets. As P2P plays the part of an
advertiser, record companies are becoming simple distributors. MP3.comhas been using their data on consumers to provide record companies infor-
mation and advice on their marketing strategies.
Other business models improve the quality of the original by offering a
number of features that could not appear on the copy: a booklet contain-
ing lyrics, pictures, or an access code to online chat rooms, forums, making-
offs, additional music and a discount on live performances, etc. Indeed, the
possibility of offering an Internet connection with login and password to a
site with additional features for purchasers of original products reduces
the value of a digital copy. In fact, the emerging new business models are
Peer-to-peer, piracy and the copyright law 77
beginning to take into account other assets than just the music. The mar-
ketplace is thus changing from a commodity market to a service one, as
illustrated by MP3.com, who offer complementary services to the music,
which represents additional value for users.
ACKNOWLEDGMENT
We wish to thank B. Caillaud, G. Llobet, P. Pereira, J. Pouyet, M. Bourreau, and session
participants of the 2nd ZEW Conference on the Economics of ITs in Mannheim and of the
2nd Annual Conference of the Society for Economic Research on Copyright Issues,
Northampton, MA, of the 29th Annual Conference of the European Association for
Research in Industrial Economics (EARIE), Madrid and of 18th Annual Congress of the
European Economic Association (EEA), Stockholm.
NOTES
1 . For instance, eMule automatically shares files that are being downloaded.2. According to an Arbitron/Edison Media Research study, over 35 per cent of American
Internet users aged 12 and older were 'streamers' in July 2002. However, the trend is flat
mainly because of the shutting down of small webcasters who were not able to pay roy-alties to copyright owners.
3. Recent survey data show that a large share of Internet users download files for sampling(69 per cent of downloaders listen to new music and 31 per cent listen to music by artists
they never heard before according to PEW Internet tracking and IPSOS-INSIGHTreports). Moreover, a survey of IPSOS of 2002 finds that 73 per cent of music down-loaders 'like being able to sample music online before making a purchase decision'.
Another survey of IPSOS of 2002 finds that 30 per cent of Internet users have changedthe genre of music they typically listened to, mainly because they were able to experi-ment with different forms of music. See Peitz and Waelbroeck (2004b). We believe that
these numbers understate the potential of file-sharing systems for sampling purposesbecause cross-recommendations and profiles of downloaders could be greatly improved.
4. The DMCA makes it a crime to circumvent anti-piracy measures built into most com-mercial software (except for research purposes, non-profit libraries, etc.). While the
DMCA limits liability of copyright infringement of Internet service providers and insti-
tutions of higher education, it requires webcasters and commercial broadcasters to paylicensee fees. Finally, the DMCA does not affect conditions of copyright infringements,
including fair use. The DMCA lays the legal foundation of pay per use even for mater-
ial that is no longer protected by copyright and so on a perpetual basis.
5. This argument is only partly validated by cross-country data analyzed by Peitz andWaelbroeck (2004a) who find that while music downloads may have reduced music sales
by as much 20 per cent in the early days of P2P networks, other factors are also playingan important role.
6. MIT Technology Review (March 2002) comments on the struggles of ContentGuardand TrustTechnologies, two important producers of DRM solutions.
7. DRM alone is not sufficient per se to secure digital content, and has to be coupled with soft-
ware or hardware solutions. Eckersley (2003), citing sources from IBM's Thomas J. WatsonResearch Center and extrapolating from the cost of a 4758 cryptographic co-processor,estimates the current cost of a complete security device around US$4000 per computer.Further R&D could reduce this figure to US$1 00- 150 on a large-scale production basis
78 Developments in the economics of copyright
and possibly to US$1 0-20 if it is possible to develop a single-chip device. In any case, the
cost of developing secure technological protection is far from being negligible.
8. Our model also applies to audio streaming where we analyze the effect of increasing roy-
alties paid by webcasters. Fewer webcasters means less music variety and less consumer's
surplus.
9. When Napster created its 'New Artist program', which enabled starting bands to set upa hyperlink to a merchant site selling their original product, more than 10000 profiles
were submitted within the first two weeks (see The Standard, 19 June, 2000).
10. Even with a strictly positive cost of creation, our results would not be affected, since we
compare profits between technologies and between different legal environments, which
do not affect this cost. Moreover, the marginal cost of distributing information goods is
negligible compared to the fixed marketing cost.
11. In another more elaborated version, we allow technological protection to decrease the
value of the original as well, through a reduction of its fair use value. The results pre-
sented in this simpler version are qualitatively similar.
12. This parametric example corresponds to the specification used by Besen and Kirby
(1989). Note however that in their model a is exogenous, while this parameter will be a
choice variable in our model.
1 3. For simplicity, we do not make a distinction between the artist and his label. The implicit
assumption is that it is possible to write a contract that eliminates vertical inefficiencies.
14. Applying the implicit functions theorem to
we get
da _ pv/2s~
Since the denominator is negative by the concavity of the profit function, da./d(f>Q.
d<f 2\ d(f> I
'**
dcp 2\*
d<p
16. The process of purchasing the original requires the availability of a connection to an
Internet site or file-sharing software, which then proposes a connection to a merchant
site. We thus assume that consumers cannot buy random products for which the fixed
costs of marketing and promotion K have not been spent. In other words, productswhich have not been distributed traditionally are not available in brick-and-mortar
stores and can only be discovered through a time-consuming online search process.
17. It is worth noting that in this section a consumer might have both versions (copy and
original) of the same product, contrary to the last section. This is due to the fact that the
informational structures are different. Since consumers are initially uninformed about
the product, they first have to test the product before knowing whether they like it or not.
Thus their actions are sequential. Moreover, since they are initially uninformed, they do
not know if the music that they download is copyrighted or not. For that reason, they
still anticipate an expected fee, although as we will show, artists using information-pull
technologies have no incentives to sue them.
1 8. At the optimum, the effect of increasing a on profit is given by
F(a,<p)= =
-_ [pv(pv(l-a)
-<p)]
-P2av2 + c'(a) .
Applying the implicit function theorem to F(a, cp)=
0, we obtain that
da. _ pv/s
d(f dF/dot
Peer-to-peer, piracy and the copyright law 79
Since the numerator is positive, it remains to show that the denominator is negative. Wehave:
= _ 2v2 -c"(a) < 0. Thus < 0.
da s dq>
19. We could not find any permissible configurations of the parameters that yielded other
qualitative results.
REFERENCES
Bakos, Y, E. Brynjolfsson and D. Lichtman (1999), 'Shared information goods,'Journal ofLaw and Economics, 42 (1), 1 17-55.
Besen, S. and S.N. Kirby (1989), 'Private copying, appropriability, and optimal
copying royalties,' Journal ofLaw and Economics, 32, 255-80.
Conner K.R. and R.P. Rumelt (1991), 'Software piracy- an analysis of protection
strategies,' Management Science, 37 (2), 125-39.
Eckersley, P. (2003), 'The economic evaluation of alternatives to digital copyright,'
working paper.
Landes, W. and R. Posner (1989), 'An economic analysis of copyright law,' Journal
of Legal Studies, 18, 325-63.
Liebowitz, S. (1985), 'Copying and indirect appropriability: Photocopying of jour-nals,' Journal of Political Economy, 93, 945-57.
Liebowitz, S. (2002), Rethinking the NetworkedEconomy, New York: Amacom Press.
Novos, I. and M. Waldman (1984), 'The effects of increased copyright protection:An analytic approach,' Journal of Political Economy, 92, 236-46.
Peitz, M. and P. Waelbroeck (2003a), 'Piracy of digital products: A critical review
of the economics literature,' CESIfo working paper #1071.
Peitz, M. and P. Waelbroeck (2003b), 'Making use of file-sharing in music distribu-
tion,' mimeo.
Peitz, M. and P. Waelbroeck (2004a), 'The effect of Internet piracy on music sales:
Cross-section evidence,' Review of Economic Research on Copyright Issues, 1 (2),
71-9.
Peitz, M. and P. Waelbroeck (2004b), 'An economist's guide to digital music,'
CESIfo working paper #1333.
Shy, O. and J.F. Thisse (1999), 'A strategic approach to software protection,' Journal
of Economics and Management Strategy, 8 (2), 63-190.
Samuelson, P. (2003), 'Digital rights management and, or, vs. the law,' Communi-cations of the ACM, 46 (4), 41-5.
Takeyama, L.N. (1994), The welfare implications of unauthorized reproduction of
intellectual property in the presence of demand network externalities,' Journal
of Industrial Economics, 42 (2), 155-66.
Takeyama, L.N. (2003), 'Piracy, asymmetric information and product quality,' in
Wendy Gordon and Richard Watt (eds), The Economics of Copyright:
Developments in Research and Analysis, Cheltenham, UK and Northampton,MA, USA: Edward Elgar, pp. 55-65.
5. 'Fair use5
as policy instrument
Timothy J. Brennan
5.1 INTRODUCTION
Battles over copyright are typically framed as between providing moreaccess for the public to created works and giving profits to those whocreated the works. To this observer of copyright policy discussions, grant-
ing any weight to the latter side of this tradeoff often seems to come grudg-
ingly. It is as if only a minimal degree of fairness justifies consideration of
creators in the balance at all.
These battles between access rights and exclusion rights take place on
many intellectual property fronts. Seemingly innocuous or obvious busi-
ness practices have been patented, leading to concerns that they will be
unnecessarily monopolized and promote collusion (Ciminello 2000).l
Cases involving MP3 music filesharing and DVD decoding cases suggest a
conflict between listeners and producers regarding access to music andfilms and on broadband deployment.
2 The DVD cases also reflect contro-
versies regarding the Digital Millennium Copyright Act's restrictions onaccess to decryption technology.
3 The Supreme Court ruled against con-
tentions that extending copyright an additional twenty years violates the
Constitutional mandate '[t]o promote the Progress of Science and useful
Arts by securing for limited Times to Authors and Inventors the exclusive
Right to their respective Writings and Discoveries.'4
One facet of these controversies is whether viewing copyrighted worksas special, e.g., as 'cultural' goods, should change copyright doctrines, par-
ticularly to expand 'fair use.' Fair use is defined statutorily as the right to
copy or use copyrighted works in ways that would otherwise lead to liabil-
ity for infringement, based on four factors: purpose and character of
the use, nature of the work, amount of it used, and effect on its marketor value. 5 The operational content of these factors continues to evolve. 6
Policy goals other than economic efficiency, e.g., maximizing access (Brown1985),
7'public knowledge, political debate, or human health' and First
Amendment considerations (Gordon 1982), or relational goals threatened
by commodification (Gordon 2002) might warrant more generous fair use
80
'Fair use'
as policy ins trumen t 8 1
standards or more radical changes in copyright. The National Research
Council's Computer Science and Telecommunications Board (2000,
pp. 137-8) has observed that fair use might be justified by 'fundamental
human rights' of freedom of expression and the press, and 'public interest
grounds' associated with schools, libraries, and the courts. 8
Models below illustrate modifying fair use to promote objectives other
than aggregate economic efficiency. Alternatives include disregarding profits
to copyright holders, recognizing only gross consumer surplus, or counting
uses of copyrighted works without regard to their different economic valu-
ations. We look at fair use when there may be 'diversion' in that some uses,
for which users could pay and could be identified, may nonetheless be
obtained for free. One might be willing to pay for a copy of a CD to carry in
one's car, but one does not need to do so. In the spirit of justifications based
on avoiding transaction costs for low value uses, we also model fair use as
allowing uses for which there is low willingness to pay.9
The motivating question is whether different policy objectives should
significantly affect copyright. The intuition as to why they might not is
as follows. Suppose, to be simplistic, that a valid proxy for looking at the
effects of copyright is the product of the number of works times the
number of uses per work. By any credible policy standard, one would want
to increase both the number of works and uses per work. One would be at
an optimum only if increasing the number of works leads to a decrease in
uses per work or vice versa. If that tradeoff is not much different under
different goals, e.g., economic efficiency vs. maximizing uses, argumentsabout the basis of copyright policy are excessively heated.
Markets for copyright involve too much detail to make precise recom-
mendations regarding fair use, compulsory licensing (e.g., as under the
Audio Home Recording Act of 1992), term limits and breadth (including
reach over derivative works). Consequently, this research is only a step
toward understanding how much policy objectives matter. Still, narrowly
structured, inherently unrealistic models may yet offer insight into how a
change in objective may lead to a change in policy.10 Such limitations may
be appropriate. Lacking more specific empirical justifications, copyright
policy rules are likely to be relatively and appropriately simple, e.g., in not
applying different fair use standards to a highly popular film compared to
one with a small audience.
5.2 PRIOR ECONOMIC ANALYSES
Formal models of the design of copyright have focused primarily on the
extent to which copying should be made more or less costly (Novos
82 Developments in the economics of copyright
and Waldman 1984; Johnson 1985; Besen and Kirby 1989; Yoon 2002).
Analyses of fair use typically attempt to justify it when market failures
impede otherwise desirable uses rather than to assess its effects. The fun-
damental insight, from Gordon (1982, pp. 1614-22), is that some con-
sumers may be willing to pay the prevailing price to obtain or use a
copyrighted work, but not be willing to pay that price plus the cost of the
transactions associated with those uses. 1 1 Brennan (1986) takes a more nar-
rowly economic approach, treating fair use as a compulsory license where
the fee is zero, building on the transactions cost-based analysis in Besen
etal. (1978).
From this standpoint, Klein et al. (2002) claim a copyright holder would
not object to fair use, since its profits would be unaffected. An importantlimitation of Klein, et al.'s insight, however, is that it ignores the other three
statutory factors, beside effect on market value, associated with fair use. 12
Copyright holders and user groups could dispute whether the nature and
use of the work in news or educational settings justify unauthorized
copying, even if there is a non-negligible effect on the work's market value.
Some analyses have taken a more strategic approach. Depoorter and
Parisi (2002) suggest that fair use can limit incentives for strategic with-
holding when copyrighted works are complements.13 In essence, each
copyright holder has an incentive to hold out for the full value of a
package of copyrighted works absent some compulsion to make the work
available. 14 In an article not specifically on fair use, Takeyama (1997) offers
a model in which fair use available to those with low reservation prices
could allow a copyright holder to credibly commit to charge high prices
in a multiperiod setting. Absent fair use, the copyright holder could not
commit not to offer a low price in later periods to low reservation price
users, reducing the amount it could extract from users with high reserva-
tion prices.
A noteworthy implication of arguments based on transaction costs is
that if those costs fall, the rationale for fair use defenses against accus-
ations of infringement also falls. As the movement of creative works
becomes increasingly digitized through the Internet, it becomes increas-
ingly feasible for copyright holders to find and charge for uses that mightnot have been worth identifying in the analog environment (NationalResearch Council 2000, p. 136). Advocates of fair use view such develop-ments as reducing free access to works below what users have come to
expect, e.g., in educational and research settings (National Council for
Higher Education, 2003). These discussions only highlight that there mightbe more than economic efficiency at stake in the design of fair use. For
example, greater fair use of general-purpose works in educational settings
might be justified as redistribution of wealth away from copyright holders
'Fair use' as policy instrument 83
of those works to help equalize educational opportunities for all. The
models below do not incorporate such distributional factors directly,
although some fair use settings we do examine, in particular those that
maximize only consumer welfare, or volume of uses of works, could speak
to these considerations.
5.3 THE ECONOMIC MODELS
We look at copyrighted works as a series of monopolies. Each has separ-
ate, independent demands, created at a fixed cost independent of actual
use. As more works are supplied, the 'marginal' work is valued less (has a
lower demand curve) then the previous one. We rule out substitutability
among works from a consumer perspective, e.g., that the availability of
work X reduces willingness to pay for work Y. 15 We also rule out comple-
mentarity in consumption or production, e.g., that availability of work Xincreases demand for Y, or that access to work X reduces the cost of pro-
ducing work Y 16 These assumptions abstract away from complications
and arbitrariness inherent in differentiated product models. As the point
of the models is to suggest the relative position of optimal amounts of
allowed (fair) use under different policy goals rather than to estimate spe-
cific quantitative fair use standards or to look at changes in external con-
ditions (e.g., new distribution technologies), these simplifications may not
be inappropriate.
Fair use policy can involve a tradeoff between the number of works
supplied and the number of uses of each work. Specifically, expanding fair
use could reduce the number of markets served by copyrighted works. In
principle, expansion could be more costly on both efficiency-based and
non-economic policy standards. For example, increasing the fraction of
uses regarded as fair could reduce the number of works produced with little
effect on number of uses per work, reducing total uses of all works. Such a
change would move the outcomes farther away from both overall economic
efficiency and non-economic objectives relating to volume of use of copy-
righted works.
We treat 'fair use' parametrically in two ways: varying the fraction k of
uses available for free at any value, and varying a reservation price ceiling r
below which uses can be obtained for free. We refer to the first as the 'equal
diversion' case, and the second as the 'reservation price' case. In both, we
show how this parameter should be adjusted to address the breadth of
policy goals that might be considered, to see which direction fair use might
be adjusted and the conditions under which substantial adjustments are
more likely to be warranted.
84 Developments in the economics of copyright
5.4 THE 'EQUAL DIVERSION' SETTING
We index the set of actual or potential copyrighted works by the param-eter n. For tractability, we assume that as n increases, demand D(p,n) for
that work at price/? declines, i.e., Z)/J<0. 17 For convenience, we assume that
each of these works can be created at a fixed cost c, constant across all,
with no marginal cost for individual uses. 18 With demand for each work
independent from the demand for any other, an owner of a copyrighted
work n would set price/? to maximize profit !!(/?,)
Tl(p,n)=pD(p,n), (5.1)
where
pDp(p,n)+D(p,n)=Q.19
(5.2)
Let /?*() be the profit-maximizing price for work n as defined by (5.2), and
let n*Cz)=II(/?*(),) be the maximum profit received by the copyright
holder for work n. 20 From (5.1), (5.2) and the envelope theorem,
11*' CO =P*(n)Dn(p*(n), n) < 0. (5.3)
Profits fall with n because Dn<Q. The number of works n* with no fair
use is determined where the profits just cover the cost of production, i.e.,
where n*(*) = c.
The Fair Use/Number of Works Tradeoff
We first look at fair use that by intent or effect allows some fraction of uses
of copyrighted works to take place without having to pay the copyright
owner. For simplicity, assume that this diversion occurs uniformly over
the demand curve. Let A;E[0, 1] be the fair use parameter, such that the
amount of uses that the copyright owner of work n can sell at price /? will
be (1-k)D(p,n). The quantity of fair uses that takes place will be kD(Q,n).
An advantage of the assumption that k is uniform over the demand curve
is that the price that maximizes profits will be determined by (5.2), inde-
pendent of k. Thus, the extent of fair use has no effect on prices of copy-
righted works that are sold. The amount of uses of work n with fair use k,
USE(n,k),is
USE(n,k) = (1 -k)D(p*(n\n) + fc/>(0,/i). (5.4)
'Fair use' as policy instrument 85
At fair use intensity k, works will be created up to the point where profits
of the marginal work N just cover creation costs, i.e.,
This implicitly defines the number of works as a function of the fair use
regime, N(k). From (5.5), (5.3), and the definition of profit, the derivative
N'(k) is given by
N'(k) =(l-k)H*'(N)
(\-k?p*(N)Dn(p*(N),N)
D(p*(N),N)
(1 -)/)(/>*(#),A^' (5.6)
The fraction is negative because demand is positive and the denominator
is negative by assumption. Equation (5.6) not only defines N'(k) but pro-
vides the basis for the assumption in the derivations below that N'(k)<Q,
i.e., that fair use reduces the number of works created. The magnitudes of
N' and Dnare inversely related. If reducing the number of works just a little
bit would show a large increase in demand for the marginal work, N will
only fall a small amount for a given increase in the degree of fair use.
Illustrating Policy Objectives
Figure 5.1 illustrates potential policy objectives for fair use for a particular
work n. The key for any case will be to recognize tradeoffs. Increasing k may
C cl(\-k)D(p*(n),n)
(\-k)D(p*(n\n) (1 -)D(0, n)
Uses of work n
, n)
Figure 5. 1 Consumer surplus, profits, and costsfor work n: 'equal diversion
86 Developments in the economics of copyright
increase the contribution to the policy objective from work n against the
reduction in the total number of works N(k) that would be supplied.
In Figure 5.1, Area 1 corresponds to the consumer surplus of those who
pay p*(n) to use work n. Area 2 is the consumer surplus from the kD(0,n)
fair uses. Area 3 is the profit to the creator of work n. Area 4 is the cost c
of producing work . Finally Area 5 is the forgone potential surplus from
those unable to obtain a fair use of work n and unwilling to pay p*(n) for
it. Were the supply of work n guaranteed, we would want to set k 1,and
set Area 2 to the entire consumer surplus, while Area 5 disappears. We can
now look at setting k to maximize different social objectives.
Number of Works
Let kN be the level of k set to maximize the number of works TV, e.g., to
encourage creativity without regard to other objectives. With this objective,
kN =Q, i.e., have no fair use, since N' <0.21
Net Economic Welfare
From Figure 5.1, the contribution to net economic welfare from work n
with fair use k is the sum of Areas 1, 2, and 3. To reduce some of the clutter
in the notation, define the consumer surplus function CS(p, n) of work n
at price /? as
CS(p,n)="JD(p,n)dp.p
Area 1 is (1-k)CS(p*(n\n)\ Area 2 is &CS(0,); Area 3 is p*(n)(\ -k)
D(p*(ri),n) c. Aggregated over all N(k) works supplied, net economic
welfare NEW(k) is
N(k)
NEW(k) =I
[(1 -k)CS(p*(n\n) + kCS(0,n)
o
+ /?*()(! -k)D(p*(n\ri)]dn-cN(k).
Recalling from (5.5) that the profits of the marginal work just cover its cre-
ation costs, the degree of fair use, kNEW,that maximizes NEW is that which
satisfies the first-order condition
J/ -\r ri TT/- \ ^(K); = f [CS(0,/i)
-CS(p*(n),n)
-P*(n)D(p*(n\n)}dn
dko
0. (5.7)
'Fair use'
as policy instrumen t 87
The integral is the increase in net economic welfare from each work supplied
with more fair use; the net gains to consumers from getting more free uses
less the lost profits to creators. The second term is the lost economic welfare
from the works not produced as fair use is expanded. Because profits from
the marginal works produced are zero, losses are borne entirely by con-
sumers of those works.
Aggregate Consumer Surplus
A third potential standard for fair use policy would be to consider costs but
to leave profits to creators out of the social benefit calculation. In Figure
5.1, this would be to focus only on Areas 1 and 2, leaving out Area 3, profits
to creators. If we let ACS(k) refer to this measure,
IN(k)
I
f p*(n)(l-k)D(p*(n),n)dn
-cN(k) .
Since profits from the marginal work just cover creation cost,
^ = /}WW^^. (5-8)
The term on the right in (5.8) is the effect of increased fair use on profits
to inframarginal works. This was subtracted out in calculating net eco-
nomic welfare. At kNElv, where NEW is maximized, ACS will thus be
increasing in k.
Accordingly and not surprisingly, the level of fair use that maximizes
consumer surplus, kACS,will be greater than kNEW. The lost profits to copy-
right work holders would not matter. The difference between kACS and
kNEW, i.e., the degree to which fair use should be increased if we focus solely
on benefits to users, depends on the magnitude of the integral in (5.8). This
is larger the greater are prices at which copyrighted works are sold. If aggre-
gate consumer surplus is the objective, fair use should be stronger the
higher are prices for copyrighted works.
Gross Benefit
A fourth potential objective might be the value of the uses to consumers of
the works, neglecting both profits and costs except to the degree that they
influence the number of works available. This would comprise Areas 1 , 2,
3, and 4 in Figure 5.1, not subtracting production costs. This gross benefit
88 Developments in the economics of copyright
GB(k) is defined by adding costs back into the expression for net economic
welfare, i.e.,
GB(k) = NEW(k) + cN(k).
Hence,
d(GB) d(NEW)dk dk
cN'(k\ (5.9)
The first-order condition for gross benefit is given by (5.7) for NEW, addingback cN', which is negative. At kNEW, the level of fair use that maximizes
net economic welfare, d(GB)/dk<0, implying that the level of fair use that
maximizes gross benefits, kGB, is less than kNEW.
This is less surprising than it may initially appear. Ignoring production
costs, the optimal number of works should increase relative to that under
a net welfare standard. The difference depends on N' . If increasing the
degree of fair use has little effect on output, the standard that maximizes
net economic welfare will be close to the standard that maximizes gross
benefit.
Number of Uses
Fifth, suppose that the object of fair use policy is to maximize the number
of uses of copyrighted works, without regard to the value of different uses.
This could follow from a belief that willingness to pay depends on wealth
and does not reflect utility (in the utilitarian, not economic sense) or anyother intrinsic social value in access to works. Letting USE reflect the
volume of access to copyrighted works under fair use regime k, we have
N(k)
USE(k) =I
[(1-k)D(p*(n),n) + kD(Q,n)]dn
o
cKfJSE)m
and
dko
+ N'[(l-k)D(p*(N),N) + kD(0,n)]. (5.10)
The integral term in (5.10) is the net increase in uses of existing works from
extending fair use. The second term is the lost uses that follow when extend-
ing fair use reduces the number of works.
No general comparison between kUSE,the degree of fair use that maxi-
mizes access, and other values of fair use found above is apparent.
'Fair use' as policy instrument 89
Equation (5.10) suggests one possibility. It is similar to the expression for
the first-order condition for maximizing aggregate consumer surplus. From
(5.7) and (5.8), we have
N(k)
I[CS(0,/i)
-CS(p*(n),n)]dn
o
+ N'[(\- k)CS(p*(N\N) + kCS(Q,N)].
The integral here equals the integral in (5.10) times the average consumer
surplus over the supplied works. The second term in the above expression
equals the second term in (5.10) times the average consumer surplus of the
marginal work. If the former average consumer surplus exceeds the latter,
the positive contribution to the condition for CS outweighs the negative
contribution at kUSE, suggesting that
and hence that kACS>kUSE, i.e., that looking at uses leads to less generous
fair use.
This result is plausible. Increasing the amount of fair use will tend to
increase uses of inframarginal works that are relatively highly valuable and
discourage creation of works that are valued less. If these relative values are
ignored and all uses counted equally, this difference will not matter and this
justification for making fair use stronger will not apply.
Summary
Recalling that the optimal fair use regimes are kN for number of works,
kNEW for net economic welfare, kcs for consumer surplus, kGB for gross
value of uses, and kUSE for number of uses,
kACS > kmw>kGB >kN = and ' probably, kACS > kUSE.
The degree of the first inequality depends on the prices at which copy-
righted works are sold. The degree of the second and third inequalities
depends on the sensitivity of the supply of copyrighted works to price. In
addition, one may be able to come up with specifications for this model in
which kNEW>kUSE, i.e., if one were concerned only with uses, one might
want a weaker fair use policy than that motivated by economic efficiency. If
increasing fair use has a sufficiently large negative effect on the number of
works produced without much of an effect on the number of uses per work,
one might get more overall uses of works with less fair use rather than more.
90 Developments in the economics of copyright
5.5 LOW RESERVATION PRICE FAIR USES
We now model fair use more in line with the transaction costs rationales
Gordon and others have described. We characterize fair use by the param-eter r, where all uses of a work valued at a reservation price below r are
free to the user. We assume throughout this section that this restriction is
enforced, i.e., that there is no diversion of higher valued uses into the fair
use regime.
As in the earlier model, n* is the copyrighted work for which profits
just cover cost, i.e., where p*(n*)=c/D(p*(n*),n*). Obtaining results in
this setting requires the further assumption that as n increases, /?*() falls, i.e.,
p*' < O.22
If/?*' < 0, fair use will not discourage any creation of works for r <
/?*(*). For all w<*,/?*(w) >/?*(*) >r, implying that the copyright holder
will lose no sales. Under these conditions, fair use will enhance all measures
of benefit based on use, without reducing the supply of works, if fair use
allows all uses where the reservation price is less than /?*(*). For notational
convenience, let r* equal p*(n*). For the rest of the analysis, we restrict our
attention to r>r*, the range where policy tradeoffs become relevant,
although with respect to actual copyright rules, r may well be less than r*.
In this setting, there are two different effects of fair use on the market for
a work, depending on whether p*(ri) > r or p*(n) ^ r. The first is displayed
in Figure 5.2.
As with Figure 5.1, Area 1 represents the consumer surplus for those who
buy work n at the copyright holder's price /?*(), Area 3 is the net profits
/*()
clD(p*(n\ n)
D(p*(n), n) D(r, n) >(0, n)
Uses of work n
Figure 5.2 Consumer surplus, profits, and costsfor work n: profit-
maximizing price >fair use price
'Fair use' as policy instrument 91
D(r, n)
Uses of work n
Z>(0, n)
Figure 5.3 Consumer surplus, profits, and costsfor work n: profit-
maximizing price <fair use price
reaped by the copyright holder of work n, and Area 4 is the cost c of creat-
ing work n. These are independent of the fair use regime as long asp*(n) > r.
Area 2 represents the consumer benefits from fair use, to those with reser-
vation prices below r obtaining the work for free. Area 5 represents con-
sumers unwilling to payp*(n) but willing to paymore than r; these consumers
do not get the work and reap no benefits. In this setting, those who value
the work a lot or a little get it; those in the middle do not. Note that r need
not be below c/D(p*(n), n) as drawn in the graph, as long as it is belowp*(ri).
The second case is where p*(n) ^ r. There, the copyright holder of work
n is unable to charge the nominal profit maximizing price, since users whovalue the work more than p*(n) but less than r will obtain it for free. The
copyright holder will charge r for the D(r,n) uses valued more than r.
Relevant effects are displayed in Figure 5.3.
Consumers make all positively valued uses of works. Area 1 represents
the surplus of those consumers who have to pay r for the work. Area 2 is
the surplus captured by consumers with reservation price less than r who
get the work freely under the fair use policy. Area 3 is the profit to the copy-
right holder; Area 4 is the cost of creating this work.
Output-maximizing Fair Use
Let us consider works for which rD(r,n) < c will not be created. Let N(r) be
the value of n for which rD(r,n)=
c. From our assumption that the equi-
librium price increases as n falls, we have that all works for which n< N(r)
92 Developments in the economics of copyright
will be created. The fair use level that maximizes output, indicated againwith the subscriptN9 is rN<r*. To avoid discouraging uses that would leave
output and prices unaffected, rN =r*. By any policy criterion, the optimalr will be no less than r*.
Net Economic Welfare
For any r, copyrighted works n>N(r) fit two categories. First, for n > N(r)such that p*(n) <r, copyright holders will set the price at r and sell D(r,),as in Figure 5.3. Define N*(r) by p*(N*) = r. Because /?*'<0, 7V'*<0;as r increases, the set of works for which owners can charge the profit-
maximizing price shrinks. Second, for n> N*(r), owners can charge /?*(),with results as seen in Figure 5.2.
Net economic welfareNEM^with fair use r is consumer surplus plus profits
aggregated over works in each of these two categories, less creation cost.
NEW(r) = [CS(p*(n),n) + CSF(r,n) +p*(n)D(p*(n),n)]dno
N(r)
i)dn-cN(r}. (5.11)+J
CS(0,)
N*(r)
As before, CS(p, n) is consumer surplus for those who pay price p to
use work n. CSF(r,), the consumer surplus from fair use for uses with
reservation prices below r obtained for free, is Area 2 in Figure 5.2. For
n G [0, N(r)], the terms in the integral are respectively Areas 1, 2, andAreas 3 and 4 together; Area 4 is subtracted off in the cost term on the left.
For n G [N(r), 7V*(r)], all uses are made of the work, so the total benefit
net of costs is just total available surplus CS(0, n). The last term above is the
subtracted cost of creating N(r) works. If r =/?*(0), N*(r)= 0. Increasing r
above p*(Q) would eliminate consumer surplus from the fewer works that
are created, and those losses would be less than the saved costs. 23 Thevalue of r that maximizes NEW, rNEW, must then be at most/?*(0), makingN*(rNEw) anc* all terms of (5.11) relevant. Consequently, rNEW is the
value of r satisfying
d(NEW) _A
+ CFS(r,N*(r)) + p*N*(r)D(p*(N*(r)\N*(r))-
CS(0, JV*(r))]
'Fair use'
as policy ins trumen t 93
Because CSFr(r,n)
= - rDr(r,n)
24and, by definition, at N*(r\p* = r, the
sum of Areas 1, 2, and 3 constitutes the entire consumer surplus CS(0,
N*(r)) available at TV*. The long second term on the right is thus zero, leaving
(r))-cNf
(r). (5.12)
The first term is the increase in surplus from allowing more uses of works
for which the selling price exceeds the fair use standard. The second term
is the lost surplus from the marginal works no longer profitable when all
uses valued below r are provided free. The last term is the reduced cost from
creating N' fewer works.
Because Drand N' are negative, it will be useful to use absolute values
and change signs as appropriate. Let
\[D(^n)-D(r,n)]dn
Nbe the average amount of fair use of works through TV at maximum reser-
vation price r. We can rewrite (5.12) as
d(NEW}^ = N*(r)rDr(r9 N*(r))-
\N'(r)\[CS(0 9 N(r))-
c]. (5.13)
The first term is, again, the gain in surplus from increasing fair use, which
equals the average increase in consumer surplus times the number of works
for which that surplus is increasing. The second term is the lost net surplusas the supply of works falls by \N'(r)\.
Net economic welfare is maximized when (5.13) equals zero. Setting it
equal to zero and multiplying through by r/N*(r)[CS(Q, N(r))-
c] gives the
equivalent first-order condition
rD(r,N*(r)) _ N(r).6 JVOF
[CS(0, N(r)) -c]~USE
N*(r)
where
^M-W^*'-"* , ....'..,..
is the elasticity of the quantity of fair uses for works where the price is
above the fair use reservation price, and
94 Developments in the economics of copyright
is the absolute value of the elasticity of supply of works with respect to the
degree of fair use.
The numerator in the fraction on the left-hand side of (5.14) is the ratio
of the product of the marginal value of fair uses times the average fair use
of works where the market exceeds the reservation price of fair uses. Thedenominator is the total net surplus of works that would be excluded if fair
use were marginally enhanced. The fraction on the left-hand side is the ratio
of total works created to works created where price exceeds the marginalfair use.
At r*, the ratio on the left-hand side is 1;the numerator and denominator
both equal *, the number of works where profits cover creation costs.
There are no created works where the profit-maximizing price is less than
the maximum fair use reservation price. The D term in the numerator will
be the average number of fair uses of all works. From (5.14), we would not
want to make fair use any more generous than r* if at r*
r*D(r,n*)LXI i . n t
<e,[CS(0,*)-c]
USE ^W
If the elasticity of the number of works created sufficiently exceeds the elas-
ticity of fair uses at r*, a more generous fair use regime above the minimumwill reduce net economic welfare.
Aggregate Consumer Surplus
As before, the difference between net economic welfareNEWand aggregateconsumer surplus ACS is that the latter does not include profits for works
created, represented by Area 3 in Figures 5.2 and 5.3. As consumer surplus
includes only Areas 1 and 2 in the cases portrayed by those figures, aggre-
gate consumer surplus over all works supplied is
N(r)
ACS(r)= [CS(p*(n\n) + CSF(r,n)]dn+ f [CS(r,n) + CSF(r,n)]dn.
Nr)
f [CS(r,n) +
N*(r)
Because CSr(r,n)
=-D(r,n), maximizing ACS with respect to r gives
25
N(r)
-D(r,n)dn
N*(r)
+ N'(r)(CS(r,N(r)) + CSF(r,N(r))}.
At N(r), the combined consumer surplus for customers with reservation
prices above r who pay for the work and for those who with reservation
'Fair use' as policy instrument 95
prices below r who use the work for free is just the total available surplus
CS(0, N(r}) less cost, as profits just cover cost at N(r) by definition. Thebracketed term in the above is CS(Q,N(r))
-c, implying from (5.12) that
d(ACS) = d(NEW)dr dr
W3[-rDr(r,n)-D(r,n)]dn. (5.16)
The expression in the integral in (5. 16) is the negative of the derivative of
the profit function at r for works where the fair use reservation price exceeds
the profit-maximizing price. For prices above the profit-maximizing price,
that derivative is negative, i.e., rDr(r,n) + D(r,n)<0. This makes the inte-
gral in (5.16) positive. At rNEW, increasing r further would increase ACS.The degree of fair use that maximizes aggregate consumer surplus, rACS,
exceeds rNEW.
The result is qualitatively the same as in the 'equal diversion' case; profits
lost when fair use is increased do not count against aggregate surplus as
they do against net economic welfare. The difference here depends on the
convexity of the profit function, which determines the degree to which the
integral in (5.16) is positive. If rNEW=r* because (5.14) holds, rACS mayequal r* as well. Changing the policy objective from economic welfare to
consumer surplus, ignoring profits to creators, need not lead to a more gen-erous fair use regime.
Gross Benefit
Define again gross benefit GB as the gross consumer surplus, neglectingcosts. For works where the fair use reservation price is less than the profit-
maximizing price, as portrayed in Figure 5.2, the gross benefit is repre-
sented by Areas 1,2,3, and 4. For works where the reservation price exceeds
the profit-maximizing price, the gross benefit will be the consumer surplusobtained if uses were free to all. In both cases, gross benefit equals net eco-
nomic welfare with creation costs added back in. Accordingly, as with
'equal diversion', (5.9) holds, adapted for the low reservation price imple-
mentation, i.e.,
c^ (r) . (5, 7)
At rNEW, where the first term on the right-hand side is zero, the derivative
of gross benefit will be negative (Nr < 0). The fair use regime that maximizes
gross benefit, rGB, will be less than rNEW. If creation costs are disregarded,the social calculus will tilt toward increasing the number of works and awayfrom uses per work relative to where net economic welfare is maximized.
96 Developments in the economics of copyright
The degree to which rGB falls below rNEW depends on the sensitivity of the
supply of works to the level of fair use. If N' is small, these will differ little.
As rNElvmay equal r*, rGB will be even more likely to equal r*, the value of
fair use that maximizes the supply of works.
Number of Uses
For works where the profit-maximizing price exceeds the fair use reserva-
tion price r, the number of uses will be those purchased at the profit
maximizing price, D(p*(n\ ), plus free uses under the fair use regime,
Z>(0,) D(r,n). For works where the fair use reservation price exceeds
the profit-maximizing price, all D(0, n) uses with a positive value will occur.
The total number of uses, USE(r\ is
7r)
N(r)
[D(p*(n),n) + 0(0, /i)-0(r,)] dn + f D(0,n)dn.
N*(r)
USE(r) is maximized at rUSE, defined by the first-order condition
MT7V1?\ N*W^^=f -D(r,n)dn + N'(r)D(Q,n)
= Q. (5.18)dr J
o
The first term is the increase in uses from making more of them 'fair;' the
second is the lost uses from the reduced output when more generous fair
use reduces the supply of works.
To compare rUSE to other fair use standards, multiply the derivative in
(5.18) through by r. Comparing the result with (5.12) gives
. N'(r)[rD(Q,n)-
(CS(0,#(r))-
c)]. (5.19)dr dr
As N' < 0, this expression is positive at rNEW> i.e., rUSE>rNEW " an(* om<y ^
r0(0,/i)<CS(0,JV(r))-c atr = rNEW.
Dividing both sides by 0(0, n) implies that rUSE> rNEW if and only if rNEWis greater than the average net surplus per use of the marginal work at rNEW.
Even if rUSE>rNEW,rUSE may be below rACS . This ambiguity is greater
than that in the equal diversion case, where we have good reason to believe
that fair use would be stronger under a consumer surplus standard than
under a total use standard. The difference arises because the added uses in
this case from increasing fair use are valued at r. In the 'equal diversion'
setting, the added uses are valued at the average willingness to pay over all
works, which is likely to be greater than the average surplus associated with
forgone uses of the marginal work.
'Fair use'
as policy ins trumen t 97
To compare rVSE and rGB , use (5.17) and (5.19) to find
d(USE) d(GB)
dr drN'(r)[rD(0,n)-CS(0 9 N(r))].
This implies that rUSE>rGB if and only if r/)(0, n)<CS(Q, N(r)) at rGB,
i.e., that r is less than the average willingness to pay for the marginal work
over all uses.
Summary
Recasting fair use to allow only uses below a target reservation price pre-
serves many of the results under equal diversion. Fair use is strongest if
aggregate consumer surplus is the objective and weakest if numbers of
works is the objective. In between those two, the level of fair use that maxi-
mizes net economic welfare exceeds the level that maximizes gross value of
uses, i.e.,
> r > r > r r*r r r ~ r
where r* is the maximum reservation price for free uses that would not
discourage creation of any works.
How a gross benefit standard compares to the net economic welfare
standard is largely the same in both cases. The degree to which rACS exceeds
rNEW depends on the extent to which increasing the level of fair use reduces
profits for works for which the fair use price exceeds the price the copy-
right owner would prefer to charge. The difference between rNEW and rNdepends on how the ratio of the supply elasticity of works compares to the
elasticity of fair uses. At rN=
r*, this ratio may be small enough so that
rNEW= rN
= r*
:>deviations above the minimum reduce net economic welfare.
If the elasticity of the number of works with respect to r is sufficiently large,
we could have strict equality among all the terms in the above expression
and thus not want to increase fair use beyond the minimum regardless of the
policy objective. The degree of fair use that maximizes uses of copyrightedworks may or may not exceed the level that maximizes net economic welfare
or gross benefit. It depends on whether the average net surplus or gross
surplus is less than the maximum value of a fair use.
5.6 CONCLUSION
Our goal is to better understand optimal fair use standards, particularly
how they depend upon criteria by which copyright policy should be judged.
98 Developments in the economics of copyright
We examined a setting in which copyrighted works are created at constant
cost and serve markets with independent demands. Fair use took two
forms, one in which some fraction of uses were freely available ('equal
diversion'), and a second in which uses below a maximum willingness to
pay were available for free ('reservation price'). Standards were evaluated
according to numbers of works supplied, net economic welfare, aggregate
consumer surplus, gross benefit, and numbers of uses.
Many results held in both contexts. From a net economic welfare per-
spective, the marginal benefit of more fair use is the increase in surplus from
added uses of works produced. This should equal the marginal cost of
more fair use, which from an efficiency perspective is lost net surplus (con-
sumer benefit less production costs) from marginal works that would no
longer be profitable to create if more uses were free. Interest in numbers of
works created would lead to weaker fair use standards; interest in consumer
surplus would lead to a more generous regime relative to net economic
welfare. Focusing on gross benefits leads to less generous fair use relative to
these other standards. Ignoring creation costs under the former standard
tilts the 'uses vs. works' tradeoff toward producing more works.
Why those differences occur and whether they are significant are influ-
enced by different factors in the two settings. Convexity of profit functions
matters more under a reservation price criteria, while price levels of the
copyrighted works play a stronger role with equal diversion. Under the
reservation price setting, if the elasticity of supply of works with respect to
the scope of fair use is sufficiently large, the choice of objective may not
matter. The optimal fair use standard under any criterion may be the
maximum consistent with no reduction in the supply of works at all.
Focusing on uses regardless of surplus led to ambiguous and different
answers in the two settings. Under equal diversion, focusing on uses alone,
regardless of differences in willingness to pay by different users, would
likely lead to a less generous fair use standard than that maximizing aggre-
gate consumer surplus. Neglecting willingness to pay uses would tend to
rebalance the relevant tradeoffs more in favor of additional, less econom-
ically valuable works and away from giving users with high willingness to
pay greater free access to works. Under the reservation price regime, the fair
use regime that maximizes uses depends on how the maximum reservation
price for fair use compares to the average surplus consumers obtain from
the marginal work. In either case, a fair use standard based upon maxi-
mizing uses of copyrighted works could be less generous than one designed
to maximize net economic welfare.
A corollary is that if the copyright policy balance is seen solely as one
between numbers of works and numbers of uses, neglecting values based
on willingness to pay (of wealthy rather than poor users, say) and profits,
'Fair use'
as policy instrument 99
fair use may be less generous than if efficiency is the standard. The optimal
level of fair use would be between that maximizing uses and that maximiz-
ing supply of works, and both standards may be less generous than that
maximizing net economic welfare.
One extension would be to consider that copyright policies today affect
the availability of copyrighted works in the future, e.g., by stimulating
creativity and fostering a sense that creators are enriching an ongoingstream of cultural works from which they have drawn (Gordon, 2003).
These effects could be represented as an external benefit based on reducing
the cost of producing future works. But how this cultural complementarityaffects fair use policy depends very much on how that external benefit is
generated. If future benefit comes from having more works produced or
gross benefit in the present, it argues for reducing fair use relative to a net
economic welfare standard, to stimulate the production of more works. If
that benefit comes from today's consumer surplus, it is likely to suggest
increasing the scope of fair use relative to that standard. Other factors,
including number of uses, are more ambiguous, as the tradeoffbetween lost
surplus from fewer uses of works (increasing fair use) and lost surplus from
having fewer works (decreasing fair use) is more relevant.
We can also investigate different settings to see whether these results are
robust. One could employ a Cournot oligopoly (Watt, 2000, pp. 37-54),
modeling works as perfect substitutes, where each copyright holder chooses
simultaneously how much of the work to make available. A second class of
models could involve competition among suppliers of differentiated prod-
ucts (Carlton and Perloff, 2000, ch. 7). As these models remain arbitrary, we
are unlikely to get precise estimates of how fair use should be designed. The
complexity and lack of generality of differentiated product models mayinhibit their ability to inform judgments on the design of copyright policy.
ACKNOWLEDGMENT
Thanks to Shubha Ghosh, Wendy Gordon, Mark Nadel, Christopher Yoo, the editors of this
volume, and other participants at the 2003 Congress of the Society for Economic Research on
Copyright Issues for comments. Errors remain the author's responsibility.
NOTES
1. State Street Bank & Trust v. Signature Fin. Group, 149 F.3d 1368 (Fed. Cir. 1998);
Amazon.com v. Barnes & Noble, 73 F. Supp. 2d 1228 (1999).
2. A&M Records, Inc., et al. v. Napster, Inc. (9th Circ., Feb. 12, 2001); Universal City
Studios, et al. v. Corley (2nd. Circ., Feb. 7, 2002).
100 Developments in the economics of copyright
3. Digital Millennium Copyright Act, Pub. L. No. 105-304, 1 12 Stat. 2860 (Oct. 28, 1998).
4. Eldred et al. v. Ashcroft (U.S. Sup. Ct. 01-618, Jan. 15, 2003), interpreting U.S.
Constitution, Article I, Sec. 8.
5. 17US.C.107.6. See Napster, n. 2 supra, Sony Corp. v. Universal City Studios, Inc., 464 U.S. 417 (1984);
Harper & Row v. The Nation Enterprises, 471 U.S. 539 (1985); Campbell v. Acuff-Rose
Music, Inc. 510 U.S. 569 (1994); American Geophysical Union v. Texaco, Inc., 60 F.3d
913(1994).7. I thank Wendy Gordon for this reference.
8. Brennan (1994) reviews the limits of markets in providing social goals.
9. This model does not exactly fit Gordon's ( 1 982) justification for fair use, that some users
value the work but are not willing to pay the transactions costs associated with the work.
In that case, the demand curve that the creator of a work sees will be net of these trans-
actions costs. We show that the fair use policy maximizing net economic welfare is typ-
ically greater than zero. A fair use policy focusing on economic considerations may then
be more generous than that which minimizes transactions costs.
10. To examine changes in one feature of copyright, it is useful to assume that the rest of
copyright law is designed efficiently, to avoid using an inferior policy instrument to
address a problem other than the one at hand (Brennan, 1988).
11. See also National Research Council (2000, p. 138). For an exclusively economic
approach, see Landes and Posner (1989).
12. From 17 U.S.C. 107, the four statutory factors courts are to consider in ascertaining
whether a use is a fair use are:
1 . the purpose and character of the use, including whether such use is of a commercial
nature or is for nonprofit educational purposes;2. the nature of the copyrighted work;3. the amount and substantiality of the portion used in relation to the copyrighted work
as a whole; and4. the effect of the use upon the potential market for or value of the copyrighted work.
Klein et al. (2002) deal only with the last of these, the potential market, and not the first
three.
13. Ramello (chapter 7 of this volume) makes a similar point from a qualitative, socio-
anthropological perspective.
14. They do not suggest that the optimal compulsory license fee need be zero; they offer only
a possible rationale for fair use.
15. We also neglect income effects, i.e., that if less money is spent on some copyrighted works,
because their price falls or because they go off the market, that demand will increase for
other works because consumers have more money to spend.1 6. We discuss allowing a work as an 'input' into the production of subsequent works in the
conclusion. Such considerations underlie concerns with parody and derivative works in
the copyright area, and innovations that build on earlier ones in the patent area.
17. We assume Dp<Q and other second-order conditions necessary to ensure unique
internal profit maximizing prices for each work. This assumption implies that the
'marginal' work has a lower demand curve than that of 'inframarginal' works.
1 8. One could regard the demand curve for work n as net of the marginal cost of producing
copies.
19. In positing a single price for the copyright work across all uses, we assume copyright
holders cannot price discriminate beyond what fair use institutes. A more complex model
could allow for some price discrimination, e.g., as in sales of hardcover and paperback
books, or theatrical release compared to video rental for films.
20. 'Profit' here refers to variable profit or producer surplus, ignoring the ex ante costs of
producing the work. The number of works is determined endogenously with entry up to
the point that this profit equals the cost of producing a work.
'Fair use'
as policy ins trumen t 101
21 . A policy to have no fair use might also maximize producer surplus, which undoubtedlywill affect the development of copyright policy in a political environment.
22. This assumption does not follow from Dn <Q. For any n, the value of p* is defined by
ry/7, n)= 0. Consequently, p*'
= - HpnIH
pp. Since the denominator is negative,/;*' <0 if
and only ifO^< 0. From (2), 11^=pD
pfl
+ Dn
. The second term is assumed negative, but
the first could be positive and outweigh it. We have/?*' < when Dpn< 0, i.e., the demand
curve becomes steeper at any price as n increases.
23. The consumer surplus of works created exceeds the creation cost c for r>r*. At r*, the
profit of work *just equals costs. As consumer surplus exceeds profits, consumer
surplus exceeds costs. It thus will be true for all <*, since profits increase as n falls.
24. From Figure 5.2, CSF(r,n) = CS(0,)- CS(r,n)-rD(r,ri). The derivative with respecttor is
CSFr(r, n)
= - CSr(r, )
-D(r, n)
- rDr(r, n).
Since CSr(r, n) D(r, n), the first two terms on the right-hand side cancel, leaving the
expression in the text. The subscript r refers to the partial derivative of demand with
respect to price. For clarity, we use r instead of p throughout this section.
25. Because p*(N*(r = r at N*, CS(p*(N*(r)),ri)=
CS(r,n). Thus, terms with TV*' cancel, as
with the derivative of NEW.
REFERENCES
Besen, S. and S.N. Kirby (1989), 'Private copying, appropriability, and optimal
copying royalties', Journal ofLaw and Economics, 32, 255-80.
Besen, S., W. Manning and W. Mitchell (1978), 'Copyright liability for cable televi-
sion: Compulsory licensing and the Coase theorem', Journal of Law and Eco-
nomics, 21, 67-95.
Brennan, T. (1986), 'Harper & Row v. The Nation: Copyrightability and fair use',
Journal of the Copyright Society of the U.S.A., 33, 368-89.
Brennan, T. (1988), 'An economic look at taxing home audio taping', Journal ofBroadcasting and Electronic Media, 32, 89-103.
Brennan, T. (1994), 'Markets, information, and benevolence', Economics and
Philosophy, 10, 151-68.
Brown, R. (1985), 'Eligibility for copyright protection: A search for principled
standards', Minnesota Law Review, 79, 579-609.
Carlton, D. and J. Perloff (2000), Modern Industrial Organization, Boston, MA:Addison Wesley.
Ciminello, D. (2000), 'Business process patents- the road to reform', Internet
Law Journal, 14 November 2000, available at http://www.tilj.com/content/
iparticlelll40001.htm.
Depoorter, B. and F. Parisi (2002), 'Fair use and copyright protection: A price
theory explanation', International Review ofLaw and Economics, 21, 452-73.
Gordon, W. (1982), 'Fair use as market failure: A structural and economic
analysis of the Betamax case and its predecessors', Columbia Law Review,
82, 1600-57.
Gordon, W. (2002), 'Excuse and justification in the law of fair use: Commodi-fication and market perspectives', in N. Elkin-Koren and N. Netanel (eds), The
Commodification of Information, Norwell, MA: Kluwer, pp. 149-92.
102 Developments in the economics of copyright
Gordon, W. (2003), 'Intellectual property', in P. Cane and M. Tushnet (eds), The
Oxford Handbook of Legal Studies, New York: Oxford University Press,
pp. 617^6.
Johnson, W. (1985), The economics of copying', Journal of Political Economy, 93,
158-74.
Klein, B., A. Lerner and K. Murphy (2002), 'The economics of copyright: "Fair
use" in a networked world', American Economic Review Papers and Proceedings,
92, 205-8.
Landes, W. and R. Posner (1989), 'An economic analysis of copyright law', Journal
of Legal Studies, 18, 325-63.
Liebowitz, S. (1985), 'Copying and indirect appropriability: Photocopying of jour-
nals', Journal of Political Economy, 93, 945-57.
National Council for Higher Education (2003), 'Of fair use, liabilities, and elec-
tronic toll booths', available at http://www.nea.org/he/nche/fairuse.html, accessed
26 April 2003.
National Research Council Computer Science and Telecommunications Board
(2000), The Digital Dilemma: Intellectual Property in the Information Age,
Washington, DC: National Academies Press.
Novos, I. and M. Waldman (1984), 'The effects of increased copyright protection:
An analytic approach', Journal of Political Economy, 92, 23646.
Ramello, G. (2005), 'Private appropriability and sharing of knowledge: Conver-
gence or contradiction? The opposite tragedy of the creative commons', in
L. Takeyama, W. Gordon and R. Towse (eds), Developments in the Economics of
Copyright: Research and Analysis, Cheltenham, UK and Northampton, MA,USA: Edward Elgar.
Takeyama, L. (1997), 'The intertemporal consequences of unauthorized reproduc-
tion of intellectual property', Journal ofLaw and Economics, 40, 51 1-22.
Watt, R. (2000), Copyright and Economic Theory: Friends or Foes?, Cheltenham,
UK and Northampton, MA, USA: Edward Elgar.
Yoon, K. (2002), 'The optimal level of copyright protection', Information Economics
and Policy, 14, 327^8.
6. Towards a differentiated products
theory of copyright
Christopher S. Yoo
6.1 INTRODUCTION
The economic analysis of copyright is largely founded on the premise that
consumption of copyrighted works is non-rivalrous, a premise that distin-
guishes markets for intellectual property from markets for other types of
goods. In the context of copyright, this is generally taken to mean that the
marginal cost associated with the consumption of an incremental unit of
a copyrighted work is effectively zero. Non-rivalry in turn gives rise to a
well-known economic conundrum. If authors are to break even, the prices
they charge must defray the fixed cost needed to produce the work in the
first place (often called 'first-copy costs') as well as cover the marginal cost
associated with producing an incremental unit. However, pricing above
marginal cost necessarily reduces welfare by excluding some potential users
from consuming the work even though the benefits they would derive
exceed the costs of permitting them to do so.
This has led scholars to frame copyright in terms of the need to balance
two opposing considerations. On the one hand are the benefits flowing
from the efficient dissemination of copyrighted works (often called the
'access' side of the tradeoff). On the other hand is the need to provide
authors with sufficient compensation to support the creation of their works
(often called the 'incentives' side of the tradeoff) (Novos and Waldman,
1984; Johnson, 1985; Liebowitz, 1985; Landes and Posner, 1989). This
tradeoff implicitly posits that markets for copyrighted works are protected
by entry barriers and that particular works do not face substantial compe-tition from substitutes.
The traditional economic approach to copyright assumes that copyright
turns authors into monopolists over their works (Novos and Waldman, 1 984;
Liebowitz, 1985). The power over price conveyed by this legal monopoly
power gives rise to the familiar efficiency losses that occur whenever prices
exceed marginal cost. The concomitant transfer of surplus from consumers
to producers has also led some scholars to raise distributional concerns
103
104 Developments in the economics of copyright
(Netanel, 1996; Boyle, 2000; Cohen, 2000). Together these considerations
have led to the emergence of a consensus that copyright protection is a nec-
essary evil and that Congress and the courts should calibrate copyright pro-
tection to the lowest level that still provides sufficient return to support
creation of the work.
The problem with the traditional approach is that copyrighted works
do in fact face competition from other works that serve as imperfect sub-
stitutes and that entry is often quite easy. In fact, the doctrine known as
the 'idea-expression dichotomy,' which limits copyright protection only to
those aspects of a work that display the author's originality and leaves
unprotected any facts or ideas contained within the work (17 U.S.C.
102(b)), has the practical effect of guaranteeing that any competitor willing
to undertake the same fixed cost investment as the original author remains
free to create alternative works with the same functional characteristics.
The growing recognition that copyrighted works generally face some
degree of competition from other works has led some scholars to turn
to dominant firm and Cournot market structures to model markets for
copyrighted works (Landes and Posner, 1989; Koboldt, 1995; Watt, 2000).
Although an improvement over the monopoly approach, these models
still fail to capture the full dynamics of entry. Other scholars have drawn
on the approach to imperfect competition best suited to capturing the
key characteristics of markets for copyrighted works; the theory of product
differentiation. Early analyses focused solely on product differentiation in
the limited context of direct copying, in which unauthorized copies serve
as imperfect substitutes for the original (Johnson, 1985; Liebowitz, 1985;
Besen and Kirby, 1989; Koboldt, 1995).1 While helpful, these models still
fail to capture the economic impact of free entry by similar works. Some
initial work applying differentiated product models to model competition
among different works has begun to appear (Lunney, 1996; Abramowicz,
2004). Previous efforts have stopped short of considering the full range of
normative and remedial implications.
This chapter offers a more complete exploration of the economic insights
provided by a shift to a differentiated products approach to copyright (see
also Yoo, 2004). Adoption of a differentiated products approach opens upthe policy space by revealing that access to creative works can be increased
by facilitating entry by new works and allowing the ensuing competition to
reduce the spread between price and marginal cost. At the same time, the
possibility of entry largely alleviates any concerns about overstimulation of
creative activity or sustainable profits generally associated with the incen-
tives side of the tradeoff. It suggests that over the long run the presence of
supracompetitive profits will only serve to stimulate entry, which improves
access via lower prices and increased product variety.
Towards a differentiatedproducts theory of copyright 1 05
Perhaps most important, the differentiated products approach suggests
that the degree of tension between access and incentives implied by the
traditional approach may be overstated. By demonstrating how facilitat-
ing entry can essentially promote both considerations simultaneously, the
differentiated products approach echoes one of the central insights of
classic property theory, which emphasizes how well-defined property rights
can promote economic efficiency (Demsetz, 1967; Hardin, 1968). Indeed, it
shows how access to copyrighted works can be served by strengthening
property rights, which would stimulate entry and drive price closer to mar-
ginal cost as well as increase product variety.2
In contrast to the traditional approach, the differentiated products
approach also provides a basis for distinguishing among the available policy
instruments. Previous models tended to represent the overall level of copy-
right protection with a single variable (Novos and Waldman, 1984; Landes
and Posner, 1989; Koboldt, 1995). Viewing copyright through the lens of
product differentiation makes it possible to isolate the impact of multiple
ways in which copyright protection can be strengthened or weakened.
What results is a reconceptualization of copyright that moves beyondthe relatively static vision of monopoly economics and captures the
dynamics of free entry. The differentiated products approach also accords
better with the institutional capabilities of governmental actors. By pro-
viding a decentralized, market-oriented means for ensuring that authors
capture only enough revenue to cover their fixed costs and no more, it
responds to the growing doubts that courts and legislatures are institu-
tionally capable of striking the proper balance between access and incen-
tives. The government is better suited to promoting access by strengthening
copyright protection than to attempting to strike the proper balance
between access and incentives through the careful calibration of the level
of copyright protection.
6.2 THE NORMATIVE IMPLICATIONS OF THEDIFFERENTIATED PRODUCTS APPROACH
The normative insights provided by applying the differentiated products
approach to copyright can most easily be understood in terms of the
theory of monopolistic competition pioneered by Chamberlin (1933).
Monopolistic competition retains most of the assumptions underlying
perfect competition, including free entry and the presence of a sufficient
number of buyers and sellers to justify ignoring strategic reactions to
pricing decisions. The key difference is that monopolistic competitionrelaxes the assumption that competing works are homogeneous. Instead,
106 Developments in the economics of copyright
product differentiation gives authors a sufficient degree of power over price
to justify modeling each work as facing a downward-sloping demand curve.
Because each author makes his or her own independent price and quantity
decisions, this approach models competition at the producer level rather
than at the industry level. When the market is analyzed at the producer level
rather than the industry level, total surplus depends not only on the amount
of surplus generated by any particular work, but also on the total number
of works created.
The market power conveyed by product differentiation leads producers of
differentiated works to set short-run prices in precisely the same manner as
a monopolist, which results in deadweight loss and short-run supracom-
petitive profits. However, the possibility of entry recognized by monopolis-
tic competition dictates that the short-run equilibrium is unstable. When
entry is free, the presence of supracompetitive profits attracts entry by
authors offering similar works. The classic Chamberlinian formulation of
monopolistic competition also assumes that consumers' preferences are
symmetric with respect to a cluster of other works. 3 The primary effect of
this assumption is to place each work in equal competition with all other
works in the group rather than in localized competition with a smaller
set of near neighbors. Because all of the works in the market are in equal
competition with one another, new entrants take business equally from each
of the incumbents, which causes the demand curve confronting each incum-
bent work to shift inward. In addition, the demand curve confronting each
author becomes more elastic as a growing number of new works increases
the number of imperfect substitutes available for each consumer.
Entry will continue until no economic profits remain, which, under
Chamberlin's original formulation, will occur when the surplus appropri-
ated by each author is just enough to cover fixed costs.4 The equilibrium
number of works is determined by the size of the market relative to the size
of fixed costs. Indeed, as the size of the market expands or the size of the
fixed costs declines, the number of works asymptotically approaches infin-
ity, and the deadweight loss approaches zero (Hart, 1979; Mankiw and
Whinston, 1986; Jones, 1987; Eaton and Lipsey, 1989).5
A New Perspective on Access
The shift to a differentiated products approach effects two significant
changes to the economic analysis of copyright. First, when products are
homogeneous, authors can compete only on a single dimension -price
-
which also greatly simplifies the welfare analysis by reducing it to total
surplus. When the competing works are differentiated, the value of product
diversity must also be taken into account, making simple price-cost
Towards a differentiatedproducts theory of copyright 107
margins incomplete indicators of economic efficiency. As a result, productdifferentiation raises the possibility that any deadweight losses caused by
non-marginal cost pricing might be offset in part by welfare gains resulting
from product variety.
Even more importantly, the differentiated products approach highlights
the existence of an alternative way to reduce deadweight loss that up until
now has largely been overlooked. Specifically, it illustrates how access can
be promoted not by lowering the degree of protection provided by copy-
right, but rather by facilitating entry by similar works. Entry by near sub-
stitutes causes the demand curve facing each work to become more elastic,
which in turn reduces the spread between price and marginal cost.6
In this manner, the differentiated products approach opens up the policy
space by identifying how entry can promote access, an insight that the
traditional approach is poorly situated to take into account. When non-
rivalrous goods are homogeneous, entry is unnecessarily duplicative and
simply wastes resources. Moreover, the tendency towards natural monopolycreated by declining-cost structures strongly suggests that no such entry
would be viable.
This analysis suggests that access may be promoted by strengthening
copyright protection, because it is the presence of profits that stimulates
entry7(although, per the subsequent discussion, the analysis may require
weakening of other aspects of copyright protection). This stands in stark
contrast to the traditional approach to promoting access, which focuses
solely on lowering the level of copyright protection. Indeed, the differ-
entiated products approach suggests that weakening copyright will only
serve to deter entry by reducing the revenue generated by each work.
Consequently, as will be discussed in greater detail later, it may have the
perverse effect of cementing an excessively concentrated market structure
into place.
The differentiated products approach also mitigates any distributional
concerns raised by increasing the amount of surplus captured by authors.
As noted earlier, entry will continue until any supracompetitive profits are
dissipated. Furthermore, any short-run transfer of surplus from consumers
to producers will largely accrue to consumers' benefit over the long run in
the form of increased product variety.
The foregoing discussion demonstrates how the differentiated products
approach opens up the policy space by revealing how access can be pro-
moted indirectly by stimulating entry rather than directly by lowering the
level of copyright protection. In the process, it reveals that under certain
circumstances, economic welfare might better be promoted by follow-
ing precisely the opposite of the policies prescribed under the traditional
approach to the economics of copyright.
108 Developments in the economics of copyright
The Formalization of Optimal Incentives
Appropriability as a determinant of optimal incentives
In addition to suggesting an alternative way to promote access, the differ-
entiated products approach also offers a solution to one of the fundamen-
tal limitations of the traditional approach, which is its inability to providea basis for formalizing the optimal level of incentives. In the absence of
such a basis, scholars have employed rough metrics to approximate the
proper balance (Fisher, 1998; Brennan, 2002). These scholars concede that
such metrics provide only a vague sense of how much creative activity
would be optimal.
The present analysis overcomes this shortcoming by offering a basis for
determining the optimal level of entry. It suggests that a work should be
produced whenever the surplus it would create exceeds the costs needed to
produce it. This condition is met whenever the total surplus generated bythe work is larger than the fixed cost (Spence, 1976a, 1976b; Dixit and
Stiglitz, 1977; Spence and Owen, 1977; Koenker and Perry, 1981).
This criterion for determining efficient levels of entry illustrates the
importance of the authors' ability to appropriate the surplus generatedfrom their works. It suggests that a reduction in authors' ability to appro-
priate the surplus created by their works can cause them to forgo creating
marginal works even though doing so would cause total welfare to increase.
The larger the reduction in the authors' ability to appropriate surplus, the
fewer welfare-enhancing works will be created.
This represents a fairly sharp departure from the view of appropriability
taken by the traditional approach, which takes the position that copyrightlaw should carefully calibrate appropriability so that works capture only
enough surplus to support creation of the work. The differentiated prod-ucts approach reveals that such fine tuning would be counterproductive,
since any limitations to authors' ability to appropriate surplus will cause a
suboptimal number of works to be created and entry by competitive works
will help ensure that no work garners excessive returns. 8 As a result, increas-
ing appropriability lacks distributional implications over the long run,
since free entry will dissipate any profits initially accrued and largely cause
them to accrue back to consumers in the form of increased product variety.
At the same time, doctrines that restrict appropriability have the inevitable
effect of exacerbating the access side of the tradeoff by decreasing the
extent to which entry can narrow the spread between price and marginalcost in the long run. Indeed, the foregoing analysis demonstrates how low-
ering the level of copyright protection, rather than promoting efficiency,
may have the perverse effect of entrenching a concentrated market struc-
ture into place by making it impossible for new competitors to enter.
Towards a differentiatedproducts theory of copyright 109
Demand diversion as a countervailing consideration
The analysis advanced thus far would appear to suggest that economic
welfare would best be promoted by maximizing authors' ability to appro-
priate surplus. Indeed, were appropriability the only relevant consideration,
copyright policy would devolve into a simple matter of allowing authors
to capture as much revenue as possible. However, complete appropriationof surplus, which would require perfect price discrimination, is a prac-
tical impossibility. One would thus conclude that markets would exhibit a
systematic tendency towards underproduction of copyrighted works that
could not be rectified no matter how much copyright law is structured to
enhance appropriability.
One countervailing consideration, however, is whether the sales captured
by a new entrant represent incremental sales to new customers or instead
represent sales cannibalized from authors already in the market. FollowingBorenstein (1985), the former effect will be referred to as 'demand creation'
and the latter effect as 'demand diversion.'9
The possibility of demand diversion allows markets to create the optimalnumber of works even when authors are unable to appropriate the entirety
of the surplus created by their works. As stated earlier, the basic welfare
criterion for evaluating the efficiency of entry offered by the differentiated
products approach dictates that a work be produced whenever the benefits
it creates exceed the costs required to produce it. Although no author is able
to appropriate the entire surplus created by his or her work, demand diver-
sion raises the possibility that the incremental surplus that an author is
unable to capture might be replaced by surplus cannibalized from other
incumbents. In other words, because of demand diversion, the fact that
perfect price discrimination is impossible need not lead to a systematic
underproduction of product variety.
In fact, demand diversion creates the possibility of excess entry, in which
authors produce new works even when the costs of doing so exceed the
benefits. As noted earlier, efficient entry requires that authors producenew works only when the surplus attributable to demand creation exceeds
the fixed costs needed to produce the work. The problem is that a profit-
maximizing author will enter whenever the total surplus it captures exceeds
the fixed costs of entry regardless of whether the surplus captured results
from demand creation or demand diversion. Such an author could finance
the fixed costs with surplus cannibalized from other producers already in
the market rather than incremental surplus generated by new consumers.
Under these circumstances, the profitability constraint does not necessar-
ily prevent the waste of resources. In such cases, it may be appropriate to
use copyright to restrict entry by increasing the degree of differentiation
required before a new work does not infringe on existing works.
1 10 Developments in the economics of copyright
6.3 THE REMEDIAL IMPLICATIONS OF THEDIFFERENTIATED PRODUCTS APPROACH
In contrast to the traditional approach, the differentiated products
approach also provides a basis for distinguishing among the available policy
instruments. As noted earlier, previous models tended to represent the
overall level of copyright protection with a single variable. Only a handful
of copyright scholars have offered some preliminary attempts to analyze the
interaction among limited aspects of copyright protection (Fisher, 1998,
Liu, 2002; Hughes, 2003).
Product differentiation theory provides a basis for modeling the impactof different aspects of copyright protection, as illustrated by the patent lit-
erature using spatial competition to analyse the tradeoff between a patent's
'length,' determined by the duration of the patent term, and its 'breadth,'
which is most usefully described for present purposes as how close a com-
peting product may come in the characteristics space to a patented productwithout constituting infringement (Klemperer, 1990).
The differentiated products approach to copyright suggests a similar
analysis, but with some important modifications. Specifically, it suggests
that the analysis might be enriched by disaggregating the concept of length
into two distinct concepts. On the one hand is what I will call the 'size' of
the right, as determined by the number of surplus-generating activities
contained within the right. On the other hand is what might be called the
'intensity' of the right, as determined by the author's ability to capture the
available surplus. An increase in the size of the right would be represented
by an outward shift of the demand curve. An increase in the intensity of the
right would be represented by an increase in the proportion of the available
surplus captured by authors rather than consumers, created for example bya change in law that facilitates price discrimination.
Expanding the analysis in this manner highlights the complex inter-
actions among these factors. Interestingly, the policy implications do not
all point in the same direction. Specifically, the differentiated products
approach indicates that the best way to render a market more competitive
is by increasing the number of surplus-generating activities encompassed
by the right, facilitating authors' ability to appropriate surplus, and by lib-
eralizing how close competing products can come without constituting
infringement. In other words, economic welfare would best be promoted bya copyright that is large and intense, but narrow. The differentiated prod-ucts approach is therefore not an unqualified endorsement for strengthen-
ing copyright protection. Instead, by providing a basis for distinguishing
among different aspects of copyright protection, it allows for a more
nuanced approach to copyright policy.
Towards a differentiatedproducts theory of copyright 1 1 1
Determinants of the Differentiated Products Equilibrium
The size of the right
An important determinant of the overall competitiveness of markets for
copyrighted works is the size of the copyright, as determined by the numberof surplus-generating activities encompassed within its scope. As noted
earlier, the overall competitiveness among differentiated products is deter-
mined by the level of entry, which is in turn determined by the magnitudeof the relevant market relative to the fixed costs of entry. Increasing the
ratio of the size of the overall market to fixed costs brings the resulting
equilibrium closer to the competitive outcome.
The primary means for policymakers to increase this ratio is to expandthe size of the copyright by increasing the number of surplus-generatingactivities that fall within the scope of each copyrighted work. One deter-
minant of the size of the right that has received a great deal of attention in
recent months is the lengthening of the copyright term effected by the
Copyright Term Extension Act. But other examples abound, such as the
extent to which copyright allows authors to retain performance rights.
Doing so has the effect of causing the demand curve confronting each
copyrighted work to shift outwards.
The differentiated products approach suggests that increasing the size
of the right can promote both access and incentives by increasing the equi-librium number of works. The suggestion that access would be promotedbest by raising rather than lowering the level of copyright protection mayseem counterintuitive. After all, it implies that the proper policy responseto markets that are too concentrated is to increase the degree of copyright
protection that authors enjoy. This apparent contradiction disappearswhen viewed in light of the traditional approach's inability to capture the
dynamics of entry. So long as entry is free, any strengthening in the level of
copyright protection will not ultimately accrue to the benefit of the incum-
bents. Instead, it will only attract more entry, which will in turn reduce
deadweight loss and bring the number of works closer to the optimum. Anyshort-term profits made possible by the expansion of the size of the right
will largely accrue back to consumers in the form of increased variety.
The intensity of the right
Another consideration that determines the market's ability to promote eco-
nomic welfare is the intensity of the right, which is determined by authors'
ability to appropriate surplus. Unlike increases in the size of a copyright,which cause the demand curve confronting a copyrighted work to shift out-
wards, increases in intensity leave the demand curve intact and simplyincrease the proportion of the area under the demand curve captured by
1 1 2 Developments in the economics of copyright
authors. Intensity is affected by copyright principles such as the first sale
doctrine, which limits authors' ability to engage in price discrimination.
The breadth of the fair use doctrine also shapes authors' ability to appro-
priate surplus.
As noted earlier, any reduction in appropriability tends to reduce the
equilibrium number of works. Such a reduction can harm the incentives
side of the tradeoff by causing the total number of works produced to dropbelow optimal levels. It also harms the access side of the tradeoff by limit-
ing the extent to which entry by new works will reduce deadweight loss in
the long run.
This suggests that increasing authors' ability to capture the surpluscreated by their works can promote both the access and the incentives side
of the copyright tradeoff. Again, this argument may seem counterintuitive
from the standpoint of the traditional approach to copyright, which views
access and incentives as being in inexorable tension. The solution lies in
understanding that access may be promoted as much by increasing the
number of works available for consumption as by mandating access to the
limited number of works that have already been created.
The breadth of the right
Finally, the extent to which consumers regard competing works as sub-
stitutes plays a natural role in determining how many firms will enter at
equilibrium and how robust the competition among those firms will be.
Copyright's breadth is a legal constraint on substitution because infringingworks cannot compete absent permission from the copyright holder. This
analysis then initially suggests that copyright should be kept narrow in
order to promote lower price-cost margins.There are, however, considerations that cut in the other direction. For
example, substitutability also determines the extent to which the surplus
captured by any particular work derives from demand diversion. The higherthe degree of substitutability between the works, the greater the propor-tion of the total surplus captured that demand diversion will represent. In
other words, the more closely related works can be without constituting
infringement, the greater the amount of surplus will come from demanddiversion.
As noted earlier, demand diversion plays a critical role in determininghow closely the total number of works produced will approximate the
optimum. Up to a point, demand diversion is beneficial, as it can replacethe surplus that authors are unable to appropriate because of their imper-fect ability to price discriminate. Beyond that point, demand diversion
creates the possibility of excess entry. The danger of excess entry is the
greatest when goods are the most substitutable.
Towards a differentiatedproducts theory of copyright 113
Policymakers may reduce the impact of demand diversion by using the
standard of copyright infringement to reduce the substitutability between
competing works. Doing so would require a delicate balance. On the one
hand, reducing substitutability limits the impact of demand diversion and
thus brings the number of works closer to optimal levels. On the other
hand, reducing substitutability also reduces the extent to which price-cost
margins are narrowed by entry. Policymakers charged with calibrating the
breadth of copyright protection may have to confront the difficult task of
balancing the welfare gains on the access side of the tradeoff against the
welfare losses on the incentives side of the tradeoff. 10
Interactions Among the Different Aspects of Copyright Protection
The differentiated products approach thus provides a framework that is able
to distinguish among three different ways in which copyright protection can
be strengthened or weakened. Although this degree of nuance enriches the
power of the analysis, it also makes copyright policy considerably more
difficult to implement.One problem is that the available policy instruments are not completely
independent. Changes in the legal regime designed to calibrate the size of
the right may also have an impact on the right's intensity or breadth. The
overlapping nature of these considerations complicates isolating the impactof any particular aspect of copyright policy.
It nevertheless may be possible to simplify the analysis with respect to
particular industries or categories of copyrighted works. If one aspect of
copyright protection can be taken as fixed with respect to certain types of
works, the problem that must be solved becomes much simpler. Empirical
studies may provide additional insights into how to balance these coun-
tervailing considerations. Some empirical studies suggest that any welfare
losses resulting from excessive entry are likely to be relatively small (Yarrow,
1985; Goettler and Shachar, 2001). Another empirical study of entry pat-
terns in the radio industry estimates that the deadweight losses attribut-
able to excess entry may be substantial (Berry and Waldfogel, 1999). The
latter study acknowledges, however, that the radio industry is somewhat
unusual in that it serves two different groups of customers - advertisers and
listeners - only one of which (advertisers) is able to make direct paymentsfor programming. What appears to be excessive entry when measured solely
in terms of benefits to advertisers may in fact be efficient when measured in
terms of both advertisers and listeners. ] l
These studies suggest the possibility of isolating the impact of each of
the three factors identified by the differentiated products approach. Even
if the empirical problem proves intractable, the differentiated products
1 14 Developments in the economics of copyright
approach should still provide useful intuitions about the way these factors
interact. It suggests, for example, that excess entry is least likely to be a
problem when a work has few substitutes. Thus, contrary to the conven-
tional wisdom, it is when a work is the most unique that the case for
strengthening the level of copyright protection in terms of the size and the
intensity of the right is the strongest.
6.4 CONCLUSION
In the final analysis, product differentiation offers significant promise as a
way to reconceptualize the economic analysis of copyright law. What
emerges is an approach that demonstrates how stimulating entry can
promote both access and incentives simultaneously. This stands in stark
contrast to the position that dominates existing copyright scholarship,
which views these two considerations as being in inexorable tension.
The differentiated products approach also suggests that the best policy
response to a highly concentrated market might well be to strengthen the
degree of copyright protection in order to stimulate new entry. There is
some irony in the fact that copyright protection might tend to be the
strongest when high fixed costs and the low degree of substitutability cause
the market to become the most concentrated, but this apparent paradoxis resolved once one understands the complex manner in which access
and incentives interact with one another. In this sense, the differentiated
products approach to copyright captures the insights of classic property
theory, which emphasizes the importance that well-defined property rights
can play in ensuring optimal investment and deployment. In so doing, it
corrects for the blind spot that results when markets for copyrighted works
are treated as monopolies, which prevents serious consideration of the role
that short-run profits can play in stimulating entry and in promoting eco-
nomic efficiency. At the same time, the differentiated products approach
incorporates the possibility of excess investment and entry stimulated bydemand diversion.
Although the theoretical implications of this analysis are clear, consid-
erable additional work remains to be done before it can be fully opera-
tionalized. As noted earlier, further work should incorporate elements of
sequential innovation that take into account the extent to which current
works serve as inputs to subsequent works, although, for reasons set forth
in the margin, such considerations are unlikely to prove problematic.12
Future work should also consider the implications of relaxing the
symmetry assumption by applying models that allow the extent to which
particular works serve as substitutes for other works to vary (Kaldor, 1935;
Towards a differentiatedproducts theory of copyright 1 1 5
Waterson, 1990). This suggests that copyright might profitably be analyzed
by using the spatial competition models pioneered by Hotelling (1929), in
which works compete by adopting locations across a characteristics space
(Yoo, 2004). Relaxing the symmetry assumption allows for the possibility
that entry by a new work will impact only some of the works rather than
being spread evenly across all of the incumbents. This localization of com-
petition has the effect of dividing the relevant market into subsegments,
with the overall competitiveness of the subsegment determined by the size
of the total surplus of the subsegment relative to the fixed cost, rather than
the size of the total surplus of the entire market relative to the fixed cost.
The lack of robust competition within a subsegment may limit the extent
to which entry can push price towards marginal cost. It can also allow the
'integer problem' to arise simultaneously with respect to multiple portions
of the overall market, as the single 'large economy' is chopped into a series
of 'small economies' that are each capable of supporting sustainable profits.
If these effects arise with respect to multiple subsegments, these adverse
effects may be quite substantial (Eaton and Lipsey, 1976). Relaxing the
assumption that works will distribute themselves evenly across the product
space creates the possibility that first movers will employ preemptive
strategies to lock in sustainable supracompetitive positions (Baumol,
1967; Hay, 1976; Prescott and Visscher, 1977; Eaton and Lipsey, 1980;
Bonanno, 1987; Neven, 1987). The analysis becomes even more complex if
one allows for the possibility of production of multiple works by a single
author (Schmalensee, 1978; Eaton and Lipsey, 1979; Brander and Eaton,
1 984; Judd, 1985).
Finally, the policy instruments that follow from the differentiated prod-
ucts approach are by their nature extremely contextual and do not lend
themselves to simple policy inferences. In addition, the interrelationships
among the available policy instruments make calibrating them simultan-
eously an extremely difficult empirical exercise. The fact that the differ-
entiated products approach is contextual and nuanced should not obscure
its basic analytical power and does not by itself justify rejecting the theory.
Indeed, the intuitions that the theory reveals about the relationship
between access and efficiency and the manner in which the various aspects
of copyright protection interrelate are sufficient to justify further inquiry.
ACKNOWLEDGMENT
This chapter is adapted from material originally published at Yoo (2004). I would like to
thank Michael Abramowicz, Mark Brandon, Tim Brennan, Andy Daughety, Paul Edelman,Shubha Ghosh, John Goldberg, Wendy Gordon, Paul Heald, Larry Heifer, Steven Hetcher,
Jack Hirshleifer, Doug Lichtman, Stan Liebowitz, Glynn Lunney, David McGowan,
116 Developments in the economics of copyright
Michael Meurer, Tom Nachbar, Mark Nadel, Richard Nagareda, Neil Netanel, Erin O'Hara,Giovanni Ramello, Bob Rasmussen, Jennifer Reinganum, Pam Samuelson, Art Snow, Lisa
Takeyama, and participants in workshops conducted at the 2003 Annual Congress of the
Society for Economic Research on Copyright Issues, the Thirty-First Annual Meeting of the
American Law and Economics Association, the 2003 Works-in-Progress Intellectual
Property Colloquium at Tulane University, Columbia Law School, and the Vanderbilt LawSchool for helpful comments on earlier drafts of this work. Financial support from the
Vanderbilt Dean's Fund is gratefully acknowledged. Responsibility for any errors rests with
the author.
NOTES
1 . Ghosh (2004) employs a differentiated products model to evaluate the related issue of
competition between originals and derivative works.
2. The differentiated products approach can thus be understood as striking a middle groundbetween the traditional analysis of increased copyright protection and the nonprotectionof copyrighted works. The possibility of entry suggests that the tradeoff between access
and incentives is not as direct as suggested by the traditional approach. Furthermore, the
differentiated products approach also redresses the central problem with nonprotection,which can provide efficient access, but which struggles to provide sufficient incentives to
support production of creative works. This conclusion is subject to an important caveat
discussed below. When fixed costs are large compared to marginal costs, it is possible that
strong property rights may induce levels of entry that are excessive.
3. Chamberlin's original formulation also made a number of other simplifying assump-tions, none of which turn out to be central to the analysis. For example, Chamberlin
posited that each producer faced identical cost and demand curves. This allowed him to
employ a single graph portraying the price-quantity response of a representative firm to
model the entire market. Allowing the cost and demand curves to vary across productswould simply cause equilibrium price and quantity to differ with respect to each firm,
which is completely reasonable given the assumption that each product is differentiated.
Firm-to-firm variations in price and quantity would not, however, change any essential
aspects of the equilibrium (Kaldor, 1935; Archibald, 1961).
4. There is, however, a well-known exception to Chamberlin's zero-profit result. It has longbeen recognized that the lumpiness of fixed costs may create a situation in which n works
might earn small profits while n + 1 works would run losses. This so-called 'integer
problem' allows for an equilibrium in which n works each earns sustainable profits. In
large economies (i.e., when n is relatively large), such profits will be negligible. This
integer problem was first identified by Kaldor (1935).
5. This conclusion depends on the convexity of consumer preferences and production sets
(Hart, 1980; Roberts, 1980).
6. Because firms must cover their fixed costs, price will not completely converge to marginalcost.
7. Acknowledging how short-run profits stimulate entry should not be confused with
Schumpeterian competition, in which competitors use innovation to obtain long-run
competitive advantage. The ease of entry dictates that any profits should be transient
rather than sustainable. This implies that there will be horizontal competition within the
market rather than vertical competition for the market. Short-run profits instead stimu-
late entry in the same manner as occurs in perfectly competitive markets that are in tem-
porary disequilibrium.
8. Subject to an important consideration discussed in the following section.
9. Other analyses use different terminology to describe the same effect (Beath and
Katsoulacos, 1991, p. 57 ('cannibalisation'); Mankiw and Whinston, 1986 ('business
stealing effect')).
Towards a differentiated products theory of copyright 1 1 7
10. To give another example, substitutability also affects the proportion of surplus that an
author can capture in another way. Monopolistic competition theory indicates that
works with relatively steep inverse demand functions capture a lower proportion of the
available surplus than do works with relatively flat inverse demand functions. Workswith steep inverse demand functions tend to be products with low own-price elasti-
cities of demand, which in turn tend to be those that have the fewest substitutes
(Spence, 1976a; Spence and Owen, 1977). For a more complete discussion, see Yoo
(2004, pp. 273^4).
11. In addition, the fact that their study assumed that the radio market is composed of
homogeneous products led them to overlook potential welfare benefits resulting from
product differentiation. The existing theoretical literature suggests that this simplifying
assumption can have a fairly dramatic effect on the welfare implications (Mankiw and
Whinston, 1986).
12. For example, just as stimulating entry by close substitutes should promote access to
readers by using increased competition to lower price, it should also promote access to
follow-on authors who seek to build on prior work. Once the market for the work becomes
sufficiently competitive, the problem of cumulative innovation, in which a copyrightedwork is licensed simultaneously to both consumers and to other authors who seek to use
the work as an input in creating other works, becomes analogous to the classic problemof transfer pricing, in which a particular good simultaneously serves as an end productand as an input used in making another product. The transfer pricing literature indicates
that welfare is maximized when the price of the good when used as an input is set equalto the price charged of the good when sold as a final product, so long as the final productmarket is sufficiently competitive (Milgrom and Roberts, 1992, pp. 79-83). Thus, so longas the total surplus in the market is sufficiently large relative to the fixed costs of entry,
there is nothing inefficient about charging the market price to follow-on authors who seek
access to a copyrighted work as an input in creating other works. Should the market for
the works not be sufficiently competitive, the differentiated products approach suggeststhat the problem might be redressed by making the market more competitive by stimu-
lating entry rather than by lowering the price paid by follow-on authors. It is true that
holdout behavior may prevent particular authors from creating particular works.
Competition policy, however, focuses on protecting competition, not particular competi-tors. Thus, unless such refusals create losses for more than just particular individuals, nointervention is warranted. And even if intervention were justified, it would take the form
of a targeted remedy and not a general revision of the scope of copyright protection.
REFERENCES
Abramowicz, M. (2004), 'An industrial organization approach to copyright',
William and Mary Law Review, 46 (1), 33-124.
Archibald, G.C. (1961), 'Chamberlin versus Chicago', Review ofEconomic Studies,
29(1), 2-28.
Baumol, W.J. (1967), 'Calculation of optimal product and retailer characteristics:
The abstract product approach', Journal of Political Economy, 75 (5), 674-85.
Death, J. and Y. Katsoulacos (1991), The Economic Theory of Product Differ-
entiation, Cambridge and New York: Cambridge University Press.
Berry, ST. and J. Waldfogel (1999), 'Free entry and social inefficiency in radio
broadcasting', RAND Journal of Economics, 30 (3), 397^420.
Besen, S.M. and S.N. Kirby (1989), 'Private copying, appropriability, and optimal
copying royalties', Journal ofLaw and Economics, 32 (2), 255-80.
Bonanno, G. (1987), 'Location choice, product proliferation and entry deterrence',
Review of Economic Studies, 54 (1), 37-45.
1 1 8 Developments in the economics of copyright
Borenstein, S. (1985), 'Price discrimination in free-entry markets', RAND Journal
of Economics, 16 (3), 380-97.
Boyle, J. (2000), 'Cruel, mean, or lavish? Economic analysis, price discrimination
and digital intellectual property', Vanderbilt Law Review, 53 (6), 2007-39.
Brander, J.A. and J. Eaton (1984), 'Product line rivalry', American Economic
Review, 74 (3), 323-34.
Brennan, D.J. (2002), 'Fair price and public goods: A theory of value applied to
retransmission', International Review ofLaw and Economics, 22 (3), 347-75.
Chamberlin, E.H. (1933), The Theory of Monopolistic Competition, Cambridge,MA: Harvard University Press.
Cohen, I.E. (2000), 'Copyright and the perfect curve', Vanderbilt Law Review,
53(6), 1799-819.
d'Aspremont, C, J.J. Gabszewicz and J.-F. Thisse (1979), 'On "Hotelling's stability
in competition" ', Econometrica, 47 (5), 1 145-50.
Demsetz, H. (1967), Toward a theory of property rights', American EconomicReview (Papers and Proceedings), 57 (2), 347-59.
Dixit, A.K. and I.E. Stiglitz (1977), 'Monopolistic competition and optimumproduct diversity', American Economic Review, 67 (3), 297-308.
Eaton, B.C. and R.G. Lipsey (1976), 'The non-uniqueness of equilibrium in the
Loschian location model', American Economic Review, 66 (1) 77-93.
Eaton, B.C. and R.G. Lipsey (1979), The theory of market preemption: The persist-
ence of excess capacity and monopoly in growing spatial markets', Economica, 46
(182), 149-58.
Eaton, B.C. and R.G. Lipsey (1980), 'Exit barriers are entry barriers: The durabil-
ity of capital as a barrier to entry', Bell Journal of Economics, 11 (2), 721-9.
Eaton, B.C. and R.G. Lipsey (1989), 'Product differentiation', in R. Schmalensee
and R.D. Willig (eds), Handbook of Industrial Organization, vol. 1, Amsterdam:Elsevier Science, pp. 725-68.
Fisher, W.W., HI (1998), 'Property and contract on the Internet', Chicago-Kent LawReview, 73 (4), 1203-56.
Ghosh, S. (2004), 'Market entry and the proper scope of copyright', SSRN working
paper available at http://ssrn.com/abstract=569 162.
Goettler, R.L. and R. Shachar (2001), 'Spatial competition in the network televi-
sion industry', RAND Journal of Economics, 32 (4), 624-56.
Hardin, G. (1968), The tragedy of the commons', Science, 162 (3859), 1243-8.
Hart, O.D. (1979), 'Monopolistic competition in a large economy with
differentiated products', Review of Economic Studies, 46 (1), 1-30.
Hart, O.D. (1980), 'Perfect competition and optimal product differentiation',
Journal of Economic Theory, 22 (2), 279-312.
Hay, D.A. (1976), 'Sequential entry and entry-deterring strategies in spatial com-
petition', Oxford Economic Papers, 28 (2), 240-57.
Hotelling, H. (1929), 'Stability in competition', Economic Journal, 34 (153), 41-57.
Hughes, J. (2003), 'Fair use across time', UCLA Law Review, 50 (2) 775-800.
Johnson, W.R. (1985), The economics of copying', Journal of Political Economy,93(1), 158-74.
Jones, L.E. (1987), The efficiency of monopolistically competitive equilibria in
large economies: Commodity differentiation with gross substitutes', Journal ofEconomic Theory, 41 (2), 356-91.
Judd, K.L. (1985), 'Credible spatial preemption', RAND Journal of Economics,
16(2), 153-66.
Towards a differentiatedproducts theory of copyright 1 1 9
Kaldor, N. (1935), 'Market imperfection and excess capacity', Economica, 2 (5),
33-50.
Klemperer, P. (1990), 'How broad should the scope of patent protection be?',
RAND Journal of Economics, 21 (1) 1 13-30.
Koboldt, C. (1995), 'Intellectual property and optimal copyright protection',
Journal of Cultural Economics, 19 (2), 131-55.
Koenker, R.W. and M.K. Perry (1981), 'Product differentiation, monopolistic com-
petition, and public policy', Bell Journal of Economics, 12 (1), 217-31.
Landes, W.M. and R.A. Posner (1989), 'An economic analysis of copyright law',
Journal of Legal Studies, 18 (2), 325-63.
Liebowitz, SJ. (1985), 'Copyright law, photocopying and price discrimination',
Research in Law and Economics, 8, 181-200.
Liu, J.P. (2002), 'Copyright and time: A proposal', Michigan Law Review, 101 (2),
409-81.
Lunney, G.S., Jr. (1996), 'Reexamining copyright's incentives-access paradigm',Vanderbilt Law Review, 49 (3), 483-656.
Mankiw, N.G. and M.D. Whinston (1986), 'Free entry and social inefficiency',
RAND Journal of Economics, 17 (1), 48-58.
Milgrom, P. and J. Roberts (1992), Economics, Organization and Management,
Englewood Cliffs, NJ: Prentice-Hall.
Netanel, N.W. (1996), 'Copyright and a democratic civil society', Yale Law Journal,
106 (2), 283-387.
Neven, D.J. (1987), 'Endogenous sequential entry in a spatial model', International
Journal of Industrial Organization, 5 (4), 419-34.
Novos, I.E. and M. Waldman (1984), 'The effects of increased copyright protection:
An analytical approach', Journal of Political Economy, 92 (2), 23646.
Prescott, E.C. and M. Visscher (1977), 'Sequential location among firms with fore-
sight', Bell Journal of Economics, 8 (2), 378-93.
Roberts, K. (1980), 'The limit points of monopolistic competition', Journal ofEconomic Theory, 22 (2), 256-78.
Schmalensee, R. (1978), 'Entry deterrence in the ready-to-eat breakfast cereal
industry', Bell Journal of Economics, 9 (2), 305-27.
Spence, M. (1976a), 'Product differentiation and welfare', American Economic
Review (Papers and Proceedings), 66 (2), 407-14.
Spence, M. (1976b), 'Product selection, fixed costs, and monopolistic competition',
Review of Economic Studies, 43 (2), 217-35.
Spence, M. and B. Owen (1977), 'Television programming, monopolistic competi-
tion, and welfare', Quarterly Journal of Economics, 91 (1), 103-26.
Waterson, M. (1990), 'Product differentiation and profitability: An asymmetric
model', Journal of Industrial Economics, 39 (2), 1 13-30.
Watt, R. (2000), Copyright and Economic Theory: Friends or Foesl, Cheltenham,UK and Northampton, MA, USA: Edward Elgar.
Yarrow, G.K. (1985), 'Welfare losses in oligopoly and monopolistic competition',Journal of Industrial Economics, 33 (4), 515-29.
Yoo, C.S. (2004), 'Copyright and product differentiation', New York University LawReview, 19 (\),2\2-W.
7. Private appropriability and sharingof knowledge: convergence or
contradiction? The opposite
tragedy of the creative commons
Giovanni B. Ramello
7.1 INTRODUCTION
One of the many Greek myths tells how Eos, goddess of dawn, fell in love
with a handsome young man of royal Trojan blood, Tithonus. The goddess
was so taken with Tithonus that she asked Zeus to make him immortal, so
that they could be happy together forever. The wish was granted but turned
out to be a double-edged sword, because the gift of immortality did not
come with perpetual youth. So as time went by, and Tithonus withered and
grew old, the passion of the goddess died away.
Now, every myth contains a metaphor designed to teach us something.
The lesson we can draw here is that, when policy design fails to accurately
grasp the nature of the reality, the outcomes of the policy can be very
different from those envisaged. In this chapter we shall argue that, due to
analytical shortcomings, the present-day policy of extending and strength-
ening copyright is liable to produce inefficient outcomes in the knowledge
domain, analogous to those described in the above myth.Note that we are not here disputing copyright as it was, say, up until the
1990s, when it seemed to adequately balance the public need for access to
information with the provision of a private incentive to creators - a state of
affairs which, for the sake of simplicity, I have defined as 'minimal copy-
right'.1 That original form of copyright, to be absolutely clear, was fully
compatible and consistent with its statutory goals, namely, promoting the
creation of new knowledge through limited private appropriation of infor-
mation items - as history has amply borne out. 2
Rather, this chapter calls into question the new form copyright has been
taking following the legal trend toward its extension, whose advent can
probably be dated to the TRIPs Agreements (1994), and which is today
120
Private appropriab ility and sharing of knowledge 121
progressively altering the equilibrium established by the previous legal
framework in balancing knowledge sharing with appropriability. I have
used the term 'maximal copyright' to denote this contemporary, extended
form of the right.
The novel legal dynamics are displacing the traditional vision of copy-
right doctrine, which recognized the public-interest goals of the law and
so constrained the property right in various ways (see for instance the
term of duration, fair use doctrine, etc. 3) in order to simultaneously allow
for both (limited) private appropriability as an instrumental device, and
the return of newly created knowledge to the public domain as its primary
goal.
Utilitarian and neoclassical economic theory have provided the general
paradigm in support of the role of copyright as an incentive to create, but
have rarely recognized the importance of balancing the appropriability
granted by copyright with the necessity of assuring broad access to the
works created as a result. The contributions generally emphasize appropri-
ability over accessibility, with one notable exception being Landes and
Posner (1989). These distinguished law and economics scholars clearly
assert, in their influential contribution, that the welfare-enhancing effects
of copyright are ambiguously dependent on a balance between two oppos-
ite forces: an increase in the supply of new works brought about by the
statutory economic incentive on one hand, and a diminution in the supply
brought about by the exclusionary effect of copyright on the other hand.
The resolution of this trade-off is therefore the key to determining what
the overall consequences of copyright will be. Unfortunately, the authors
do not further pursue the analysis in this direction, but merely rely on some
specific assumptions in this respect.4
This chapter shall take the route of attempting to gain further insight
into this ambiguous relationship, paying special attention to the peculiar-
ities which characterize the production side. Using a wide selection of lit-
erature drawn from the various social science disciplines, it will attempt to
expose some of the shortcomings of traditional economic theory as it is
applied to copyright. The resultant perspective should lead to a better
understanding of the incentive/exclusion trade-off, showing how the
overall effect is likely to be greater than expected. This is so because
maximal copyright affects not just the cost of accessing each individual
protected work, but equally and significantly, it impacts on the productive
contexts (those which we have termed creative commons) by leading to a
depletion of knowledge and its dynamic productivity.
It follows, by mere straightforward application of neoclassical economic
reasoning, that the new extended copyright is liable to determine dynam-
ically inefficient outcomes analogous to those presaged by Hardin (1968) in
122 Developments in the economics of copyright
his landmark paper, and which, in tribute to this seminal author, we have
termed 'the opposite tragedy of the creative commons'.
The chapter is organized as follows: section 7.2 presents the tragedy of
the commons and describes the idiosyncratic feature of creative commons,in which knowledge is created and exists. The asserted peculiarity is then
supported with a discussion of the special nature of knowledge, which is
intrinsically dependent on the sharing process. Section 7.3 expands further
upon the structure of knowledge, whose two complementary constituent
elements, the tacit and codified dimensions, are equally necessary for
knowledge to exist in its entirety, and both have a productive role. It then
illustrates how real-world policies for promoting knowledge creation are
generally consistent with this representation, thus implicitly acknowledg-
ing the limited ability of proprietary systems to stimulate production of
optimal levels of new knowledge. Section 7.4 provides a test for the sound-
ness of the preceding arguments, through a simple check on whether the
proposed policy indications are consistent with the Lockean paradigm
widely invoked in justification of property rights. Finally, section 7.5 con-
tains the concluding remarks.
7.2 COMMONS, COPYRIGHTS AND TRAGEDIES
The economics literature on intellectual property rights generally presents
information as a public good, and consequently treats copyright as a device
designed to stimulate optimal levels of its production (see, for example,Landes and Poster, 1989; Watt, 2000). In other words, the property right is
introduced to avert a specific market failure, i.e., the production of a sub-
optimal level of new expressions of ideas. From this perspective, copyright-
protected information is considered to be an output, and the market failure
to be avoided is its underproduction.However there exists another, distinct type of market failure, arising
from the fact that information also serves as an input to the production
process and represents a common resource.
From Commons to Creative Commons
Summing up, in the knowledge domain, intellectual property rights should
enable individuals to reap the benefits of the new ideas, or expression of
ideas in the case of copyright, which they contribute to create. But, at the
same time, they must also avoid an altogether different kind of market
failure - namely, the depletion of the special type of common resource that
is knowledge.
Private appropriability and sharing of knowledge 1 23
These are two distinct and in a sense contrasting goals, because, as weshall see below, property rights on outputs can have negative effects on the
preservation of inputs.
In order to explain this assertion, though, we need to examine in greater
depth the nature of the knowledge as a commons. A 'commons' is gener-
ally defined to be a resource held in common, that is to say in joint use or
possession. Typical examples are uncultivated pastures, hunting grounds,
fishery stocks, etc. Now the problem here is that allowing open access to all
members of the community which owns the commons can lead to overex-
ploitation of the resource and its consequent depletion. This dynamic has
been extensively investigated in the scientific literature of complementarydomains. 5
The crucial element of all these contributions is that the production func-
tion of the commons exhibits diminishing returns with increasing number
of users. Some examples of this are pastures grazed by increasing numbers
of animals, or fishery stocks exploited by increasing numbers of fishing
boats. Hardin (1968, p. 1243) dramatically describes the previous situation
in his seminal 'tragedy of the commons': 'Ruin is the destination toward
which all men rush, each pursuing his own best interest in a society that
believes in the freedom of the commons. Freedom in a commons brings
ruin to all'.
The pivotal element of the above reasoning is, as always, the exhaustibil-
ity of the commons, which in turn refers us back to the basic postulates of
'scarcity of resources' in economic theory. In other words, the pasture,
fishery stock, etc. are all scarce resources, and as such need to be allocated
efficiently. This is generally achieved through rationing of access.
However, the specific instance of a knowledge commons is characterized
by some markedly different and contrasting traits. In fact, though much of
the above theory still holds true, knowledge is a special type of collective
resource with a peculiar feature that substantially alters the economic
evaluation - namely, the presence of increasing returns to number of users.6
This specific attribute depends, as we shall see, on the inherently collective
nature of knowledge, which means that an excessive rationing of access
through copyright can, ceteris paribus, produce an inverse tragedy to that
prefigured by Hardin -essentially an 'opposite tragedy of the creative
commons'. In the remainder of this work we shall therefore consider refer-
ences to knowledge as an input- which exists within collective contexts -
as corresponding to the concept of 'creative commons'. 7
Naturally, all the above does not seek to negate the economic method,which retains its validity; it merely challenges the results of scientific work
that fails to factor in the peculiar characteristics of creative commons, and
uncritically applies the hypothesis of diminishing returns. Whereas the
124 Developments in the economics of copyright
novel feature of the context under study concerns precisely the traditional
concept of scarcity, which as we have seen is no longer brought about by
over-exploitation of the collective resource, but rather by its under-
exploitation.8
The Social Nature of Knowledge Production
This important difference in the nature of knowledge is determined by its
social provenance and the sharing process that unavoidably characterizes
its existence. Knowledge in fact belongs intrinsically to the collective con-
texts where it is created. It is brought to fruition in the symbolic sphere
defined by society and renewed through sharing, which is thus an indis-
pensable feature (one might even say it constitutes an essential and not sub-
stitutable input in its own right) for individual creative activities.
Generally speaking, there is a widely held consensus, within the discip-
lines that study creative and inventive activities, that 'intellectual activity in
general is not ex nihilo creation. Given this vital dependence of a person's
thoughts on the ideas of those who came before him, [consequently] intel-
lectual products are fundamentally social products' (Hettinger 1989,
p. 38).'
Note that, in the domains of scientific and technological knowledge,even mainstream economic theory has explicitly adopted the thesis set forth
by Suzanne Scotchmer (1991, p. 29) in a celebrated article: 'Most innov-
ators stand on the shoulders of giants, and never more so than in the
current evolution of high technologies, where almost all technical progress
builds on a foundation provided by earlier innovators'. This view is sup-
ported by the observation that innovation 'consists to a substantial extent
of a recombination of conceptual and physical materials that were previ-
ously in existence' (Nelson and Winter, 1982, p. 130).10
It is no coincidence that the generic term denoting creative activity is
'composition' (we speak in effect of literary composition, musical compos-
ition, etc.) whose Latin etymology specifically invokes the above-described
process: 'cumponere' means literally 'to put together', suggesting that cre-
ation is first and foremost a novel arrangement of existing elements.
However, this viewpoint only seizes upon the incremental aspect of the
production of ideas,11 whereas there is also a non-hierarchical dimension to
consider, namely that of the sharing process which engenders the cognitive
value of knowledge. This is closely bound up with the semantic nature of
knowledge, which originates and exists exclusively within collective con-
texts. Paraphrasing Geertz (1973, p. 11), knowledge 'is public because
meaning is'.12 Thus there can be no knowledge without meaning. And,
accordingly, there can be no meaning without a human group to share it.
Private appropriability and sharing of knowledge 1 25
Therefore, even though knowledge fragments are often created by
individuals, this can only happen embedded within the broader context of
the collective semantic space to which the knowledge fragments are inextric-
ably tied.
Anthropologists (Romney, 1999, p. 104) provide us with some further
insight into this idea:
knowledge, found mostly in humans, arises from human inventions, is learned
and handed down from one generation to the next, and usually varies from one
society to another ... it is shared among relevant participants and ... it is
learned as part of our social heritage. The word 'relevant' alerts us to the idea
that there may be small specialized subgroups, such as medical practitioners,
whose members share esoteric knowledge not possessed by the wider cultural
group. In short, careful reflection reveals that the very notion of [knowledge]involves sharing of ideas, concepts, behaviours, etc., by more than one person.
In other words, citing again the enlightening definition formulated byWeber and more recently by Geertz (1973, p. 5), we can say that 'man is an
animal suspended in webs of significance that he has himself spun', where
knowledge constitutes the webs that connect him to other individuals. This
definition will also aid in understanding the following section.
Note that the above idea can be easily transposed into the categories of
economics, by thinking of it as an extreme instance of network externalities;
the semantic, and hence the cognitive and the resultant economic value of
knowledge or its fragments (as copyrighted works) is, at least to a certain
extent, an increasing function of the numbers of individuals that share it.13
Hence, if creative commons are deprived of knowledge sharing- not just
in the reductive sense of static shared consumption of information goods,which is normally termed 'information sharing' and sends back just to the
public goods market failure, but rather the sharing of signs, symbols, cre-
ative processes and, in short, all those elements which constitute knowledgein its broadest sense - then they will inevitably dry up in terms of meaningand, by consequence, of value to individuals. 14
It follows that a necessary existence condition for a knowledge creative
commons is the possibility of sharing and it presents value as an output byreason of this feature. Let us look at some examples of this. Consider,
for instance, a simple fragment of knowledge as a literary text or book.
It is attributed value by individuals - who consequently have a certain
willingness-to-pay for its purchase- due to its profound relation with the
social context in which the individuals live and have developed. A text
written in an incomprehensible or invented language would not have any
meaning for the reader, nor any cognitive utility. A similar observation can
be made at the semiotic level: a collection of words which have no referent
126 Developments in the economics of copyright
in a shared symbolic universe would have the same effect upon the reader
as a text in an unknown language. It cannot be understood, and therefore
has no meaning and no value. 15 This argument can be applied to all the
semantic dimensions of literature.
If we now move on from text to sounds, the above reasoning proves just
as valid: musical language is related to the social context which produces it,
and has value for the individuals as a result of the social references which
it creates at various levels. This is true for both the syntax and the distinc-
tion made between 'sound' and 'noise'. In fact, to a listener from one
cultural context, the music of another culture may sound tuneless, incom-
prehensible or even distasteful. In this case too, not-sharing of the seman-
tic content conveyed through a musical piece translates into a feeling of
strangeness, or even aversion, in the listener (note that in this case seman-
tic does not necessarily mean textual, but rather functional; in other words,
relating to some social function, for example recreational, religious, hedon-
istic, etc.).16
It is worth noting that a growing body of work in economic theory is
today acknowledging the collective and networked dimension of know-
ledge, albeit in an unconscious and indirect form that focuses on the pro-
duction of specific goods, as in the case of industrial clusters (Koepp, 2002;
David and Foray, 2002), as well as scientific and technological knowledge.With respect to the latter, for example, Nelson and Nelson (2002, p. 732)
observe that 'to students of the subject, there is no escaping the fact that,
while particular individuals may know different things, what they know
they have drawn largely from the culture in which they live, and to be used
effectively that know-how often needs to be part of a coordinated activity
involving several persons. And while particular individuals play a key role
in the advance of knowledge, they do so from a largely common platformof know-how, and a large part of what they have achieved sooner or later
becomes part of the knowledge shared among professionals in the field'.
It should be pointed out, again with reference to the traditional eco-
nomic categories, that sharing therefore plays a twofold role in the produc-tion process, first as a constituent and inalienable feature of knowledge as
an input (no sharing= no meaning = no knowledge), and second because
it brings the added benefit of reducing the transaction costs of interper-
sonal relations by creating a common ground in which relations can more
easily and quickly be conducted. This is tantamount to opening a commu-nication channel that enables the various individual members of social
groups to engage in dialogue and learn more easily.17
The above arguments highlight an additional feature of creative
commons, which is that they constitute a social locus rather than a physical
one as in the case of ordinary commons (i.e. a hub for interactions between
Private appropriability and sharing of knowledge 1 27
individuals).18 To put it another way, the creative commons largely exists
embedded within the sharing process and in association with a social
grouping; whereas ordinary common resources are spatially identifiable,
separated by individuals and divisible into finite portions whose sum will
correspond to the original amount. An uncultivated pasture, for example,
occupies a physical location, exists irrespective of its potential users, and
can be parcelled into smaller lots, each of which is subject to specific prop-
erty rights and whose sum will always add up to the original resource.
So a creative commons is a special case, as we have seen from the above
arguments, which exists as a shared entity, inextricably bound up with the
existence of the individuals that make use of it, and can only partially be
broken up into fragments (i.e. expressed in tangible media) whose sum will
therefore not be equal to the whole. These characteristics are discussed
further in the next section.
7.3 THE ATOMIZATION PROCESS
This failure to appreciate the peculiar features of creative commons is due,
to a large extent, to the fragmentary view of knowledge taken by the con-
tributions dealing with the economic analysis of copyright, though not bythe economic theory tout court, as we shall see below. In fact, the represen-
tation implicitly adopted by this literature is that, even if each copyrightedwork constitutes a public good in itself, it is also, from a productive stand-
point, a well-defined entity clearly separable from the others. The incentive
mechanism is consequently designed to act on individual units - n times n
copyrightable works - in order to stimulate the creation of an optimalnumber of new expressions of ideas. In other words, the production processis atomized, with each creator likened to a standard production function,
acting individually and in isolation. The problem then becomes, at most,one of guaranteeing access to previous copyrights to be used as static
inputs, in order to enable an efficient number of creators/production
processes to function and produce their output. However such a represen-
tation is inconsistent with the true nature of knowledge, and fails to take
into account the productive role of the creative commons, which taken as
a whole constitutes an input in its own right.
If we instead adopt a more correct analytical perspective, which takes
into account the productive and constituent functions of the sharing
process within the knowledge domain, we find that rationing of access to
individual information items through copyright will in reality have muchmore far-reaching effects than those traditionally envisaged by the conven-
tional economics literature.
128 Developments in the economics of copyright
The Dyadic Structure of Knowledge
The arguments presented in the previous section, on the whole, bear out the
unavoidably (or inherently) collective nature of knowledge, and the social
indebtedness and embeddedness of the knowledge, or the commodified
fragments thereof, that are ultimately transformed into copyrighted works.
This knowledge therefore possesses productive potential only by virtue of
being created within, and connected with, the broader sphere of knowledge
sharing.
This is consistent with the observation, long consolidated in the other
social sciences, that 'cultural creation cannot be totally integrated into an
industrial production system' (Morin 1962, p. 25).19 In this connection we
notice how many 'creative' industries, unable to fully incorporate creative
activities into the organization, rely upon continual monitoring of the
social context for new ideas and knowledge that can subsequently be inte-
grated into their new products.20
Now, the analytical weakness in the economic theory of copyright lies in
its failure to recognize this significant feature, adopting instead an 'atom-
ized, undersocialized conception of human action' which 'disallows by
hypothesis any impact of social structure and social relations on produc-
tion, distribution, or consumption' (Granovetter 1985, p. 483). In particu-
lar, this procedure of atomization, according to its critics, has the effect of
cancelling from view all that cannot be directly attributed to single indi-
viduals, and which resides instead in the relationships between them. If
such a simplification can have distorting effects on the economic and social
analysis in general, as some scholars maintain, the misrepresentation
becomes dramatic in the knowledge domain, where it can significantly alter
the outcomes and the attendant policy design.21
Once again, it is the nature of knowledge that justifies this assertion. Anextensive body of scientific literature, including economics, has in fact
pointed out the peculiar structure of knowledge, which simultaneously
draws from and feeds back into the individual and social dimensions.
However this peculiar feature is, as we have seen, entirely discounted by the
traditional economic analysis of copyright and intellectual property rights.
In fact, in all the contributions of that literature, the treatment of infor-
mation as an output which can later serve as an input is based on the
assumption that knowledge corresponds to the sum of information items
produced by creative processes, viewed as discrete entities. 22 Hence, accord-
ing to this assumption, a proper incentive will produce the optimal level of
information, and therefore of knowledge.However such a simplification does not accurately reflect the reality,
because knowledge actually has a different and dyadic structure, made up
Private appropriability and sharing of knowledge 1 29
of two distinct but complementary components. On one side, there is the
so-called 'codified knowledge', which is able to be articulated and encoded;but on the other side there is also an immanent form of knowledge, not
clearly articulable or codifiable, but which is nevertheless communicated
and therefore exists within relations between individuals, consequentlytermed 'tacit knowledge'.
23
The former is what we usually term information (i.e., a fragment of
knowledge) encoded in any one of the various media made available by
society (and which are themselves a subset of knowledge) such as language,
writing, reproduction technologies and so forth. Because it can be encoded,it is then also able to be 'commodified', giving rise to markets of informa-
tion goods, such as the markets for copyrighted works. On the other hand,the latter (tacit) form of knowledge is dynamic in nature, and as such can
never be entirely encoded or commodified. Consequently, as we shall see
below, it cannot be produced and exchanged on the markets, but dependsinstead upon interpersonal relationships for its production and dissemina-
tion. In effect, tacit knowledge grows out of the dynamic sphere of the com-munication process, whose existence is a necessary precondition for
knowledge to exist, as well as for the sustenance of human groupings.24
Therefore, in view of the multidimensional nature of knowledge, as
opposed to the one-dimensional nature of information, we can say that the
knowledge set encompasses the information set, so that the sum of all infor-
mation is only a subset of knowledge.25
Economic theory, with the exception, of course, of that applied to intel-
lectual property, has long recognized the dyadic nature of knowledge.
Beginning in the middle of the last century, von Hayek (1945, pp. 521-2),
for example, contrasted the portion of scientific knowledge that can be
encoded with 'unorganized knowledge . . .,the knowledge of the particu-
lar circumstances of time and place' which is essentially inalienable anddemands the participation of several individuals in order to be exploited.
26
In more recent times, an analogous concept has been again advanced in
economic theory by Nelson and Winter (1982, p. 1 34) for the case of a pro-
ductive organization (i.e., a relevant group of individuals such as a firm)
where in addition to the knowledge owned by individual workers, there
exists a tacit knowledge, embodied within the organization itself and its
procedures, that is 'not consciously known or articulable by anyone in par-
ticular', but effective and crucial for the functioning of the organization.
Transmission of this latter, tacit form of knowledge can only take place,
as we have said, within a context of social interaction and sharing. Theabove argument explains, for example, the rationale behind the educational
systems of different societies: books are generally used to provide informa-
tion, but the presence of a teacher is necessary to impart learning, and in
1 30 Developments in the economics of copyright
general to communicate those aspects of knowledge which cannot be stat-
ically encoded. Accordingly, Nelson (2003, p. 917) asserts that any 'class-
room equipment ... [is a] complement not a substitute for an effective
teacher working with students'. What is missing, in fact, is the imparting of
that tacit knowledge which books alone are unable to convey.27
In the light of this perspective, we can reinterpret all those institutions
and events that bring individuals together, for scientific or cultural pur-
poses, as serving to promote, in more or less stable form, the dissemination
and circulation of tacit knowledge which could not be accomplished by anyother means.
Enforcing Private Rights and Financing Social Activities: Resolving the
Apparent Ambiguity
If public policies in support of the production and dissemination of know-
ledge are to be effective, they must take both its tacit and codified dimen-
sions into account. However, the distinction between these two components,
though possible at the theoretical level, is in practice liable to be misleadingbecause the two are not so easily divided. The decoding of codified know-
ledge generally relies upon possession of its complementary component,which is the corresponding tacit knowledge (Rooney et al., 2003).
28
Therefore, any action exerted on one component may have indirect
effects on the other, while the availability of one does not guarantee acqui-
sition of the other, and hence of knowledge in its entirety. This assertion is
backed up by the findings of various studies on technology transfer. These
studies have underlined the difficulty of transferring the tacit portion of
knowledge as compared with its codified portion, precisely due to the
necessity of moving not just physical goods but also individuals (i.e., estab-
lishing ad hoc social relations). At the same time, these studies have pointedout the unavoidable necessity of promoting both types of transfers, if the
policy undertaken is to have a favourable outcome.29
Now, from this perspective too, the economic theory of intellectual
property displays some analytical flaws that need to be addressed. On one
side, these rights appear effectively able to stimulate the production and
exchange of optimal levels of new information items; they in fact avert a
market failure, which would seem to imply that the market, with the appro-
priate legal correctives, is a sufficient instrument for achieving an efficient
equilibrium in the knowledge domain. But, notwithstanding this, private
and public institutions are not content to rely upon the market for optimumknowledge production, and instead finance significant numbers of collect-
ive initiatives which, in a truly efficient market, would have no real reason
to exist.
Private appropriability and sharing of knowledge 131
Within the scientific community, for example, conferences, workshopsand the like are considered to be indispensable to the work and intellectual
growth of scholars, and for this reason are allocated public or private
funding, may sometimes benefit from tax subsidies, etc. In the same vein,
all the governments of countries with sufficiently developed economies sub-
sidize universities, research centres and other such institutions, in the firm
belief that this expenditure will contribute to the nation's technological
advancement and competitiveness.30 We can therefore say that, alongside
the support provided by property systems, there generally exists a public
and private attentiveness toward activities which promote interaction
between individual members of specific scientific and cultural groups.
At first glance, such an observation might seem at odds with the trad-
itional economic analysis of copyright: if intellectual property rights are
able to secure efficient results in the domain of the production and
exchange of new knowledge, the remaining activities are in effect not only
superfluous, but also constitute a needless waste of production resources.
This apparent schizophrenia, which on the one hand causes private
rights to be instituted, whilst on the other promotes and funds socializing
activities, can instead be resolved precisely in light of the twofold nature of
knowledge. The private appropriability, guaranteed ex lege and which
underpins the reward mechanisms of copyright and intellectual property
rights in general, is only effective with respect to codified knowledge, and
can at best stimulate the production of an efficient level of new items of
information. It functions optimally in markets where it is possible to
produce and exchange items comparable in nature to private goods, such as
those derived from codified knowledge. However, this solution is unable to
also provide a correct incentive for the production and exchange of tacit
knowledge, whose existence relies upon non-market substrates connected
with the process of communication and sharing; in other words, on social-
izing frameworks. So, public and private organizations quite correctly
respond to this need by promoting activities designed to bring together
groups of individuals in a more or less stable capacity.
The two different policies are in effect complementary, and work jointly
towards achieving a common aim. However, it is essential to strike a properbalance in combining these two aspects
- and it is this balance, which the
original copyright protection laws appear to have achieved, which is grad-
ually being eroded by maximal copyright.
For example, consider again educational processes. As mentioned earlier,
these rely on a mix of codified knowledge (i.e., teaching materials such as
textbooks, etc.) and tacit knowledge (i.e., direct teaching), which exist, so
to speak, in a symbiotic relationship, with neither of the two components
being able to fully substitute for the other. As a result, the non-availability
132 Developments in the economics of copyright
of the one will have the indirect effect of also preventing transmission of
the other - it is not possible to teach 'in practice' something to which access
is precluded. To put it another way, restricting access to a book will also
curtail opportunities for circulation of the tacit knowledge indissolubly
tied to that book and its content. 31Naturally, this effect will be magnified
in direct proportion to the degree of appropriation permitted by the law.
Once again, these arguments are backed up by contributions from the
various social science disciplines. Communication studies, for example,
have shown that knowledge 'as a whole' is what enables us to assign
meaning to experience, shaping our perception of it, as well as contribut-
ing to storing the knowledge of the past and perpetuating that of the
present (McQuail, 1983). Therefore, any action on one dimension of know-
ledge will have inevitable repercussions on the other. Ultimately, rationing
of access to codified knowledge by means of copyright will generally have
more far-reaching consequences than those envisaged by the traditional
economic literature, because it also has effects on the transmission of tacit
knowledge.32 This situation can be described with the metaphor of an
opposite tragedy of the creative commons, as suggested in the preceding
paragraph.It should be emphasized that the previous argument is consistent with the
original rationale behind minimal copyright laws, which consider know-
ledge to be a public resource and put a limit on the duration of property
rights, precisely to avoid exaggerating the privatization process and thereby
impoverishing the creative commons. For the same reason, the above laws
provided for various exceptions to exclusive rights; for instance, in fair use
doctrine, when such rights were deemed to be economically inefficient from
a cost-benefit standpoint (Gordon, 1982), or to constitute an impedimentto related creative activities which imperatively demand access to a specific
copyrighted work for incremental creation, without bringing a significant
economic benefit to the holder (e.g., critical essays, scientific works).33
The structure of minimal copyright was designed to grant individual
incentives through private appropriability, but without significantly
impacting on the public dimension connected with the sharing process.
Because relations between individuals are characterized by a certain degree
of randomness, arising from the dynamic nature of human behaviours, a
weak proprietary system will not appreciably affect the sharing of know-
ledge and the productivity of creative commons; as a matter of fact, a
certain degree of entropy must always by definition characterize knowledgeand communication among human beings.
34 Provided the creative
commons system is sufficiently broad, the interference of a minimal form
of copyright- limited in time and space
- will be marginal, with only imper-
ceptible effects on human relationships and hence on tacit knowledge.
Private appropriability and sharing of knowledge 1 33
In contrast, the increasingly exclusionary effects of maximal copyright,which continues to dilate the private space at the expense of the public
sphere, are having ever more serious repercussions on human interactions,
on tacit knowledge, and thus on the creative commons as a productiveresource.
Naturally, these effects are not immediately apparent, but emerge grad-
ually over a fairly long-term time horizon. As a matter of fact, the effects
of this new legal regime have up until today been mitigated (though admit-
tedly in an improper manner) by the phenomena of piracy and private
copying, which have preserved a 'de facto'', even if not 'de jure\ form of
minimal copyright.35
Nevertheless, as the balance between tacit knowledge/codified know-
ledge gradually shifts in favour of the latter, under the effects of appropri-ation mechanisms, there will be a gradual impoverishment of the sharing
process and hence of the creative commons.One (extreme) example of such a shift in balance would be the total elim-
ination of social relationships which occurs when all the individuals
belonging to a specific society/culture die. In such an eventuality it wouldbe possible to preserve the codified knowledge (objects and relics, cultural
products and literary texts, and information of all kinds); whereas, the tacit
knowledge (i.e., those elements pertaining to interactions among individ-
uals) would be irretrievably lost, as would be the culture and its knowledgeas a whole.
7.4 BACK TO THE BEGINNING: POLICYGUIDELINES AND LOCKEAN PRESCRIPTIONS
Under the proposed representation of creative commons, the social
dimension is central to the creation and existence of knowledge, in the
sense that sharing is an integral and characterizing part of that knowledge.It follows that the same criterion of dynamic efficiency invoked by the eco-
nomic theory of copyright in justification of the right must also challengethe existence of any institutions that significantly impair the sharing
process.
This is not, however, to cast doubt on intellectual property rights across
the board, but only to question the efficiency of their present-day structure
and enforcement. There is no doubt that, in many instances, the incentive
provided by minimal copyright can truly be effective. But one can with
equal validity maintain that creative commons have throughout human
history demonstrated their dynamic efficiency, even in the absence of copy-
right, with the production of quantitatively and qualitatively elevated levels
1 34 Developments in the economics of copyright
of information and knowledge.36 A comparative analysis in this domain
therefore risks raising a difficult debate, which in any case falls outside the
scope of the present work. However, it could justifiably be argued that,
because a market of information goods does exist today- and does succeed
in regulating the behaviour of creators and allocating their energies with
respect to other markets - its total elimination might not only prove unfeas-
ible, but also possibly endanger social stability. For the present purposes,
therefore, it is sufficient to treat copyright as a given- an institutional
outcome of western society, which has the attendant virtue of potentially
providing an incentive for the production of new expressions of ideas. The
real issue, though, is to examine how its extension might interfere with the
endogenous dynamics of creative contexts, and in this light attempt to for-
mulate appropriate policy indications.
One additional interesting test of the soundness of the above conclusions
is to see whether they are consistent and compatible with the totemic para-
digm formulated by John Locke (1690), which has been widely relied upon
by the literature for the justification of property rights (Drahos, 1996). The
reference is particularly apposite because this part of Locke's work focused
precisely on the problem of fairly regulating exploitation of the commons,so as to safeguard both the original resource and private appropriability,
defining rights and limits accordingly.
Now one might say that the solution proposed by the philosopher seems
almost to prefigure, at least from an instrumental perspective, the efficiency
objectives prescribed by modern economic theory, and that it is consistent
with the arguments presented here. In fact, alongside the legitimization of
property rights, the English philosopher formulated two complementary
provisos designed to simultaneously guarantee preservation and access to
the commons.
The first clause, the so-called 'sufficiency proviso', recognizes private
appropriability in reason of the individual contribution, but only admits
appropriation 'at least where there is enough and as good left for others'
(Locke, 1690, section 27). By so doing it places a first, weak limit upon indi-
vidual rights, acknowledging in practice the original collective ownershipof any common resource, and the precedence which this takes over private
appropriability.
The second clause, contained in the so-called 'spoilage proviso', reasserts
with greater force the necessity of preserving the integrity of commonresources, clearly calling for the limitation of property rights when these
can have significant effects leading to the impoverishment and depletion of
said resources. 'The same law of nature, that does by this means give us
property,' Locke wrote,'does also bound that property too. . . . Nothingwas made by God for man to spoil or destroy' (Locke, 1690, section 27).
Private appropriability and sharing of knowledge 1 35
An appropriate transposition of the above two criteria to the context of
copyright confirms the guidelines suggested by our analysis: a minimal
copyright which remunerates individual contributions can be acceptable
and useful for dynamic efficiency, but by this same criterion, given the pecu-
liar architecture of creative commons, it must leave 'enough and as goodfor the others', carefully avoiding their depletion. It follows that a transi-
tion to a maximal form of copyright, which significantly impairs the know-
ledge sharing process, is not to be desired.
In stark contraposition to the above, the recent international legislative
and jurisprudential trends (which should in theory be informed by the
Lockean categories) have instead been moving toward a strengthening of
exclusive rights through the extension of their duration (see US CopyrightTerm Extension Act 1998 and EU Duration Directive, 93/98/EEC),
through the international extension of the rights and their enforcement (see
TRIPs Agreements), through their extension to new subject matters
(see the EU Directive 96/9/EEC on databases), and through a progressive
erosion of the scope for derogation from the exclusive right (i.e., fair use
doctrine and analogous doctrines in the various national legislations).
In light of the preceding discussion, the present-day trend toward boost-
ing the exclusionary effect, without any counterbalancing correctives, runs
the risk of significantly interfering with the sharing process, and could in
the long run bring about conditions of social inefficiency, both in the
markets of ideas and for the creation of knowledge. What is more, because
sharing and communication are of such crucial importance to human
groupings,37 who may react in various ways to interference in these spheres,
society may well also incur additional and neglected costs of social and
legal conflicts. 38
7.5 CONCLUSIONS
Copyrighted works consist of signs which acquire significance by virtue of
belonging and referring to a broader semantic universe - the knowledgedomain - which is by its nature collective. Even when knowledge fragmentsare able to be encoded and occasionally transformed into commodities (i.e.,
codified knowledge) they remain indissolubly linked to that componentwhich exists and is reproduced within the sphere of human relationships
(i.e., the tacit knowledge).These two dimensions can never be entirely separated, and are equally
necessary for the transmission and production of knowledge. So given this
perspective, and even allowing due importance to individual contributions,
knowledge production emerges as being inextricably embedded within
1 36 Developments in the economics of copyright
social systems, with the locus of its production being the creative
commons - a special case of common resource which has the peculiar and
distinguishing trait of providing increasing returns with increasing
numbers of users, thus implying a pivotal role of the sharing process.
In other words, the collective setting is the only possible ecosystem for
the creation of knowledge, and as such needs to be preserved. The alterna-
tive hypothesis, rooted in a romantic view of the artist creating in splendid
isolation from the world, is not backed up by the history of art, nor by the
sciences that study knowledge and its production.
In general, the economic analysis of copyright (though not the economic
theory as a whole) falters in its handling of the above-described dynamic,
because application of its method relies upon a representation of reality
which simply does not hold in the case of knowledge. Nevertheless, this
analytical misrepresentation has led some of the literature to justify and
propose policies that are not applicable to the knowledge domain, and
which may have the unintended consequence of impoverishing knowledgeand creative commons.
The original architecture of copyright, which we have here termed
minimal copyright, appeared to take into account the peculiar features of
the knowledge domain, and was accordingly designed as a bounded prop-
erty right in order to avoid any significant interference with the sharing
process. History has also testified to the possibility of peaceful coexistence
between creative commons and bounded proprietary systems, provided the
latter do not significantly compromise the sharing process and hence, the
efficiency of knowledge production.
Today, however, the indiscriminate extension of copyright at the expense
of the social dimension of knowledge, which we have termed maximal
copyright, is liable instead to have the undesired effect of depleting the cre-
ative commons, to the detriment of knowledge.Given this perspective, it would therefore be socially desirable, as a
general rule, to avoid any policy designed to bolster private appropriability
at the expense of knowledge sharing, and instead maintain appropriability
at a minimal level and adopt (or at least preserve the existing) criteria for
derogation from strong exclusive rights that move in the direction of assur-
ing adequate communication flows.
This is not to imply a negation of the usefulness of copyright and intel-
lectual property in general in stimulating individual creation, but rather to
acknowledge that they play an ancillary role, not with respect to the
market, but certainly with respect to the production of knowledge and
culture.
Private appropriability and sharing of knowledge 1 37
ACKNOWLEDGMENT
This chapter was originally drafted during a visiting scholarship at SIMS, University of
California, Berkeley. I would like to thank Hal Varian for his kind hospitality there, and Jim
Stockinger for having introduced me, on that occasion, to the literature on the sociology of
culture. I am additionally indebted to a number of colleagues and friends for their helpful
comments, suggestions and criticisms, and in particular Jiirgen Backhaus, Alberto Cassone,Alberto Chilosi, Michael Einhorn, Angela Fraschini, Wendy Gordon, Dennis Khong, AnnaMaffioletti, Carla Marchese, Alain Marciano, Federico Revelli, Francesco Silva, Frank
Stephen, Ruth Towse, and an anonymous referee. I also wish to thank all the participants of
SERCIAC 2003 in Northampton, MA, the SIEP Conference 2003 in Pavia, the Alessandria
MIUR Workshop 2003 and the Erfurt Workshop in Law and Economics 2004. The usual dis-
claimers apply. This work was supported by the MIUR research grant on The Governance of
Intellectual Property'.
NOTES
1 . The adjective 'minimal' is a concise way to describe the nature of copyright in its ori-
ginal form -i.e., a heavily bounded property right designed to take into consideration
and pursue the public interest in the broadest sense, according to the precepts of the
standard theory of property and intellectual property (see Drahos 1996).2. However, this is not a particularly useful argument in favour of the efficiency of copy-
right, because human history has also shown that creative activities flourish equally in
the absence of copyright.3. See, for example, Posner (1992). On fair use doctrine, see Gordon (1982).4. On the subject of inventive activities in the patent domain, Arrow (1962, p. 129) seems
to have also intuited this peculiarity, warning that the production of knowledge can con-
stitute not only a productive output, but also an input for other inventive activities. If
this is the case, Arrow observes, the private appropriation procured by intellectual prop-
erty rights may seriously compromise the incremental accrual of knowledge and, conse-
quently, the collective well-being. Unluckily, here again, the observation is not further
followed up in his analysis.
5. In addition to the general contribution of Hardin (1 968), see for example, in economics,
Dasgupta and Heal (1979) and Comes and Sandier (1983).
6. A similar hypothesis in philosophical debate is advanced by Drahos (1996).7. It should be pointed out that there is also an opinion movement called 'creative
commons', whose aims are to promote access to knowledge and limit the exclusion
brought about by copyright through the introduction of novel contractual forms (see
http://creativecommons.org/). Now, although the working goals of this movement are
certainly consistent with the arguments presented here, there is no specific ideologicalconnection.
8. This idea is also supported by some contributions in the legal literature (see Rose, 1986)
dealing with public policy aspects (Rooney et al., 2003).9. This thesis finds wide support today in the economics literature, in Landes and Posner
(1989), Scotchmer (1991) and Foray (2000). Note that sociology, starting with the
contributions of Max Weber (1967) and Merton (1973), has remarked on the collective
dimension of inventive activities.
10. This assertion is also consistent with the Schumpeterian theory of innovation putforward in The Theory of Economic Development (1934).
1 1 . This is in essence also recognized by Landes and Posner ( 1 989).12. Note that Geertz, and anthropologists in general, often refer to the knowledge of a
specific social group as 'culture'. In this chapter we shall always be using the term
'knowledge'.
1 38 Developments in the economics of copyright
1 3. On network externalities see for example Liebowitz and Margolis (1 998). For the appli-cation of network externalities to the knowledge domain with specific reference to the
value of a language, see Church and King (1993).
14. The assertion is supported by growing consensus in contributions dealing with the eco-
nomic aspects of knowledge. Rooney et al. (2003, p. 8), for example, observe that '[a]
connected characteristic of knowledge is that it tends to grow through sharing. As
people exchange ideas in conversation the parties to the conversation tend to increase
their knowledge and in all likelihood create new knowledge. This is particularly so in the
case where ideas are shared between people with complementary knowledge working in
related areas. Furthermore, as knowledge diffuses its social value tends to be enhanced'.
Drahos (1996) also agrees on the peculiarity of creative commons.15. It could of course be appreciated for some other physical characteristic, but in this case
would no longer be an information good.16. On these topics see for instance, Schafer (1977).
17. This, in essence, is analogous to establishing a 'common language' (see Nahapier and
Ghoshal, 1998).
18. Of course this does not mean that the sharing process cannot take place in a physical
space. Universities and research institutions, for example, are physical spaces set up pre-
cisely to encourage the development of creative commons, but they are naturally not cre-
ative commons merely by virtue of being spaces.
19. Some authors, dealing for the most part with innovation, have similarly argued that the
production of knowledge does not follow a conventional linear production process, but
instead exhibits a non-linear and collaborative dynamic that is 'not well understood bypolicymakers' (Rooney et al., 2003, p. 69). Cowan et al. (2000), in emphasizing the speci-
ficity of knowledge production, also speak of an 'enculturation model'.
20. This is the case in the fashion industry, but also in the music industry where the majors
generally profit from external Schumpeterian innovators more closely tied to the social
context (see Silva and Ramello, 2000).21 . Note that, though it was subsequently forgotten, this is a criticism that was first levelled
against the traditional economic approach as early as the middle of the twentieth
century. Paul Sweezy, in 1942 (p. 3), already observed that: 'Society is more than a
number of individuals. It is a number of individuals among whom certain definite andmore or less stable relations exist. The form of society is determined by the character andform of these relations'.
22. In this connection, Cowan et al. (2000, p. 217) point out that '[fjor some considerable
time, economists took little if any interest in the question of separating the notion of
knowledge from their idea of information'.
23. For an overview, see Cowan et al. (2000). The fact that tacit knowledge is physically held
by single individuals does not negate its social nature. In effect, because there is no such
thing as a physical entity called 'social relationships', it follows that individuals must be
the 'bearers' of this type of knowledge. But the expression and transmission of tacit
knowledge nevertheless requires interaction between individuals to take place.24. This statement is consistent with the ideas in the literature on communication theory (see
for example, for the 'Palo Alto' school, Watzlawick et al. 1967).
25. The above-described dichotomy between knowledge and information, which attributes
to the former a social dimension that cannot be completely owned by a single individ-
ual, is widely accepted in the social sciences (see Geertz, 1973; but also Weber, 1967, and
Durkheim, 1973).
26. This theory was subsequently taken up, among others, by Polanyi (1958, 1966), Cowanet al. (2000), David and Foray (2002), and Nelson (2003).
27. Nelson and Winter (1982, p. 78) also argue that 'a trait that distinguishes a good instruc-
tor is the ability to discover introspectively, and then articulate for the student, much of
the knowledge that ordinarily remains tacit'.
28. For example, the execution of a musical score requires the ability to read and interpret
it, as well as to perform it on the instrument (Rooney et al., 2003).29. See for example Williams and Gibson (1 990) and Takii (2004).
Private appropriability and sharing of knowledge 1 39
30. In general, looking at OECD data, we can easily observe that, in developed countries,
the public sector funds at least one-third of R&D expenditure. Additional public con-
tributions then go to educational systems and cultural activities.
31. Consider, in this connection, a simple example: the preparation of a dish involves two
components: the written recipe (codified knowledge) and the practical ability to execute
it (tacit knowledge). If the opportunity to cook a dish is precluded- in other words, if
its recipe is appropriated in an exclusive manner - this will also preclude handover of the
technical ability to execute it, because the cook who knows how to make it will not be at
liberty to teach it.
32. Cowan et al. (2000, p. 225): 'But information written in a code can only perform those
functions when people are able to interpret the code'. So we are once again referred back
to the social dimension.
33. Note that, in this case too, the law appears to take an instrumental view of copyright. It
therefore provides for derogations designed to avoid static inefficiency (negative balance
of welfare without a significantly increased incentive for the copyright holder) or
dynamic inefficiency (endangering certain incremental creations, e.g., satire which plays
an important social role in warranting freedom of expression).
34. Some relatively recent scientific work dealing with social networks has noted how human
relationships are characterized by an exogenous degree of instability, and that this state
of affairs promotes efficient results (see Jackson and Wolinsky, 1996; Dutta and
Mutuswami, 1997).
35. For a taxonomy of unauthorized reproduction, see for example Silva and Ramello
(2000).
36. For western society, see Weiner (2000). For China, see the intriguing book by Alford
(1995).
37. See note 25 above.
38. Consider for example the Napster case, which, albeit indirectly, involved university stu-
dents, the future intelligentsia of US society, and high school students. If, in a similar
case in the future, a massive law enforcement action were to directly target these cat-
egories of infringers, it could lead to unsustainable social consequences.
REFERENCES
Alford, William P. (1995), To Steal a Book is an Elegant Offense, Stanford, CA:Stanford University Press.
Arrow, Kenneth J. (1962), 'Economic welfare and the allocation of resources for
invention', in NBER, The Rate and Direction of Incentive Activity: Economic andSocial Factors, Princeton, NJ: Princeton University Press, pp. 609-24.
Church, J. and I. King (1993), 'Bilingualism and network externalities', Canadian
Journal of Economics, 26, 337-45.
Cornes, R. and T. Sandier (1983), 'On commons and tragedies', American Economic
Review, 73, 787-92.
Cowan, R., PA. David and D. Foray (2000), 'The explicit economics of knowledgecodification and tacitness', Industrial and Corporate Change, 9, 21 1-53.
Dasgupta, Parta and Geoffrey Heal (1979), Economic Theory and Exhaustible
Resources, Cambridge: Cambridge University Press.
David, PA. and D. Foray (2002), An introduction to the economy of the knowledge
society', International Social Science Journal, 54, 9-23.
Drahos, Peter (1996), A Philosophy ofIntellectual Property, Dartmouth: Aldershot.
Durkheim, Emile (1973), On Morality and Society: Selected Writings, Chicago andLondon: University of Chicago Press.
140 Developments in the economics of copyright
Dutta, B. and S. Mutuswami (1997), 'Stable networks', Journal ofEconomic Theory,
76, 322^4.
Foray, Dominique (2000), L'economie de la connaissance, Paris: La Decouverte.
Geertz, Clifford (1973), The Interpretation of Cultures, New York: Basic Books.
Gordon, WJ. (1982), 'Fair use as market failure: A structural and economic analy-
sis of the Betamax case and its predecessors', Columbia Law Review, 82, 1600-57.
Granovetter, M. (1985), 'Economic action and social structure: The problem of
embeddedness', American Journal of Sociology, 91, 481-510.
Hardin, G. (1968), 'The tragedy of the commons', Science, 162, 1243-8.
Hayek, F. von. (1945), 'The use of knowledge in society', American Economic
Review, 35, 519-30.
Hettinger, B.C. (1989), 'Justifying intellectual property', Philosophy and Public
Affairs, 18,31-52.
Hippel, Erik von (1988), The Sources of Innovation, Oxford and New York: Oxford
University Press.
Jackson, M.O. and A.Wolinsky (1996), 'A strategic model of social and economic
networks', Journal of Economic Theory, 71, 4474.
Koepp, R. (2002), Clusters of Creativity: Enduring Lessons on Innovation and
Entrepreneurshipfrom Silicon Valley and Europe's Silicon Fen, Chichester: Wiley.
Landes, W.M. and R.A. Posner (1989), 'An economic analysis of copyright law',
Journal of Legal Studies, 18, 325-63.
Liebowitz, Stan J. and Stephen E. Margolis (1998), 'Network effects and external-
ities', in P. Newman (ed.), The New Palgrave's Dictionary of Economics and the
Law, vol. II, London: Macmillan, pp. 671-5.
Locke, John (1690), Second Treatise on Government, URL: http://www.ilt.colum-
bia.edu/academic/digitexts/locke/second/locke2nd.txt.
McQuail, Dennis (1983), Mass Communication Theory: An Introduction, London:
Sage.
Merton, Robert K. (1973), The Sociology of Science. Theoretical and Empirical
Investigations, Chicago and London: Chicago University Press.
Morin, Edgar (1962), L 'esprit du temps, Paris: Grasset.
Nahapier, J. and S. Ghoshal (1998), 'Social capital, intellectual capital and the
organizational advantage', Academy of Management Review, 23, 253.
Nelson, R.R. (2003), 'On the uneven evolution of human know-how', Research
Policy, 32, 909-22.
Nelson, K. and R.R. Nelson (2002), 'On the nature and evolution of human know-
how', Research Policy, 31, 719-33.
Nelson, Richard R. and Sidney G. Winter (1982), An Evolutionary Theory ofEconomic Change, Cambridge, MA: Belknap Press.
Polanyi, Michael (1958), Personal Knowledge: Towards a Post-Critical Philosophy,
London: Routledge.
Polanyi, Michael (1966), The Tacit Dimension, New York: Doubleday.
Posner, Richard A. (1992), Economic Analysis ofLaw, Toronto: Little Brown & Co.
Romney, A.K. (1999), 'Culture consensus as a statistical model', Current
Anthropology, 40, 103-15.
Rooney, David, Greg Hearn, Thomas Mandeville and Richard Joseph (2003), Public
Policy in Knowledge-Based Economies, Cheltenham, UK and Northampton,MA, USA: Edward Elgar.
Rose, C. (1986), The comedy of the commons: custom, commerce, and inherently
public property', University of Chicago Law Review, 53, 71 1-81.
Private appropriability and sharing of knowledge 141
Schafer, Murray R. (1977), The Tuning of the World, Toronto: McLelland and
Stewart.
Schumpeter, Joseph (1934), The Theory of Economic Development, Cambridge,MA: Harvard University Press.
Scotchmer, S. (1991), 'Standing on the shoulders of giants: Cumulative research
and the patent law', Journal of Economic Perspectives, 5, 29-41.
Silva, F. and G.B. Ramello (2000), 'Sound recording market: the ambiguous case of
copyright and piracy', Industrial and Corporate Change, 9, 415-42.
Sweezy, Paul M. (1942), The Theory of Capitalist Development, Oxford and NewYork: Oxford University Press.
Takii, K. (2004), 'A barrier to the diffusion of tacit knowledge', Review of
Development Economics, 8, 81-90.
Watt, Richard (2000), Copyright and Economic Theory: Friends or Foes?,
Cheltenham, UK and Northampton, MA, USA: Edward Elgar.
Watzlawick, Paul, Janet Bavelas Beavin and Don D. Jackson (1967), Pragmatics ofHuman Communication, New York: Norton.
Weber, Max (1967), // lavoro intellettuale come professione, Einaudi, Turin (origi-
nally published as Wissenschaft als Beruf).
Weiner, Robert P. (2000), Creativity and Beyond, New York: State University of
New York Press.
Williams, Frederick and David V. Gibson, (eds) (1990), Technology Transfer: ACommunication Perspective, Newbury Park, CA: Sage.
8. IMS Health or the questionwhether copyright still deserves
a specific approach in a market
economy?Alessandra Narciso and
Paul L.C. Torremans
8.1 INTRODUCTION
The ongoing IMS Health saga has yet again highlighted the apparentclash between copyright and competition law and its essential facilities
doctrine. The IMS Health case is a complex one that is still unfoldingbefore the EU Court of Justice. IMS is a major supplier of marketingdata to pharmaceutical and other healthcare companies. In the German
market, it has established a structure of local geographic segments, 1860
in total, called bricks, each containing a comparable number of phar-macies. Such a structure enables the collection of standardized data
without violating data protection laws that prohibit the identification of
individual pharmacies' data. That structure is protected by copyrightunder German copyright law. The brick structure has de facto becomethe industry standard and IMS's competitors had asked for it to be
licensed to them in order for them to be able to compete. The grant of a
licence was refused, which subsequently prompted IMS's competitors to
file a complaint with the European Commission for abuse of dominant
position.
The facts and developments in the IMS Health case will be introduced
subsequently, as will some of the detailed legal issues that arise in the case,
with a focus on the essential facilities doctrine. However, the detailed issues
need to be seen against the background of the apparent clash between
copyright and competition law.
142
IMS Health or does copyright deserve a specific approach? 143
8.2 COPYRIGHT VS. COMPETITION LAW
Copyright
The starting points
Two leading studies on the economic analysis of copyright support the con-
tinued existence of copyright. The starting point indicates therefore that the
debate should not develop along the lines copyright or competition law, but
should rather focus on the role played by both sets of rules in a market
economy (Lemarchand et al., 2003, p. 17).
Two studies highlight the conundrum we are trying to solve, the solution
of which we believe should underpin any detailed ruling in cases such
as IMS. In their study, Cooter and Ulen (1988, p. 145) summarize it as
'[p]ut succinctly, the dilemma is that without a legal monopoly too little
of the information will be produced, but with the legal monopoly too little
of the information will be used'. Landes and Posner (1989, p. 326) empha-size the matter in even clearer terms when they state that 'Copyright
protection- the right of the copyright's owner to prevent others from
making copies- trades off the costs of limiting access to a work against the
benefits of providing incentives to create the work in the first place. Striking
the correct balance between access and incentives is the central problem in
copyright law.'
Historically, copyright is concerned with the protection of literary and
artistic works such as books and sculptures. It is important to note in this
respect that copyright never set out to protect ideas. What was protected
was the way in which the individual author used these ideas creatively and
the resulting expression of his creativity. Only that expression, the individ-
ual non-copied expression by the author of a certain idea was protected.
A lot of emphasis was therefore put on the link between the author and
his work (Torremans, 2001, p. 176).
Justifying copyright
This approach can in part be justified by means of John Locke's labour
theory, which is in essence the combination of two concepts. According to
the first concept, everyone has a property right in the labour of his own
body and brain; the second concept builds on that by adding that the appli-
cation of human labour to an unowned object gives whoever applies that
labour to it a property right in the previously unowned object. If one
applies this to copyright (Nozick, 1974, pp. 181-2), one can see why the
starting point of any copyright law is that the author gets the copyright in
the book or sculpture. However, the theory does not go a lot further than
the issue of the allocation of the (copy)right. It does not explain why an
144 Developments in the economics of copyright
immaterial right needs to be created and why such a right is necessarily an
object for the purposes of the theory. There is therefore a need for an add-
itional element that justifies the existence of copyright and explains why an
immaterial right in the author's individual expression of an idea needs to
be created (Spector, 1989, p. 273).
That other element is purely economic in nature (Torremans, 2001,
pp. 1421). In an efficient free market structure, competition takes place on
both the production and creative/innovation levels. As was noted above, in
the copyright sphere we are dealing with works that are the expression of
ideas. Starting from these ideas, one has to recognize that they are by their
nature public goods and can therefore freely be accessed and used by
anyone. The way these ideas enter the public domain is through their
expression by individual authors, as such expression is required for the
transmission of the ideas. From an economic point of view, it is also
important to keep in mind that such access is non-exhaustive in nature. The
consumption of the expression does not necessarily make the expressionand its material support unsuitable or unavailable for further consumption.It is also the case that in light of modern (digital) technological advances,
reproduction and distribution of the expression of ideas can be readily and
quickly achieved, and at a very nominal cost. There is therefore plenty of
room for free-riders. In the absence of copyright, the situation is therefore
entirely in favour of competition at the production level. At the innovation
and creation level, there is very little incentive to create. The creator maynot be able to recoup the cost of production, as there is no tool to reap anysubstantial benefit from such creative activity. There is therefore no efficient
market for the author's expression of ideas (Ramello, 2003, p. 8).
Copyright is the tool that is created to give authors a right in their expres-sion of ideas, hence securing for them appropriate profits deriving from the
act of creation. Copyright leads to the creation of an immaterial property
right in the expression of an idea by the author, a right which the author
can use to secure appropriate profit from his or her act of creation on the
market (Maskus, 2000, pp. 28-32). This will enhance creation by providing
incentives; therefore, competition on the innovation and creation level will
be stimulated, whilst any such right will inevitably limit competition at the
production level, as competitors are no longer free to copy the copyrightwork. A restriction on competition is put in place in furtherance of com-
petition (Lehmann, 1989, p. 12).
Competition Law
Competition law aims to maximize social welfare by avoiding or correcting
specific market failures. Consumer welfare should, as a result, be enhanced
IMS Health or does copyright deserve a specific approach? 145
and innovation should be encouraged. Competition law attempts to
achieve this objective through the elimination of anti-competitive practices
and forms of behaviour, as these lead to inefficient outcomes.
The agreement between lawyers and economists in this area clearly
covers the point that competition law is an instrument to combat monop-olistic behaviour that results in the exclusion of a segment of consumers
from the market, relative to the competitive context. But when it comes to
the practical application of these principles to the facts of a specific case,
economic theory is incapable of providing the legal community and the
courts with specific tests to detect and correct such anti-competitive behav-
iour (Neumann, 2001), especially in relation to the new information
economy. Economic theory is still largely focused on static efficiency and
the dynamics of the manufacturing industries. The information economyhas quite different dynamic processes. Since copyright operates increasingly
in such an information economy, when it is used to protect information
goods, it is evident that the application of competition law in such cases
may not be clear-cut.
Copyright vs. Competition Law
As copyright confers an exclusive right, there is by definition a possibility
that such an exclusive right will be used to restrict competition. IMS Health
shows this clearly. The exclusive right in the brick structure can be used to
restrict access to the information supply market by refusing to license the
vital elements to potential competitors. There may therefore be a conflict
between competition law and copyright law in a given situation. Copyright
preserves the exclusive right and competition law prima facie has a problemwith whatever restricts competition.
In order to understand the nature of such a potential conflict better, one
needs to see how both regulatory systems, i.e., copyright and competition
law, operate. Copyright operates at a structural level. It puts in place a
rewards and incentives structure that applies to all relevant cases and aims
to enhance competition at the creation and innovation level by granting an
exclusive right in the copyright work, which, at the same time, restricts to
some extent competition at the production level. Copyright, therefore,
defines rights. Competition law, on the other hand, operates at the behav-
ioural level and it does so on a case by case basis. It is therefore not the case
that there is a conflict between two sets of rules that operate in the same
way and at the same level (Ramello, 2003, p. 28). Copyright and competi-
tion law share the same objective, but they have different ways of achiev-
ing it and they operate at different levels (Lemarchand et al., 2003,
pp. 17-19).
146 Developments in the economics of copyright
Competition law does not operate at the structural level and as such, it
does not interfere with the structure created by copyright to enhance
creation and innovation. However, competition law may interfere at a later
stage with some of the behaviour that flows from the grant of copyright.
What can be derived at this stage is that a system that is designed to regu-
late behaviour should not operate in a way to undo the structural system
put in place by a system such as copyright. The legislator has made a policy
decision to put a structure such as copyright in place to enhance and reward
creativity and innovation and that decision should be respected. In other
words, competition law should restrict itself to its proper role which is to
address anti-competitive behaviour, such as any use of copyright that is not
pro-competitive as was envisaged when the structural system was set up(Torremans, 2001, pp. 302-9).
8.3 THE IMS HEALTH CASE
Background
TheIMS Health case is a complex saga that is still unfolding before the Court
of Justice. IMS is developing along two parallel tracks on the basis of the fol-
lowing facts. IMS is a major supplier of marketing data to pharmaceuticaland other healthcare companies. In the German market, it has established a
structure of local geographic segments, 1 860 in total, called bricks, each con-
taining a comparable number of pharmacies. Such a structure enables the
collection of standardized data without violating data protection laws that
prohibit the identification of individual pharmacies' data. That structure is
protected by copyright under German copyright law. The brick structure has
de facto become the industry standard and IMS's competitors had asked for
it to be licensed to them in order for them to be able to compete. The user is
indeed only interested in data that are presented in a comparable fashion.
The grant of a licence was refused and that resulted in a complaint to the
European Commission for abuse of dominant position. In due course, the
Commission issued its decision, in which first-stage interim measures were
imposed. That Commission decision was appealed, and resulted in the two
Orders of the President of the Court of First Instance 1 and one Order from
the President of the Court of Justice2 that suspended those interim measures.
At the same time though, IMS's competitors had decided to make use of
the brick structure anyway, without waiting for the grant of the licence.
IMS therefore brought copyright proceedings against them in the Germancourts. The German courts referred the case to the Court of Justice for a
preliminary ruling.3
IMS Health or does copyright deserve a specific approach? 147
The Commission Decision
General points
When IMS's competitors NDC and AzyX applied for a licence covering the
brick structure, the grant of such a licence was refused by IMS. NDC and
AzyX subsequently filed a complaint with the European Commission for
abuse of dominant position.
The Commission issued a decision4 in which it argued that there was
indeed an abuse of a dominant position under Article 82 and that the
urgent nature of the matter required the imposition of interim measures to
avoid irreparable damage being done. 5
Let us consider the process followed by the Commission in giving this
decision and try to find the justifications (if any) for the need to introduce
interim measures to address the abuse of dominant position by IMS. The
Commission first of all identifies the relevant geographical market,
Germany,6 whereas the relevant product market is the regional sales data
services. 7
As regards the position of dominance, IMS is found to be dominant after
examining several factors, but the Commission primarily refers to the fact
that IMS holds a high market share not only in Germany but also in
Europe;8Germany, being the country that has 'the largest market for
regional sales data services in Europe'9 may therefore be regarded as a 'sub-
stantial part of the common market' in this respect.10
In order to understand the background to the case, it is useful to know
what regional sales data services are. Pharmaceutical companies rely on the
brick system because it enables them to gather data on sales of a particular
drug and compare sales figures for each of their products with those for the
products of their competitors. It also offers them a tool to measure the per-
formance of their sales representatives.] l In the brick structure, these data
are collected in each brick for a small number of pharmacies in a local area.
This allows the pharmaceutical companies to obtain a detailed picture
without infringing German data protection laws, under which the use of
data relating to individual doctors or pharmacies would be prohibited.
The Commission concludes that IMS occupies a position of quasi-
monopoly12 and this is, according to the Commission, due to the fact that
IMS owns the copyright in a unique structure for recording sales data for
pharmaceutical products and services and that in the relevant market there
was simply no competition before the arrival of NDC and AzyX. The
Commission argues that IMS obtained this quasi-monopoly position
through a close collaboration with the German pharmaceutical industry,
which led them to the creation of the brick structure. 13 The exact shape of
the brick structure is therefore largely the outcome of that collaborative
148 Developments in the economics of copyright
effort. That involvement of the primary customer in the establishment of
the structure also helps to explain why this structure seems to have been
indispensable for other competitors that found it impossible to build a new
structure; the pharmaceutical industry was unwilling14 to switch to another
structure, given the significant additional costs they would incur from
doing so. 15
The Commission is further sustaining that the IMS brick structure is
becoming a 'de facto' industry standard. For the Commission, the 'I860
brick structure is becoming a common language for communicating the
information' 16 and all the pharmaceutical companies which assisted with
its creation are in the end, effectively 'locked in' that structure on a volun-
tarily basis. The 1860 brick structure is especially a great and unique source
of information 17(and here the Commission sees parallels with the Magill
case, where broadcasters' TV listings were a unique source of informa-
tion 18).
The Commission concluded, therefore, that IMS should grant all its
competitors a licence upon payment of a reasonable royalty for the use of
its copyright work. The Commission justifies its conclusion as follows:
there is an abuse that is due to the dominant position occupied by IMS and,
moreover, IMS owns an essential facility that is impossible to substitute.
Reliance on the Oscar Bronner case
In its decision, the Commission refers repeatedly to the essential facility
doctrine, as set out by the Court of Justice in Oscar Bronner. 19
Bronner was not a case concerned with intellectual property rights, but it
may be important because it is clearly a case about essential facilities.
Bronner could therefore shed light on the correct interpretation of the
essential facilities doctrine in European competition law. Bronner dealt with
a system for the house-to-house distribution of newspapers in Austria set
up by the major national newspaper. A smaller competitor argued that it
should be allowed to make use of the system in order to be able to compete.In its view, the system had become an essential facility. The Court of Justice
followed the suggestion of its Advocate General to interpret the essential
facilities doctrine restrictively and then ruled that it did not apply here.
Important for our current purposes is the establishment of a test that a firm
violates Article 82 where:
'the refusal of access to the facility is likely to eliminate all competi-tion in the relevant market;'
'such refusal is not capable of being objectively justified;' and
'the facility itself is indispensable to carry on business, inasmuch as
there is no actual or potential substitute in existence for that facility'.20
IMS Health or does copyright deserve a specific approach? 149
However, the Commission forgets to mention that Bronner is a case that
can only show how the essential facilities doctrine can be applied in a
European competition law context in general, being a case of pure compe-tition rules application. It is indeed vital to recall that in a normal situation
competition rules are the only set of rules used to regulate market behav-
iour. In copyright cases, there is not just one set of rules that affects the
market behaviour regulation point. Both copyright and competition rules
set out to regulate competitive conditions on the market. The presence of
such a second set of rules must have an impact on the test that is used,
otherwise its vital role is completely ignored. It is therefore submitted that
it is not possible to transfer the test set out in the Bronner (Treacy, 1998 and
Stothers, 2001) case to copyright cases without modifying it. This must
mean that any strict reliance by the Commission on the Bronner case in its
decision must be misguided.
Reliance on the Magill case
To be fair to the Commission, it also argues that its decision follows from
the judgments in Magill21 and Ladbroke 22
It is by now well known that the Magill case started when Magill was
refused a licence covering the copyright in the TV listings. In its decision
the Commission23 ruled that the broadcasters had abused their dominant
position in exercising the exclusive right awarded to them by copyright.
The Commission ordered that advance programme listings be supplied to
Magill. The broadcasters then took the case to Court of First Instance, but
they were disappointed.24
By then the case had become one of principle. On appeal,25 the Court of
Justice's judgment did not, however, change a great deal, the Court reach-
ing the same conclusion as the Court of First Instance. Moreover, the
wording of the judgment was oriented almost exclusively towards compe-tition law principles and the resemblance with the ICI and Commercial
Solvents vs. Commission26approach, which relies on elements of the US
essential facilities doctrine. One should be reminded though, that in Magill,
the Court first of all confirmed that the mere ownership of an intellectual
property right, such as the copyright in the programme listings, does not as
such confer a dominant position.27
Furthermore, even a firm holding a
dominant position can refuse to grant a licence without it constituting an
abuse of the dominant position. The Court then went on to say that in
exceptional circumstances, the exercise of an exclusive intellectual property
right by a proprietor, e.g., a refusal to grant a licence, can involve abusive
conduct that is prohibited by Article 82 of the Treaty.28 Such a prohibition
can be enforced by the Commission through the imposition of an obliga-
tion to grant a licence.29
1 50 Developments in the economics of copyright
The exceptional circumstances referred to by the Court in relation to the
broadcasters' refusal to licence the programme listings in Magill comedown to two elements. By refusing to make programme listings available to
Magill under licence, the broadcasters prevented the appearance of a new
product that they did not offer, but for which there was a clear consumer
demand. It is important to add that as a result of their exclusive copyright
in the listings, the broadcasters were the only source for any such informa-
tion. The second element is that by denying access to that vital material,
without which the new product could not see the light of day, the broad-
casters were effectively reserving the secondary market for weekly TVguides to themselves. The Court clearly emphasizes the secondary market
point by indicating that the broadcasters' main activity is broadcasting.
At this stage, one clearly sees the appearance of the essential facilities
doctrine. Magill can therefore be seen as an application of the essential
facilities doctrine in a copyright case, whereas Bronner applies the same
doctrine to a 'normal' non-intellectual property case, as discussed above
(Woolridge, 1999; Treacy, 1998; and Stothers, 2001). It is also important to
highlight the fact that the Court emphasized that there was a clear and
unmet demand on the consumers' side. Like all US essential facilities cases,
Magill also links the essential facilities doctrine to the benefit that arises for
the consumer from its application (Areeda, 1989).
When the Commission refers to the Court's judgment in Magill, it sum-
marizes the position adopted by the Court as follows:
The Court therefore recognised that in exceptional circumstances the exercise of
an exclusive right deriving from an intellectual copyright may be abusive even in
the absence of additional abusive conduct when, inter alia, it prevents the
appearance of a new product.30
This summary is causing problems in as far as the use of the term 'inter alia'
is possibly a misrepresentation of the Court's judgment in Magill. Indeed,
the Commission seems to suggest that the appearance of a new product is
not a necessary element and that other elements might replace it. It is sub-
mitted that it is wrong to use the term 'inter alia? in this way and to give this
interpretation to thejudgment of the Court. Instead, it is submitted that the
'inter alia' type of exceptional circumstances which the Court withheld in
Magill amounts to the fact that the information concerned had become an
essential facility for the publishers of TV guides (Torremans, 2001, p. 307).
The judgment in Magill supports the point of view that the application of
the essential facilities doctrine in relation to intellectual property (i.e., for
there to be an abuse of the dominant position, the possession of which is
not an abuse per se), requires the presence of all elements set out above.
IMS Health or does copyright deserve a specific approach? 1 5 1
Reliance on the Ladbroke case
The Commission also argues that its view is supported by the judgment of
the Court of First Instance in Ladbroke. 31 This case dealt with the question
of whether a refusal by PMU to license sound and pictures of French horse
races to Ladbroke's betting shops amounted to an abuse of a dominant
position to which the approach adopted in Magill should be applied
(Fitzgerald, 1998).
The Court of First Instance came to the conclusion that the facts of the
Ladbroke case did not warrant the application of the rules set out in
Magill. That conclusion was based almost entirely on the fact that the
images of the horse races were not an essential element for Ladbroke.
Ladbroke did not need the images to enter the market. Despite the fact that
PMU copyright controlled the only access to the images, the essential facil-
ities doctrine did not operate here. PMU could also not be accused of
reserving any market for themselves, as they were not even active on the
betting shop market. The real issue was that Ladbroke wanted to offer its
clients these images as an additional service to its main betting activity in
its shops.
One could argue that the main reason for the Court of First Instance not
to examine the point of the emergence of a new product is the fact that if
one condition is not met, the case was bound to fail as the conditions apply
cumulatively. But one could go further, as the Commission seems to be
doing in its IMS Health decision, and highlight the Court of First Instances
statement that the:
refusal to supply the applicant could not fall within the prohibition laid down
by Article [82] unless it concerned a product or service which was either essen-
tial for the exercise of the activity in question, in that there was no real or poten-tial substitute, OR [our emphasis] was a new product whose introduction mightbe prevented, despite specific, constant and regular potential demand on the part
of the consumer.32
Taken on its own, this quote could suggest that there is no need to establish
that the abuse takes place on a downstream market before Magill can be
applied and, even more importantly, that the conditions found in Magill
might be alternatives, which excludes the need for their cumulative appli-
cation. It is submitted though, that this is not a correct approach. The
whole approach hinges on the literal interpretation out of context of the
single word 'or'. If one puts it in context, that interpretation becomes
difficult. The Court of First Instance felt hardly any need to investigate the
second point (Hull et al., 2002, p. 37). They dismissed the case on the first
ground. Anything they said about the second point is therefore obiter (i.e.,
only incidental).
152 Developments in the economics of copyright
The Commission's interpretation also disregards the economic analysis
set out above. The approach may well work in a non-intellectual property
context, such as Bronner. Here, the requirements do not need to be applied
cumulatively. But we already explained that the presence of intellectual
property rights and their competition regulative function change the
picture dramatically. It is therefore submitted that the Commission is mis-
guided in its reliance on Ladbroke for its views in the IMS decision (Hullet al., 2002, p. 37).
The Commission's overall conclusion in its IMS decision - that there wasindeed an abuse of a dominant position, based on the fact that it saw the
brick structure as an essential facility before applying its liberal interpreta-
tion of Magill to this case, must therefore become open to serious doubts
as to its correctness.
8.4 IMS IN COURT
The proceedings before the president of the Court of First Instance andbefore the President of the Court of Justice in the IMS case33 were up to
now primarily concerned with the interim measures issue.
From the discussion above it has become clear that it is at least arguablethat Magill had left open several issues. 34 These were primarily open to
argument because it was not clear whether the decision on that point in
Magill depended entirely on the facts of the case or whether they were nec-
essary elements from a legal point of view. In fairness to the Commission,all it was trying to do was to fill in the gaps and clarify the matter, albeit in
a somewhat peculiar way.
One of the issues left open is the question as to whether or not excep-tional circumstances required cumulatively the fact that the intellectual
property right was linked to essential inputs for secondary markets and that
a new product had to be introduced in that market for which there was sig-
nificant and unmet demand. The Commission answers that question clearly
in the negative, widening the application of Article 82 significantly.35 The
Commission flatly stated, in addition, that 'there is no requirement for a
refusal to supply to prevent the emergence of a new product in order to be
abusive'. 36Similarly the Commission dismisses the requirement that the
dominant firm's refusal to licence and grant access to the copyright workmust enable it to restrict competition on a second (downstream) market. 37
IMS counters these arguments with references to other aspects of the
Court's case law. It points out that in order for the use of intellectual prop-
erty rights to become an abuse, there must be exceptional circumstances.
These can be found where the refusal to licence is combined with other
IMS Health or does copyright deserve a specific approach? 153
unlawful behaviour or where there is an essential facility at stake. 38 The
notion of essential facilities can, however, in its view only apply where two
distinct markets are involved and the product or service supplied in one
(usually an upstream) market is a necessary input for the production of
goods or services in the second (usually downstream) market.
In his Order of 26 October 2001, the President of the Court of First
Instance summarizes IMS Health's argument further as follows:
There are two essential aspects to the claim advanced by the applicant on the
basis of this case-law. First, it contends that the essential facility doctrine pre-
supposes the use by a dominant undertaking of its market power on the market
where it is dominant in order to preclude or undermine competition on a down-stream or neighbouring market where it already operates or wishes to operate,
or which it wishes simply to prevent, as in Magill, from emerging as a newmarket. Secondly and referring specifically to Magill ... IMS contends that it
was the rightholders' use of the copyright acquired as a result of their activities
on the market where they were dominant (broadcasting) in order to extend that
dominance into a downstream market (weekly television guides) that amountedto the 'exceptional circumstances' justifying the characterisation of their refusal
to license as an abuse. IMS Health's refusal, in this case, to license its competi-tors to use the 1860 brick structure is justified because it is not seeking to exploit
its market power in a separate market. It cannot be abusive, the applicant insists,
for a copyright holder to refuse to license competitors who wish to have access
to its right in order to compete against it on the very market where the exclusiv-
ity granted by that right is used as the central feature of the rightholder's busi-
ness and where its dominance may, at least potentially, turn on the maintenance
of its exclusive right.39
This is indeed important. In Magill, the broadcasters were the sole suppli-
ers of the information on the TV programmes, as only they knew what they
were going to broadcast. Magill could not develop alternative listings. In
IMS, however, competitors in the data services market can develop other
formats. Here too, the Commission's approach would not simply be an
application of Magill, but a massive extension of it.
Additionally, the Commission is unable to point towards any new
product for which there was a significant unmet demand. The President of
the Court of Justice approved, in this respect, of the ruling by his counter-
part in the Court of First Instance. Serious consideration was therefore
given to IMS Health's view; that is, existing case law required that the
refusal to licence must prevent the emergence of a new product in a market
that is separate from the one in which the dominant undertaking refusing
to licence is operating.40
This suggests that doubts still surround the Commission's ultra-liberal
interpretation of Magill, Ladbroke and Bronner in relation to copyright.
These doubts are re-enforced by the fact that the President of the Court of
154 Developments in the economics of copyright
First Instance confirms the importance of copyright, which points towards
the conclusion that intellectual property rights may deserve and require
some form of special treatment.
Sticking with the latter case for a second, one can also raise doubts as to
whether the Commission interpreted it correctly on another point. ThePresident of the Court of First Instance ruled that:
[i]t cannot ... be excluded that the balance of interests effected in the contested
decision by the Commission, which seems to equate the interests of NDC and
AzyX with the interests of competition . . . ignores the primary purpose of
Article 82 EC, which is to prevent the distortion of competition, and especiallyto safeguard the interests of consumers, rather than protect the position of par-ticular competitors.
41
8.5 CONCLUSION
Where does all this lead us and what should be the way forward? It is hard
to say where the Court of First Instance and the Court of Justice will even-
tually take European law when they render theirjudgments in the IMS case.
We feel a lot of sympathy for the approach taken and defended by IMS.
This sympathy is based on sound economic arguments. But this debate is
clearly not limited to the facts of the case at hand. What we are really con-
cerned with is the fundamental interaction between copyright and compe-tition law, which, as could be derived from the structure of the chapter,
requires a heavy reliance on fundamental principles in order to understand
both concepts. That interaction cannot be simplified to the question as to
which of the two prevails. That is exactly the wrong approach and cannot
be justified. Intellectual property rights have a positive role to play in a
modern free market society. Like competition law, they serve the purposeof promoting innovation and enhancing consumer welfare. It is therefore
imperative that the presence of an exclusive right, such as copyright or anyother intellectual property right, be taken into account as an additional
factor when cases such as Magill and IMS (and other cases where an intel-
lectual property right allegedly leads to competition law infringements) are
examined. Competition law has indeed as one of its characteristics that all
elements of the market in which competition takes place are taken into
account when its rules are applied. All elements of the competitive struc-
ture of the market are relevant in this respect.
How is that balance to be achieved and how are all these elements to be
taken into account? It is submitted that the approach in Bronner cannot be
applied unreservedly, as it only looks at competition law when establishing
a test for essential facilities cases. Bronner addresses cases where there are no
IMS Health or does copyright deserve a specific approach? 1 55
additional elements of the competitive structure of the relevant market that
are to be taken into account. That is not the market situation with which we
are concerned here. As an additional right that regulates competition by
imposing restrictions at one level of competition in furtherance of competi-
tion at another level, copyright changes the competitive framework; a more
complex approach is therefore warranted. As such, intellectual property
rights as exclusive rights justify a degree of monopolization of a primarymarket. After all, it is their role to protect intellectual creations embodied in
certain articles. Articles such as books, CDs and databases should be pro-
tected in relation to their commercial exploitation on the primary market, as
copyright is given to the works contained in them to reward the creative
activity of the author. That reward does not come as a lump sum, but as a
chance to market any such articles on their normal market. Things are
different when downstream markets are also affected. The reward for the
positive contribution made through the creation of intellectual property is
situated on the primary market and does not justify its abusive use for other
purposes on a secondary market. Monopolization of a secondary market
must therefore be required to apply the essential facilities doctrine.
But that is not all; an abuse in market terms also involves the fact that
another positive development is stopped. This means that stopping the
emergence of a new product for which there is an unmet demand is also
required in a cumulative way. Otherwise, there is no justifiable need to
curtail an intellectual property right that in general fulfils a positive role.
On the other hand, copyright exists to protect the article in which the copy-
right work is included. That protection is not served and does not include
therefore the blocking of the emergence of a new product or article that is
not in direct competition with the first article, and especially not if there is
no alternative for the producer of the new product due to the dominant
position that is occupied by the copyright-holder (Stothers, 2001, p. 93).
By adopting this stricter approach in applying competition law and the
essential facilities doctrine, both areas of law can fulfil their proper role and
serve their common purpose even better. One should also insist on evidence
of the blocking of a new product for which there is a demand and on the
monopolization of a secondary market if the President of the Court's views
are to be followed, and the consumer needs to benefit as well before the
essential facilities doctrine is to be put into operation.
The question is therefore by no means who wins: copyright or competi-
tion law. That is a false debate. The relevant question is instead whether or
not we are able to strike the correct competitive balance by combining all
tools and elements. We have suggested in this chapter that that is indeed
possible and the components and requirements of such a test have been set
out above. That test is necessarily different from the standard competition
156 Developments in the economics of copyright
law case, as the presence of copyright adds another relevant element to the
competition law equation. The test needs to reflect that, and just as any
competition law test, needs to take account of all the competitive circum-
stances on the relevant market.
NOTES
1 . Order of the President of the Court of First Instance of 1 August 200 1 in case T- 1 84/0 1
R and Order of the President of the Court of First Instance of 26 October 2001 in case
T-l 84/01 R both available at www.Curia.eu.int.
2. Order of the President of the Court of Justice of 1 1 April 2002 in case C-481/01 P(R),
available at www.Curia.eu.int.
3. Case C-48 1/01 IMS Health vs. NDC Health, pending.4. Commission Decision 2001/165/EC of 3 July 2001 in Case COMP D3/38.044 - NDC
Health/IMS Health: Interim Measures [2002] OJ L59/18.
5. The Commission refers in paragraph 38 of its decision to the judgement of the Court in
Case T-44/90 La Cinq [1992] ECR II- 1, where at paragraph 28 the power of the
Commission to impose interim measures finds its justification on the basis that 'protect-
ive measures may be granted only where the practices of certain undertakings areprima
facie such as to constitute a breach of the Community rules on competition'.
6. Decision, paragraph 55.
7. Decision, paragraph 5 1 .
8. Decision, paragraph 59, referring to Case C-62/86 AKZO Chemie BV vs. Commission
[1991] ECR 1-3359.
9. Decision, paragraph 60.
10. Ibid.
1 1 . Decision, paragraph 93.
12. Decision, paragraph 58.
1 3. Decision, paragraphs 83, 86, 87 and 1 85.
14. Decision, paragraph 86 infine.
1 5. Especially, it has been argued by the competitors of IMS that a change to the brick struc-
ture involves a change in working conditions under German labour law. The additional
costs therefore include the fact that the contract of service of the sales representatives must
be renegotiated using the German system of co-determination. Any such change to the
brick structure used will therefore mean a long and costly procedure, due to the involve-
ment of the workers' council in the renegotiation procedure. Decision, paragraph 115.
16. Decision, paragraph 89.
17. Decision, paragraphs 102, 103 and 104.
18. The crucial fact in Magill is that the monopoly was a monopoly over information and
that information happened to be the raw material required by a third party. Therefore,
Magill is now not any longer the only case to be remembered for its unusual facts; that
is, the unusual protection of copyright of factual information.
19. Case C-7/97 Oscar Bronner GmbH & Co KG vs. Mediaprint Zeitungs- und Zeitschriften-
verlag GmbH & Co KG [1998] ECR 1-7791, [1999] 4 CMLR 112.
20. Ibid.
21. Case T-69/89 Radio Telefis Eireann vs. Commission (Magill TV Guide Ltd intervening)
[1991] 4 CMLR 586; case T-70/89 British Broadcasting Corporation andBBC Enterprises
Ltd vs. Commission (Magill TV Guide Ltd intervening) [1991] 4 CMLR 669; and case T-
76/89 Independent Television Publications Ltd vs. Commission (Magill TV Guide Ltd
intervening) [1991] ECR 11-575, [1991] 4 CMLR 745.
22. Case T-504/93 Tierce Ladbroke SA vs. Commission (Societe d'Encouragement et des
Steeple-Chases de France intervening) [1997] ECR 11-923, [1997] 5 CMLR 309.
IMS Health or does copyright deserve a specific approach? 157
23. Magill TV GuidellTP, BBC and RTE decision [1989] OJ L78/43.
24. See above, note 21.
25 . Joined cases C-24 1 /9 1 P and C-242/9 1 P Radio Telefis Eireann and Independent Television
Publications Ltd vs. Commission [1995] ECR 1-743, [1995] 4 CMLR 718.
26. IC1 and Commercial Solvents vs. Commission [1974] ECR 223.
27. See also Philips Electronics NV vs. Ingman Ltd and the Video Duplicating Company Ltd
[1999] FSR 112, where Laddie J. applied the Magill approach in a UK case.
28. Joined cases C-241/91 P and C-242/91 P Radio Telefis Eireann and Independent Television
Publications Ltd vs. Commission [1995] ECR 1-743, [1995] 4 CMLR 718, at paragraphs46, 49, 50 and 54.
29. Ibid.
30. Decision, paragraph 67.
31. See above, note 22.
32. Decision, paragraph 68.
33. Order of the President of the Court of First Instance of 10 August 2001 in case T- 184/01
R, Order of the President of the Court of First Instance of 26 October 2001 in case T-
1 84/01 R and Order of the President of the Court of Justice of 1 1 April 2002 in case C-
481/01 P(R), all available at www.Curia.eu.int.
34. See the discussion in paragraphs 88-105 and the conclusion in paragraph 106 that there
is 'a very serious dispute' concerning these in the Order of the President of the Court of
First Instance of 26 October 2001 in case T-l 84/01 R. This Order was upheld by the
Order of the President of the Court of Justice of 1 1 April 2002 in case C-481/01 P(R).35. Paragraph 100 of the Order of the President of the Court of First Instance of 26 October
2001 in case T-l 84/01 R and paragraph 17 of the Order of the President of the Court of
Justice of 1 1 April 2002 in case C-481/01 P(R).36. Paragraph 180 of Commission Decision 2001/1 65/EC.
37. Paragraphs 85, 86 and 102 of the Order of the President of the Court of First Instance
of 26 October 2001 in case T-l 84/01 R.
38. Paragraph 79 of the Order of the President of the Court of First Instance of 26 October
2001 in case T-l 84/01 R.
39. Paragraph 8 1 of the Order of the President of the Court of First Instance of 26 October
2001 in case T-l 84/01 R.
40. Paragraph 1 7 of the Order of the President of the Court of Justice of 1 1 April 2002 in
case C-481/01 P(R) with reference to paragraph 105 of the Order of the President of the
Court of First Instance of 26 October 2001 in case T-l 84/01 R.
4 1 . Paragraph 1 45 of the Order of the President of the Court of First Instance of 26 October2001 in case T-l 84/01 R.
REFERENCES
Areeda, P. (1989), 'Essential facilities: An epithet in need of limiting principles',
Antitrust Law Journal, 58, 841-53.
Cooler, R. and T. Ulen, (1988), Law and Economics, 1st edn, New York: HarperCollins (3rd edn, 2000), Harlow and Reading, MA: Addison-Wesley.
Fitzgerald, D. (1998), 'Magill revisited: Tierce Ladbroke SA v The Commission',
European Intellectual Property Review, 20 (4), 154-61.
Hull, D., J. Atwood and J. Perrine (2002), 'Intellectual property- Compulsory
Licensing', European Antitrust Review, 36-9 (A Global Competition Review
Special Report).
Landes, W. and R. Posner (1989), 'An economic analysis of copyright law', Journal
of Legal Studies, 18, 325-63.
158 Developments in the economics of copyright
Lehmann, M. (1989), 'Property and intellectual property-
property rights as
restrictions on competition in furtherance of competition', International Review
of Industrial Property and Copyright, 20 (1), 1-15.
Lemarchand, S., O. Freget and F. Sardain (2003), 'Biens informationnels: entre
droits intellectuels et droit de la concurrence', Proprietes Intellectuelles, 6, 1 1-23.
Maskus, K. (2000), Intellectual Property Rights in the Global Economy,
Washington, DC: Institute for International Economics.
Neumann, M. (2001), Competition Policy. History, Theory and Practice,
Cheltenham, UK and Northampton MA, USA: Edward Elgar.
Nozick, R. (1974), Anarchy, State and Utopia, Oxford: Basil Blackwell.
Ramello, G. (2003), 'Copyright and antitrust issues', in Wendy Gordon and
Richard Watt (eds), The Economics of Copyright: Developments in Research and
Analysis, Cheltenham, UK and Northampton, MA, USA: Edward Elgar,
pp. 118-47.
Spector, H. (1989), An outline of a theory justifying intellectual and industrial
property rights', European Intellectual Property Review, 11 (8), 270-3.
Stothers, Ch. (2001), 'Refusal to supply as abuse of a dominant position: Essential
facilities in the European Union', European Competition Law Review, 22 (7),
256-62.
Torremans, P. (2001), Holyoak and Torremans Intellectual Property Law, 3rd edn,
London: Butterworth.
Treacy, P. (1998), 'Essential facilities - is the tide turning', European CompetitionLaw Review, 19(8), 501-5.
Watt, R. (2000), Copyright and Economic Theory: Friends or Foes!, Cheltenham,
UK and Northampton, MA, USA: Edward Elgar.
Woolridge, F. (1999), The essential facilities doctrine and Magill II: The decision
of the ECJ in Oscar Bronner', Intellectual Property Quarterly, 2, 256-64.
9. The basics matter: at the peripheryof intellectual property
F. Scott Kieff and Troy A. Paredes
9.1 INTRODUCTION
Controversies often arise at the interfaces where intellectual property ('IP')
law meets other topics in law and economics, such as property law, contract
law, and antitrust law. }
Participants in the debates over how to mediate these
interfaces often view each interface as a special case deserving unique treat-
ment under the law. 2 The doctrines of copyright and patent misuse are cases
in point: they graft select antitrust principles onto copyright or patent law,
even though there is an entirely distinct body of law - antitrust law -
designed to deal with the putative concerns about competition that
allegedly give rise to misuse. We argue that such specialized approaches to
IP are built by selectively exalting and ignoring particular aspects of the
positive and normative frameworks from distinct substantive areas of law -
IP law, antitrust law, property law, and contract law. Overlooking the total-
ity of these frameworks frustrates the nuanced equilibria to which they each
have evolved, as well as the full complement of important dynamic forces
that affects each legal framework as it continues to develop.3
Instead, we argue that the better approach focuses on the 'basics' - or
core principles and features - of each distinct area of law.4 Our approachavoids specialized frameworks for analysing IP law and the interfaces it
shares with other bodies of law. To do so, the 'basics approach' has both a
procedural and a substantive component.On procedure, our approach emphasizes that the analysis in any particu-
lar case should apply carefully whatever legal regimes the issue at hand
implicates- IP law, antitrust law, contract law, etc. In short, courts should
not create new doctrines and approaches unique to IP when other bodies
of law already apply. At a minimum, courts should take better care to con-
sider accurately and fully the framework of the existing positive law
regimes, as well as the history of the normative debates leading to their evo-
lution. Courts, then, should expressly identify the perceived failures of
these regimes, if any, before creating new doctrines and approaches to
759
1 60 Developments in the economics of copyright
resolve the putative failures. The mere formality of identifying and rigor-
ously considering the different bodies of law that apply to some IP-related
matter can result in greater respect for the dignity of such bodies of law and
ultimately can bring important discipline and restraint to judicial decision
making.On substance, our approach emphasizes judicial adherence to the full
range of established positive and normative frameworks within each bodyof law that the court applies. Consequently, as we understand the basics of
the various bodies of law we consider in this chapter, our proposed
approach will in some cases yield different substantive outcomes than if
judges take what we see as a more activist stance toward IP.
When it comes to IP law in particular, it is regrettable that courts and
commentators have demonstrated a surprising willingness to abrogate, if
not ignore, the express language of the statutes Congress has passed in this
area, as well as the reasons for these legislative enactments. Two examples
from patent law are demonstrative: the doctrine of misuse and the doctrine
of non-obviousness. 5 Before the present institutional framework for the
patent system, the 1952 Patent Act, both of these doctrines of patent law
were at best unpredictable, and at worst so predictably anti-patent that no
patent benefited from either, causing the expected value of patents to
plummet.6 The 1952 Act statutorily overruled both of these aspects of
patent law. 7
Yet, on the issue of non-obviousness, over ten years passed after imple-
mentation of the 1952 Act before the Supreme Court in the famous Graham
case instructed lower courts to apply the framework of the Act's new
Section 103 requirement of non-obviousness. 8 And then soon afterwards,
the Court re-injected confusion into the non-obviousness doctrine by con-
triving new requirements for 'synergism' and 'combination' patents.9It then
took until the creation of the Federal Circuit, a full 30 years after the
passage of Section 103 in the 1952 Act, before these 'innovations' in apply-
ing the law of Section 103 were eliminated and the decisional framework of
the 1952 Act was applied consistently according to its own terms. 10
Even more strikingly, on the issue of misuse, almost 30 years passed after
the 1952 Act before the Supreme Court issued an opinion instructing the
lower courts to apply the framework of the new Section 271 provisions
about what does not constitute misuse and what does constitute indirect
infringement.11 And even after this Supreme Court case, Congress acted
again in 1988 to add subsections 4 and 5 to Section 271(d) to be emphat-
ically clear that subsections 1 through 3 were to be applied according to
their terms. 12
Although this chapter emphasizes judicial decision making, the 'basics
approach' provides guidance for policymakers deciding what laws and
The basics matter 1 6 1
regulations to promulgate.13 Similar to courts applying existing legal
regimes, legislators and regulators setting new policy also should carefully
take into account existing positive law regimes and their normative and his-
torical underpinnings before adopting new laws or regulations that mightrun afoul of the substantive basics of existing legal institutional frameworks.
Fundamentally, we argue that the dignity of each separate and distinct
area of law can and should be respected and applied on its own terms to
settle disputes involving IP. Under our approach, the basics matter, to a
very large extent. The basics matter in the sense that they are where the
analysis of any dispute or transaction involving IP should begin. The basics
also matter in that they are where the analysis should in fact, end.
For example, in our 'basics approach,' there is no need to create special
doctrines or approaches to address matters such as price discrimination or
restrictive licensing arrangements involving IP.14
Rather, analyzing the
legality of such arrangements simply requires one to look to the basics of
each applicable substantive law regime: IP law, antitrust law, and what some
people call the 'general law' - property law, contract law, and the like. The'basics approach' gives us a workable - and more predictable
- frameworkof analysis than creating one-off doctrines that are unique to IP at the
periphery of the law of IP where it intersects with other areas of the law.
The 'basics approach' yields rules for resolving disputes that are easier to
apply and that transacting parties can better understand and rely on in
advance than using more specialized approaches tailored for IP - such as
the doctrines of copyright or patent misuse. Misuse doctrines are unpre-dictable in several respects. First, they include various limitations onrestrictive licensing arrangements beyond what antitrust law or contract
law would prohibit. Second, misuse doctrines do not even impose such
additional limitations in a predictable fashion because the decisional
frameworks themselves for misuse are unpredictable-except, of course, to
the extent the doctrines become so firmly entrenched as essentially to evis-
cerate entire areas of IP law. 15By reducing legal uncertainty, the 'basics
approach' facilitates the ex ante coordination necessary to promote innov-
ation through the commercialization of the inventions, symbols, and cre-
ative works that are protected by patents, copyrights, and trademarks - the
primary goal of IP law and an important goal of antitrust law.
We proceed in section 9.2 to discuss the broad framework of the 'basics
approach,' using the topic of price discrimination as a representative
example. Section 9.3 reviews the basics of the core substantive areas of law
that IP typically implicates: IP law itself, as well as antitrust law andso-called general law, which includes property law and contract law. Sec-
tion 9.4 shows how to solve various problems at the periphery of IP law
by employing the 'basics approach,' as opposed to an approach, such as
162 Developments in the economics of copyright
copyright or patent misuse, that selectively emphasizes or alternatively
ignores particular features of various legal disciplines in crafting special-
ized doctrines for IP. By focusing on the basics, our approach suggests an
important way to reconceptualize IP law with important implications for
bringing new ideas to market. Section 9.5 concludes.
9.2 THE THEORETICAL FRAMEWORK
IP rights generally operate as rights of exclusion. 16 As a result, many worry
that their enforcement will result in too little use of whatever they cover.
Further, the subject matter IP rights cover generally is understood to show
prototypical attributes of public goods in that it is non-rival and non-
exclusive. Classic work by Demsetz, however, has shown that private pro-
ducers can produce and sell an efficient level of public goods under
appropriate conditions, and that price discrimination can advance a com-
petitive equilibrium outcome for public goods, resulting in little, if any,
deadweight loss.17 When an owner of IP rights is permitted to price dis-
criminate, the owner may adopt a pricing regime and licensing scheme that
increases output, eating into any deadweight loss otherwise associated with
market power and the underproduction of public goods.18
Yet, an IP owner's use of price discrimination may not always lead to this
welfare-enhancing outcome. Recent works by Gordon, Lunney, and
Meurer have shown that while price discrimination by IP owners might lead
in theory to more use in certain instances, in practice some price discrim-
ination strategies-
depending, in part, on the licensing arrangements
employed to discriminate among users - can result in less output than if
such price discrimination were prohibited.19 Put simply, price discrimin-
ation has its own shortcomings, and sometimes results in less, not more, use.
The indeterminate results of price discrimination caused us to think
more critically, and more broadly, about IP and price discrimination and
ultimately about the interface IP law shares with other disciplines, such as
antitrust law and the general law, including property law and contract law.
More specifically, there are different types and strategies of price discrim-
ination with different potential consequences- both positive and negative
from the perspective of social welfare. Price discrimination can be done by
the explicit use of different stated prices, in the extreme case by charging
each user his or her reservation price. Price discrimination can also be
done through more complex licensing arrangements, like tying, which can
allow each user more specifically to reveal his or her own demand for
the tying good by how much of the tied good they use. The 'basics
approach' is particularly useful for analyzing the legality of each form of
The basics matter 163
price discrimination since each implicates aspects of IP law, antitrust law,
contract law, and property law.
As discussed more fully below, approaching IP from the basics of IP law,
antitrust law, and the general law of property and contracts enables trans-
acting parties to know better ex ante how to structure transactions that will
be enforced later. In addition to reducing legal uncertainty, when a court
disciplines itself to using an analysis that applies each body of law on its
own terms, there is less opportunity for courts to fashion new and uniquedoctrines of IP law that undercut private ordering by effectively rewritingex post the parties' contract, let alone the legislature's actions. 20 Courts are
ill-equipped to second-guess the substance of contracts entered into by
sophisticated parties merely because the courts believe that some different
arrangement would promote better the use of the underlying IP rights.
Such judicial meddling is particularly troublesome when its downstreamincentive effects on parties
-including owners of IP rights and financiers,
such as venture capitalists- are taken into account. Not only do special-
ized doctrines such as misuse and preemption create uncertainty, but moretimes than not they have the effect of eroding the legislatively created prop-
erty rule protection for IP rights, further compromising commercialization
and private ordering by restricting an IP holder's rights both to use his IP
and to exclude others from having access to the subject matter it covers. 21
There are at least three additional advantages to the 'basics approach,'besides facilitating private ordering and predictability. First, each substan-
tive area of law provides a more informed forum for debate of the issues
that arise in that field. Courts, for example, should not reach out to 'solve'
perceived shortcomings in antitrust law or contract law through the law of
IP, which itself has specific statutory components passed to overturn
similar court action in the past.22
Second, as the product of a long historyof adjudication, lawmaking and academic debate, each area of law pre-
sumably reflects a relatively efficient framework and set of principles that is
actually workable, having stood the test of time. Such longstanding bodies
of law are in contrast to special approaches thatjudges certainly can employto deal with IP, but that are untested and that might simply reflect a particu-lar normative viewpoint that is not satisfied when more appropriate legal
regimes are applied. Third, good cases can be made for each legal regime to
continue to evolve, and they certainly will. However, the one-off, sui generis,
or specialized approaches courts have used at the interface where IP law
meets these other regimes have the effect of skirting many of the diverse
views present in the vibrant debates that persist over how each such regimeshould develop. Put differently, these special judicial approaches to IP
subvert the open and constructive debate that exists within each body of law
regarding whether and how it should evolve further.
164 Developments in the economics of copyright
Given our view of the present substantive basics of antitrust law, prop-
erty law, and contract law, the 'basics approach' shows greater respect for
private ordering than the IP-specific approaches that we question.
However, to be clear, we would urge courts to follow the 'basics approach'
so as to apply each applicable body of law on its own terms when consid-
ering matters involving IP, even if each such body of law was more restrict-
ive presently or in the future than we understand it or might prefer it to be.
That is, the virtues of the procedural component of the 'basics approach'
are independent of our or any other particular interpretation of the sub-
stantive basics of each body of law that courts would apply.23
Courts that adopt special approaches to address matters at the periphery
of IP law run the risk of crafting judicial doctrines that inappropriately
override well-established bodies of law that are informed by longstanding
judicial and scholarly thought and consideration of each area. Put simply,
when considering disputes and transactions at the periphery of IP law
where it intersects other bodies of law, courts often take select principles
from each body of law out of their larger context and legal framework,
while ignoring other basic features and principles of relevant legal regimes.
For example, the misuse doctrine overlooks a number of considerations
involving vertical restraints of trade that drive the present conclusion under
antitrust law that few vertical restraints are anticompetitive and that manyare in fact procompetitive. Such selective picking and choosing not only
creates uncertainty, but, as suggested, often gets it wrong. In part, the
'basics approach' reflects humility toward the complexity and values
embodied in each area of law.
The kind of respect for private ordering associated with the 'basics
approach,' together with the corresponding benefit of greater predictabil-
ity, promotes the commercialization of IP and the subject matter it
protects. Our approach is in contrast to the approaches offered elsewhere
by academics such as Baxter, Bowman and Kaplow, who each offer
analytical tools that only can be applied ex post to evaluate the validity
of any particular licensing arrangement, and as a result have limited
utility ex ante for parties seeking to structure their affairs in a mutually
beneficial way.24
9.3 THE BASICS
Antitrust law, IP law and the general law of property and contracts are each
well-established disciplines and bodies of law. To be sure, numerous debates
exist within each field, and the law continues to develop. But we believe that
general consensus can be found on the broad positive legal frameworks of
The basics matter 165
each field and the core principles that undergird them. Although further
development within each discipline may be advantageous, it will be realized
best if reached through a debate that is fully informed of all diverse views
by occurring within the context of that field.
Approaching from the basics embraces the established frameworks and
principles, even as they may evolve in the future, and affords each area of
the law equal dignity. The 'basics approach' applies each area of the law
according to its own terms, and leaves the debates within each legal field to
be had and resolved within such field. In other words, questions about
restraints of trade are left to the field of antitrust law, and questions of con-
tract validity are left to the field of contract law. More to the point, focus-
ing on the basics avoids the fashioning of new doctrines within IP law that
skirt the basics of IP law, antitrust law, or contract law, such as happenswhen some licensing arrangement that does not violate the antitrust laws
or that is otherwise a valid contract is held invalid as a matter of some form
of sui generis IP law like misuse.
To help frame the 'basics approach,' the following discussion highlights,
at a general level, what we understand to be the core of each discipline-
antitrust law, IP law, property law, and contract law. The discussion is
designed to be a summary, by nature; and so does not attempt to review
fully the entirety of each discipline, which in each case fills volumes.
Nevertheless, this summary discussion does endeavor to represent fairly the
basic principles and positive framework of each body of law.
Antitrust Law
Antitrust law is designed to root out unreasonable restraints of trade and
transactions that substantially lessen competition or tend to create monop-oly.
25 But it is well established that antitrust law does not prohibit market
power as such. Nor does antitrust law prohibit a monopoly, if it is achieved
by having lawfully outcompeted other competitors. As Judge Learned
Hand famously put it, The successful competitor, having been urged to
compete, must not be turned upon when he wins.'26 And increasingly,
antitrust law takes account of dynamic efficiency, as well as allocative
efficiency. Even specific types of conduct that are often associated with
restraining trade and that partly drove the passage of the federal antitrust
laws - such as price discrimination, tying, and exclusive dealing- are not
prohibited in every instance. Rather, such conduct generally is prohibited
only to the extent it unreasonably restrains trade. Indeed, many such prac-
tices are procompetitive. The usual test for unreasonableness in this context
is highly fact-dependent and generally is based on a 'rule of reason' analy-
sis as opposed to treating such conduct as an antitrust violation per se.27
166 Developments in the economics of copyright
Furthermore, antitrust law generally allows unilateral refusals to deal. 28 AsJustice Holmes and then-attorney Rich, who later was the chief architect
of the Patent Act of 1952 and a Federal Circuit judge, also pointed out, it
makes no sense to tell a property owner that she can absolutely exclude
others on the one hand but that she cannot on the other hand be more gen-erous and allow limited access to her property, without giving away the
entire store.29 Accordingly, restrictive licensing arrangements also generally
are permitted.30 To use a simple analogy, as a homeowner, I have the right
to exclude you entirely from my house or to sell you my house, lease you a
room for a limited period of time, or grant you a limited easement across
my front yard.31 Even though refusals to deal and restrictive licenses might
technically restrain trade, they generally do not do so unreasonably and
may by procompetitive.
Intellectual Property Law
Intellectual property law is designed to and indeed does facilitate the
downstream commercialization or realization of the protected subject
matter. 32 While intellectual property law does positively reward, and
thereby encourage, invention and innovation,33
it is not adaptable to being
finely tuned to this goal. It is quite difficult to figure out how appropriately
to reward invention and innovation, and it turns out that a great deal of
inventive and innovative activity would not predictably be responsive to
direct rewards. 34 In practice, IP law facilitates commercialization by
forcing parties to negotiate with each other under the threat of suits for
infringement.35
IP law recognizes that limiting the property owner's causes of action to
be against only those who directly infringe, would unduly undermine or
even eviscerate the role of IP rights in important cases. As a result, the doc-
trines of indirect infringement- induced and contributory
- arose to
capture those activities that, at the time conducted, clearly cause essentially
the same economic effect as direct infringement.36 In the patent context, for
example, by requiring the IP owner to prove not only that his IP rights have
been directly infringed by the one induced, but also that the alleged inducer
intended to induce the infringement, the inducement doctrine captures onlythose who clearly intend to induce infringement and who are successful in
doing so. 37 The contributory infringement doctrine operates similarly. It
requires proof of direct infringement and proof that the alleged contribu-
tor knew that the allegedly contributing conduct was 'especially made or
especially adapted for use in an infringement,' although broad safe harbor
is given to those who provide something that is 'a staple article or com-
modity of commerce suitable for substantial non-infringing use.' 38 Indirect
The basics matter 167
infringement is not accidental. If so desired, it can be avoided relatively
easily through ex ante consideration of known patent rights in view of these
basic legal rules, which are expressly provided by statute and thereby know-
able relatively easily.39
The ability of an IP owner to elect to sue or license those who would
otherwise be guilty of direct or indirect infringement facilitates both price
discrimination and coordination among complementary users. For this
reason, the 1952 Patent Act expressly provides- and, indeed, this was a
major impetus for the writing of the Act - that neither efforts to price dis-
criminate nor the granting of a restrictive or unrestrictive license to a
potential infringer shall constitute misuse.40 This provision was ignored by
many courts until the 1980 Dawson decision,41 which finally recognized its
impact. To be certain this was clear, Congress acted again in 1988 by adding
subparts 4 and 5 to Section 271(d) of the Patent Act to provide expressly
that neither a refusal to license nor a tying arrangement in the absence of
market power is patent misuse.42
Importantly, because the doctrines of copyright misuse43 and trademark
misuse44 are based on the doctrines of patent misuse and patent law's indir-
ect infringement, our discussion has focused on patents. The lessons
learned from the 'basics' view of patents are also applicable throughoutIP law.
Under the basics of IP law, it is clear that contracts facilitating price dis-
crimination or imposing restrictions on a licensee are allowed - indeed,
they are contemplated - at least to the extent they are otherwise properly
formed and enforceable under the general law of contracts, as explored
below. IP rights only give IP owners rights of exclusion, not rights to use.45
The uses to which an IP owner can put their IP or the subject matter pro-
tected by it is (or at least should be) determined by other areas of law. IP
law does not limit the rights of an IP owner to use his or her IP or the sub-
ject matter covered by it in any way that otherwise would be permissible
under other areas of law, including antitrust law, property law, and contract
law. To rely on the express statutory rights of exclusion against others that
IP law grants to IP owners as a basis for restricting the IP owner's rights to
use, conflicts with the basics of IP law; an owner of IP should enjoy similar
rights to use as an owner of tangible property enjoys. On the other hand,
the ownership of IP rights does not magically immunize the owner from
whatever limitations on use of IP or the subject matter it covers are imposed
by other areas of law, including antitrust law and the general law of prop-
erty and contracts. For example, an IP owner's exercise of his IP rights
should (and does) remain subject to the antitrust laws, and a restrictive
licensing arrangement should not be enforced if it is not validly entered into
under contract law.46
168 Developments in the economics of copyright
The General Law: Property and Contracts
Property law and contract law operate to facilitate private ordering, a keyto commercialization of IP assets and to the exploitation of their value.
While property law generally eschews restraints on alienation and, throughits numerus clausus principle, seems to recognize only certain estates in land,
these doctrines only operate as default rules in practice, and a nearly infin-
ite range of dealings can be carried out through contract.47 Moreover, even
within the traditional forms of property, transferors and transferees have a
great deal of flexibility to carve up interests in property along the dimen-
sions of time, use, and the number of property owners.48 For example, whenit comes to real property, highly particularized defeasible fees can be
created and will be enforced, and a real property owner can create anynumber of leasehold interests in his property. All of these transactions are,
of course, facilitated by a general regime of property rule protection, as
opposed to liability rule protection, for rights in both real and personal
property.49
To be sure, when parties order their affairs through contract, this requires
a valid contract, which in turn requires compliance with proper formation
details, such as consideration and no unconscionability. With very few
exceptions, positive contract law does not regulate the substance of the
parties' arrangement, focusing instead on the contracting process.50
Whatever strictures property law and contract law impose on private
ordering, parties are generally free to order their affairs and to carve uprights, duties and obligations as they see fit. These basics of the general law
of property and contracts should extend to the use and licensing of IP
rights, just like they do to other types of property. Nothing under propertyor contract law provides any particular reason to be skeptical about IP con-
tracts that facilitate price discrimination, exclude certain parties from
having access to IP rights, or impose restrictions on licensees. What is more,
special approaches to disputes and transactions involving IP rights often
ignore or intentionally override purposeful normative and positive features
of antitrust law, IP law, or the general law and, in so doing, risk upsetting
well-developed frameworks without adequately accounting for competingconsiderations.
9.4 WHEN APPLIED, THE BASICS SOLVETHE PROBLEM
Applying the 'basics approach' to prototypical cases at the periphery of IP
law, including price discrimination, restrictive licensing arrangements, and
The basics matter 1 69
suits against indirect infringers, provides a set of rules that are useable
ex ante by all market participants in a way that helps them order their
affairs while at the same time being more fair and efficient. The 'basics
approach' has important normative implications, in that judicial fidelity to
the basics ultimately allows market actors to have greater freedom and
ability in structuring their interactions in welfare-enhancing ways, in addi-
tion to reducing legal uncertainty.
The cases we explore are appropriately viewed as prototypical for several
reasons. They involve fact patterns that are representative. They have actual
historical significance through their particularly important roles in the
body of case law. And the primary architect of the present patent system-
the 1952 Patent Act - wrote a five-part series of articles about these cases
before drafting the statute designed to change fundamentally the waycourts applying the law would look at the issues raised by the cases. 51
As discussed more fully below, the cases can be divided into two sets.
Importantly, a review of both sets of cases shows that the 'basics approach'is not merely a veiled effort to promote pro-patent or pro-copyright
- or
more generally, pro-business-
positions. Rather, the basics framework, as
a method of judicial decision making, is offered as a coherent approachthat more predictably can be engaged ex ante and that reflects fidelity to,
and respect for, separate areas of the law. Although we focus on patents,
since the core features of other areas of IP law largely derive from patent
law, the basics framework and the essence of the following analysis extend
to copyrights and other forms of IP as well.
Indirect Infringement vs. Breach of Contract
The first set of cases involves the tension between indirect infringement and
indirect participation in a breach of contract. Indirect infringement may be
actionable as a matter of IP law, as discussed earlier. Indirect participation
in a breach of contract may be actionable as a matter of contract law under
doctrines such as tortious interference with contract, as in the famous
multi-billion dollar judgment from the Texaco, Inc. vs. Pennzoil, Co. litiga-
tion. 52However, the happenstance that a contract relates to patents should
not transform interference with that contract into patent infringement. The
facts that need to be proven are different under these different frameworks.
The potential remedies are different as well. 53
Wallace, the classic case of indirect infringement, involved a patent on
an oil lamp having a new burner, together with a standard fuel reservoir,
wick, and chimney.54 In the case, a competitor of the patent owner had sold
a rival product, which included the new burner and other lamp parts but
not the chimney. The court reasoned that the defendant had contributed to
170 Developments in the economics of copyright
infringement on the part of its customers, because they would inevitably
add a chimney. A judgment of contributory infringement makes sense
under the 'basics approach' because the intended and actual impact of the
competitor's efforts were to make sure that its customers acted in an
infringing manner. Indeed, Wallace is the case that gave rise to the entire
doctrine of indirect infringement throughout all of IP law. 55
By way of comparison, if the plaintiff-patentee in Wallace instead had
entered into arrangements with its customers obligating them to buy chim-
neys from the patentee, the analysis under the basics of IP law would be
different. A rival seller of chimneys might be liable for tortious interference
with contract, or the tying arrangement might violate the antitrust laws.
However, the competing chimney seller would not be liable for contributory
infringement under the basics of IP law. 56
The Heaton case involves an example of just this type of arrangement.
Heaton involved a patentee who sold a patented machine with a label
license57 under the patent that restricted the machine's use to certain
unpatented inputs (staples, literally)- a tying arrangement
- and a defend-
ant who sold competing inputs.58 The court seemed to reason that by pro-
viding its staples for use in the machine, the defendant was contributing to
breach of the label contract, which had given the permission through the
label patent license to use the machine. Once that license under the patent
was gone, the use of the machine became infringing. Rather than sue for
interference with the contract, the plaintiff sued the competing seller of
staples for indirect infringement of the patent under patent law. The court
decided that the defendant was, indeed, committing contributory infringe-
ment of the patent. But this turned a case about indirect participation in
breach of contract into patent infringement. By deciding the case the wayit did under IP law, the court in effect extended inappropriately the scope
of IP rights. A collateral inappropriate consequence of the court's reason-
ing in Heaton, of course, would be to immunize potentially anticompetitive
licensing arrangements from the antitrust laws.
The 'basics approach' rejects the analysis of Heaton. Under the 'basics
approach,' and as pointed out by Rich, this decision was inappropriate
because it 'transformed the law of contracts into "patent law".' 59It may
have been appropriate for the plaintiff to consider an interference with con-
tract argument, if sufficient facts could be proven to substantiate the claim
under contract law. 60 It may even have been appropriate for the defendant
to consider an antitrust tying argument, if the case could be proven under
antitrust law.61
By not addressing these contract and antitrust arguments head on, cases
like Heaton allow parties, and judges, selectively to mix features of various
bodies of law and to extrapolate from them to forge new hybrid doctrines
The basics matter 171
of law that run afoul of the basics of each area. In many instances, such
selective application of the law leads to doctrines, such as misuse, that erode
IP rights. In other cases, such as Heaton or those cases in which courts have
subjected transactions involving IP to less scrutiny under antitrust law, the
new doctrines can work to expand IP rights. What is more, in all cases, the
courts fail to give any meaningful test for determining when those IP rights
should be so eroded or expanded.The Supreme Court applied the same approach as Heaton in the A.B.
Dick case, which involved a patent on a mimeograph machine sold with a
label restriction limiting the brand of unpatented ink that could be used in
the machine. 62 As in Heaton, the Court agreed with the plaintiff-patentee
in A.B. Dick, and held that there was contributory infringement of the
patent. Because this was a Supreme Court case, its reasoning had a longer
lasting impact in pushing IP law in a direction contrary to the 'basics
approach.'63
The 'basics approach' rejects the reasoning of A.B. Dick for the same
reason it rejects the reasoning of Heaton. Indeed, eventually, these cases
were effectively overturned.64 As Rich pointed out later in his testimony
before Congress concerning the provisions he drafted on indirect infringe-
ment in the 1952 Patent Act, any effort to follow this inappropriate body of
law 'would kill itself in time.'65
An understanding of the basics suggests why Heaton, A.B. Dick, and
their progeny were not sustainable over the long run. The problem is not
merely one of courts going too far one way (e.g., effectively extending the
scope of IP rights to anything connected to IP and simultaneously immun-
izing all transactions involving IP from serious antitrust scrutiny) or the
other (e.g., eliminating the doctrine of indirect infringement, thereby
eroding IP rights). The problem is more fundamental. Namely, cases like
Heaton and A.B. Dick ignore the basics of each implicated body of law -
IP law, antitrust law, and the general law of property and contracts. As a
result, they lead to unpredictable results and, in the name of IP law,
encroach upon the boundaries of other well-established bodies of law that
reflect more nuanced and time-tested doctrines and rules that have staying
power and that are perfectly capable of resolving the disputes on their ownterms.
Infringement under IP Law vs. Sui Generis Law
The second set of prototypical cases involves the question of what body of
law should govern determinations of infringement: the body of organic
IP law - patent, copyright, or trademark - or some special sui generis bodyof law. In many of the cases involving charges of indirect infringement
172 Developments in the economics of copyright
and misuse - which are admittedly somewhat difficult doctrines - too manycourts and commentators have not followed the 'basics approach' and have
instead tried to re-hash the normative case for IP to develop new special-
ized approaches in these doctrinally difficult cases that they hope will get
IP scope just right. The fundamental problem with these specialized
approaches is that they re-cast the entire legal institutional framework for
IP in a way that has pernicious ripple effects throughout IP law by ignor-
ing the many choices that have been made over IP law's development.One basic trap into which these courts and commentators have fallen
when adopting such sui generis approaches to IP is focusing on the wrongparty when considering whose behavior should matter in cases of possible
indirect infringement. The behavior of the putative indirect infringer to
facilitate or encourage direct infringement is relevant to the analysis under
both inducement of infringement and contributory infringement. Thebehavior of the patentee
- in the sense of the patentee engaging in conduct
that leverages his IP rights with the goal of extracting value - is not rele-
vant to inducement or contributory infringement. Indeed, in such
instances, the patentee is simply exercising his rights to exclude and to use,
as the basics of IP law and the general law anticipate. Put simply, the ques-tion of a putative defendant's infringement should not turn on whether or
not the patentee was trying to get as much out of the patent as possible
through some restrictive licensing arrangement, tie-in, or otherwise. If the
patentee, or any property owner for that matter, behaves in a way that
antitrust law or contract law properly prohibit, then that is a matter of
antitrust law or contract law.
The modern trend towards sui generis analysis of infringement- as com-
pared with an analysis based in the basics of IP law - has its most visible
roots in the Supreme Court's Leeds & Catlin decision. Just like the classic
indirect infringement case of Wallace, discussed previously, Leeds & Catlin
involved a patentee's competitor selling something that had no substantial
non-infringing use. 66 In Leeds & Catlin, the defendant-infringer sold spe-
cially grooved records that could only be used in a patented record player
known as a 'Victrola.' The Supreme Court reasoned that the defendant's
selling of the records was infringement because the records were the 'dis-
tinction [or key element] of the invention.'67 This reasoning is flawed.
Although the 'basics approach' might reach the same result - a finding of
contributory infringement- it would do so for an entirely different reason
than offered by the Court. Under the basics of patent law, there is no 'dis-
tinction,' or key element, of subject matter claimed under the patent. The
patent system operates using what is known as 'peripheral claiming'- as
distinct from 'central claiming'- in which the function of the patent claim
is not to set forth the heart of the protected subject matter but rather to set
The basics matter 1 73
forth its outer bounds. 68 Direct infringement is measured against these
outer bounds. Indirect infringement is premised upon some occurrence of
direct infringement. But the reach of indirect infringement does not turn
on whether the putative defendant is targeting some key element of the
claim. Rather, as discussed earlier, for a proper analysis of contributory
infringement under the basics, a key question is instead whether there were
any substantial non-infringing uses for the grooved records. Because there
were no such uses in Leeds & Catlin, and because the other elements of
contributory infringement were established (i.e., direct infringement and
knowledge of the patent), applying the basics would have resulted in a
finding of contributory infringement.
Hanging determinations of indirect infringement on the factors outlined
earlier in our discussion of the basics - such as intent for induced infringe-
ment and absence of non-infringing substitutes for contributory infringe-
ment - may seem like an effort to exalt form over substance. After all, the
reasoning the Court adopted in Leeds & Catlin seems to strike at the heart
of substance by focusing on the key element. But the Court fails to give anyinstruction on how to determine which element is key, and neither has anyother court or commentator of which we are aware. The tests for indirect
infringement have the essential advantage of being comparatively easy to
administer. They look to facts well within the control of the putative
infringer and are strongly biased in favor of the putative infringer in the
types of errors one would expect the tests to generate.
The intent requirement under an inducement analysis and the broad and
readily identifiable safe harbors under a contributory analysis ensure these
important biases and that the doctrines are relatively easy to administer. 69
Importantly, the improper reasoning of the Court in Leeds & Catlin is not
mere harmless error. It matters which approach is adopted by courts - espe-
cially the Supreme Court, even if the results are the same in a particular
case. By suggesting in Leeds & Catlin that the case turned on the heart of
the invention, the Court advanced a line of precedent that focused on the
wrong issues in patent cases. One of the most pernicious cases in this line
of precedent was Carbice, in which the Court denied relief to a patentee
after reasoning that the patentee was trying to extend the patent beyond the
key elements of the claim. 70 The plaintiff-patentee in Carbice had a patent
on a packaging method that used dry ice. What troubled the Court was that
the patentee had a practice of entering into licensing arrangements oblig-
ating the licensee to use only certain containers for packaging productswith the dry ice. The facts of Carbice are somewhat similar to those of
Leeds & Catlin with one important difference: the defendant sold a product-dry ice - that was a staple article of commerce usable in many non-infring-
ing manners other than in the patented method of ice-cream packaging.
174 Developments in the economics of copyright
The 'basics approach' would again yield the same result as the Court's
analysis- in this case, no contributory infringement
- but again for a
different reason. Instead of focusing on the patentee's alleged extension of
the patent beyond its key elements, the 'basics approach' would turn on the
many non-infringing uses for dry ice.
As Rich emphasized, it is the behavior of the putative contributory
infringer that is relevant to a determination of contributory infringement,
not that of the patentee.71 Under the 'basics approach,' it makes sense that
the organic IP law- in this case patent law
- evolved so that as implementedin the 1952 Act's provision of Section 271, focus is on the behavior of the
putative infringer precisely because it is comparatively easy to judge.
Furthermore, an IP holder should not be denied relief for contributory
infringement- or even direct infringement
-simply because the IP holder
is exercising his rights to exclude and to use through a tying arrangementor restrictive license. Such conduct is properly a subject for antitrust law
and contract law, but should have no bearing on a court's analysis of indir-
ect (or direct) infringement under patent law. Courts should not recast such
conduct as an effort by the IP holder to 'extend' his patent rights for the
purpose of transforming a matter for antitrust and contract law into a
matter for some new version of IP law.
To be sure, the Court did not always reach the right result, as it did in
Leeds & Catlin and Carbice. Because the Court continued to misplace its
focus on the putatively key elements of patent claims, by the time of the
Mercoid cases, the doctrine of indirect infringement had been almost entirely
eliminated as a result of judicial reasoning that precluded any action for
indirect infringement. In essence, because by its nature every indirect
infringement case involves a defendant who is not triggering at least one
element of the patent claim - direct infringement occurs when all elements
are satisfied - the focus on 'key element' in the Court's reasoning allowed
every putative indirect infringer to argue that the missing element was the one
that was 'key' and therefore no action for indirect infringement could lie.72
In response, Rich drafted what became Section 271 of the 1952 Patent
Act to overrule statutorily cases like Mercoidand to revive indirect infringe-
ment. 73 Under this established basic framework of patent law after the 1952
Act, the essential inquiry for indirect infringement is on the comparatively
easy to administer framework discussed earlier. 74 While it may be appro-
priate to debate the benefits and costs of allowing actions for indirect
infringement, the above review is designed to show at least two important
things. First, sui generis attempts to re-hash the proper scope of an organic
IP right when addressing cases of misuse or indirect infringement will yield
a test that is comparatively more difficult to administer, that eliminates the
doctrine, or both. Second, unlike prior approaches commentators have
The basics matter 1 75
offered for addressing issues at the periphery of IP law - many of which
urge a nearly impossible ex post balancing of dynamic and allocative
efficiency that inappropriately emphasizes trying to achieve some optimal
reward to inventors as opposed to commercialization - the 'basics
approach' provides a set of clearer rules and doctrines that market partici-
pants can rely better on ex ante in structuring their affairs.
Other Pernicious Ripple Effects
The 'basics approach' has important implications for resolving matters
involving at least three current and controversial issues found at the periph-
ery of IP law: patent and copyright misuse, restrictive licensing arrange-
ments, and preemption. Applying the basics to these and other tough cases
that simultaneously implicate IP law, antitrust law, and contract law avoids
a host of pernicious ripple effects - namely, undercutting innovation and
the commercialization of IP - that arise from more specialized approachesto disputes and transactions involving IP.
Concerning the misuse doctrine, the 'basics approach' is not compatiblewith the Federal Circuit's present view of patent misuse, which seems to
leave a broad and vaguely defined space for misuse. 75 In Virginia Panel, the
Federal Circuit suggested the following test for determining whether a
patentee has misused his patent: 'When a practice alleged to constitute
patent misuse is neither per se patent misuse nor specifically excluded from
a misuse analysis by Section 271(d) [of the Patent Act], a court must deter-
mine if that practice is reasonably within the patent grant.'76 But import-
antly, the patent statutes make no provision for per se misuse. 77Rather,
Section 271(d) provides specific safe harbors for conduct that is not misuse.
Further, it is inappropriate to suggest that some use of a patent is not within
its scope, since patents only give a right to exclude. The right to use is
derived from sources external to IP law.
When the 'basics approach' is employed, other bodies of law, such as
antitrust law, provide the proper legal lens through which to inspect a paten-
tee's use of a patent and the subject matter it covers. For example, the basic
thrust of putative misuse is that an IP holder should be denied relief for
infringement when he has used his IP in some allegedly anti-competitive
way. Yet, if the challenged conduct is indeed anti-competitive, it ought to
trigger the antitrust laws. As discussed earlier, patentees and copyright
holders, like other property owners, are subject to antitrust law, because
patents and copyrights give only a right to exclude, not a right to be free
from the constraints of other laws. In brief, the pernicious effect of the
misuse doctrine is that it erodes IP rights, at least at the margin, and risks
rooting out procompetitive and competitively-neutral behavior that the
176 Developments in the economics of copyright
antitrust laws recognize as such and permit.78 If the antitrust laws are too
lax, the appropriate remedy is to fix the antitrust laws. As Rich pointed out
in commenting on the unfortunate habit of courts to treat potential
antitrust concerns as some how more serious and in greater need of polic-
ing when IP is involved:
The patent right is not the only form of property subject to such misuse. But it
is so little understood, as compared to other forms of property, that much
mystery attaches to it and much confusion surrounds it. ... [Practices that
restrain trade are] not due to the patent law .... [They are] due to failure to
enforce the anti-monopoly laws. The advocates of reform would do well to
restrict the attack to the latter aspect and not confuse the issue by abortive
attempts to emasculate the patent law . . . ,
79
Concerning restrictive licensing arrangements, the 'basics approach' sug-
gests that courts generally should enforce restrictive licenses involving IP as
long as they are enforceable under contract law and do not run afoul of the
antitrust laws. Indeed, affording IP holders the right to carve up interests
in their IP and the subject matter it covers is consistent with the basics of
property law and the right to use enjoyed by owners of tangible property.
Courts adopted the 'basics' reasoning in considering the validity of restrict-
ive licenses of copyrights in the ProCDSQ case and of patents in the
Mallinckrodt81 case. 82 Even when a potential or actual IP owner tries to
extract payments for activities that fall outside the protection of IP,83 courts
should enforce these contracts to pay as long as the arrangement- which
may amount to little more than an effort to ease either the risk burden or
the financing burden of the transaction - is properly enforceable under con-
tract law. 84
By way of contrast, in practice, courts are skeptical of contracts that
happen to be tied to royalty payments beyond patent term, even though the
economic justification for this skepticism is lacking.85
Further, courts that
do not stick to the basics will often err by finding that restrictive licensing
arrangements, including tie-ins, constitute some sort of impermissible
extension of IP rights.
Finally, concerning preemption, the 'basics approach' is not compatible
with the approaches adopted by the Supreme Court, which generally can
be 'seen as efforts to place limits on the ability for [IP owners] to avail them-
selves of various State laws.' 86 The Court's approaches make no sense in
part because IP rights confer rights of exclusion on IP owners, not add-
itional restrictions on use on IP owners and not additional rights to use on
non-owners. 87
For example, in Bonito Boats, the Court's approach is premised on the
contrivance that federal patent law creates a right to copy.88 The Court in
The basics matter 1 77
Bonito Boats decided that this right to copy would be frustrated by the state
law at issue, which regulated one particular form of copying boat hulls,
called 'plug molding.'89 The case arose because a party seeking to engage
in plug molding in violation of the state statute argued that federal patent
and copyright law preempted the state statute based on the SupremacyClause of the US Constitution. The Court invalidated the state statute
under the doctrine of conflict preemption.90
There are several problems with the Court's reasoning. There is no right
to copy - indeed, no affirmative right at all - that is conveyed on the public
by patent law, or for that matter by copyright or trademark law. 91 These IP
regimes only create under certain situations specific rights of exclusion
vested in the IP holder, as explained above. Although the plug molding
activity was not covered by any of these federal IP rights, the mere absence
or expiration of any such right of exclusion says nothing about a third
party's affirmative right to use the subject matter such right of exclusion
might have covered. Indeed, use of IP rights, whether by the IP holder or
some third-party licensee, is often restricted, if not outright blocked, byother IP rights and by regulation, but this should not be grounds for finding
that such restrictions are preempted by IP law.
The Bonito Boats Court essentially rejected, or at least glossed over, these
arguments by suggesting that the purpose of the state statute somehow con-
flicted with a purported 'strong federal policy favoring free competition in
ideas which do not merit patent protection.'92 But this analytical framework
is unworkable in that it would seem to extend to block any state law or legal
enforcement of contracts that interferes with a right to use or copy.
Consider, for example, a state law against cheating on exams or, to be closer
to the case, a contract term against plug molding. Also consider a contract
term making a promise to do or to abstain from doing any activity in a waythat allegedly conflicts with the IP law regime putatively doing the pre-
empting.93 Under the reasoning of Bonito Boats, each such state law or
contract term would not be enforceable. In addition, the Bonito Boats rea-
soning may eviscerate the rights of exclusion that the patent, copyright, and
trademark statutes are designed to create and that are both properly justi-
fied and authorized.94 That is, every IP right and every form of market reg-
ulation or other exercise of police power will impact, to some extent,
competitive economic concerns of the type that also underlie each of the
federal IP regimes, thereby providing under the Bonito Boats rationale an
extensive basis for preemption and interference with private contracting.
What is more, to anyone informed by public choice theory, every IP right,
market regulation, and other exercise of police power can be seen as motiv-
ated at least in part by its impact on these same competitive economic con-
cerns. Put differently, the reach of the preemption analysis in Bonito Boats
1 78 Developments in the economics of copyright
would allow any potential defendant to select a federal IP regime that does
not reach such party-say patent law - and then use that federal IP regime
to justify a finding of preemption against any other state (or even federal)95
law that impacts competitive economic concerns. What is more, if the
federal IP regime that is said to have this preemptive effect is itself weaker -
such as when it does not reach indirect infringement or is limited bymisuse - then the overall power of the Bonito Boats preemption approachis even greater.
In contrast, the 'basics approach' recommends the Federal Circuit's
alternative analytical framework for conflict preemption called the 'extra
element test,' which does not suffer the shortcomings of Bonito Boats and
which, indeed, facilitates the smooth operation of each IP and competitionlaw regime, federal and state alike.96 That test simply asks whether the basic
legal elements of the cause of action that is putatively preempted are
exactly the same as those of the cause of action that is putatively doing the
preempting; the presence of extra elements means no preemption. The state
statute at issue in Bonito Boats would be analyzed easily under this test
because liability under it turns on a host of elements that are unconnected
to patent law, including copying via the plug mold technique. Even closer
cases, such as those involving state laws regulating statements about patents
themselves, can be decided using the extra element test.97
The 'basics approach' also provides guidance for policymakers evaluat-
ing whether to promulgate IP laws that interface with other state law
regimes. For example, in derogation of principles of state contract law and
trusts and estates law, the copyright regime gives authors a non-transferable
right to terminate transfers98 of their copyrights and even sets forth its owntrusts and estates provisions governing who gets this termination right
upon the author's death so as to expressly preempt other arrangementsauthors might make by testamentary will as well as state default rules
of intestacy.99By way of another example, the anti-circumvention provi-
sions of the recently promulgated Digital Millennium Copyright Act
('DMCA'),100 which prohibits manufacture or distribution of any technol-
ogy, product, service, or device that circumvents copy protection technol-
ogy, does not make sense under the substantive component of the 'basics
approach' because the transactions it regulates are better governed byeither state contract law or by the IP law of indirect infringement, discussed
earlier. 101 In short, the 'basics approach' would urge policymakers consid-
ering each such positive law framework instead to leave these issues to be
decided as a matter of state law. That having been said, even if the wisdomof some legislative or regulatory action can be challenged under the 'basics
approach,' once policymakers have acted, courts should apply the law on
its terms and refrain from crafting new doctrines or from engaging in other
The basics matter 179
creative decision making, which would have the effect of end running the
legislative or regulatory body.
9.5 CONCLUSION
Like the Supreme Court in Dawson and earlier work by commentators such
as Baxter, Bowman, Gordon, Kaplow, Lunney, Meurer, and Rich, we strike
a balance between a view of IP that is too restrictive and one that goes too
far. But following more closely the writing of Rich, who after all drafted
the statutory framework that Congress adopted for patent law, we look not
only to the direct impact that applying the basics at the periphery of IP law
has on the commercialization of the subject matter it protects. We sepa-
rately believe that it is important to respect the different legal institutional
frameworks of the various bodies of law that are involved - IP law, antitrust
law, and the general law - each of which strikes its own balance among the
competing needs of those who interact with these disciplines. We do not
dispute here that it is important for doctrine to develop over time toward
more efficient and equitable outcomes. Such doctrinal developments,
though, should occur within the context of the applicable body of law. IP
law, for example, should not be a vehicle for restructuring contract or
antitrust law collaterally from outside those legal regimes. Although in
practical terms, the 'basics approach' often reaches results that are similar
to the outcomes of other approaches, we provide a normative justification
for a positive law framework that is more predictable and that captures the
distinct and important balances that are struck within each separate bodyof law that is implicated.
We offer, in the end, a framework for understanding IP law and the
broader interfaces that IP law shares with a number of bodies of law, such
as antitrust law, property law, and contract law. Our framework is in large
part animated by a property rights perspective that places priority on
ensuring the appropriate ex ante incentives to facilitate the complex trans-
actions needed to ensure wider use of the subject matter IP rights cover,
such as through information dissemination and commercial sales of
embodiments. Our framework is equally motivated by attention to the
basics of each other body of law we discussed with an understanding that
only through coherent discussion of each area in a piece-wise fashion can
the right progress be made on both positive law and normative fronts. Afurther advantage of the 'basics approach' is that it should reduce legal
uncertainty, which itself is a source of inefficiency.
Finally, the 'basics approach' reflects a general skepticism about the insti-
tutional capability of courts to make ex post determinations regarding how
1 80 Developments in the economics of copyright
to facilitate the complex commercialization process that must occur for the
public to derive the benefits of the various works protected by IP rights. Put
differently, we believe that private ordering and markets are more effective
than courts, all things considered, at solving what in essence are industrial
organization matters.
ACKNOWLEDGMENT
The authors gratefully acknowledge financial support from the John M. Olin Foundation, the
Hoover Institution, and the Washington University School of Law, as well as intellectual con-
tributions from participants in the 2003 Society for Economics Research on Copyright Issues
Annual Conference held June 19-20, 2003, in Northampton, Massachusetts. The authors also
gratefully acknowledge more detailed comments provided by Michael Abramowicz, John
Barton, Tim Bresnahan, Richard Epstein, Paul Goldstein, Wendy Gordon, Mark Lemley,
Larry Lessig, Stanley Liebowitz, Glynn Lunney, Charles McManis, Michael Meurer, PamSamuelson, Joel Seligman, Henry Smith, Polk Wagner, and John Witherspoon. A revised
version of this chapter has been published in the George Washington Law Review (2004).
NOTES
1 . See, e.g., Symposium, The Interface Between Intellectual Property Law and Antitrust Law,87 Minn. L. Rev. 1695 (2003); William F. Baxter, Legal Restrictions on Exploitation of the
Patent Monopoly: An Economic Analysis, 76 Yale L.J. 267 (1966); Ward S. Bowman, Jr.,
Patent and Antitrust Law: A Legal and Economic Appraisal, at xii (1973); Michael A.
Carrier, Unraveling the Patent-Antitrust Paradox, 150 U. Pa. L. Rev. 761 (2002); Wendy J.
Gordon, Intellectual Property as Price Discrimination: Implicationsfor Contract, 73 Chi.-
Kent L. Rev. 1367 (1998); Louis Kaplow, The Patent-Antitrust Intersection: AReappraisal, 97 Harv. L. Rev. 1813 (1984); Glynn S. Lunney, Copyright and the SupposedEfficiency of First-Degree Price Discrimination (2002) (working paper); Michael J.
Meurer, Copyright Law and Price Discrimination, 23 Cardozo L. Rev. 55 (2001).2. Indeed, these debates often take on status as their own specialized disciplines bearing
new 'and'-based names, such as 'intellectual property and antitrust,' which in turn spawnnew sub-specialties, such as 'copyright and antitrust.'
3. For earlier articulations of the views at the core of the 'basics approach' developed below,
see, e.g., Donald S. Chisum, Craig Allen Nard, Herbert F. Schwartz, Pauline Newmanand F. Scott Kieff, Principles of Patent Law 1066-1 155 (2001); Troy Paredes, CopyrightMisuse and Tying: Will Courts Stop Misusing Misuse?, 9 High Tech. L.J. 271 (1994).
4. We do not use the word 'basics' pejoratively, such as in the sense of an unduly simplecharacterization of the law or legal process. In addition, we recognize that there is
sufficient path dependency and context dependency to the development of the 'basics'
that our discussion here is most applicable to the regimes that have evolved in the US.That being said, we do think that the comparative institutional analysis offered here maybe useful in elucidating relative strengths and weaknesses of different strategies to
shaping IP and other commercial laws outside the US as well.
5. See generally Chisum et al. supra note 3, at 514-19, 1066-99 (discussing evolution of the
non-obviousness and misuse doctrines).
6. Although there was some variation in the way courts treated patents under these doc-
trines, a sufficient number of powerful courts (including the Supreme Court) were treat-
ing the patents that came before them so poorly under these doctrines that the expected
The basics matter 1 8 1
value for all patents plummeted. On non-obviousness, courts applied a tautological and
unpredictable subjective decisional framework then called the 'requirement for inven-
tion': to be patentable, an invention had to constitute an 'invention.' This standard
became so vague and yet so difficult to satisfy that Justice Jackson remarked, '[T]he only
patent that is valid is one which this Court has not been able to get its hands on' Jurgensenvs. Ostby & Barton Co., 335 U.S. 560, 572 (1949) (Jackson, 1, dissenting). On misuse,
courts applied such a broad definition of misuse that for all intents and purposes patents
could no longer be asserted against indirect infringers. See generally infra notes 40-41
and accompanying text (discussing evolution of the misuse doctrine in relation to the
doctrines of contributory infringement and inducement of infringement).
7. On non-obviousness, Congress passed Section 103 of the statute in the 1952 Act to
replace the requirement for 'invention' with the requirement for 'non-obviousness.' Far
more than a different word, this test for patentability set forth a much clearer and more
objective decisional framework. For a comparative institutional analysis of these deci-
sional frameworks, see F. Scott Kieff, The Casefor Registering Patents and the Law andEconomics of Present Patent-Obtaining Rules, 45 B.C. L. Rev. 55 (2003) (comparativeinstitutional analysis of patent-obtaining rules). On misuse, Congress passed Section
271 of the statute in the 1952 Act to revive the doctrines of contributory infringementand inducement of infringement and to make clear what does not constitute misuse. See
infra notes 40^1 and accompanying text (discussing evolution of the misuse doctrine in
relation to the doctrines of contributory infringement and inducement of infringement).
8. Graham vs. John Deere Co., 383 U.S. 1, 17-18 (1965) (describing new framework).
Although there is some language in the Graham opinion to suggest that the 1952 Act did
not change the law, it is important to note that the opinion ties the statutory objective
standard of non-obviousness to eighteenth century case law that employed a similar
objective standard while specifically rejecting the nineteenth century case law that
employed a subjective standard. See Kieff, supra note 7, at 88-95.
9. See Sakraida vs. Ag Pro, Inc., 425 U.S. 273, 282 (1976) (holding patent invalid because it
was a mere combination of old elements and had no synergistic effect); Andersons-Black
Rock, Inc. vs. Pavement Salvage Co., 396 U.S. 57, 61 (1969) (holding patent invalid
because '[n]o such synergistic result is argued here'). Of course, the problem with treat-
ing so-called 'combination' patents differently is that all patent claims in the present USpatent claiming system can be viewed as combinations of 'old elements.' See Kieff, supranote 7, at 1 1 1 (explaining how a claim operates as a simple logical list of elements andthat infringement is only found when each and every element on that list is present in the
allegedly infringing product or process).
10. See George M. Sirilla, 35 U.S.C 103: From Hotchkiss to Hand to Rich, the Obvious
Patent Law Hall-of-Famers, 32 J. Marshall L. Rev. 437, 543 (1999) (describing import-ance of the Federal Circuit's creation for application of the Section 103 framework).
11. Dawson Chem. vs. Rohm and Haas Co., 448 U.S. 176 (1980) (setting forth history of
Section 271 and then applying the statute to hold no misuse where the holder of a patenton a method of using a chemical as a herbicide charges customers of the herbicide above
market price for the chemical itself and sues competing chemical company for contribu-
tory infringement).1 2. See infra note 42 and accompanying text.
13. That is, we do not suggest that legislative promulgation is itself infallible. See also infra
note 23.
14. For the most influential articulation of the 'basics approach' we explore in this chapter,
and the earliest we could find, see Giles S. Rich, The Relation Between Patent Practices
and the Anti-Monopoly Laws (pts. 1-5), 24 J. Pat. Off. Soc'y 85, 1 59, 241 , 328, 422 (1942).
15. See infra notes 70-71 and accompanying text (discussing evisceration of the areas of
contributory infringement and inducement of infringement by the doctrine of patent
misuse).
16. IP rights are rights to exclude others from doing something. IP rights are not rights to
do that something. Their impact is more precisely viewed as being exclusionary than
exclusive. The impact of IP rights is only properly viewed as being exclusive in those cases
1 82 Developments in the economics of copyright
where the one exercising the right to exclude happens otherwise to be free (such as from
other rights of exclusion or other regulations) to do the excluded activity.
17. Harold Demsetz, Toward a Theory of Property Rights, 57 Am. Econ. Rev. 347, 354
(1967); Harold Demsetz, The Private Production of Public Goods, 13 J.L. & Econ. 293
(1970). For a basic overview of the economics of price discrimination, see Jean Tirole,
The Theory of Industrial Organization 133-68 (1997).
18. See generally F. Scott Kieff, Property Rights and Property Rules for Commercializing
Inventions, 85 Minn. L. Rev. 697, 727-32 (2001).
19. See generally Gordon, supra note 1; Lunney, supra note 1; Meurer, supra note 1. For
more on the debate over the impact of imperfect price discrimination on output, see
Richard A. Posner, Antitrust in the New Economy, 68 Antitrust L.J. 925, 933 n. 10 (2001):
Perfect price discrimination would bring about the same output as under competi-
tion, because no customer willing to pay the seller's marginal cost would be turned
away. But perfect price discrimination is infeasible, and imperfect price discrimina-
tion can result in a lower or higher output than under competition, or the same
output. See F.M. Scherer and David Ross, Market Structure and Industrial
Performance 494-96 (3rd ed. 1990); Paul A. Samuelson, Foundations of Economic
Analysis 42-45 (1947); Joan Robinson, The Economics of Imperfect Competition188-95 (1933). Many economists believe that even crude discrimination is more likely
to expand than to reduce output, see, e.g., Robinson, supra, at 201 ;Scherer and Ross,
supra, at 494-96; Peter O. Steiner, Book Review, 44 U. Chi. L. Rev. 873, 882 (1977),
but there does not appear to be a firm basis for this belief. See Hal R. Varian, Price
Discrimination, in Handbook of Industrial Orgnanization, at 597, 629-33 (RichardSchmalensee and Robert D. Willig eds, 1989).
20. The court must discipline itself in several respects. In part, this means limiting, or at least
connecting, judicial analysis to established positive and normative decisional frame-
works. In part, this means expanding the analysis sufficiently to include the panoply of
established positive and normative decisional frameworks of both IP and non-IP areas
of law.
21 . For an example of the powerful effect of one such specialized doctrine - patent misuse-
see infra notes 70-71 and accompanying text. For more on the importance of propertyrule protection of intellectual property and the legislative history of the present statu-
tory regimes, see Kieff, supra note 18.
22. See infra notes 70-71 and accompanying text.
23. Although, again, we do urge those promulgating IP policy at any level to integrate the
procedural aspects of our approach into their decision making processes. See supra note
13 and accompanying text (describing procedural aspect of our approach when applied
to substantive analysis of potential new policies).
24. Baxter, supra note 1, would require that the licensing arrangement be confined 'as nar-
rowly and specifically as ... technology . . . and . . . administration permit.' Baxter,
supra note 1, at 313. Bowman, supra note 1, would endeavor to determine the extent to
which the arrangement deals with something that a court later determines to be com-
petitively superior to other available options-presumably rejecting the views of parties
to the particular arrangement under scrutiny who must have elected to enter into it over
other options available at the time of entering the arrangement. Kaplow, supra note 1,
would examine the ratio between the reward the patentee receives if the arrangement is
enforced and the monopoly loss that would result. More recently, Carrier, supra note 1,
has argued to look even more broadly to ex post data about how particular industries
have experienced innovation to determine whether it has tended to be driven more by
competition or by innovation, without offering devices for measuring any of these manyfactors.
For an interesting comparative institutional analysis that stresses the importance of
certainty and predictability in judicial decision making, see Cass R. Sunstein and Adrian
Vermeule, Interpretation and Institutions, 101 Mich. L. Rev. 885 (2003).
The basics matter 1 83
25. See generally Phillip Areeda and Louis Kaplow, Antitrust Analysis: Problems, Texts,
Cases 174-250, 447-77, 785-806 (1997).
26. United States vs. Aluminum Co. of America, 1 48 F.2d 4 1 6, 430 (2d Cir. 1 945).
27. For more on the rule of reason generally, see Areeda and Kaplow, supra note 25, at
203-50.
28. See generally id. at 663-784.
29. I suppose that a patentee has no less property in his patented machine than any other
owner, and that, in addition to keeping the machine to himself, the patent gives him
the further right to forbid the rest of the world from making others like it. In short,
for whatever motive, he may keep his device wholly out of use. So much being undis-
puted, I cannot understand why he may not keep it out of use unless the licensee, or,
for the matter of that, the buyer, will use some unpatented thing in connection with it.
Generally speaking, the measure of a condition is the consequence of a breach, andif that consequence is one that the owner may impose unconditionally, he may imposeit conditionally upon a certain event. . . . The domination [over a material used in a
patented device] is one only to the extent of the desire for the [patented device].
Motion Picture Patents Co. vs. Universal Film Mfg., 243 U.S. 502, 519-20 (1917) (Holmes,
J., dissenting) (citations omitted); Giles S. Rich, The Relation Between Patent Practices
and the Anti-Monopoly Laws (pt. 4), 24 J. Pat. Off. Soc'y 328, 330 (1942) (citing same and
providing English translation from Latin for the Justinian Maxim cited by Holmes: '[one]
to whom the greater is lawful ought not to be debarred from the less as unlawful').
30. See generally Areeda and Kaplow, supra note 25, at 413-44, 686-784.
3 1 . What is particularly troubling about the approaches we criticize is that they would have
striking implications if applied in analogous fashion to the real estate transactions men-tioned here, which, of course, they are not. As discussed infra in Section 9.4 (third sub-
section), we criticize in the IP context approaches that treat restrictive contractual
arrangements as illegal. As a result, they are not only unenforceable, but also efforts to
use them would be viewed as misuse and so would lead to the property to which those
transactions relate to be essentially forfeited. See Morton Salt vs. Suppiger, 314 U.S. 488
(1942) (a finding of misuse renders the IP right unenforceable). Consider the implication
of this reasoning for a real estate transaction involving the sale of half of a parcel in
which the half that is sold is encumbered by a negative easement - such as a promise not
to build a factory that produces smelly emissions. The reasoning of the preemption
approach we criticize would allow the buyer to argue that the proper domain of restric-
tions on emissions is the body of federal environmental law and that, therefore, under
the doctrine of conflict preemption the contract term to limit use, which is a matter of
state law, is preempted and thus not enforceable. What is more, the reasoning of the
misuse approach we criticize would further allow the buyer to argue that it is a misuse of
that property right to attempt to extract - or extort - such a promise, and as a result the
property right in the entire parcel itself is forfeited. Put simply, the one-two punch of the
approaches we criticize would allow even a buyer who is sophisticated, not resource con-
strained, advised by counsel, and fully possessed of contractual intent (and therefore not
a good candidate for the contract law defenses to formation of unconscionability, adhe-
sion, duress, mistake, etc.) effectively to take possession of the entire parcel of land
without paying a cent by simply waiting for the seller to offer half the parcel encumbered
by the negative easement at a price lower than for the whole. For an example of this typeof one-two punch in the case of IP, see infra note 85 and accompanying text.
32. Although there are a number of incentive-based theories for IP that are mentioned in the
literature - including 'incentive to invent,' 'incentive to disclose' or 'teach,' 'incentive to
innovate,' and 'incentive to design around' - there are essentially three dominant theories
today: (1) some version of the 'incentive to invent' and 'incentive to disclose' theories
treated together under the rubric of 'reward;' (2) the 'prospect' theory; and (3) the com-mercialization theory. IP law certainly does have a number of important effects, and each
of these theories of IP is useful in elucidating these effects. We emphasize here the com-mercialization theory and its associated focus on coordination because at a minimum
1 84 Developments in the economics of copyright
this theory did motivate the shaping of present IP regimes as a historical fact andbecause we see the commercialization effect as the most important in that the regimescan be and in many cases are most easily and effectively adapted to achieve that goal. For
a recent review of the patent literature on incentive theories and a collection of sources,
see Chisum et al., supra note 3, at 58-90 (reviewing various incentive theories for the
patent system); Rebecca S. Eisenberg, Patents and the Progress of Science: Exclusive
Rights and Experimental Use, 56 U. Chi. L. Rev. 1017, 1024-46 (1989) (same); A. Samuel
Oddi, Un- Unified Economic Theories Of Patents - The Not-Quite-Holy Grail, 71 Notre
Dame L. Rev. 267 (1996) (same). For recent reviews of the copyright literature on incen-
tive theories and a collection of sources, see Michael Abramowicz, Copyright
Redundancy, George Mason Law & Economics Research Paper No. 03-03, available at
http://papers.ssrn.com/sol3/results.cfm (reviewing and collecting sources and highlight-
ing the opportunity cost issues discussed by Lunney as well as showing how additional
works on the margin may contribute little while at the same time causing rent dissipa-
tion); Glynn S. Lunney, Jr., Reexamining Copyright's Incentives-Access Paradigm, 49
Vand. L. Rev. 483 (1996) (reviewing and collecting sources and suggesting that incentives
may draw efforts away from other productive activities). It should be noted, though, that
the 'basics approach' would hold even if the incentive-based theories for IP were stressed.
33. Innovation is a broader term than invention and is generally understood to include the
downstream dissemination of inventions. It is sometimes also called commercialization.
34. For a discussion of the problems with efforts to reward inventive activities, see, e.g.,
Chisum et al., supra note 3 at 70-72 (reviewing so-called 'incentive to invent' theory of
patents and criticisms thereto); Kieff supra note 18, at 707-17 (reviewing problems with
reward alternatives to patents).
35. For a thorough model of the commercialization goals of IP law, see Kieff, supra note 18.
36. For an overview of contributory and induced infringement and their history in the
patent context, which is representative for the rest of IP, see, e.g., Giles S. Rich,
Infringement Under Section 271 of the Patent Act of1952, 35 J. Pat. Off. Soc'y 476 (1953).
37. See, e.g., Hewlett-Packard Co. vs. Bausch & Lomb Inc., 909 F.2d 1464 (Fed. Cir. 1990)
(inducement of patent infringement requires proof of both intent to induce and actual
direct infringement by the one induced) (citing 35 U.S.C. 271(b)).
38. 35 U.S.C. 271(c) (contributory patent infringement).39. To be sure, the ease of predicting outcomes of indirect infringement is attenuated by the
uncertainties in other aspects of IP law on which indirect infringement may depend, such
as the basic scope of IP subject matter. For example, in patent law the basic scope of the
patent right to exclude hinges on the body of law governing the field called 'claim con-
struction,' which is presently the topic of substantial debate because it is considered by
many to be too uncertain. For more on claim construction, see the recent important
empirical work by Polk Wagner at www.claimconstruction.com. For another example,the extent of the home recording and person sharing exemptions in copyright law caused
a great deal of the uncertainty surrounding the indirect infringement claims in the
famous Napster and Aimster cases. See A&M Records, Inc. vs. Napster, Inc., 239 F.3d
1004, 1020 (9th Cir. 2001); In re Aimster Copyright Litigation, 334 F.3d 643 (7th Cir.
2003).
40. Before the 1 952 Act, courts had used the misuse doctrine to erode the ability for intel-
lectual property owners to price discriminate or engage in restricting licensing. Section
271(d) expressly states that such conduct shall not be misuse. See 35 U.S.C. 271(d)(l-3)
(added by the 1952 Patent Act); see also Kieff, supra note 18, at 736-38 (discussing
history of Section 271 of the 1952 Patent Act).
41. Dawson Chem. vs. Rohm and Haas Co., 448 U.S. 176 (1980).
42. See 35 U.S.C. 271(d)(4^5) (added by Pub. L. No. 100-703, 201, 102 Stat. 4676 (1988)).
In its entirety, augmented Section 271(d) provides:
No patent owner otherwise entitled to relief for infringement or contributory
infringement of a patent shall be denied relief or deemed guilty of misuse or illegal
extension of the patent right by reason of his having done one or more of the
The basics matter 1 8 5
following: (1) derived revenue from acts which if performed by another without his
consent would constitute contributory infringement of the patent; (2) licensed or
authorized another to perform acts which if performed without his consent would
constitute contributory infringement of the patent; (3) sought to enforce his patent
rights against infringement or contributory infringement; (4) refused to license or use
any rights to the patent; or (5) conditioned the license of any rights to the patent or
the sale of the patented product on the acquisition of a license to rights in another
patent or purchase of a separate product, unless, in view of the circumstances, the
patent owner has market power in the relevant market for the patent or patented
product on which the license or sale is conditioned.
43. See, e.g., Video Pipeline, Inc. vs. Buena Vista Home Entertainment, Inc., 342 F.3d 191,
203-04 (3rd Cir. 2003) (noting that although '[njeither the Supreme Court nor this Court
has affirmatively recognized the copyright misuse doctrine . . . [tjhere is ... a well-
established patent misuse doctrine, and . . . other courts of appeals have extended the
doctrine to the copyright context.').
44. See generally Carl W. Schwarz, The Intellectual PropertyIAntitrust Interface, in 1 No. 6
Andrews Intell. Prop. Litig. Rep. 15 (2000), available at WL 7 No. 6 ANIPLR 15 (copy-
right and trademark misuse are each derived from the law of patent misuse) (citing Juno
Online Servs., L.P. vs. Juno Lighting, Inc., 979 F. Supp. 684 (N.D. III. 1997) (trademark
misuse)). For a review of intellectual property misuse, including trademark misuse, see
generally American Bar Association Antitrust Section Intellectual Property Misuse:
Licensing and Litigation (2003).
45. Patents give the patentee the right to restrict use of what is claimed in the patent.
Copyrights give the copyright holder the right to restrict copying of the creative expres-
sion embodied in the protected work. Trademarks give the trademark owner the right to
restrict use of symbols that are confusingly similar to (and in some cases also those that
dilute) the protected mark. For none of these IP systems does the IP right give its holder
some affirmative right to use. Indeed, rights to use are entirely controlled by other areas
of law. For example, a patent on a drug does not allow the patentee to avoid FDA or
EPA restrictions on the drug's use. Similarly, various criminal and other public safety
laws would restrict the holder of a patent on a gun's right to use that gun. See generally
F. Scott Kieff, Patentsfor Environmentalists, 9 Wash. U.J.L. & Pol'y 307, 307-08 (2002)
(invited symposium piece for National Association of Environmental Law Societies'
annual meeting entitled 'Sustainable Agriculture: Food for the Future,' held 15-17
March, 2002, at Washington University School of Law) (discussing how the right to
restrict use conferred by IP law does not interfere with other restrictions on use).
46. See supra note 45.
47. See Thomas W. Merrill and Henry E. Smith, Optimal Standardization in the Law of
Property: The Numerus Clausus Principle, 1 10 Yale L.J. 1 (2000).
48. Id.
49. For more on property rules versus liability rules in the context of IP, see, e.g., Kieff, supranote 18, at 732-33.
50. See generally Russell Korobkin, Bounded Rationality, Standard Form Contracts, and
Unconscionability, 70 U. Chi. L. Rev. 1203 (2003) (reviewing debates in contract law
about the applicability of the unconscionability doctrine).
5 1 . See Giles S. Rich, The Relation Between Patent Practices and the Anti-Monopoly Laws
(pis. 1-5), 24 J. Pat. Off. Soc'y 85, 159, 241, 328, 422 (1942). As suggested supra note 14,
this is one reason why Rich's views have been so influential.
52. Texaco, Inc. vs. Pennzoil, Co., 729 S.W2d 768 (Tex.App. 1987).
53. The transformation of breach of contract into patent infringement is significant. At least
one essential difference between patent infringement and breach of contract is that the
remedies for infringement include a right to exclude (i.e., property rule protection),
whereas a contract is generally viewed as little more than a promise either to perform or
to breach and pay actual damages (i.e., liability rule protection).
54. Wallace vs. Holmes, 29 F.Cas. 74 (No. 17,100) (C.C.D. Conn. 1871).
1 86 Developments in the economics of copyright
55. See Chisum et al., supra note 3 at 950-55 (discussing history of contributory infringe-
ment doctrine and the role of the Wallace case).
56. Because the chimneys are usable with non-infringing lamps and are not specially adaptedfor infringing uses, their sale falls within the safe harbors of Section 271(c). See supranotes 36-38 and accompanying text (discussing safe harbors of Section 271(c)). Put
differently, the patent could not be asserted against the sale of the chimneys as a matter
of direct or indirect infringement.57. That is, the contract for sale of the machine included a set of contract terms relating to
the patent that were written on the label that was affixed to the machine itself.
58. Heaton-Peninsular Button-Fastener Co. vs. Eureka Specialty Co., 77 F. 288 (C.C.A. 6
1896) (opinion by Lurton, C.J.) (also known as the 'Button Fastener Case').
59. Rich, supra note 51, at 251. The successful argument in Heaton - offered by Frederick
P. Fish, founding partner of the law firm formerly known as Fish, Richardson, & Neave,which later became the firms of Fish & Richardson and Fish & Neave - held out the sales
of the staples as proxies, or counters, for measuring use of the patented machine. Theymay have been, and such an arrangement would likely have been efficient. But the cause
of action against the defendant, if any, would then be some form of interference with
contract, not patent infringement. Depending on the ultimate interpretation of the label
contract, the plaintiff may have had a cause of action against the party who was a cus-
tomer of both the plaintiff and the defendant for both breach of contract and patent
infringement.60. The court opinion suggests there may have been sufficient facts to mount such an
argument.61. The court opinion does not discuss these facts, but it is likely there was no evidence of
market power. It is curious that the court did not discuss the antitrust argument, because,
as Rich pointed out, the opinion was written against a background in which antitrust
law was recently enacted: The Sherman Act had been passed six years before!' Rich,
supra note 51, at 254 (punctuation emphasis in original).
62. Henry vs. A.B. Dick Co., 224 U.S. 1 (1912).
63. These cases supported the improper view that causes of action for patent infringementcould be maintained in situations where the basics would only allow a cause of action
for some form of contractual business tort, at most. This led pro-patent courts to undulystretch the reach of patent law and to the inevitable response by anti-patent courts that
the entire body of indirect infringement should be eliminated. See infra note 65 and
accompanying text.
64. See infra note 65 (reviewing history of these cases).
65. As the Supreme Court later pointed out in Dawson, when Rich was testifying in supportof what became Section 271 of the 1952 Patent Act, 'Rich warned against going too far
[and] took the position that a law designed to reinstate the broad contributory infringe-
ment reasoning of [A.B. Dick] "would kill itself in time."' Dawson Chem. vs. Rohm and
Haas Co., 448 U.S. 176, 208 (1980) (citing Hearings on H.R. 3866 before SubcommitteeNo. 4 of the House Committee on the Judiciary, 81st Cong., 1st Sess., 17 (1949) (testi-
mony of Giles Rich)).
As the Court also pointed out in Dawson, A.B. Dick 'was followed by what may be
characterized through the lens of hindsight as an inevitable judicial reaction.' Dawson,448 U.S. at 191 (citing Motion Picture Patents Co. vs. Universal Film Mfg. Co., 243 U.S.
502 (1917) (reaching a result opposite to A.B. Dick on similar facts involving a patent ona film projector and a restrictive label contract limiting use to certain film)). CompareMotion Picture Patents, 243 U.S. at 519-21 (Holmes, J., dissenting) (arguing that the
patentee should be entitled to capture all the market generated by the invention and
expressing concerns about the transactions that had been entered in reliance on the rule
of A. B. Dick).
The law continued to fluctuate after Motion Picture Patents. In United States vs.
United Shoe Machinery, 247 U.S. 32 (1918) ('Shoe Machinery P), a case also argued for
the patentee by Frederick P. Fish, the Court returned to reasoning similar to that in
A.B. Dick to permit a complex leasing arrangement. Soon thereafter, the Clayton Act
The basics matter 187
was passed, in part, in response to cases like A.B. Dick and Shoe Machinery /, and its
Section 3 was directed to sales and leases of articles of commerce 'whether patented or
unpatented.' 15 U.S.C. 14. Not surprisingly, in United States vs. United Shoe
Machinery, 258 U.S. 451 (1922) ('Shoe Machinery //), the Court found that the leases
violated the Clayton Act. Similarly, in International Business Machines Corp. vs. United
States, 298 U.S. 131 (1936) (75AT), the Court found a set of complex leasing arrange-ments accompanied by sales of punch cards to violate the Clayton Act.
This brief review of the evolution from A. B. Dick to IBM is provided here only for his-
torical context. A significantly more complete treatment is provided in Rich, supra note
51, at 24 1-283.
66. Leeds & Catlin Co. vs. Victor Talking Mach. Co., 213 U.S. 325 (1909).
67. Id. at 335.
68. For more on peripheral claiming, see F. Scott Kieff, Perusing Property Rights in DNA, in
F. Scott Kieff, Perspectives on Properties of the Human Genome Project 135 (2003). Adetermination of infringement under a central claiming system requires the court to
determine the heart of the invention and whether the putative infringement is close
enough to that heart to justify a judgment of infringement. A determination under
peripheral claiming requires the court to determine only the outer bounds of the claim.
Anything within those bounds infringes and anything outside does not. The so-called
'doctrine of equivalents' ('DOE') that exists under the present patent system, even
though not provided for in the statute, is an odd exception to the peripheral nature of our
present peripheral claiming system because it allows the patentee to capture somethingoutside the claim. Although some commentators like this doctrine because it gives some
flexibility, they fail to see how the patentee can achieve this same flexibility in a mannerthat is not only less costly to the patentee but also to all third parties by simply draftinga better patent disclosure at the outset. F. Scott Kieff, Property and Biotechnology, in
Chisum et al., supra note 3, at 318-323 (showing how as a matter of positive law and
practice the disclosure rules of Section 1 12 of the Patent Act can operate better than the
DOE for both patentees and third parties and citing F. Scott Kieff, The Case for
Registration and the Law and Economics of Present Patent-Obtaining Rules, 45 B.C. L.
Rev. 55, 99-105, 109-1 14 (2003) (discussing the normative case for the disclosure rules
and showing how they are a better institutional choice - in terms of minimizing social
costs - for allowing both patentees and third parties to manage the problem of claim
breadth than other institutional approaches such as the DOE)).69. Again, as discussed supra note 39, the relative crispness of these doctrines can be
muddied in practice by their interaction with other, fuzzier doctrines of each IP law
regime.
70. Carbice Corp. vs. American Patents Corp., 283 U.S. 27, 33 (1931). A similar approach wasfollowed in Lietch Mfg. vs. Barber Co., 302 U.S. 458 (1938) (also known generally as
'Barber'} (Brandeis, J.) (patentee 'attempting ... to employ the patent to secure a limited
monopoly of unpatented material').
7 1 . Rich supra note 5 1, at 345 (describing the opinions of the Court in Carbice and Barber
as revealing 'a very significant preoccupation by the Court with the objective of the
plaintiffs rather than with the doings of the defendant') (emphasis in original).
72. Mercoid Corp. vs. Mid-Continent Investment Co., 320 U.S. 661 (1944) ('Mercoid T), andMercoid Corp. vs. Minneapolis-Honeywell Regulator Co., 320 U.S. 680 (1944) ('Mercoid
//') (patent on new furnace stoker switch). The same approach was used earlier in
American Lecithin Co. vs. Warfield Co., 105 F.2d 207 (C.C.A. 7, 1939) (also known gen-
erally as'
Warfield} (patent on use of lecithin as an emulsifier in chocolates to improveits properties by, for example, preventing 'whitening' after only a few days).
73. See, e.g., Dawson Chem. vs. Rohm and Haas Co., 448 U.S. 1 76, 214 (1980) ('Respondent'smethod of doing business is thus essentially the same as the method condemned in the
Mercoid decisions, and the legislative history reveals that 271(d) was designed to retreat
from Mercoid in this regard.'). Section 271 achieved this result by codifying in subsec-
tions (a), (b), and (c) those acts that would constitute direct, induced, and contributory
infringement, respectively; while at the same time codifying in subsection (d) that it
188 Developments in the economics of copyright
would not be misuse for a patentee to sue or license anyone who could be sued undersubsections (a), (b), or (c).
74. See supra notes 36-39 and accompanying text.
75. The US Court of Appeals for the Federal Circuit has jurisdiction over most appeals in
patent cases. See Federal Courts Improvement Act of 1982, P.L. 97-164, 96 Stat. 25
(Apr. 2, 1982) (creating a uniform forum for patent appeals in the Federal Circuit bymerging the Court of Claims with the Court of Customs and Patent Appeals and trans-
ferring to the new court jurisdiction over appeals from patent cases that were tried in the
district courts). Patent cases for purposes of making this jurisdictional decision are those
in which the well-pleaded complaint alleges a claim arising under federal patent law.
Holmes Group, Inc. vs. Vornado Air Circulation Systems, Inc., 535 U.S. 826 (2002).76. Virginia Panel Corp. vs. MAC Panel Co., 1 33 F.3d 860, 869 (Fed. Cir. 1997) (internal cita-
tions omitted).77. According to the Federal Circuit in Virginia Panel'.
The courts have identified certain specific practices as constituting per se patent
misuse, including so-called 'tying' arrangements in which a patentee conditions a
license under the patent on the purchase of a separable, staple good, see, e.g., MortonSalt Co. vs. G. S. Suppiger Co., 314 U.S. 488, 491 (1942), and arrangements in whicha patentee effectively extends the term of its patent by requiring post-expiration roy-
alties, see, e.g., Brulotte vs. Thys Co., 379 U.S. 29, 33 (1964). Congress, however, has
established that other specific practices may not support a finding of patent misuse.
See 35 U.S.C. 271(d) (1994); Dawson Chem. Co. vs. Rohm & Haas Co., 448 U.S.
176, 202 (1980) (construing earlier version of 271(d)). A 1988 amendment to
271(d) provides that, inter alia, in the absence of market power, even a tying arrange-ment does not constitute patent misuse. See 35 U.S.C. 271(d)(5) (1994) (added byPub.L. No. 100-703, 201, 102 Stat. 4676 (1988)).
133 F.3d at 869 (internal citations shortened).78. For an expanded discussion of this point in the context of copyright misuse, which
derives from patent misuse, see Paredes, supra note 3.
79. Rich, supra note 29, at 245.
80. ProCD, Inc. vs. Zeidenberg, 86 F.3d 1447 (7th Cir. 1996) (non-commercial use restriction
in shrink-wrap copyright license for computer program held valid and enforceable as acontractual limit on use).
81 . Mallinckrodt, Inc. vs. Medipart, Inc., 976 F.2d 700 (Fed. Cir. 1992) (single use restriction
in label license held valid and enforceable limit on grant of authority so that
unauthorized acts may support suit for infringement).82. Efforts to respond statutorily to these cases and others at the interface between contract
law and IP law, such as the proposed Article 2B of the Uniform Commercial Code andthe Uniform Computer Information Transactions Act ('UCITA'), should be conducted,if at all, along the lines of the procedural aspects of our 'basics approach' discussed
supra notes 1 3 and 23 and accompanying text.
83. For example, the payment may be for an activity that is not protectable by IP generally,
happens not to have been protected by any particular piece of IP, or was formerly pro-tected by some particular piece of IP.
84. Chief Justice Burger, writing for the majority of the Court, even allowed a promise to
pay royalties to reach activity that was never patented so long as at the time the contract
was executed it reflected both parties' reasoned assessment of the likelihood and payoffof the different states of the world under which patent rights might or might not materi-
alize. Aronson vs. Quick Point Pencil Co., 440 U.S. 257 (1979) (contract to pay royalty ona technology was enforceable even though no patent ever issued on the technology whereat the time the contract was entered into the technology might have been patented andthe contract provided a low royalty rate for the case where no patent issued and a higherrate for the case where a patent did issue).
The basics matter 189
85. See, e.g., Scheiber vs. Dolby Labs., Inc. 293 F.3d 1014 (7th Cir. 2002) (Posner, J.) (dis-
cussing at length the strength of the reasoning of the dissenting opinion of Justice
Harlan in Brulotte vs. Thys Co., 379 U.S. 29, 34 (1964) (Harlan, J., dissenting), but
nonetheless following the majority opinion in that case in refusing to enforce a properly
formed IP licensing contract - indeed, a settlement agreement from prior litigation-
among commercial parties simply because some payments happened to extend beyond
patent term at the request of the licensee). The case at the root of this line of precedent,
Brulotte, involved a patentee who sold a hop-picking machine to farmers and who had
several patents that would be infringed by such a machine. The machines were not sold
for a simple one-shot price. Instead, payment was to be made over time and based on the
actual economic advantage the machine generated for the farmer over alternative hop-
picking approaches. Because this meant that payment would extend beyond the last of
the patent terms, the Court held the contract to be unenforceable beyond that term in an
opinion written by Justice Douglas, who was well known for his dislike of patents. In
dissent, Justice Harlan pointed out that this holding would make unenforceable deals
that were actually advantageous to farmers who either were liquidity constrained at the
time of purchase or who were skeptical of the economic value of such capital equipment.86. Chisum et al., supra note 3 at 1 1 55. See generally id. at 1 1 55-96 (reviewing preemption).87. Nor do IP rights give IP owners any affirmative right to use.
88. For more on the conflict between preemption and the basics of IP law, see generally
F. Scott Kieff, Contrived Conflicts: The Supreme Court vs. the Basics of Intellectual
Property Law, 30 Wm. Mitchell L. Rev. 1717(2004) (invited piece for symposiumentitled 'The United States Supreme Court's Effect on Intellectual Property Law This
Millennium' at William Mitchell College of Law held April 24, 2004).
89. See Bonito Boats, Inc. vs. Thunder Craft Boats, Inc., 489 U.S. 141, 144-^5 (1989) (holding
that state law against so-called 'plug molding' of boat hulls was preempted by federal
patent law and citing Fla.Stat. 559.94 (1987)). To be sure, the reasoning explored in this
chapter is not new and indeed was more thoroughly set forth in the opinion by Judge Rich
in the case that was in conflict with the decision by the Florida Supreme Court in Bonito
Boats. See Interpart Corp. vs. Italia, 111 F.2d 678 (Fed. Cir. 1985) (Rich, J.) (no preemp-tion because patent law says nothing about a right to copy and because the state statute
did not even prevent copying - it merely prevented one form of copying). Similarly, the
reasoning of Bonito Boats is not new either and its roots can be found in the earlier cases
of Sears and Compco. Sears, Roebuck & Co. vs. Stiffel Co., 376 U.S. 225 (1964); CompcoCorp. vs. Day-Brite Lighting, Inc., 376 U.S. 234 (1964). As reviewed in detail throughoutChisum et al., supra note 3, Sears and Compco did not raise as many alarms as Bonito
Boats because these earlier cases came so soon after the 1952 Patent Act. It took the
Court until 1980, for the most part, to recognize the total overhaul in the framework of
patent law that was implemented by the 1952 Act.
90. Bonito Boats, 489 U.S. at 168.
91 . Compare, e.g., Bonito Boats, 489 U.S. at 152 ('[T]he federal patent laws must determine
not only what is protected, but also what is free for all to use. We have long held that after
the expiration of a federal patent, the subject matter of the patent passes to the free use
of the public as a matter of federal law.').
92. 489 U.S. at 168.
93. Also consider the real estate analogy discussed supra note 3 1 .
94. It is well recognized that Congress has the power to promulgate the statutes that create
the institutional framework for the positive law IP regimes. Patent and copyright laws
are promulgated pursuant to express authorization in Article 1 of the U.S. Constitution,
while the trademark laws are promulgated under the general commerce clause power of
Article 1 that is now recognized to be quite expansive. Compare In re Trade-Mark Cases,
100 U.S. 82, 94 (1879) (holding trademark laws to be improper exercise of the power to
promulgate patent and copyright laws and of the commerce clause power because they
regulate activity that is not sufficiently interstate) with Wickard vs. Filburn, 317 U.S. 1 1 1
(1942) (holding that even growing wheat for personal consumption in one's own back
190 Developments in the economics of copyright
yard has sufficient nexus to interstate commerce that it may be regulated by Congress
using commerce clause power).95. Even though the federalism and Supremacy Clause concerns of the US Constitution are
not applicable, the Court essentially used this same preemption approach against federal
IP law in Dastar and to a lesser extent in TrafFix. See Dastar Corp. vs. Twentieth CenturyFox Film, 539 U.S. 23 (2003) (holding that the Lanham Act does not prevent unaccred-
ited copying of uncopyrighted work and expressing concerns that otherwise the LanhamAct would interfere - or conflict - with the Copyright Act); TrafFix Devices, Inc. vs.
Marketing Displays, Inc., 532 U.S. 23 (2001) (holding that the existence of expired utility
patents in which certain design elements were mentioned created sufficiently strong evi-
dentiary inference of design's functionality that the design was not eligible for trademark
or trade dress protection and suggesting that otherwise there might be conflict between
the Lanham Act and the Patent Act); see also, Kieff, supra note 88, at 7-9 (discussing
Dastar and TrafFix cases).
96. See, e.g., Dow Chemical Co. vs. Exxon Corp., 139 F.3d 1470, 1473 (Fed. Cir.1998)
(exploring interaction between patent law and a state law providing a business tort for
interference with contract).
97. See, e.g., Hunter Douglas, Inc. vs. Harmonic Design, Inc., 153 F.3d 1318, 1336-37 (Fed.
Cir.1998) (no conflict-type preemption of various state law claims based on publicizingan allegedly invalid and unenforceable patent in the marketplace as long as the claimant
can show that the patent holder acted in bad faith in publication of the patent, which is
the 'extra element' beyond patent law).
98. Before the 1 976 Copyright Act, copyright term was shorter, but the copyright owner was
given a non-transferable renewal right. Since the implementation of the 1976 CopyrightAct, copyright term has been very long, but the original copyright owner is given a right
to terminate any transfers of that right during a statutorily-defined period within this
longer term.
99. See 17 U.S.C. 203 (governing grants executed on or after January 1, 1978); see also id.
304 (governing grants executed before January 1, 1978).
100. Digital Millennium Copyright Act, PL. 105-304, 1 12 Stat. 2860 (1998), implemented at
17 U.S.C. 1201(a)(lH2), (b)(l) (2000).
101. See supra Section 9.3, second and third subsections (discussing indirect infringementand contract, respectively).
Index
A&M Records, Inc. vs. Napster, Inc. 5
A.B. DickcasQ 171
Abramowicz, M. 104
access
and differentiated products theoryof copyright 106-107, 114
side of tradeoff 103, 105
advertising revenues 55
agency rule-making 9-10
Agreement on Trade Related Aspectsof Intellectual Property Rights
(TRIPS) 9, 15, 120,135American Society of Composers,
Authors and Publishers (ASCAP)38
ancillary products, tax on 41, 45, 51-2,
56
antitrust law 159, 161, 165-6, 175-6
appropriability 108, 134
and knowledge sharing 121, 131
Areeda, P. 150
atomization, of the production process127-33
audio equipment, tax on 5 1
Audio Home Recording Act (1992) 41,
81
audio-streaming technologies 60, 61
authors
as monopolists 103
payments to journals 47
rights of 4
'vanity' press 47
authorship 1,2
Bakos, Y. 64
Baumol,W.J. 115
Berry, S.T. 113
Besen, S. 26, 32, 64, 82, 104
blanket licenses 39, 40, 50, 53
comparison to copyright collectives
54-7
radio 54
television 54, 56
Bonanno, G. 1 1 5
Bonito Boats case 176-8, 189
book market 47
Borch, K. 31
Borenstein, S. 109
Boyle, J. 104
Brander,J.A. 115
breach of contract, vs. indirect
infringement 169-71
Brennan, D.J. 108
Brennan, T. 82
broadband 80
Broadcast Music, Inc. (BMI) 38
Brown, R. 80
cable operators 40, 41
cable transmission 55
Canada 14
Carbicecase 173
Carlton, D. 99
CDs 24, 42, 45, 51-2
and Internet piracy 60
market 50, 52
sales 47
Chamberlin, E.H. 105
Ciminello, D. 80
clubs 64
Coase theorem 23
codified knowledge 129, 130, 132, 133
Cohen, I.E. 104
commons 134
knowledge as 123
'tragedy of the commons' 122, 123
competition, and copyrighted works
104
competition law 144-5
vs. copyright 142
complementary domains 123
'composition' 124
797
192 Developments in the economics of copyright
compulsory licensing 2, 38, 39, 40, 41
Computer Associates vs. Altai, Inc. 1 1
computer programmesapplication programming interfaces
(APIs) 21
and copyright law 8, 10-14
'structure, sequence, and
organization' ('SSO') 10-1 1
comScore/Media Metrix 37
Conner, K.R. 65
consumption, non-rivalrous 103
Cooler, R. 143
copyright 1
differentiated products theory103-19
duration of 2-3, 135
extended 120, 121-2
foreign 3
impact of economics 2, 3-6
imperfections of system 43^4
justification 143^4-
length of term 45
and Locke's labour theory 143-4
markets 47
maximal 121, 133, 135
minimal 120, 132, 133, 136, 137
misuse 167, 175
models of design 81-2
nexus between industry and policymakers 4
open-ended rules 10
perpetual term 7
and policy 3-4
special exemptions 2
vs. competition law 142, 143-6
weakness in economic theory 128
welfare-enhancing effects of 121
Copyright Act 62
Copyright Arbitration Royalty Panel
(CARP) 40, 52, 54, 61
copyright collectives 2, 39
advantages of 24
amount of revenue to be raised
48-50
as coalitions 25
comparison to blanket licenses 54-7
distribution mechanisms 25
division of revenue 52^1
and efficient consumption 45-6
functions of 23
income 23-36
mispricing by 57
model 25-8
and MP3s 37-59
placement of tax 51-2
risk sharing distribution
mechanisms 28
to replace current system 40^1copyrighted works
and competition 104
as monopolies 83
copyright law 62-3
and computer programmes 8, 10-13
and database development 9
general applicability 5
literature on 2
Copyright License Board (CLB) 48,
49-50
copyright protectioninteraction among different aspects
113-14
letters and diaries 2
Copyright Term Extension Act
(CTEA)6, 111
and Eldred 6-8
Cournot market structures 104
Cournot oligopoly 99
Court of First Instance 151, 152, 153,
154
creation costs 98
'creative commons' 123, 125-7, 132,
133, 135, 136, 137
productive role 127
cultural differences 4
cultural economics 2
database development, and copyrightlaw 9
David, P.A. 126
Dawsoncase 167, 186
deadweight losses 106-107, 162
demand diversion 109
Demsetz, H. 105, 162
Depoorter, B. 82
derivative work right 2
differentiated products theory of
copyright 103-19
and access 106-107, 114
appropriability 108, 109, 112
breadth of rights 112-13
Index 193
and deadweight losses 106-107
demand diversion 109
determinants of equilibrium 111-13
early analyses 104
incentives 114
intensity of rights 111-12
and market power 106
normative implications 105-109
and optimal incentives 108-109
remedial implications 110-14
and sequential innovation 1 14
size of rights 1 1 1
ways of strengthening or weakening
rights 113-14
Digital Millennium Copyright Act
(DMCA) 6, 6 1,62, 77, 178
anti-circumvention provisions 9, 63
and decryption technology 80
digital music distribution 60-61
digital product, music as 61-2
'digital rights management' (DRM) 57,
63, 64, 77-8
fingerprinting 63
and market power 64
standards 64
vertical inefficiency 64
watermarking 63
distributional concerns 104
Dixit, A.K. 108
dominant firm structures 104
Drahos, P. 134
duration of copyright 2-3, 135
Duration Directive, EU 135
DVD decoding cases 80
Eaton, B.C. 106, 115
Eaton, J. 115
economic analysis 45differences in 5
economic aspects 1 3
economic literature, language problems4
economics, impact on copyright 2, 3-8
educational processes 131-2
Eldred vs. Ashcroft 3
and Copyright Term Extension Act
(CTEA) 6-8
effect on economics in copyright 6-8
envelope theorem 72-3
equal-share rule 31-2
European Union (EU)Court of Justice 142, 146, 149, 153
Directive on Competition Policy 1 3
Directive on sui generis 9
Duration Directive 135
exemptions from copyright 2
extended copyright 121-2
'fair use' 62, 63, 64, 121, 135
aggregate consumer surplus 87, 94-5
definition of 80
doctrine, and US copyright law 2
'equal diversion' setting 849gross benefit 87-8, 95-6
net economic welfare 92-4
number of uses 96-7
number of works trade offs 83,
84-5
optimal standards 97-8
as a policy instrument 80-102
reservation price 90-97, 98
and transactions costs 82-3, 90-97
file-sharing technologies 60, 61, 65, 71
legal actions against 62
network effects 65-6
fingerprinting, 'digital rights
management' (DRM) 63
Fisher, W.W. Ill 108, 110
Fitzgerald, D. 151
Foray, D. 126
foreign copyright 3
free entry 105, 106, 108
Geertz, C. 124, 125
'general law' 161
property and contracts 168
Goettler, R.L. 113
Gordon, W.J. 2, 80, 82, 99, 132
Graham case 160
Granovetter, M. 128
hacking 63
Hardin,G 105,121
Hart, O.D. 106
Hay, D.A. 115
Hayek, F. von 129
Heatoncase 170-71
Hettinger, E.C. 124
Hollander, A. 23
Hotelling,H. 115
123
194 Developments in the economics of copyright
Hughes,! 110
Hull, D. 151, 152
IBM 13
1CI and Commercial Solvents vs.
Commission 149
'idea-expression dichotomy' 2, 104
illegal copies, welfare implications 67
IMS Health case 142-58
background 146
Commission decision 147
court proceedings 1 52-4
licensing of brick structure 142, 146
reliance on Ladbroke case 151-2
reliance on Magillc&SQ 149-50, 151,
153
reliance on Oscar Bronner case
148-9, 154-5
incentive/exclusion trade-off 103, 105,
121
independent creation defense 2
indirect infringement 1 84
vs. breach of contract 169-71
information, as a public good 122
informational intermediaries 76
information transmission technologies67-8
infringement under IP law, vs. sui
generis\aw 171-5
innovation 161, 184
'integer problem' 115
intellectual creations, as public goods1,23,43,58,127
intellectual products 43^intellectual property
economic research on 5
and general law 168
incentive based theories 183-^
and price discrimination 162
Intellectual Property Clause, USConstitution 6
intellectual property law 159, 166-7
basics approach 161, 163-8
application 168-79
and the Beaton case 170-71
economic justification 1-2
infringement under vs. sui generis
law 171-5
periphery claims 1 87
theoretical framework 162^
intellectual property rights 130, 133
as rights of exclusion 162
Internet piracyand CDs 60
digital music distribution 60-61
effect of increased protection on
profits 74-6
with increased copyright protection,
model 66-76
information-pull technology 71-3,
74,76
information-push technology 69-7 1,
74
legal protection 66, 68, 70-71, 72, 76
technological and legal aspects60-64
technological protection 66, 68, 72
ISPs, tax on 52, 56, 59
Johnson, W. 82, 104
Jones, L.E. 106
Judd, K.L. 115
Kaldor, N. 114
Kazaa 76
Kirby, S.N. 64, 82, 104
Klein, B. 82
Klemperer, P. 110
knowledge
anthropologists on 125
codified knowledge 129, 130, 132,
133
collective and networked dimension
126
as a 'common' 123
dyadic structure 128-30
economic aspects 1 3
sharing process 124-5
social dimension 133
social nature of production 1 24-7
tacit knowledge 3, 129-30, 131, 132,
133, 135, 138
value of 125
knowledge creation, collective setting
136
knowledge sharing 125, 126, 138
and appropriability 121, 136
Koboldt, C. 104, 105
Koenker, R.W. 108
Koepp, R. 126
Index 195
Ladbrokecasc 149, 151-2
Landes, W.M. 64, 103, 104, 105, 121,
122, 143
learn-then-distribute rule 28, 29, 30, 33
Leeds & Catlin case 1 72-3
Lehmann, M. 144
Lemarchand, S. 143, 145
Lessig, L. 40
letters and diaries, copyright protection2
Liebowitz, S.J. 64, 65, 103
Lipsey,R.G. 106, 115
Liu,J.P. 110
Locke, John 1 34
labour theory 143^1
totemic paradigm 1 34
Lunney, G.S., Jr. 104
McQuail, D. 132
Magillcase 148, 149-50, 151, 152
use of term inter alia 1 50
Mallinckrodt case 176
Mankiw,N.G. 106
market failure 122, 130
market inefficiencies 44
market power 106
markets 46
copyright 47
Maskus, K. 144
maximal copyright 121, 133, 135
MercoidcasQs 174
minimal copyright 120, 132, 133, 136,
137
misuse 160, 161, 175-6
monopolies, as copyrighted works 83
monopolistic competition theory 105
monopolists, authors as 103
moral hazard 33
Morin, E. 128
MP3.com 76, 77
MP3sand album sales 49
cases involving 80
and CD sales 60
and copyright 57
and copyright collectives 37-59
downloads as a proxy for the market
50-51
measurement of downloads 53-4, 58
quality of files 65
multiple works 1 1 5
music
as an experience good 66, 68
as a digital product 61-2
musical composition 245, 25-8
heterogenous and correlated
lotteries 32-3
mutuality principle 3 1
Napster 37, 58,62-3,78, 139
National Commission on NewTechnological Uses of
Copyrighted Works (CONTU) 8
National Council for HigherEducation 82
National Research Council 82
Computer Science andTelecommunications Board 8 1
Nelson, K. 126
Nelson, R.R. 124,126,129,130Netanel, N. 41, 104
network effects, file-sharing
technologies 65-6
network externalities 4, 138
Neumann, M. 145
Neven,D.J. 115
Nielsen NetRatings 37
non-authorized audio-streaming
technologies 60, 61
non-obviousness 160, 181
non-rivalrous consumption 103
non-rivalrous good pricing 44
Novos, I. 65,69,81, 103, 105
Nozick, R. 143
open-ended rules 10
optimal incentives 108-109
Oscar Bronner case 148-9, 150, 152
Owen, B. 108
ownership 2
Parisi, F. 82
Patent Act (1952) 160, 166, 167, 169
patents 5, 12, 80, 137, 170, 180-81,185
'combination' 160
misuse 160, 167, 175
'non-obviousness' 160
and spatial competition 1 10
payola 46-7, 59
196 Developments in the economics of copyright
peer-to-peer (P2P) networks 5, 37, 60,
61,62,67-8
monitoring 65
opportunity cost 69
penalties for use 63
and record companies 76
Peitz, M. 64, 66
perfect price discrimination 44, 58
performing rights
as an ancillary market 55-6
pricing efficiency 545performing rights societies 39^0, 53,
54,56
performing rights tariffs 38, 54-5
Perloff, J. 99
perpetual term copyright 7
Perry, M.K. 108
PEW Internet project 37
piracy 133
see also Internet piracy
policy, and copyright 3-4
policy guidelines 133-5
Posner, R.A. 64, 103, 104, 105, 121,
122, 143
preemption 176-7
Prescott, E.G. 115
price discrimination 162, 167
pricing 47
privacy 63
private rights, enforcement 130-33
ProCD case 176
production, atomization of 127-33
profits, short-run 107, 116
property rights 105, 114, 121, 122
duration of 1 32
and Locke's totemic paradigm 134
on outputs 123
public choice problems 5-6
public goodinformation as 122
intellectual creations as 1, 23, 43, 58,
127
quid pro quo systems 42
radio
blanket licenses 54
excess entry to industry 113
performing rights tariff 55
and record sales 58-9
Ramello, G. 144, 145
Rate Courts 54
real-world difficulties 46-8
record industry 37-8, 42, 57
album revenues 49
and peer-to-peer (P2P) networks
76
royalty payments 55
sales and radio 58-9
Recording Industry Association of
America (RIAA) 37, 53, 62-3
research and development (R&D) costs
1,10restrictive licensing arrangements
176
Review of Economic Research on
Copyright Issues (RERCI) 15
risk sharing 23-36
Romney, A.K. 125
Rooney, D. 130
royalties 48
rates 54, 55
record industry 55
Rumelt, R.P. 65
Samuelson, P. 63
scarcity 124
Schmalensee, R. 115
Scotchmer, S. 124
Sega Enterprises, Ltd vs. Accolde, Inc.
13-14
Shachar, R. 113
sharing process, knowledge as 124
short-run, profits 107, 116
Shy, O. 65
social activities, financing 130
Society for Economic Research on
Copyright Issues (SERCI) 15
software copyright law 8, 10-14
songs, on record 40
Sony Betamax case 2
sound, social context of 126
spatial competition models 1 1 5
Spector, H. 144
Spence, M. 108
status quo systems 42
Stiglitz, I.E. 108
Stothers, Ch. 150, 155
strategic withholding 82
substitutability 112-13, 117
Index 197
sui generis 1 \ -2
European Union (EU) Directive 9
infringement under IP law 171-5
tacit knowledge 3, 129-30, 131, 132,
133, 135, 138
Takeyama, L.N. 65, 66, 82
tax, on ancillary products earlier twice
41,45,51-2,56
tax-and-subsidy system 41
television
blanket licenses 54, 56
performing rights tariff 55, 59
Texaco, Inc. vs. Pennzoil, Co. 169
Thisse, IF. 65
Torremans, P.L.C 143, 144, 150
'tragedy of the commons' 122, 123
transactions costs 23, 24, 25, 27, 32, 34,
82
and fair use 82-3, 90-97
Treacy,P. 149,150TRIPS Agreement see Agreement on
Trade Related Aspects of
Intellectual Property Rights
Ulen, T. 143
unauthorized copying, literature on64
USCongress
constitutional powers 7
extension of copyright 8
Constitution, Intellectual PropertyClause 6
copyright law, fair use doctrine 2
Copyright Office 3, 9-10
Copyright Terms Extension Act
(1998) 135
House and Senate JudiciaryCommittees 4
National Commission on NewTechnological Uses of
Copyrighted Works (CONTU) 8
'vanity' press 47
Virginia Panel case 175
Visscher, M. 115
Waelbroeck, P. 64, 66
Waldfogel, J. 113
Waldman, M. 65, 69, 82, 103, 105
Wallace case 169-70
watermarking, 'digital rights
management' (DRM) 63
Waterson, M. 115
Watt, R. 9, 104
Weber, M. 125
welfare-enhancing effects of copyright121
Whelan Associates vs. Jaslow Dental
Labs, Inc. 10-11
Whinston, M.D. 106
willingness to pay 98-9, 125
Wilson 32
Winter, S.G. 124, 129
Woolridge, F. 150
work made for hire rule 2
World Trade Organization (WTO) 9, 15
Yarrow, O.K. 113
Yoo,C.S. 104,115
Yoon, K. 82
Stanford Law Libre
3 LIDS Db3 ^37 Iflb