The Cambridge Handbook of Investment-Driven Intellectual Property 1108839193, 9781108839198

Citation preview

the cambridge handbook of investment-driven intellectual property This handbook challenges the conventional wisdom that intellectual property is the law of creativity. Traditionally, IP has been instrumental for protecting creations of the mind, with only inventors of original works enjoying exclusive rights. Related, sui generis, and quasi-IP rights, which protect monetary investments and efforts rather than originality and inventiveness, were considered exceptions to the general principles of IP. But increasingly, IP rights are being granted to safeguard corporate investments. This handbook brings together an international roster of contributors to explore this emerging trend. Why are investments the primary driver of legal protection, and often the main requirement to obtain it? Who benefits from such new forms of protection? What should the scope of these new rights be? And are they desirable in the first place? In doing so, the volume is the first to highlight and systematically critique the move from ‘intellectual’ to ‘investment’ property. enrico bonadio is a Reader in Law at City, University of London. He researches and teaches in the area of intellectual property law. patrick goold is a Reader in Law at City, University of London. He is a legal philosopher with interests in IP, property, and private law theory.

Published online by Cambridge University Press

Published online by Cambridge University Press

The Cambridge Handbook of Investment-Driven Intellectual Property Edited by

ENRICO BONADIO City, University of London

PATRICK GOOLD City, University of London

FOREWORD BY GRAEME DINWOODIE

Published online by Cambridge University Press

Shaftesbury Road, Cambridge cb2 8ea, United Kingdom One Liberty Plaza, 20th Floor, New York, ny 10006, USA 477 Williamstown Road, Port Melbourne, vic 3207, Australia 314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre, New Delhi – 110025, India 103 Penang Road, #05–06/07, Visioncrest Commercial, Singapore 238467 Cambridge University Press is part of Cambridge University Press & Assessment, a department of the University of Cambridge. We share the University’s mission to contribute to society through the pursuit of education, learning and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781108839198 doi: 10.1017/9781108989527 © Cambridge University Press & Assessment 2023 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2023 A catalogue record for this publication is available from the British Library. A Cataloging-in-Publication data record for this book is available from the Library of Congress isbn 978-1-108-83919-8 Hardback Cambridge University Press & Assessment has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

Published online by Cambridge University Press

Contents

page vii

List of Figures and Tables List of Contributors

ix

Foreword by Graeme B. Dinwoodie

xi

Introduction

xv

I

II

Creativity, Pluralism, and Fictitious Narratives: Understanding IP Law through Karl Polanyi Uma Suthersanen

1

part i science, technology and industry

25

Sui Generis Protection of Non-creative Databases Caterina Sganga

27

III Test Data Exclusivity: An Elusive Pursuit to Strike a Balance between Affordable Drugs and Investment Incentives Daria Kim

54

IV Copyright in Works Created by Artificial Intelligence: Between Creativity and Investments Enrico Bonadio, Luke McDonagh, and Plamen Dinev

73

V

86

Plant Variety Protection and Investment Viola Prifti

VI Software Protection under Copyright Law: Interoperability and Protection of Program Interfaces Noam Shemtov

98

VII Bilski and the Information Age a Decade Later Michael J. Meurer

114

VIII Pharmaceutical Patents and Evergreening Hazel V. J. Moir

133

v Published online by Cambridge University Press

vi

IX

Contents

part ii culture and entertainment

153

The Press Publishers’ Right under EU Law: Rewarding Investment through Intellectual Property Stavroula Karapapa

155

X

Copyright in Published Editions: What Lessons Does It Teach Us? Rita Matulionyte

167

XI

Protecting Sound Recordings Between Investments and Creativity Enrico Bonadio

182

XII

Copyright in Broadcast Transmissions and the Investment-Protection Rationale 203 Bryan Khan

XIII

Copyright Protection of Previously Unpublished Works Patrick Masiyakurima

XIV

Cinematographic Works and Copyright in Nollywood: The Clog in the Wheel of Creativity and Originality Ayoyemi Lawal-Arowolo

246

part iii signs, images and designs

259

XV

The Investment Function of Trade Marks Richard Arnold

261

XVI

The Protection of Well-Known Trade Marks as a Way to Protect Investment? Ilanah Fhima

XVII Ambush Marketing and Protection of Investments Arul George Scaria and Varsha Jhavar XVIII EU Geographical Indications and the Protection of Producers and Their Investments Andrea Zappalaglio XIX

Design Right: From Investment to Creativity for ‘Industrial Copyright’ Phillip Johnson

XX

The Philosophical Foundations of Investment-Driven IP: On Reason, Faith, and Pluralism Patrick Goold

Published online by Cambridge University Press

227

275

291

308

329

341

Figures and Tables

figures 5.1 Plant patent applications filed worldwide by main companies in the seed sector 5.2 Patent applications on plant-related patents

page 92 93

tables 5.1 Classification of the potato plant/Solanum tuberosum 14.1 Comparing the evolution of investments in Nollywood and Hollywood

vii Published online by Cambridge University Press

page 87 257

Published online by Cambridge University Press

Contributors

Richard Arnold (Royal Court of Justice, London) Enrico Bonadio (City, University of London) Plamen Dinev (Goldsmiths University) Ilanah Fhima (University College London) Patrick Goold (City, University of London) Varsha Jhavar (National Law University of Delhi) Phillip Johnson (University of Cardiff ) Stavroula Karapapa (University of Essex) Bryan Khan (University of The West Indies) Daria Kim (Max Planck Institute for Innovation and Competition, Germany) Ayoyemi Lawal-Arowolo (Babcock University, Nigeria) Patrick Masiyakurima (University of Leicester) Rita Matulionyte (Macquarie University, Australia) Luke McDonagh (London School of Economics) Michael J. Meurer (Boston University) Hazel V. J. Moir (Australian National University) Viola Prifti (independent researcher and conference manager for the European Academy for Taxes, Economics and Law, Berlin) Arul George Scaria (National Law University of Delhi) Caterina Sganga (Sant’Anna School of Advanced Studies, Pisa) Noam Shemtov (Queen Mary University of London) Uma Suthersanen (Queen Mary University of London) Andrea Zappalaglio (University of Sheffield)

ix Published online by Cambridge University Press

Published online by Cambridge University Press

Foreword Graeme B. Dinwoodie

The justifications for different forms of intellectual property protection have always been somewhat ecumenical. In particular, as many chapters in this wonderful book note, advocates have often toggled between the moral obligation to reward personal creativity (or, alternatively, labour) on the one hand, and more instrumentalist arguments based upon incentives to behave in particular socially useful ways, on the other. Without intruding too far into the semantics of the recent debate between Rob Merges and Mark Lemley,1 adherence to all of these approaches depends upon some first-order commitments. Faith in facts is no less a faith. A devotion to supposed empirical instrumentalism is simply fealty to a different core tenet than that from which deontologists draw support for the same or similar rights. We each have our own imperfect deities. Intellectual property law (and lawyers) is no different. Moreover, any assessment of the underlying basis for intellectual property protection must also reckon with the relatively late provenance of the term ‘intellectual property’ both internationally and domestically. As Lionel Bently commented a decade ago: ‘“intellectual property” has rapidly moved from being a category of laws developed from the convenient presentation of distinct legal regimes into a legal concept – a category in law’.2 At the international level, the systems of ‘literary and artistic property’ and ‘industrial property’ have merged into ‘intellectual property’. But the resonance and normative foundations of those concepts differs, as they always have. Most consequentially, trademark and unfair competition law sometimes appear in different ways to be outsiders at the familial sibling get-together. But even at the national level, there are longstanding unresolved debates over whether ‘trade secrets’ or ‘publicity rights’ or ‘unfair competition’ can really be called ‘intellectual property’.3 One might argue not so much that the nature of intellectual property law has shifted with the expansion of the forms of protection canvassed in this volume, but that the entire legal category has lost any coherence. With all that said, the editors of this volume have identified an important shift in the way that we (and especially legislators and policymakers at the international level) think and talk about intellectual property. Increasingly, emphasis is placed upon the importance of protecting 1

2 3

See Mark A Lemley, ‘Faith-Based Intellectual Property’ (2015) 62 UCLA Law Review 1328; Robert P Merges, ’Against Utilitarian Fundamentalism’ (2015) 90 St. John’s Law Review 681. Lionel Bently, ‘What Is Intellectual Property?’ [2012] Cambridge Law Journal 501, 502 (emphasis added) Cf eg, Lionel Bently, ‘Trade Secrets “Intellectual Property” But Not “Property”?’ in Helena Howe and Jonathan Griffiths (eds) Concepts of Property in Intellectual Property Law (CUP 2013) 60, with Emily Hudson, ‘Phillips v Mulcaire: A Property Paradox’ in Simon Douglas, Robin Hickey and Emma Waring (eds) Landmark Cases in Property Law (Hart 2017).

xi https://doi.org/10.1017/9781108989527.001 Published online by Cambridge University Press

xii

Foreword

‘investment’. This enhanced prominence for the concept of investment takes a number of different forms. First, it clearly drives the stated justifications for several new national laws. Second, it represents the protectable res in certain new sui generis schemes, such as the database rights created in the European Union. Third, at the international level, treating intellectual property as ‘investment’ offers the possibility of new forms of international protection under investor–state dispute resolution mechanisms. This shift is therefore consequential, even if describing it as a fundamental reorientation of intellectual property requires greater certitude about the coherence of what went before than I think history allows. How should we measure the significance of this shift? First, in order to quantify it, it is important to establish a baseline. Historically, not all intellectual property law has been conceptualised as protecting creators (despite the label given to the field). In particular, trademark law confers rights on the person with whom a sign is associated, not upon the person who came up with the sign. This is not new. In some countries, such as the United States, the work-for-hire doctrine in copyright law has long conferred the status of authorship on the person who paid for the creation of the work rather than upon the individual creator – what some have called an ‘economic conception’ of authorship. And international treaties in the field have for decades recognised and regulated what civil-law countries would call neighbouring rights; rights conferred on, for example, the producers of sound recordings and broadcast organisations. Such regimes simultaneously recognise the important claim of certain types of investment in contributing to the creative ecosystem while granting lesser rights than available to ‘true creators’ such as artists and authors. Second, even when doctrinal terminology has long talked about the protection of creativity, sometimes that masks a policy commitment to investment. Thus, patent protection turns on the importance of novelty and inventiveness, but those standards work to tell us that it is concerned with encouraging investment in activity that meets those thresholds. The EU Database Directive, a modern exemplar of the new commitment to investment, made this doctrinally explicit,4 when the Court of Justice of the European Union in William Hill refused to confer rights as a result of investment in activities not the focus of the legislation.5 Third, unmasking and making transparent these embedded assumptions, as this book does repeatedly, is immensely valuable. Many of the doctrinal concepts in intellectual property law have little intrinsic content. To make sense of them, and give them practical meaning, requires stripping them down to reveal underlying policy motivations or concerns. This can sometimes be done close to the surface, making explicit what doctrine barely conceals. But sometimes this requires more intense analysis, exploring how neutral doctrines nominally unconcerned with investment privilege it in important ways. A recent example is the work of Jeanne Fromer, who argues that the doctrines surrounding the establishment of trademark rights via the concept of secondary meaning effectively protect investment in marketing.6 Judicial declarations that secondary meaning turns on the effects of marketing investment rather than the investment as such offer only a thin rebuttal. Fourth, recognizing the shift affects outcomes in concrete cases. Most obviously, this occurs in case where investment is textually the sine qua non of a protectable right, as in the EU database rights regime. But in a system where purposive interpretation is important (and it

4 5 6

Football Dataco v Sportradar [2013] EWCA Civ 27 at [44]. British Horseracing Board v William Hill, Case C-203/02 [2004] ECR I–10415 (CJEU Grand Chamber). Jeanne C Fromer, ‘Against Secondary Meaning’ (2022) 98 Notre Dame Law Review (forthcoming).

https://doi.org/10.1017/9781108989527.001 Published online by Cambridge University Press

Foreword

xiii

normally is, even in jurisdictions that might tout alternative interpretive methodologies), it can help guide decisionmakers. Thus, a UK court stressed the importance of assessing infringement of copyright in films or broadcasts by reference to the rationale for protection of such works, namely, protection of the investment, rather than whether there has been taking of an author’s own intellectual creation as might be relevant for authorial works.7 Finally, the excavation that the book performs should start a broader conversation about the relationship between investment and creativity, perhaps beyond the many illustrations discussed in the pages that follow. For example, we tend to frame the contest between those competing lodestars in terms of the contribution of the right holder. But a fuller appreciation of the relative power of creativity and investment to shape the contours of protection might also take into account the ways in which those tropes are deployed by defendants as well. Thus, in Google v Oracle, an example of the recent US Supreme Court decision on fair-use defence, the majority opinion of Justice Breyer placed weight on the fact that much of the value of the plaintiff’s computer code derived from the investments made by third-party computer programmers.8 (In some respects, this was an inversion of the ‘reaping where you have not sown’ metaphor that grounded liability in INS v Associated Press).9 In dissent, Justice Thomas suggested such considerations were irrelevant. And in an earlier US Supreme Court case on fair use, the Court (while affirming that a defendant cannot escape liability simply by showing how much of his work was not copied) recognised that the creative contribution of a defendant might still help tip the fair-use scales in its favour.10 This book is a clear-eyed exploration of what drives contemporary intellectual property law across a number of jurisdictions. It draws back the curtain in ways that make one see several existing forms of protection in a different light, but will also provoke the reader into rethinking a range of broader assumptions about how and why we protect intellectual property. The book beautifully embodies the pluralistic values at the heart of its inquiry: the investment of time by the contributors and their creativity of thought are both evident throughout. Readers will be richly rewarded by investing their own time in reconsidering whether and to what extent intellectual property law is about more than the protection of creativity. Chicago, October 2022

7 8 9 10

English & Wales Cricket Board Ltd v Tixdaq Ltd and Fanatix Ltd [2016] EWHC 575 (Ch), paras [65]–[66] (Arnold J). Google LLC v Oracle Am Inc 141 S Ct 1183 (2021). Int’l News Serv v Associated Press 248 US 215 (1918). Campbell v Acuff-Rose Music Inc 510 US 569, 589 (1994).

https://doi.org/10.1017/9781108989527.001 Published online by Cambridge University Press

https://doi.org/10.1017/9781108989527.001 Published online by Cambridge University Press

Introduction

On the surface, intellectual property (IP) law is a body of rules and principles that grant legal rights in the products of the mind. The law’s central doctrines – for example, originality and non-obviousness – restrict IP rights to the outputs of the mental faculties of creativity, ingenuity, and imagination. But do not take our word for it. The idea of IP law as intellectual property, properly so called, can be found in a vast range of legal and non-legal sources. Judges say that copyright protects the ‘creative powers of the mind’,1 and originality is its ‘sine qua non’.2 American presidents say that patents add the ‘fuel of interest to the fire of genius’.3 British prime ministers claim that we live in ‘the most creative and imaginative country on earth’ and ought to have the ‘gumption to exploit our intellectual property’.4 And lexicographers define IP as property in the ‘work of the mind or intellect’.5 But scratch below the surface and a different picture emerges. The lofty idea of IP law as property of the mind has long been more rhetoric than reality. In the nineteenth century, IP lawyers increasingly justified IP protection through appeals to the romantic author or genius inventor and their natural rights.6 But during the same period, the very same lawyers sought to extend IP protection by excising these mythical figures from the realities of legal doctrine. On the one hand, jurists like Justice Oliver Wendell Holmes Jr extolled the virtues of originality and creativity, while, on the other, they decided that almost any work (such as circus lithographs) would be sufficiently creative to enjoy copyright protection.7 With a straight face, patent lawyers wrote that the ‘design of the patent laws is to reward those who make some substantial discovery or invention’ and that such inventors ‘are worthy of all favour’,8 while at the same time they embraced employer ownership of inventions. In America, what emerged was a new ideological scheme that was full of contradictions and built-in tensions.9 In Great Britain, the result was a

1 2 3

4

5 6 7 8 9

Trademark Cases 100 US 82, 94 (1879) (USA). Feist Publications Inc v Rural Telephone Service Co 499 US 340, 345 (1991) (USA). Abraham Lincoln, ‘Lecture on Discoveries and Inventions, Jacksonville, Illinois (February 11, 1859)’ in Lincoln: Speeches and Writings: 1859–1865 (Literary Classics of the United States 1989) 11. Boris Johnson, ‘Now Here’s a Wizard Idea: Why Not Bring Harry Potter Home?’ The Telegraph (7 June 2010) . Merriam Webster’s Dictionary. Oren Bracha, Owning Ideas: The Intellectual Origins of American Intellectual Property, 1790–1909 (CUP 2017). Bleistein v Donaldson Lithographing Company 188 US 239 (1903) (USA). Atlantic Works v Brady 107 US 192, 200 (1882) (USA). Bracha (n. 6).

xv https://doi.org/10.1017/9781108989527.002 Published online by Cambridge University Press

xvi

Introduction

shift away from a pre-modern regime that focused on the mental creativity that produced the intangible and towards a modern IP law that focused on the intangible itself as a commodity, regardless of the amount of creativity required for its genesis.10 As Uma Suthersanen in her opening chapter to this book (‘Creativity, Pluralism, and Fictitious Narratives: Understanding IP Law through Karl Polanyi’) explains, while IP is a pluralist concept which protects a vast variety things for a vast variety of reasons, increased participation of corporations since the nineteenth century has led to an increasingly commodified and industry-driven IP law. This book is concerned with the latest phase in this historical tension. While IP law is still described and justified in terms of creativity and ingenuity, legal reality is becoming ever more distanced from such concerns. In the late twentieth and early twenty-first centuries, IP law is increasingly composed of sui generis rights, related or neighbouring rights, and quasi-IP rights, which explicitly protect non-creative and non-inventive intangibles. At the same time, the extension of traditional IP rights, such as copyright and patents, to trivially and insignificantly creative intangibles has picked up pace. And while the subject matter of these rights – the databases, the phonograms, the pharmaceutical test data, and so on – may not be the product of creativity, it is all the product of monetary investment, and so, it is claimed, it ought to be protected. Such ‘investment-driven’ IP rights began life as an exception to the principle (or myth?) that IP protects the works of the mind or intellect. But today, they are the exception that increasingly characterises the entire field. This pattern is also evident in trademark law. One might potentially think of trademarks as the original investment-driven IP right. So familiar are we of including trademarks within the field of IP that we sometimes forget how historically and conceptually contentious it is to think of trademarks as property,11 let alone intellectual property.12 Yet today, investment-driven extensions to trademark rights have resulted in a creeping propertisation of signs and symbols. As confirmed in several decisions by the Court of Justice of the European Union, the ‘investment’ function of trademarks has become of paramount importance.13 Increasingly, the public can no longer use trademarked signs when to do so would interfere with the sign’s ability to build up a reputation.14 This is particularly important in the case of well-known trademarks, where ‘anti-dilution’ provisions do not aim at avoiding consumers’ confusion in the marketplace, but have the purpose of protecting the investments and efforts made by the trademark owner to make the sign popular among the public. Against this background, the essays in this collection explore the increasingly non-creative and investment-driven nature of IP law in a variety of jurisdictions and from a variety of methodological perspectives. The book is divided into three parts. Part I examines the rise of investmentdriven IP in relation to technology, science, and industry. Some of the chapters in this part consider the new investment-driven IP rights lawmakers enacted in the late twentieth and early twenty-first centuries. Caterina Sganga (in ‘Sui Generis Protection of Non-creative Databases’) explains how and why the EU Database Directive came to protect non-creative databases through the sui generis database right. Daria Kim (in ‘Test Data Exclusivity: An Elusive 10 11 12 13

14

Brad Sherman and Lionel Bently, The Making of Modern Intellectual Property Law (CUP 1999). Mark A Lemley, ‘The Modern Lanham Act and the Death of Common Sense’ (1999) 108 Yale Law Journal 1687. Trademark Cases (n. 1). See the CJEU judgment of 22 September 2011 in Interflora Inc and Interflora British Unit v Marks & Spencer plc and Flowers Direct Online Ltd (C-323-09). See the CJEU judgment of 23 March 2010 in Google France SARL and Google Inc v Louis Vuitton Malletier SA (C236/08); Google France SARL v Viaticum SA and Luteciel SARL (C-237/08); and Google France SARL v Centre national de recherche en relations humaines (CNRRH) SARL and Others (C-238/08).

https://doi.org/10.1017/9781108989527.002 Published online by Cambridge University Press

Introduction

xvii

Pursuit to Strike a Balance between Affordable Drugs and Investment Incentives’) considers the economic case for providing legal exclusivity over pharmaceutical test data. Similarly, Enrico Bonadio, Luke McDonagh, and Plamen Dinev (in ‘Copyright in Works Created by Artificial Intelligence: Between Creativity and Investments’) consider whether works created by artificial intelligence also ought to enjoy sui generis protection, and Viola Prifti (in ‘Plant Variety Protection and Investment’) argues that countries must have flexibility in how they provide protection to plant varieties. Meanwhile, some of the chapters focus on the investment-driven extensions to existing IP rights. Noam Shemtov (in ‘Software Protection under Copyright Law: Interoperability and Protection of Program Interfaces’) focuses on the adequacy of intellectual property in offering appropriate protection against copyists, especially where copying occurs in relation to inter-operability – allowing facets of computer programs which are crucial for a thriving software market capable of attracting investment. And Michael Meurer (in ‘Bilski and the Information Age a Decade Later’) critiques the claim that patents are required to encourage firms to invest in the development of new business methods. Hazel V. J. Moir (in ‘Pharmaceutical Patents and Evergreening’) concludes this section by exploring how pharmaceutical companies claim patents on insignificantly inventive modifications to drugs to delay generic competition in the market. Part II turns to culture and entertainment. Once again, some chapters consider newly added investment-driven rights, while other chapters consider investment-driven extensions to old rights. In the former category, Stavroula Karapapa (in ‘The Press Publishers’ Right under EU Law: Rewarding Investment through Intellectual Property’) explores the rationale and history of the sui generis press publishers’ right created by the EU Directive on the Digital Single Market. In the latter camp, Rita Matulionyte (in ‘Copyright in Published Editions: What Lessons Does It Teach Us?’), Enrico Bonadio (in ‘Protecting Sound Recordings: Between Investments and Creativity’), and Bryan Khan (in ‘Copyright in Broadcast Transmissions and the InvestmentProtection Rationale’) explore what British copyright lawyers call the ‘entrepreneurial works’, that is, unoriginal works which enjoy copyright protection (or related rights protection). Meanwhile, Patrick Masiyakurima (in ‘Copyright Protection of Previously Unpublished Works’) and Ayoyemi Lawal-Arowolo (in ‘Cinematographic Works and Copyright in Nollywood: The Cog in the Wheel of Creativity and Originality’) illustrate how copyright is frequently used not to encourage creativity, but to encourage investments into publication and dissemination. Part III analyses investment-driven IP rights in relation to signs, symbols and designs. The contributions to this part explore how recent extensions to trademark rights encourage investment in businesses and goodwill. Lord Justice Richard Arnold (in ‘The Investment Function of Trademarks’) pins down the nature of the ‘investment function’ and what it means for comparative advertising and debranding. Much like Suthersanen, Ilanah Fhima (in ‘The Protection of Well-Known Trade Marks as a Way to Protect Investment?’) reminds us that IP is a pluralist concept through an examination of European dilution law. While the protection of investments is a common theme in this area, there can be no single normative basis for such an expansive power. Meanwhile, Arul George Scaria and Varsha Jhavar (in ‘Ambush Marketing and Protection of Investments’) consider how a range of IP rights protect the investments necessary to bring about major sporting events, and Andrea Zappalaglio (in ‘EU Geographical Indications and the Protection of Producers and Their Investments’) consider the sui generis protection of geographical indications. Penultimately, in a helpful counter-thesis, Phillip Johnson (in ‘Design Right: From Investment to Creativity for “Industrial Copyright”’) explains how the British design right was born out of a desire to protect investments, but over time has gradually become more

https://doi.org/10.1017/9781108989527.002 Published online by Cambridge University Press

xviii

Introduction

focused on creativity. One wonders what causes some rights to abandon investment-driven concerns, when others move in the opposite direction. Finally, Patrick Goold (in ‘The Philosophical Foundations of Investment-Driven IP: Reason, Faith, and Pluralism’) concludes by evaluating the philosophical arguments underpinning investment-driven IP rights and reflecting on what the growth of investment-driven IP in the Information Age tells us about contemporary IP.

https://doi.org/10.1017/9781108989527.002 Published online by Cambridge University Press

I Creativity, Pluralism, and Fictitious Narratives Understanding IP Law through Karl Polanyi Uma Suthersanen

Patent law has been built around the mental activity of invention and copyright upon that of aesthetic creation. In both cases it has been the commercial investors – the factory owners and the book sellers – behind these legal figureheads – the inventors and authors – who have driven the campaigns to rid markets of direct imitators.1

I.I INTRODUCING THE CHAPTER’S AIM

In the 2021 best-seller Exponential,2 Azhar explores social evils that will invariably result from exponential technological leaps such as excessive resource extraction, destabilization of social, economic and political institutions, and the dramatic rise of corporate power in societal and economic terms. Like Polanyi, he fixes the industrial revolution as the point of genesis of current social problems. And, akin to the Polanyian ‘double movement’, Azhar suggests that a dystopian world is not inevitable if we accept the vital importance of institutional norms – such as the rule of law or international intellectual property (IP) agreements – in protecting the social and economic fabric of humanity. Reform, he argues, should lie in forging new customs and norms based on collective ownership and commonality principles such as interoperability. Or even in instituting a global data body which creates ‘a consistent approach towards artificial intelligence, citizens’ data and intellectual property’.3 This chapter is a simpler narrative which nevertheless suggests, as Azhar does, that the solution to IP conundrums lies in the recognition that global IP norms are vital and that their reformation must be in accordance with societal motivations. Purists argue that the sole beneficiaries of IP law should be those individuals or companies that are responsible for outputs resulting from identifiable human-led creativity. Progressives would counter with a vision of post-anthropocentric creativity mixed with human labour and capital investment. It would be argued in the latter camp that IP rights should vest in communal groupings, legal persons and corporations, non-creative collaborators, virtual and geographical entities, non-humans, non-sentient and AI-based machines. For over a century, jurists have debated as to whether IP rights should be extended to this or that output such as photographs, smells, computer programs, indigenous knowledge, products of data aggregation, personalities, 1 2 3

WR Cornish, Intellectual Property: Omnipresent, Distracting, Irrelevant? (OUP 2004) 42. A Azhar, Exponential: How Accelerating Technology Is Leaving Us behind and What to Do about It (Random House 2021. Azhar, Exponential 187; also see 79–83, 249–57. Azhar does not directly cite Polanyi, except by reference to James Boyle’s reference to Polanyi’s The Great Transformation as a precursor to a discussion on IP enclosures.

1 https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

2

Uma Suthersanen

virtual objects, natural and organic materials including medicines, plants, wine, ‘creative’ algorithms and genes. No doubt, fresh conundrums will continue to arise. What is surprising is the continuing and bifurcating discourse landscape: ‘IP law protects and rewards creativity’ versus ‘IP law protects investment’. This collective angst then tumbles into the perennial conundrum of ‘what is IP law’ and identifying the right justification to rationalize the different subspecies of rights. From my perspective, legal protection of some sorts has extended, over a quintennial period, to all types of intangible res for various rationales. It should be entirely unsurprising that IP law has been accorded to various beneficiaries for various types of products and services, emanating from various motives. Corporate entities, sole and collective inventors and authors, social or trade communities and local groups will labour and manufacture for a variety of needs from creative impulses to ensuring livelihoods, to market demands, to rentseeking behaviour. However, value could, and still does, attach to any intangible res. IP law indiscriminately and simultaneously protected all sorts of res emanating from creative and entrepreneurial labour as well as from capital investment. This is precisely how and why this species of law was constructed. And continued attempts to discuss or adjudicate IP law, policy and reform through strict conceptual silos, such as sole author/inventor, creativity, labour and investment, are doomed narratives. Instead, I invite readers to appreciate that early modern rules in sensu lato were about mixed motivations of reward, investment, recognition, competition, honour, public interest and more. The implicit suggestion in this chapter is that the root cause of continuing legal conundrums today is the failure to appreciate the pattern set out by two related socio-legal mappings: (i) motivations for early rules in relation to creative and entrepreneurial labour and products in Europe, and (ii) the extent to which current IP norms and laws evolved to absorb such myriad motivations over 500 years. For conceptual guidance, I employ Karl Polanyi’s theory on the commodification of labour, including investment-backed labour, into tradeable, fictitious commodities. Law and property were viewed as vital institutional tools which disembedded the economy from societal considerations. Polanyi’s motivation ethos (discussed below) is effective in synthesizing the different mapping points to create a pointilliste landscape which, in turn, shows why IP law only makes juridical sense if it is viewed as a pluralistic concept. Furthermore, to locate and socio-legally map motivations, I employ three questions which may not necessarily be answered here, but serve to steer the landscaping exercise: (i) Can we identify a set of clear and repeated motivations for early IP rules, privileges, entitlements and laws? (ii) Do different motives attach to the different rationales for current IP law? (iii) Were such IP laws continuously recalibrated to protect creative or derivative outputs emanating from labour and/or capital? Section I.I maps out the pluralistic dynamics which flow when employing Karl Polanyi’s vision of society and market systems. I do not set out to prove that Polanyi is factually accurate, but to distil his key concepts. Section I.II applies Polanyi’s concepts to map the literature on the genesis and institution of IP-type rules in Europe from the fifteenth century. The legal landscaping focuses on the diverse motivations underpinning early rules in relation to different IP res. Section I.III shows why the legal mapping, based on Polanyi’s motivation ethos, concludes that IP must be a pluralist notion and, as such, is being constantly reconstructed based on varying values and interests. The chapter concludes by exploring how a pluralist concept of IP can further embed Polanyi’s ‘counter-movement’ concept as a means of preventing dire societal consequences.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

Creativity, Pluralism, and Fictitious Narratives

3

I.II POLANYI’S THE GREAT TRANSFORMATION

Why Polanyi? In reviewing the historical transformation of humanity from a feudal society to a market society, Polanyi contends that history records an economy that was, and should be, embedded in non-economic institutions and social relations.4 A 2016 survey as to the ten most important twentieth-century writings by economists placed Polanyi’s The Great Transformation second only to Keynes’s General Theory of Employment, Interest and Money.5 But Polanyi’s book remains influential also in the fields of anthropology, economic archaeology, and economic sociology and law.6 Cotula, for example, employs the Polanyian perspective to examine ‘land grabs’ and international investment law as it highlights the commodification of land and labour.7 Ebner claims that ‘Karl Polanyi remains a most influential proponent of an institutionalist perspective in the social sciences, combining socio-economic analysis with a normative drive for social reform’.8 And I find a theoretical companion in Peukert who urges that a Polanyian perspective of IP law will ensure that ‘market-based transactions coexist with non-market modes of accessing and sharing information so that authors, inventors, and other entrepreneurs have as many options as possible at hand, and all members of society possess adequate possibilities to acquire knowledge’.9 Polanyi’s theoretical critique of the capitalist society, distilled in The Great Transformation,10 declaims the reductionist practice of consolidating land and human labour as an exchange value whereby what has no price appears to have no value in society. His thesis is that the selfregulating market economy is pernicious because it assumes that all actions and behaviours in everyday life are led by motives of gain.11 This is shown through his focus of land, labour and money as constituting ‘fictitious commodities’. Though the commodity concept is not new (having been part of classical economic theory from Adam Smith to Karl Marx),12 Polanyi’s novel approach was to distinguish between genuine and fictitious commodities and to link the commodification process to his ‘double movement’ theory. The narrative in The Great Transformation ends optimistically with Polanyi’s assertion that capitalism fails in its impossible utopian goal of totally disembedding societal considerations. A wholly self-regulating market

4 5

6

7 8 9 10 11 12

See discussion below. Schumpeter’s Capitalism, Socialism and Democracy and Galbraith’s The Affluent Society were placed at third and fourth positions, respectively – see Kari Polanyi Levitt, ‘The Return of Karl Polanyi: From the Bennington Lectures to Our Present Age of Transformation’ in R Desai and K Polanyi Levitt (eds) Karl Polanyi and Twenty-First-Century Capitalism (Manchester University Press 2020) 22 (hereafter Desai and Levitt). M Hudson, ‘Debt, Land and Money: From Polanyi to the New Economic Archaeology’ in Desai and Levitt 61–77; for legal scholarship, see A Perry-Kessaris, ‘Reading the Story of Law and Embeddedness through a Community Lens: A Polanyi-Meets-Cotterell Economic Sociology of Law?’ (2011) 62(3) Northern Ireland Legal Quarterly 401, and the essays in C Joerges and J Falke (eds), Karl Polanyi, Globalisation and the Potential of Law in Transnational Markets (Hart 2010). Lorenzo Cotula, ‘The New Enclosures? Polanyi, International Investment Law and the Global Land Rush’ (2013) 34 (9) Third World Quarterly 1605. A Ebner, ‘Transnational Markets and the Polanyi Problem’ in Joerges and Falke, Globalisation and the Potential of Law in Transnational Markets. Alexander Peukert, ‘Fictitious Commodities: A Theory of Intellectual Property Inspired by Karl Polanyi’s “Great Transformation”’ (2019) 29 Fordham Intellectual Property, Media and Entertainment Law Journal 1151, 1200. Karl Polanyi, The Great Transformation: The Political and Economic Origins of Our Time (1944); this chapter refers to the edition by Beacon Press 2001, with a new foreword by Joseph E. Stiglitz) – hereinafter referred to as ‘Polanyi’. Polanyi 31. Polanyi defines the market economy as implying ‘a self-regulating system of markets; in slightly more technical terms, it is an economy directed by market prices and nothing but market prices’, 46. N Sammond, ‘Commodities, Commodity Fetishism, and Commodification’ in The Blackwell Encyclopedia of Sociology (George Ritzer ed., Blackwell 2007) 607–12.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

4

Uma Suthersanen

just cannot exist as it will annihilate human and society. Instead, there is a dynamic countermovement against reducing the planet and humanity into value-laden commodities. And law and property are key characters in his narrative. I.II.A Technology and Disembeddedness In reviewing the historical transformation of humanity from a feudal society to a market society, Polanyi contends that history records an economy that was, and should be, embedded in noneconomic institutions and social relations.13 Polanyi sets out a pre-industrial revolution English landscape whereby economic activities were regulated through determined state intervention, including craft guild regulations and feudal privileges, as opposed to self-regulating markets.14 Adopting an Aristotelian approach to what constituted ‘good society’, Polanyi pointed to 2,000year-old norms and traditions that ensured provision for everyone, where markets and monies were mere accessories to an otherwise self-sufficient household and where production was for use.15 His account of such traditions, from tribal to mercantilist societies, including laws and regulations (which can now be appreciated as codifications of community responsibility and different motivations), maps economic activities that were not automatically framed by motives to maximize utility profit or other acquisitive motives.16 Economic activities, such as manufacture, exchange, distribution and consumption of goods and services, were subordinated to nonmarket means of social integration. Such activities were by-products of social relationships framed by non-economic institutions and which reflected societal interests, collective conventions (including kinship) and shared norms. The industrial revolution, on the other hand, wrought profound changes on the way human societies existed. Conceding that it did herald an era of technological progress which undoubtedly led to miraculous improvements in the tools of production, Polanyi also describes how, with the advent of technology, non-profit motivations and communitarian traditions were swept away by an ethos which demanded production for gain: people moved from fields, farms and workshops into factories with deplorable working conditions, greater working hours and looser regulatory oversight. Although technology is not necessarily the sole cause of the bifurcation between society and the economy, it is clear that Polanyi views machines and technology as an essential feature of the causative chain which led to the great transformation of human society: from regulated to selfregulating markets; from the embedded society to a disembedded society; from a holistic human and natural world to one where capitalist theorists and entrepreneurs demanded an institutional separation of this embedded society into economic and political spheres.

13

14

15 16

It is disputed as to whether Polanyi himself placed such a great significance on the concept of embeddedness since he used the term sparingly in The Great Transformation – see chapter 6, and 135. Nevertheless, his use of the concept of embeddedness has been continuously discussed within economic, political and sociological circles with the most popular explanation being that the concept refers to either actors’ social networks or to social institutions. See M Beckert, ‘The Great Transformation of Embeddedness: Karl Polanyi and the New Economic Sociology’ in C Hann and K Hart (eds), Market and Society: The Great Transformation Today (CUP 2009) 38–55. Polanyi 56, 68–69, 73. This is not to state that there were no self-regulated competing markets at all; some markets were non-competitive (functioning as barter and trade activities), but other markets were competitive (in the instance of local or village markets, 63). Polanyi 56. Polanyi 56.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

Creativity, Pluralism, and Fictitious Narratives

5

I.II.B Labour and Investment as Fictitious Commodities Polanyi viewed a commodity as something that has been produced for sale on a market. Nature, society and humanity were commodified and entered the market as land, money and labour – entirely fictitious commodities that could be bought and sold freely on the market.17 This account explains not only why the new artificial commodities were to become the main mechanism for sourcing goods and services, especially through technology, but also why there is a change of motivations on the part of the members of society: the motive of subsistence, or State interest, was substituted to that of gain.18 Polanyi was particularly scathing as to the commodification of humans and went so far as to equate the creation of a labour market to ‘an act of vivisection performed on the body of society’.19 As Stiglitz accurately notes, part of Polanyi’s argument is based on morality: namely, it is simply wrong to treat nature and human beings as objects which are priced by the market. This violates societal norms on human dignity and sacred principles.20 Indeed, Polanyi’s warnings on the eventuality and consequences of global land commodification will appeal to the environmentalists and Greta Thunbergs among us today, though they must have presented a startling, if not mad, vision in 1941: The mobilization of the produce of the land was extended from the neighbouring countryside to tropical and subtropical regions – the industrial-agricultural division of labor was applied to the planet. As a result, peoples of distant zones were drawn into the vortex of change the origins of which were obscure to them, while the European nations became dependent for their everyday activities upon a not yet ensured integration of the life of mankind. With free trade the new and tremendous hazards of planetary interdependence sprang into being.21

It is also within this angst-ridden discourse that we find an explanation as to why capital investments occupy a central role in economic development. The ‘invention of elaborate and therefore specific machinery and plant’ called for ‘long-term investment with corresponding risks’.22 This, in turn, called for safeguards to all elements of supply, namely labour, land and money. Otherwise, investment into the production is simply ‘too risky to be undertaken both from the point of view of the merchant who stakes his money and of the community as a whole which comes to depend upon continuous production for incomes, employment and provisions’.23 Thus, in a self-regulated commercial society, such supply could only be ensured if such elements were commodified artificially and organized for sale on the market. Law is viewed as a vital institutional tool in The Great Transformation. It transforms land and labour (including investment-backed labour) into tradeable, fictitious commodities, thus enabling disembeddedness. Detrimental laws included rules that abolished Elizabethan legislation on minimum wages (likened to ‘a right to live’), or rules that disassembled privileges and laws in 17

18 19 20 21 22 23

Polanyi 75. ‘Labor is only another name for a human activity which goes with life itself, which in its turn is not produced for sale but for entirely different reasons, nor can that activity be detached from the rest of life, be stored or mobilized; land is only another name for nature, which is not produced by man; actual money, finally, is merely a token of purchasing power which, as a rule, is not produced at all, but comes into being through the mechanism of banking or state finance. None of them is produced for sale. The commodity description of labor, land, and money is entirely fictitious’; also see 42–43, 45. Polanyi 43–44. Polanyi 132. Stiglitz’s Foreword in Polanyi xxv. Polanyi 190. Polanyi 78. Polanyi 43.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

6

Uma Suthersanen

England and France, which had previously restricted land enclosures but now allowed unregulated enclosures or that accelerated the commodification of land and labour.24 All such rules, Polanyi asserted, were due to the eagerness of entrepreneurs to purchase not just more land but also human labour, which was now unleashed into an unregulated marketplace to be dealt with as a commodity ‘which must find its price in the market’.25 And for the proper functioning of the market system, property rights are seminal as it is ‘the legal aspect . . . of capitalism’: ‘While the actual content of property rights might undergo redefinition at the hands of legislation, assurance of formal continuity is essential to the functioning of the market system.’26 I.II.C Counter-Movements However, such a self-regulating market cannot exist infinitely as it will annihilate humankind and society. When land, labour and money are transformed into fictitious commodities, and when the market economy becomes too exploitative, society reacts with counter-movements which are a reaction against the ‘dislocation which attacked the fabric of society, and, which would have destroyed the very organization that the market called into being’.27 At this point, there is an institutional re-assertion of regulation, laws, policies and measures to check the actions of the market in relation to labour, land and money. These would also morph with the establishment of rules to protect the liberty and rights of human beings, and to guarantee the continued existence of man and nature.28 This shift from first disembedding to then (re)embedding is Polanyi’s ‘double movement’ concept. The notion of a continuous dialectical tension between moving towards commodifying everything within society, and the counter-movement away from markets towards social protection or social goals is also espoused by other economists and sociologists: Hayek, for instance, adopts this ‘double movement’ notion, albeit coming to an opposing conclusion, namely, that market liberalism expands humanity and life.29 As stated above, legal reform is a key agent in enabling and intervening in both types of movements: first to facilitate the creation of new fictitious commodities and the securing of investment; second to rein-in market forces and to adjust the more pernicious effects of a completely unregulated/self-regulated market economy.30 We return to the value of employing the counter-movement concept in relation to IP reform in Section I.IV. I.II.D IP as a Motivations Landscape: Labour, Creativity and Investment In Peukert’s examination of Polanyi’s theory as applied to IP, he argues that there is no correlation between property rights and technological progress.31 This is true to a certain extent 24 25 26 27 28

29 30 31

Polanyi 82–86, 122. Polanyi 122. Polanyi 178 and 243; see also Peukert 1161. Polanyi 3–4, 136. Polanyi 71–80. See also chapter 7‘s discussion of the ‘Speenhamland Law’ which Polanyi construes to have been such a social measure which introduces a minimum wage and a ‘right to live’. See also Peukert 1189–1190, ‘The countermovements Polanyi studies are self-protective measures of society at large against the destructive effects of the fictitious commodification of labor and land. Their purpose is to re-embed these commodities and the respective markets into society and the environment, with the ultimate aim to guarantee the continued existence of man and nature.’ FA Hayek, The Fatal Conceit: The Errors of Socialism (University of Chicago 1988) 134. Cotula 1609. Peukert 1163

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

Creativity, Pluralism, and Fictitious Narratives

7

if we are seeking to situate the current IP legal system in the past, and if we use Polanyi to discover ‘property’. I do not. Instead, we are on a hunt as to whether early motivations for rules on certain types of res were transformed into the nineteenth-century framework of national and international IP laws. By using the Polanyian concept of an embedded society as it existed prior to the Industrial Revolution, we encounter specific rules which affected the production and exploitation of things, labours, industries and cultures. Employing the Polanyian measure of ‘motivation’ as to what constitutes an embedded society, we encounter similar mixed motivations for pre-modern quasi-IP rules, that is, individual, societal, communitarian, State and market motivations. And this exploration of mixed motivations within an embedded society does explain to a certain extent why IP is a contested concept without a uniform inherent meaning even before the commodification wave instilled by the reform of IP law in the nineteenth and twentieth centuries. I.II.E Early Modern Era The genesis of modern IP rules derives from a recognizable corpus of legal instruments within Western society in the last five centuries. This is not to say that these initiatives were the earliest exemplars of IP conventions and rules. Wengrow’s archaeological insights on fourth millennium BC Mesopotamia indicate the development of commodity marks, seals and labels in relation to terroir and standardization – thereby signalling the existence of value-added things and labour in relation to mass consumption products.32 Roman publishers operated a quasi-unfair competition regime whereby their investments were protected from misappropriation.33 Shao’s seminal scholarship on Chinese legal history indicate early trademark regulations during the Han Dynasty (206 BC–AD 220), and a well-established authorial rights system by the Ming Dynasty (1368–1644).34 Long has argued that it is ‘clear that within medieval cities the attitude developed that craft processes constituted intangible property with commercial value subject to conditions of ownership’.35 Ford’s recent scholarship casts new light on the links between the Carolingian encouragement of Christian learnings, the legal immunities for monasteries and the eventual Venetian rules on inventions.36 And there are now attempts to uncover early Islamic understandings on property in intellectual labour.37 Although it has been argued that the Venetian regulations were cross-cultural hybrids aimed specifically to encourage and protect skills and technologies from the Eastern Mediterranean, Eurasia and the Islamic Empire, the Venetian (and other European) privileges and regulations in the early modern era do stand, nevertheless, as stark examples of pre-IP rules.38 Peukert deals 32 33

34

35 36 37

38

D Wengrow, ‘Prehistories of Commodity Branding’ (2008) 49(1) Current Anthropology 7, 9–11. FA Mumby, Publishing and Bookselling: A History from the Earliest Times to the Present Day (4th ed. Jonathan Cape 1956) 15–17; F Reichmann, ‘The Book Trade at the Time of the Roman Empire’ (1938) 8(1) The Library Quarterly 42, 43. K Shao, ‘Alien to Copyright?: A Reconsideration of the Chinese Historical Episodes of Copyright’ (2005) 4 Intellectual Property Quarterly 400–31; K. Shao, ‘The Promotion of Learning in Chinese History: To Discover the Lost Soul of Modern Copyright’ (2010) 24(1) Columbia Journal of Asian Law 63–85; K Shao, ‘Look at My Sign: Trademarks in China from Antiquity to the Early Modern Times’ (2005) 87 Journal of the Patent and Trademark Office Society 654. P Long, ‘Invention, Authorship, Intellectual Property and the Origin of Patents: Notes Toward a Conceptual History’ (1991) 32(4) Technology and Culture 846, 876. Laura R Ford, The Intellectual Property of Nations (CUP 2021) ch. 6. M El Said, ‘Rethinking the Foundations of Intellectual Property: Applying Islamic Principles on Selected Contemporary IP Challenges’ in P Drahos, G Ghidini and H Ullrich (eds) Kritika: Essays on Intellectual Property (Edward Elgar 2015) 94–130. LR Bradford, ‘Inventing Patents: A Story of Legal and Technical Transfer’ (2015) 118 West Virginia Law Review 267.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

8

Uma Suthersanen

with these early privileges quickly as they do not qualify in his quest for property or commodification.39 It is true that these early regulations had no uniform regional or state policy as to what would constitute the nature of the right or res protected.40 Nonetheless, IP jurists and scholars will detect familiar language in these instruments as they reflect the mixed motivations of authorities, creators and investors including compulsory licensing, incentivization, unjust enrichment and the protection of investment/labour/honour. For example, the general Venetian State laws of 1453 and 1474 gave temporary monopolies to importers and inventors of innovations while personal privileges were granted for all sorts of artisanal and technological processes and products.41 Scholarship, as pointed out above, claims that State privileges and guild practices in this period were tools for harnessing and regulating foreign technologies as well as censoring domestic printing. However, it is also accepted that such privileges and practices (in the forms of municipal favour [gratiae] or exceptions to law [priva lex]) worked to secure and protect the investment, and in some cases, acted as positive entitlements or licences to practice trades to foster domestic competition with the then dominant Venetian trade guilds.42 The Venetian 1474 Statute, for example, incorporated several policies reflecting individual and communitarian needs. The law enumerated defensive mechanisms to prevent misappropriation of the machine as well as the technical manuals and trade secrets attached thereof.43 It incorporated rhetorical language as to the protection of the inventor’s honour. And public interest is ensured with the linking of the regulation to the promotion of competition and the incentivization of more technology, as well as to the express provision allowing government use.44 The pattern of encountering mixed motivations which include protecting investment of capital, reward and recognition of labour, and public or societal benefits continue as we shift timelines and jurisdictions. Ginsburg’s study on sixteenth-century Vatican printing privileges, for instance, show that these were granted on the bases of investment (for the labour and expense invested in producing work), incentivization and reward (to the author or printer for ensuring 39 40

41

42

43 44

Peukert, 1163 (‘Windmills, mining technologies, and the printing press were not immediately regulated by freely transferable property rights but for a long time by privileges.’) J Kostylo, ‘From Gunpowder to Print: The Common Origins of Copyright and Patent’ in R Deazley, M Kretschmer and L Bently (eds) Privilege and Property: Essays on the History of Copyright (Open Book Publishers 2010) 21–50 (all references hereafter are to numbered paragraphs) 2–6; see also O Bracha, ‘The Commodification of Patents 1600–1836: How Patents Became Rights and Why We Should Care’ (2004) 38 Loyola of Los Angeles Law Review 177. For copyright-like privileges, see A J Mann, ‘“A Mongrel of Early Modern Copyright”: Scotland in European Perspective’ in R Deazley, M Kretschmer and L Bently (eds) Privilege and Property: Essays on the History of Copyright (Open Book Publishers 2010). Including one for introducing a Byzantine device into Venice (1416 privilege to Petri), another for the art of printing in Venice for five years (1469 privilege to Johannes of Speyer), and most famously, the 1486 privilege to Marc Antonius Sabellico for the author’s right to choose a printer, and to prevent reprinting. See G Dutfield and U Suthersanen, Dutfield and Suthersanen on Global Intellectual Property Law (2nd edn, Edward Elgar 2020) 81; G Mandich, ‘Venetian Patents (1450–1550)’ (1948) 30 Journal of the Patent and Trademark Office Society 166; J Loewenstein, The Author’s Due: Printing and the Prehistory of Copyright (University of Chicago Press 2002) 71; cf G Mandich, ‘Venetian Origins of Inventor’s Rights’ (1960) 42 Journal of the Patent and Trademark Office Society 378; E W Hulme, ‘The History of the Patent System under the Prerogative and at Common Law’ (1896) 12 Law Quarterly Review 141; and Terrell on the Law of Patents (19th edn, Sweet & Maxwell) para 1–07 et seq. C May and S Sell, Intellectual Property: A Critical History (Lynne Rienner 2005); J Kostylo, ‘From Gunpowder to Print: The Common Origins of Copyright and Patent’ (n 12); T Sichelman and S O’Connor, ‘Patents as Promoters of Competition: The Guild Origins of Patent Law in the Venetian Republic’ (2012) 49 San Diego Law Review 1267. Kostylo 45. Translations differ. G Mandich and FD Prager, ‘Venetian Patents 1450–1550’ (1948) 30(3) Journal of the Patent Office Society 166, 176–77; cf. J Kostylo (2008) ‘Commentary on the Venetian Statute on Industrial Brevets (1474)’ in Primary Sources on Copyright (1450–1900) eds L Bently and M Kretschmer . Also see J Phillips, ‘The English Patent as a Reward for Invention: The Importation of an Idea’ (1982) 3(10) Journal of Legal History 71.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

Creativity, Pluralism, and Fictitious Narratives

9

the accuracy of the work), fair competition (as the privileges forbade unfair or unjust enrichment of another’s fruits), and public interest. She further suggests that these privileges show nascent concepts of derivative works, first publication rights and the French droit de divulgation.45 The 1479 Episcopal Privilege, granted by the Bishop of Würzburg to three printers for the printing of the Bishop’s breviary book, controlled several aspects of the tangible and the intangible res including (i) defensive mechanisms such as censorship, and sanctions for the theft of the printers’ goods and chattels involved in the manufacture of the books, and (ii) positive mechanisms such as the exclusive right to print, and to use the Bishop’s coat of arms as a proto-trademark designating the origin of the book.46 Both the papal and bishopric privileges espoused public interest rationales: the former was to ensure the dissemination of accurate religious tracts and to enhance the likelihood of the creation of future beneficial works with some petitions pursuing future rights for translation; the latter was to enable a market structure which protected the printing investment by preventing competition (e.g., the emergence of derivative and new versions of the book). Identical motivations, namely investment, labour and public interest arose within early English laws and jurisprudence. In the 1602 Case of Monopolies, the court’s biblical basis was not merely against anti-competitive monopolies but also on the traditional principle of ensuring protected labour for the welfare of the commonwealth.47 The 1624 English Statute of Monopolies juxtaposes patents against the competing public interests of state welfare and regulated competition.48 The element of public interest was further seen in later English patents which incorporated compulsory working and revocation clauses.49 Many motivations were further rooted to emerging Renaissance and common law doctrines on humanism, equity and culture instancing these examples: the moral interests of inventors, the economic concerns of investors and the wider societal benefits (such as consumer protection against counterfeiting and deceit) were treated as having complementary values; notions of dignity and conscience were assumed to be inherent in some types of labour by English equity courts which eschewed utilitarian bases to adopt deontologically biased notions of ‘good conscience’.50 These show, perhaps, the validity of Polanyi’s theoretical idea of an embedded society. The example that encapsulates Polanyi’s idea that laws, as institutional regulators, were vital in upholding an embedded society with its pluralist concerns was Brunelleschi’s application for a privilege on a newly improved cargo boat. The application for a monopoly was claimed on the

45

46

47

48

49 50

JC Ginsburg, ‘Proto-Property in Literary and Artistic Works: Sixteenth-Century Papal Printing Privileges’ in I Alexander and H Tomas Gomez-Arostegui, The History of Copyright Law: A Handbook of Contemporary Research (Edward Elgar 2015). F Kawohl, (2008) ‘Commentary on the Privilege Granted by the Bishop of Würzburg (1479)’ in Primary Sources on Copyright (1450–1900) eds L Bently and M Kretschmer ; E Armstrong, Before Copyright: The French Book-Privilege 1498–1526 (CUP 1990). ‘No man shall take the nether or the upper millstone to pledge: for he taketh a man’s life to pledge’ (Deut. 24:6); it is reported that Edmund Coke concluded that the scripture showed that ‘a man’s trade is accounted his life, because it maintaineth his life; and therefore, the monopolist that taketh away a man’s trade, taketh away his life and therefore is so much the more odious’. See Darcy v Allen, 77 Eng. Rep 1620 (K.B. 1603) (known also as the Case of Monopolies), discussed in B Malament, ‘The Economic Liberalism of Sir Edward Coke’ (1967) 76 Yale Law Journal 1321, 1343–44. Statute of Monopolies, 1624, 21 Jam. 1, c. 3 (Eng.). Early decisions are: Davenant v Hurdis, Moore 576, 72 Eng. Rep. 769 (K.B. 1599) (known also as the Merchant Tailor’s Case); Cloth Workers of Ipswich 78 Eng. Rep. 147 (K.B. 1615) – for a synopsis, see Terrell on the Law of Patents (n 10) paras 1–11 to 1–22. Also note Malament, ibid ‘The Economic Liberalism of Sir Edward Coke’ (1967) 1345 et seq. Bracha 12–13. H Macqueen, ‘The War of the Booksellers: Natural Law, Equity, and Literary Property in Eighteenth-Century Scotland’ (2014) The Journal of Legal History, 35(3) 231–57; A Hudson, ‘Equity, Confidentiality and the Nature of Property’ in H Howe and J Griffiths (eds) Concepts of Property in Intellectual Property Law (CUP 2013) 94–115.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

10

Uma Suthersanen

grounds of intellect, industry, invention, fear of theft of the ‘fruit of his genius and skill’, a desire to ‘disclose’ his knowledge and a need to incentivize further ‘higher pursuits’.51 So, when do we see the transformation of privileges into Polanyian fictitious commodities? Is it true, as Polanyi asserts, that property is the institutional tool whereby labour is transformed into commodities? In Peukert’s view, what eventually led to the commodification of activities and artefacts into IP is the technological advances from the mid-eighteenth century which made it clear that ‘publishers, authors and inventors require property right that enables them to recoup their sunk investments in the first prototype’. His eighteenth-century examples are the transformations which occurred in the German book trade (switching from a barter in books model to a sale of their productions model), and the French revolutionary decrees of 1791/1793.52 Peukert further notes that ‘the evaluation of works and inventions became more and more detached from arguments of traditional aesthetics and public benefits. In the end, the only relevant value remaining was the market-conception of value: every use value, however motivated or characterized, has to translate into an exchange value.’53 Peukert declares that the ‘history of IP from the nineteenth to the 21st century is replete with examples of this Polanyian logic of commodification’.54 I agree and based on Peukert’s analysis, I extrapolate his observations to map out several global consequences which, in turn, frame our current IP norms in the next section. Nevertheless, I would also argue that one cannot easily ignore the earlier system of privileges and rules which granted not only manufacturing, printing, publication and quasi-trademark rights but also rights of importation into the territory which show to some extent Polanyian concepts of why an embedded society reflected pluralist motivations. First, at least in Europe, this ad hoc system reflected the main concerns of the ruling elites, that is, to attempt a balance between censorship, regulation of technologies, barriers to market entry and an ethos of ‘good values’ or ordre public. Creators appeared as beneficiaries in early rules, but it was not until the early eighteenth century when they were regarded (at least rhetorically) as primary stakeholders, rather than secondary or even minor stakeholders in the world of technologies and arts. Second, there were no boundaries as to the themes for which petitions were granted though from the eighteenth century onwards, IP concepts, rules and rationales appeared. But such concepts were cross-hybridized versions which grew from a corpus of medieval European legal institutions that protected both creator and investor, mental labour and private investment, in widely marketable objects. I.II.F Res Becomes Fictious Commodities First, Peukert is indeed right to locate the eighteenth century as being the approximate start of discourses on works and inventions as ‘labour’. Hegel, for instance, refers to ‘mental aptitudes, erudition, artistic skill [. . .], inventions’ as mental activities which constituted ‘mental property’ the moment they became expressed and alienated and subsequently ‘put into the category of ‘things’. Hegel is fully cognizant of the role of property and the law in relation to mental labour: it is ‘the primary means of advancing the sciences and arts’ so as to guarantee scientists and artists against theft, and to enable them to benefit in a similar way that property is used to advance trade and 51 52 53 54

Kostylo 41–42 Peukert 1161–65. Part of this supposition arises from Immanuel Kant’s essay ‘On the Wrongfulness of Unauthorized Publication of Books (1785) – see Dutfield and Suthersanen, 45–47. Peukert 1165. Peukert 1167.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

Creativity, Pluralism, and Fictitious Narratives

11

industry against robbery. Ever sensible of balance, Hegel can also be read as arguably justifying the rights of users to appropriate others’ mental property to create derivative works.55 Bently and Sherman also note, in relation to British IP law, the importance of the notion of ‘mental labour’ as the basis of property rights in the eighteenth century especially as there was no law of copyright, patents, designs or trademarks. Used to distinguish such labour from manual labour, the term ‘mental labour’ formed a common link between different forms of mental or intellectual activities and know-how and may have been wide enough to encompass personality.56 Second, there is an abundance of literature on the subsequent transformations in national law and policy from the nineteenth century onwards whereby ‘mental labour’ which is embedded within physical objects (for example, recipe books or manufactured objects) is extended to cover ancillary outputs in the form of confidential information, trade indicia, and know-how (to name a few). These were slowly wrenched away from their physical manifestations, as if that is all there was, and aggregated with the incorporeal part of the ‘thing’. The latter existed thus as discrete value-laden commodities which were, in time, recognized by courts and legislators as intangible res. Peukert points to the dis-embeddedness of signs, for example, from medieval guild practices to stand-alone accessories attached to commodities traded on the competitive market. I would add that the expansion of trademark protection to three-dimensional shapes has been complemented by expansive justificatory rhetoric while simultaneously dealing with the possible anticompetitive effects of IP protection of shapes.57 Indeed, the function of trademark law in the EU has clearly extended beyond origin to communication, investment and advertising.58 Unsurprisingly, the first international patent and copyright conventions absorbed the then dominant national positions on the nature and ambit of intangible property, ambiguous as they were.59 One notes the origins of the French wine appellation system which grew around concerns as to genericide of place names (a cultural value), consumer protection as to fraud and product origin (a traditional trademark function), and rent seeking (a very modern trademark rationale for the protection of investment).60 These concerns were subsequently embedded in later eras so that the

55 56 57

58

59

60

GWF. Hegel, Philosophy of Right ( TM Knox trs and ed, Clarendon Press 1967) §43, §69. L Bently and B Sherman, The Making of Modern Intellectual Property Law (CUP 1999) 2–3, 15–18. For the problem of shape protection under trade dress, passing off, design, copyright, unfair competition and trademark laws, see A Kur, ‘Unité De L’Art is Here to Stay: Cofemel and its Consequences’, Max Planck Institute for Innovation and Competition Research Paper No. 19-16; U Suthersanen and M Mimler, ‘An Autonomous EU Functionality Doctrine for Shape Exclusions’ [2020] GRUR International 567; I Simon, ‘Technical Functionality in European Trade Mark Law’ (2021) 137 Law Quarterly Review 113–40; J Reichman, ‘Legal Hybrids between the Patent and Copyright Paradigms’ (1994) Columbia Law Review 94, 2432. See U Suthersanen, ‘The European Court of Justice in Philips v Remington: Trade Marks and Market Freedom’ (2003) 7 Intellectual Property Quarterly 257; WM Landes and RA Posner, The Economic Structure of Intellectual Property Law (Belknap Press/Harvard University Press 2003) chapter 7; A Kur, ‘TradeMarks Function, and Don’t They? CJEU Jurisprudence and Unfair Competition Principles’ (2014) 45 International Review of intellectual Property and Competition Law 434; Art 3, EU Trade Marks Directive 2015/2436/EU; Art 4, Community Trade Marks Regulation (EC) No 207/2009; L’Oreal v Bellure [2010] RPC 1 (wherein the CJEU arguably extended the trademark protection to brand value); for a critique of L’Oreal v Bellure, see, for instance: D Gangjee and R Burrell, ‘Because You’re Worth It: L’Oreal and the Prohibition on Free Riding’ (2010) 2 Modern Law Review 282–304; and L McDonagh, ‘From Brand Performance to Consumer Performativity: Assessing European Trademark Law after the Rise of Anthropological Marketing’ (2015) 4 Journal of Law and Society 611–36. It has been argued that the Paris Convention 1883 on industrial property and the Berne Convention 1886 on copyright followed the quantitative, incentive-based, approach which had moulded national IP laws. See R Dreyfuss and S Frankel, ‘From Incentive to Commodity to Asset: How International Law Is Reconceptualizing Intellectual Property’ (2015) 36 Michigan Journal of International Law 557, 561. A Stanziani, ‘French Collective Wine Branding in the Nineteenth–Twentieth Centuries’ in D Gangjee (ed) Research Handbook on Intellectual Property and Geographical Indications (Edward Elgar 2016) 13–45.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

12

Uma Suthersanen

protected or regulated res would not only include tangible artefacts and human labour but natural phenomena (including scientific, medical principles and discoveries), and geographical culture (as in terroir in regulating beer and wine in Germany and France).61 Thus, in the true Polanyian spirit, the commodification process within the nascent international IP regime extended not just to labour but products of the land with which labour was arguably mixed.62 This was a logical extension of the ‘let’s commodify all’ spirit. This may clarify the extent of the term ‘industrial property’ as used in the earlier texts of the Paris Convention. The term successfully captured industrial and agricultural extractive outputs including wines, grain, fruit, cattle and mineral waters; by 1925, the provision extended to include tobacco leaf, and by 1934, to such staples as beer and flour.63 All these outputs would be successfully repurposed under different IP sub-constructs, that is. patents, industrial designs, trademarks, appellations of origin, etc.64 Since the 1970s, the type of intangible res which has been categorized as IP has accelerated: patent protection has extended to computer programs, medicines, biotechnological inventions, life forms and business methods.65 A similar phenomenon occurred under the Berne Convention 1883,66 while national and regional laws continued to press for expanded subject matter: from photographs, to films to press publications, and to controlling dissemination of works on all sorts of media and platforms.67 Perhaps the most notable example of Polanyi’s concerns as to the vagaries of market-led regulation is how the principle of unfair competition was harnessed to secure fictitious commodities trading in the open market. The principle of unfair competition is principled on the notion that in accordance with ‘fair practices’, a competitive market will comprise actors which conform to certain types of acceptable market behaviour. And yet Art.10bis of the Paris 61

62

63

64 65 66 67

The 1791 French law was in relation to ‘Useful Discoveries, and to the Means of ensuring the Property thereof’. See F Savignon, ‘The French Revolution and Patents’ (1989) Industrial Property 391–400. Also see D Gangjee, ‘From Geography to History: Geographical Indications and the Reputational Link’ in I Calboli and W Ng-Loy (eds) Geographical Indications at the Crossroads of Trade, Development, and Culture: Focus on Asia-Pacific (CUP 2017) 44 et seq (for the absorption of early branding regulations into trademark and geographical indication laws); Kostylo, 42–44 (for the privileges sought for medical and technical manuscripts in fifteenth- and sixteenth-century Venice, in contrast to the usual practice of secrecy) and A Pottage and B Sherman, ‘On the Prehistory of Intellectual Property’ (arguing that early notions were focused on material embodiments of the intangible) in Concepts of Property in Intellectual Property Law op cit 13–15. This recalls Locke’s labour-based justification of property which has been challenged by many, the most notable being Nozick’s parable as to whether a person standing on a beach emptying a can of tomato juice into the ocean can claim ownership based on the mixing of his thing with another. Ownership claims based on mixing labour with terroir (as in geographical indication rights in food and wine) can become slightly absurd when the mixed labour is imported. Of note was the French Government’s intervention to prop the legitimacy of the French wine brand ex post la crise du vin caused by the phylloxera ravage of the French vineyards and the subsequent questionable re-plantation of cultivated hybrid vines from the USA and North Africa into French terroir. For challenges to Locke, see J Waldron, ‘Two Worries about Mixing One’s Labor’ (1983) 33 Philosophical Quarterly. 37; and R Nozick, Anarchy, State and Utopia (Basic Books 1974) 174–75. For the French saga, see D Gangjee, Relocating the Law of Geographical Indications (CUP 2012) chapter 3. Convention of Paris for the Protection of Industrial Property: Final Protocol of 20 March 1883; Act of The Hague of 6 November 1925; Act of London of 2 June 1934. Also see the Law of 28 July 1824 (prohibiting use of place names except by authorized wine producers), Law of 1857 (on marks); Law of 11 January 1892. Art 1, Paris Convention (Act of Stockholm 1967). The extent of expansion is still a sovereign matter although the TRIPS Agreement, and other international instruments set a minimum framework of obligations. The Convention begun with all expressions and fixations of speech and text, performances, music, images, dance and expanded into industrial design and photographs – Art 2, Berne Convention. See Dutfield and Suthersanen op cit 55. Berne Convention for the Protection of Literary and Artistic Works, 9 September 1886; Trade-Related Aspects of Intellectual Property (TRIPS) Agreement, 15 April 1994; Copyright in the Digital Single Market Directive (EU) 2019/ 790.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

Creativity, Pluralism, and Fictitious Narratives

13

Convention on Industrial Property can be read as converting workable competition rules (i.e., fair practices) into a positive entitlement for IP proprietors.68 Thus, the historical and current legislative narratives as to why we need sui generis database rights, plant variety rights, geographical indications, sound recording rights and pharmaceutical test data exclusivity manifest as confused and extensive discussions on whether ‘entitlements’ were justified at all within copyright or patent or trademark or unfair competition laws.69 I.II.G Rationales, Typology and Lexicon: More Fictitious Narratives The third point is that, as all sorts of labour and objects became qualified as tradeable commodities, Procrustean solutions were adopted in order to accommodate the expanding res either by accretion (which involves re-purposing an existing category to encompass new commodities) or emulation (whereby new res is brought within the law by creating new rights which emulate existing rights and rationales).70 Not only was mental labour and its ancillary outputs commodified into tradeable assets and captured by existing legal constructs (such as copyright, patents and trademarks), but the typology and legal lexicon of IP law burgeoned, with ever growing and contradictory rationales.71 The dynamism of rationales and discourses in the past two hundred years – when the res to be regulated and protected expands beyond traditional justifications of reward, compensation, incentivization, dignity or honour – is not new. Burrell and Kelly show that justifications in

68 69

70 71

Art 10bis, Paris Convention for the Protection of Industrial Property, 14 July 1967 (1971) (preventing ‘any act of competition contrary to honest practices in industrial or commercial matters’). See for example, D Gangjee, ‘From Geography to History: Geographical Indications and the Reputational Link’ op cit 59; O Bracha and PR Goold, ‘Copyright Accidents’ (2016) 96(3) Boston University Law Review 1025 (on the view that copyright is closely related to tort under US law); and Art 39(3) TRIPS Agreement. WR Cornish, ‘The International Relations of Intellectual Property’ (1993) 52(1) Cambridge Law Journal 46, 54–5. For economic rationalizations of IP rights: F Maclup, ‘An Economic Review of the Patent System, Subcomm. On Patents, Trademarks, and Copyrights, Senate Comm’. On the Judiciary, 85th Cong (Study No 15) (often cited for Machlup’s wry summary: ‘If we did not have a patent system, it would be irresponsible on the basis of our present knowledge of its economic consequences to recommend instituting one. But since we have had a patent system for a long time, it would be irresponsible based on our present knowledge to recommend abolishing it’ 45–46); R Merges, ‘Intellectual Property Rights and the New Institutional Economics’ (2000) 53 Vanderbilt Law Review 1857, 1877; AB Jaffe and J Lerner, Innovation and Its Discontents: How our Broken Patent System Is Endangering Innovation and Progress, and What to Do about It (Princeton University Press 2004) (accepting the economic incentives for the patent system but criticizing legislative and institutional reforms as being contrary); U Suthersanen, G Dutfield and KB Chow, Innovation without Patents: Harnessing the Creative Spirit in a Diverse World (Edward Elgar 2007) 3–69 (arguing from a mixed economics and development platform); Sir A Plant, The New Commerce in Ideas and Intellectual Property (Athlone 1953); HL Macqueen, ‘Law and Economics, David Hume and Intellectual Property’ in Argument Amongst Friends: Twenty-Five Years of Sceptical Enquiry (Nick Kuenssberg ed, David Hume Institute, Edinburgh 2010) 9–14. For social and philosophical rationalizations: A Chander and M Sunder, ‘Is Nozick Kicking Rawls’s Ass? Intellectual Property and Social Justice’ (2007) 40(3) UC Davis Law Review 563; H van den Belt, ‘Robert Merton, Intellectual Property and Open Science: A Sociological History for our Times’ in The Commodification of Academic Research: Science and the Modern University (Hans Radder ed, University of Pittsburgh Press 2010). For mixed rationalizations (including cultural, distributive justice, ethics and development): W Fisher, ‘Theories of Intellectual Property’ in Stephen Munzer (ed) New Essays in the Legal and Political Theory of Property (CUP 2000) 168–99 (American scholars tend to adopt Fisher’s four justifications for intellectual property law: (i) utilitarian / economic; (ii) Lockean/labour dessert; (iii) personhood/personality; (iv) promotion of cultural development); S Parthasarathy, Patent Politics: Life Forms, Markets and the Public Interest in the United States and Europe (Chicago University Press 2017) (intriguing exploration on how culture, history, politics and ideology shape the way the public interest is framed and is translated into national or regional intellectual property policies, rules and institutions); and R Merges, Justifying Intellectual Property (Harvard University Press 2011).

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

14

Uma Suthersanen

nineteenth-century patent law had to include ‘public access’ as this would allow patent law to underwrite the investment costs incurred in harnessing foreign technologies: the law could treat the importer and the inventor as equally deserving – the benefit to the British public, who would gain access to an invention for the first time, was the same in either case. Such equivalence would have been much more difficult to justify had patent law ultimately been centred on a narrative of creativity, genius, and a natural right to property.72

The perplexing purposes of patent law have also been criticized by Adcock and Beyleveld. They question the need for ‘inventiveness’ as a criterion of protection if the purpose of patent law is to protect investment. And in their view, the purpose has slowly been transformed from reward and natural right to unjust enrichment to incentivization of investment: Thus, if the purpose of patent law is taken to be this, there should be no need to fictionalise certain discoveries as inventions because discoveries should be patentable. But conversely, if the law insists on inventiveness as a condition for protection, then the rationale cannot simply be the need to create incentives for investment.73

The accretion/emulation approach was adopted so that copyright law could expand to accommodate soulless productions, devoid of human author or persona. The emulation method produced sui generis IP regimes which mix a dizzying array of concepts borrowed from copyright, patent or trademark laws. National design laws reflect an ambivalence in rationales when one examines the different national and international criteria: eye appeal, pattern, shape, configuration, form, model, ornamental, functional, original, commonplace, individual character, and novelty.74 If financially dictated labour could not be labelled as creative authored works, international law was adjusted to include sound recordings and broadcasts and integrated circuit layouts, and to set out new rights.75 We know that jurists and legislators theorized deeply on new rationales, rules and categories (‘neighbouring rights’, ‘related rights’, or even sui generis rights) to stretch protection to computer programs,76 sound recordings,77 photographs/films78, computer-generated works and now AI-derived inventions, music and art.79 72 73 74 75

76

77 78 79

R Burrell and C Kelly, ‘Parliamentary Rewards and the Evolution of the Patent System’ (2015) 74(3) Cambridge Law Journal 423, 441. MD Adcock and D Beyleveld, ‘Purposive Interpretation and the Regulation of Technology: Legal Constructs, Legal Fictions, and the Rule of Law’ (2007) 8(4) Medical Law International 305, 316. U Suthersanen, ‘Cross-border Copyright Protection in Europe’ in H Hartwig, Research Handbook in Design Law (Edward Elgar 2021) 482, 506. Peukert highlights how the logical amplification of commodification results in further fictitious rights whenever technologies change: if one reproduction technology replaces another, IPRs are expanded to accommodate the new technology though the justification is often couched in economic or market-biased language; namely, we need new rights to stem the loss of income or to prevent the dilution of a trade name in new markets: Peukert 1172. See TRIPS Agreement, arts 1(3) and 2 enumerating the different conventions absorbed into WTO rules. See, for instance, the early debate in P Samuelson et al. (1994) ‘A Manifesto Concerning the Legal Protection of Computer Programs’ Columbia Law Review 94, 2308, and P Samuelson, ‘The Strange Odyssey of Software Interfaces as Intellectual property’ in M Biagioli, P Jaszi and M Woodmansee (eds) Making and Unmaking Intellectual Property: Creative Production in Legal And Cultural Perspective (University of Chicago Press 2011). R Fleischer, ‘Protecting the Musicians and/or the Record Industry? On the History of “Neighbouring Rights” and the Role of Fascist Italy’ (2015) 5 Queen Mary Journal of Intellectual Property 327, 336. B Edelman, Ownership of the Image (Routledge & Kegan Paul 1979) 44–7; J Gaines, Contested Culture: The Image, the Voice, the Law (University of North Carolina Press 1991) 46; and Burrow-Giles v Sarony (111 US 53, 1884). For representative cases related to AI authorship and inventorship, see, for instance, Shenzhen Tencent v Yinxun (‘Dreamwriter’ case, recognizing an article generated by an AI program as a literary work under copyright law); RAGHAV (recognizing an AI painting app as a co-author in a copyrighted work); and DABUS (South African Patent Office issued a patent to an AI tool as patent inventor). The cases are available from: WM Schuster, ‘Artificial Intelligence and Patent Ownership’ (2018) 75 Washington & Lee Law Review 1945; E Bonadio and L McDonagh,

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

Creativity, Pluralism, and Fictitious Narratives

15

As in the early modern era, the rules for new commodities embraced a variety of motivations which underpinned such works or underpinned rationales for property rights – banal, subsistence, educational, creative, competitive, trade indicia, investment or state welfare. The different thresholds may also suggest that the rationales for subaltern IP rights were ill-thought out and emanated not so much from protecting the labourer or their creativity but the sunk investment by manufacturers of ‘industrial’ commodities. I.III REIMAGINING IP AS A PLURALIST NARRATIVE

I.III.A Returning to Basics Cornish, in his first textbook edition, defined IP law as the ‘branch of the law which protects some of the finer manifestations of human achievement’; he cynically added in a later edition that ‘it also shields much that is trivial and ephemeral’.80 Bently and Sherman suggest that intellectual property law ‘regulates the creation, use and exploitation of mental or creative labour’.81 Vaver argues that the phrase ‘intellectual property’ starts ‘from the premise that ideas are free as the air – a common resource for all to use as they can and wish. It then proceeds systematically to undermine that principle.’82 Spence’s view is that an IP right is ‘a right: (i) that can be treated as property; (ii) to control particular uses; (iii) of a specified type of intangible asset’.83 Dutfield and Suthersanen push the boundaries slightly further to propose that IP is ‘a type of property regime whereby creators and entrepreneurs are granted a right, the nature of which is entirely dependent on the nature of the creation on the one hand, and the legal classification of the creation on the other’.84 If we turn to stipulative definitions as laid down in international laws, we encounter divergent representations of IP as a legal collective noun for the classifications of artefacts and activities. In its most turgid and tautologous form, the WTO–TRIPs Agreement uses the term ‘intellectual property’ to refer ‘to all categories of intellectual property that are the subject of Sections 1 through 7 of Part II‘,85 namely copyright and related rights, trademarks, geographical indications, industrial designs, patents, layout-designs of integrated circuits and the protection of undisclosed information. And evidently nothing more. The WIPO Convention, on the other hand, theoretically embraces every conceivable labour output of man and industry with an open list of IP as including the rights relating to ’literary, artistic and scientific works, performances relating to artists, phonograms and broadcasts, inventions in all fields of human endeavour, scientific discoveries, industrial designs, trademarks, service marks, commercial names and designations, protection against unfair competition, and all other rights resulting from intellectual activity in the industrial, scientific, literary or artistic

80 81 82 83 84 85

‘Artificial Intelligence as Producer and Consumer of Copyright Works: Evaluating the Consequences of Algorithmic Creativity’ (2020) 2 Intellectual Property Quarterly 112–37 and B.Zhou, ‘Artificial Intelligence and Copyright Protection: Judicial Practice in Chinese Courts’ . In an earlier edition, WR Cornish and D Llewelyn, Intellectual Property: Patents, Copyright, Trademarks and Allied Marks (5th edn, Sweet & Maxwell 2003) 3. In an earlier edition, L Bently and B Sherman, Intellectual Property Law (2nd edn, OUP 2004) 1. D Vaver, Intellectual Property Law (Irwin Law 2011) 1. M Spence, The Concept of Intellectual Property (OUP 2007) 12–13. G Dutfield and U Suthersanen, Dutfield and Suthersanen on Global Intellectual Property Law (2nd edn, Edward Elgar 2020) 14. TRIPS Agreement 1994, art 1(2).

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

16

Uma Suthersanen

fields’.86 On the other hand, the Universal Declaration of Human Rights states that everyone has a right to interests resulting from ‘any scientific, literary or artistic production of which he is the author’. Besides reflecting a gender bias, the provision also reflects the literature which situates IP law within a rights-based discourse.87 Turning next to theoreticians, we encounter Harris’s view of IP law as constituting trespassory rules which restricts uses of the ideational entity thereby preserving ‘to an individual or group of individuals an open-ended set of use-privileges and powers of control and transmission characteristic of ownership interests over tangible items’.88 Drahos explains IP within his own sociopolitical justificatory context: ‘We shall say that intellectual property rights are rule-governed privileges that regulate the ownership and exploitation of abstract objects in many fields of human activity.’89 Thus far, we encounter current notions of IP law as being concerned with human achievement or activity, intangible (or abstract) objects (or assets), mental or creative labour, creators and perhaps entrepreneurs. So, does the law grant rights in intangible objects or does it accord privileges to control certain uses or activities ad rem? Both Drahos and Spence veer towards the latter perspective. The former dramatically inflates IP rights as ‘liberty-intruding privileges of a special kind’ which ‘promote factionalism and dangerous levels of private power’.90 The latter argues that the right does not vest in the intangible asset per se, but rather focuses on the ownership of the asset as a gateway to control of usage – the object of an IP right is the legal right itself.91 Pila’s lexical account of IP skirts close to this view of IP as a ‘type of property conferring temporary exclusionary rights in respect of the objects it exists to protect’. However, she also concedes that one cannot understand the function of the law as a right without a detailed exploration of the nature and individuation of the subject matter protected: ‘however widely or narrowly the scope of IP protection is cast, its primary reference point remains the individual subject matter to which the property rights attach, conceived as a subject matter of relevant (protectable) type’.92 George’s theoretical approach adopts a similar approach in exploring the definition of IP, but to no avail, as she concludes IP is symbolically important but an essentially contested concept without an inherent meaning; instead her suggested approach is to reconstitute a metaphysical notion of IP based on society’s internalization process.93 Okediji notes the success of the commodification of mental labour and investment in relation to the rebranding of ‘IP goods’ on the global trading market as investment assets. She also notes, affirming the Polanyian cautionary narratives, that when IP is cast as an investment, the focus shifts from ‘a conditional grant to incentivize creativity and innovation to a guaranteed right justified by reference to the amount of capital committed, the expenses incurred, the profits anticipated, and the

86 87

88 89 90 91 92 93

WIPO Convention 1967, art 2(viii) UDHR 1948, art 27(2). In this I disagree with Pila’s view of the provision as supporting a conception of IP subject matter as artifacts – see Pila, The Subject Matter of Intellectual Property (2017) 64–65. Her view is perhaps more supported by the UNESCO Convention for the Safeguarding of the Intangible Cultural Heritage, art 2 – where the tangible and intangible aspects of heritage and property are intertwined. See also Protocol 1 of the European Convention on Human Rights, art 1; Charter of the Fundamental Rights of the EU, 2000, art 17(2) (intellectual property clarified as a human right); Dutfield and Suthersanen, op cit, 328–48 (for the literature on IP as a human right). JW Harris, Property and Justice (OUP 1996) 44; see also ch. 8. P Drahos, A Philosophy of Intellectual Property (ANU eText 2016) 7. Drahos 7. Spence 13; Drahos, 7. J Pila, The Subject Matter of Intellectual Property (OUP 2017) ch. 1. A George, Constructing Intellectual Property (CUP 2012) 31–136.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

Creativity, Pluralism, and Fictitious Narratives

17

risk assumed’.94 Such consequentialist platforms are also part of the current narratives for EU IP laws: the need for copyright to ‘protect and stimulate the development and marketing of new products and services’; the desirability for trademark law to promote ‘the development of economic activities’; and the need for the patent system to ‘encourage research into and production of biotechnological medicines’.95 Okediji further argues that to prevent this consequential shift in IP function, state intervention is essential as it incorporates other state obligations, for example: ‘the protection of fundamental freedoms, national security, environmental or public health needs could be compromised’.96 I.III.B (Intellectual) Property as a Pluralist Concept Dagan’s wider treatise on property urges that we adopt a more pluralist notion of property which not only allows ‘the coexistence of a diverse set of social institutions’, but also allows for differing emphases depending on whether the property type arises from a more utilitarian basis (then the emphasis is on investment and economic success) or from a more social basis (then the emphasis may be on active participants and social cohesion). Indeed, he forbids any liberal thinker to adopt a ‘one size fits all’ understanding of property as ‘[s]uch a straitjacket is not only misleading but also unfortunate because property’s structural pluralism enables diverse forms of the good to flourish’.97 And there is an increasing tide of scholarship advocating such a pluralistic concept in relation to IP law. For example, Merges argues for an IP model which comprises ‘a commitment to individual ownership as a primary right, respect for third-party interests that conflict with this right, and, from the philosophy of John Rawls, an acceptance of redistributive policies intended to remedy the structural hardships caused by individual property rights’.98 An earlier work by Resnik expresses a similar approach by exposing six bases of IP rights: Locke’s libertarian property theory, utilitarianism (relying on Mill and the US Constitution), Hegel’s theories on self-expression and freedom, privacy, egalitarianism including Marxist notions of distributive justice, and finally Rawls’s ‘reasonable’ pluralistic conception of justice.99 A pluralistic approach provides ‘the best account of IP’, he declares, as it acknowledges the disparate types of IP with inherently different values. More importantly, Resnik points to Rawl’s supra-factor as to why a pluralistic account of IP is vital: modern democratic societies are pluralistic, reflecting different ethnic and cultural backgrounds which encompass different moral, philosophical and religious beliefs.100 This accords with David Hume’s view that there is nothing natural about property rights as 94

95

96 97 98

99 100

RL Okediji, ‘When Is Intellectual Property an Investment?’ in C Geiger (ed) Research Handbook on Intellectual Property and Investment Law (Edward Elgar 2020) 94, 101–02. See also R Dreyfuss and S Frankel, ‘From Incentive to Commodity to Asset: How International Law Is Reconceptualising Intellectual Property’ 36 Michigan Journal of International Law (2015) 557; H Ruse-Khan, ‘Litigating Intellectual Property Rights in Investor-State Arbitration: From Plain Packaging to Patent Revocation’ (2014) – a working paper available at Recital (2), EU Directive 2001/29/EC on copyright in the information society; Recitals (3) and (5), Regulation (EU) 2017/2001 on Community trademark; Recitals (16)–(18), EC Directive 98/44 EC on the legal protection of biotechnical inventions, respectively. RL Okediji, ‘When Is Intellectual Property an Investment?’ in C Geiger (ed) Research Handbook on Intellectual Property and Investment Law (Edward Elgar 2020) 94, 101–02. H Dagan, A Liberal Theory of Property (CUP 2021) 22, 85. RP Merges, Justifying Intellectual Property (Harvard University Press 2011) 5, 13; for a rather critical review of Merges’s theory, see J Pila, ‘Pluralism, Principles and Proportionality in Intellectual Property’ (2014) 34 Oxford Journal of Legal Studies 181. D Resnik, ‘A Pluralistic Account of Intellectual Property’ (2003) 46 Journal of Business Ethics 319. Resnik, ibid 331, relying on principles drawn from J Rawls, A Theory of Justice (Harvard University Press 1971), and J Rawls, Justice as Fairness: A Restatement (Harvard University Press 2001).

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

18

Uma Suthersanen

they are merely historically accepted societal practices that arises and derives its significance out of the scarcity of the object in question.101 This chapter joins the body of scholarship that advocates that pluralism is structurally inherent and necessary for any IP right to be understood.102 The smorgasbord of definitions clearly evidences the implausibility of uncovering a single, uniform rationale or meaning of IP. And what the law has done – if the motivation landscape is to be believed – is to assiduously follow the varied stakeholder interests. IP law is a beast wearing a palimpsest of the many narratives and rules that have been re-written and re-purposed throughout the centuries to its current state. We should not be too flustered about the continuous expansion of protected subject matter, or the continuous lack of coherent rationales as to why something is or should be considered ‘IP’. What the tangled circularity of explanations does is emphasize that the law is a dynamic construct reflecting the socio-economic interests and values of specific periods and peoples. My suggestion is that we need to accept that IP – as a property right or as an entitlement – is a socio-legal fiction. In which case, the false dichotomy between ‘property as a thing’ and ‘property as a bundle of normative relationships concerning the use of the thing’ should be dismissed especially as justifications for IP are nebulous. Instead, I assume that property rights can attach to any tangible and intangible object conceptually as the law creates any res or thing, whether corporeal or not, through the notion of property rights. An IP right – as a property right – should be considered to be an individual property object or ‘thing’ or ‘res’ to which a person’s use-rights are attached.103 With this understanding of what constitutes IP, we can continue to argue whether the law protects human creativity, capital, cultural geography or heritage. But should we? I.IV ‘DOUBLE MOVEMENT’: OUR NARRATIVE FOR IP REFORM

We may, instead, wish to focus more resolutely on the institutional and societal demands for IP law to be continuously re-structured and re-evaluated to not only allow for expansion (to cover both creative and investment-led works) but also to allow for creative re-use, for access to essential public goods, for cultural or educational use, and so on. There is some movement towards this type of reasoning though I feel the debate is still distracted by the creativity–investment bifurcation. Geiger, for example, rightly argues that we should look to re-constitutionalizing IP law within fundamental freedoms. And yet he embraces the fiction of ‘IP is only about creativity’ narrative: Providing for a constitutional clause that is capable of demonstrating, by its wording and use, that intellectual property is intrinsically linked to the interests of society would also bring IP

101

102

103

D Hume, A Treatise of Human Nature [1739] ( LA Selby–Bigge and PH Nidditch eds, Clarendon Press 1978) 488; HL MacQueen, ‘Law and Economics: David Hume and Intellectual Property’ in N Kuenssberg (ed) Argument Amongst Friends: Twenty-Five Years of Sceptical Enquiry (David Hume Institute 2010) 9–14. See, for instance, D Gervais, ‘TRIPS Pluralism’ (2021) World Trade Review 1–22; S Frankel (ed) Is Intellectual Property Pluralism Functional? (Edward Elgar 2019); O Bracha and T Syed, ‘Beyond Efficiency: ConsequenceSensitive Theories of Copyright’ (2014) 29 Berkeley Technology Law Journal 229, 245–46. A Rahmatian, Copyright and Creativity: The Making of Property Rights in Creative Works (Edward Elgar 2011) 10–12, relying on Roman law, Jeremy Bentham and Lord Kames for this perspective. The literature on IP as property is vast; for a collection of essays which usefully catalogues the literature and arguments for and against protecting mental and intellectual labours through property rights and even within equity, see H Howe and J Griffiths (eds) Concepts of Property in Intellectual Property Law (CUP 2013). In exploring property, the following American authors offer intriguing perspectives: H Demsetz, ‘Towards a Theory of Property Rights’ (1967) 57 American Economic Review 347; G Calabresi and AD Melamed, ‘Property Rules, Liability Rules, and Inalienability: One View of the Cathedral’ 85 Harvard Law Review 1089 (1972); S Munzer and K Raustiala, ‘The Uneasy Case for Intellectual Property Rights in Traditional Knowledge’ (2009) 27 Cardozo Arts & Entertainment Law Journal 37.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

Creativity, Pluralism, and Fictitious Narratives

19

rights closer to people [. . .] Investments are needed and are important in the creative ecosystem. But in IP, investments are not protected as such, only indirectly: an exclusive right is granted as a counterpart to the enrichment of society with a new creative output, right which then can be exploited to recoup the investments made to allow the creative process. If the investment (however substantial) does not lead to a creative output, no IP is granted.104

The problem with this type of romantic and purist reasoning is that it guillotines the law from origins and society: from its European socio-historical origins which covered multifarious motivations; and from the harsh socio-economic reality of how a capitalist and technologydriven society will invariably gobble and commoditize any intangible res. And we have accepted these legal fictions before – ‘computer programs are a literary work’ springs to mind. So, why not continue? In my opinion, Polanyi’s The Great Transformation predicted the consequences of global commodification and fictional narratives, especially if we recall his passage on the environmental consequences of land commodification. He would have supported the call for greater state intervention through his final concept – the ‘double movement’. His argument, after all, is that societies are constituted by two opposing movements: the expansion of the scope of the market, and the protective ‘double-movement’ to resist the disembodiment of the economy. Polanyi’s theory is robust and realistic enough to suggest that markets are a vital aspect of our society – and different types of market embeddedness occurs at different historical movements, with some types of embedded market societies fostering innovation at all costs, with other markets tempering this with socialist and humanist/global concerns.105 The debate on whether vaccinations for pandemics should constitute global public goods with no IP rights attached or with waivers attached is the prime 2021 example.106 If we accept IP as a pluralistic doctrinal concept, then we have to accept that the law as it stands is a product of many motivational bases including: labour, reward, dignity, public welfare, control of competition, direct protection of investments and innovation.107 In this way, we can see why IP constitutes not just patent, copyright and trademark laws but also serves as the bases for rights and entitlements in relation to: personality, resale royalties, plant variety, utility model, neighbouring, derivative, topography, geographical indications, indigenous and traditional knowledge, database, technological measures and trade secrets. And it follows that we have seen and will continue to witness counter-movements occurring within the law. Here are a few examples of how we can gauge such counter-movements within IP law which seek to re-embed IP res within societal control, and thus form a bases for legal reform. I.IV.A Humanist, Communitarian and Stakeholder Aspects The argument, that IP is a natural and exclusive property right subsisting in human-based intellectual creations, was an obvious one historically to displace other competing notions of privilege, monopoly, investment and the prevailing patronage systems. Aligned with this ethos is 104 105 106

107

C Geiger, ‘Intellectual Property and Investment Protection: A Misleading Equation’ (14 November 2021) Peukert finds Polanyi inapplicable to IP law as the latter’s primary aim ‘is not to control market forces but to integrate innovative and entrepreneurial activity into the market system’, Peukert 1190. See for example, WHO, ‘WHO SAGE values framework for the allocation and prioritization of COVID-19 vaccination’ 14 September 2020 See note 86.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

20

Uma Suthersanen

the romantic rhetoric of authors and inventors pursuing an egotistical agenda by advocating individual rights in valuable intangibles arising from the natural and just result of the creator’s persona or her individual labour.108 And indeed, not every eighteenth-century jurist or reformer blindly strove to transform mental labours, investment-led activities or intellectual knowledge or skills into commodities detached from the object’s motivation, the creator or the community. Eighteenth-century jurists such as Lord Mansfield or Le Chapelier advocated a rather pluralistic perspective of property rights for creative mental labour justified either on the principle of rewarding or incentivizing labour or on the more humanist principle of a person’s right to property, personality and dignity – but definitely a property right to be countered by public interests!109 An ancillary dimension is most, if not all, justifications holds that IP rights should be curtailed – by duration, by third party rights of use, by competition and by public or global interest. Examples of such doctrines within IP laws include: copyright law’s fair use or fair dealing, patent law’s compulsory licensing or research use, trademark law’s exclusion of generic marks. These exceptions or limitations are often posthumous recognition of societal rights overlooked by industry. The same balance is found in arts. 7 and 8 of the TRIPS Agreement which, on the one hand, espouses a utilitarian view of IPRs (Intellectual Property Rights) and states that IPRs are private rights. The agreement also structures the understanding of IP law, on the other hand, in terms of social goals. This is in tandem with the movement we see of equating IP as a human right, though not without controversy.110 Thus, recognizing IPRs as investment-driven property does not entirely deprive IP law of its welfare and social contours; it still should be acknowledged whether we adopt a utilitarian or more human rights perspective. Okediji, Geiger, et al. are right from a Polanyian perspective in insisting for greater state intervention to make counter-movements if we avoid the creativity– investment bifurcation.111 I.IV.B Technology, Environment and Health as Public Goods Technology has systematically made inroads into our societal and economic relations which in turn has led to law reform within IP law. It is argued that our interaction with peoples, governments and companies from interactive experiences to grocery orders to medical 108

M Woodmansee, ‘The Genius and the Copyright: Economic and Legal Conditions of the Emergence of the Author’ (1984) 17 Eighteenth-Century Studies 425. 109 Lord Mansfield in Sayre v Moore (1785) – ‘we must take care to guard against two extremes equally prejudicial; the one, that men of ability, who have employed their time for the service of the community, may not be deprived of their just merits, and the reward of their ingenuity and labour; the other, that the world may not be deprived of improvements, nor the progress of the arts be retarded’; the full case is discussed in Cary v Longman 1 East 359 (reprinted in 102 Eng. Rep. 138 (1378–1865)). The French advocate of copyright law, Le Chapelier, noted that once the author had disclosed the work to the public, ‘when the work is in the hands of everyone, when all educated men know it and have imbued it with happy memories, it seems that since then, the writer has affiliated the public with his property, or rather has fully transmitted his property to the public’, in Le Chapelier’s Report (1791), Primary Sources on Copyright (1450–1900), eds L Bently and M Kretschmer, . The US Supreme Court in Sony Corp. v Universal City Studios 464 US 417 (1984) (citing United States v Paramount Pictures, Inc 334 US 131 (1948): ‘The copyright law, like the patent statutes, makes reward to the owner a secondary consideration. [. . .] The sole interest of the United States and the primary object in conferring the monopoly lie in the general benefits derived by the public from the labors of authors.’ 110 Art 15(1)(c) of the International Covenant on Economic, Social and Cultural Rights, and art 27(2) of the Universal Declaration on Human Rights. See Dutfield and Suthersanen on Global Intellectual Property Law, op cit 337, 343 discussing IP as a human right. 111 See discussion in notes 00 and 00.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

Creativity, Pluralism, and Fictitious Narratives

21

appointments works through the vast profusion of networks enabled by machine-learning algorithms: ‘Our education and healthcare is increasingly delivered through AI-enabled technologies. Our manufactured products, be they household conveniences or houses, will soon be produced by 3D printers.’112 And all this can be further harnessed by the fact that previous local and regional trade networks have evolved into our current ‘just in time’ global supply chains.113 On the other hand, farmers worry as to their rights to knowledge, seeds and repair,114 consumers worry about their right to ‘own’ digital music and e-books,115 scholars and citizens are concerned as to the increasing costs of accessing academic publishing especially in science and technology,116 religious and environmental groups view patented plants and green technologies as abhorrent to morality and social welfare,117 societies are concerned as to world trade rules which countenance the commoditization of drugs and vaccines;118 and we wonder as to infiltration of AIenabled technologies in all aspects of human living.119 Okediji’s fear revolves around the idea that we are in danger of reducing the IP res to one simple category, namely the capital risked or invested. In such a scenario, we would have lost all other functions of IP as housed in current justifications. This view accords with those which argue that our current crises (in public health or educational or technology) have been brought about by our governments who have delegated responsibility for the ‘public interest’ to corporations in the specific area of producing goods and services; and in return corporations have the right to make profit by any means whatsoever.120 The latter argument is certainly prevalent in writings of those who interrogate whether pharmaceutical companies, as non-state actors, should be held accountable in tandem with governments, for fulfilling public health obligations.121 Once again, all these concerns represent the nascence of counter-movements within health, trade, environmental and technological fora which will hopefully form the bases for legal reform, thereby re-embedding IP res within a holistic societal control.

112 113 114

115 116 117

118

119 120 121

A Azhar, Exponential: How Accelerating Technology Is Leaving Us Behind and What to Do about It (Random House/ Penguin 2021) 88. Azhar, ibid 43–44, 62–66. Bowman v Monsanto Co 133 S.Ct. 1761 (2013) (Sup Ct (US)); Australian Productivity Commission Inquiry Report, Right to Repair, No 97, 29 October 2021; K Wiens, ‘We Can’t Let John Deere Destroy the Very Idea of Ownership, Wired, 04.21.2015. Péter Mezei, Copyright Exhaustion: Law and Policy in the United States and the European Union (CUP 2018) 92–165 (chap 4). M Skladany, Copyright’s Arc (CUP 2020) 24–64 (chap 2); Dutfield and Suthersanen, op cit 349–71 (chap 12). O Tur-Sinai, ‘Patents and Climate Change: A Skeptic’s View’ (2018) 48 Environmental Law 211 (sceptical about whether there is any correlation between the need for environmentally sound technologies, and the core rationales of patent law, namely incentivization and reward). See for example ‘A Patent Waiver on Covid Vaccines Is Right and Fair’, Editorial, Nature 593, 478 (2021) doi: ; S Thambisetty, A McMahon, L McDonagh, HY Kang and G Dutfield, ‘The TRIPS Intellectual Property Waiver Proposal: Creating the Right Incentives in Patent Law and Politics to end the COVID-19 Pandemic’ (24 May 2021). LSE Legal Studies Working Paper No 06/2021 . Ana Correia and Irina Reyes, ‘AI Research and Innovation: Europe Paving Its Own Way’, European Commission, 16 November 2020 . S. Mansell, Capitalism, Corporations and the Social Contract: A Critique of Stakeholder Theory (Business, Value Creation, and Society), (CUP, 2013) 46–47. J Tasioulas, ‘The Minimum Core of the Human Right to Health’, World Bank, 2017 License: CC BY 3.0 IGO; H Hestermeyer, Human Rights and the WTO: The Case of Patents and Access to Medicines (OUP 2008) chapters 3 and 4.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

22

Uma Suthersanen

I.IV.C Lines for Analyses for Reform? Polanyi’s suggestions are that things should be re-embedded within society. His caution is simple: irrespective of what is being protected, the property aspect of IP law should be restrained for society and humanity. We all know of the various writings and policies which try to ensure a balanced IP law.122 I leave it to others to suggest what specific reforms are needed to counter the relentless push and commodification of IP res, acknowledging that ‘investment’, ‘labour’ and ‘creative’ will remain uneasy companions within the IP narrative. Nevertheless, I offer this list to gently prod the reader to either declaim my choices (some of which have been espoused by others) or to add their wish list: • An international agreement to curtail the durations of protection and/or to introduce registration to copyright systems. • An international public interest rule within all Preambles or Objectives to state ‘An equitable or fair mandatory remuneration is ensured for all human creators.’ • An international public interest rule mandating access to IP res for users exercising their own fundamental rights (to education, science and culture, health, self-development, freedom of expression, etc.) OR: Re-classification of limitations and exceptions within IP laws as constitutional rights or as fundamental rights of others vis-à-vis the rights to health, etc. • Introducing the notion of ‘essential’ or ‘staple’ IP res (akin to essential goods doctrine within competition law) with a legal presumption of reasonable pricing or compulsory licensing. • A UNESCO–WIPO Register of declared public domain (as with orphan works in EU). • Exploratory studies on the viability of environmentally derived ‘rights’ in relation to repair, farmers, exhaustion doctrine and reverse engineering for innovations which impact agriculture and food security. I.V CONCLUSION: THE POLANYIAN VISION

The journey in this chapter focused at locating a set of clear and repeated motivations underpinning early European rules in relation to IP res. It was envisaged that a legal mapping would reveal a landscape that holds the key in determining whether IP law was and is continually being re-calibrated to fulfil two functions. First, did different motives attach to the different rationales under current IP law to protect subaltern IP rights? Second, did the different motives accommodate creative, entrepreneurial and investment-led labour and investment? And does Polanyian theory help? First, over a quintennial period, rights and privileges grounded in property, tort and criminal laws have vested in all sorts of beneficiaries in relation to all types of intangible subject matters under various rationales. Second, the literature review mapping the genesis and institution of IPtype rules shows prototype IP laws emerging from fifteenth-century European states extending 122

J Reda, ‘Legislative Perspective on the Development of the EU Copyright Law’, in C Geiger (ed) The Intellectual Property System in a Time of Change: European and International Perspectives (LexisNexis 2016); R Giblin and K Weatherall, What If We Could Reimagine Copyright? (ANU Press 2017); M Biagioli, ‘Weighing Intellectual Property: Can We Balance the Social Costs and Benefits of Patenting?’ (2019) 57(1) History of Science 140–63.

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

Creativity, Pluralism, and Fictitious Narratives

23

entitlements or privileges to sole individuals, collective groups and corporate entities comprising inventors, authors, importers, entrepreneurs and localized groups. These stakeholders represented a pluralist agenda ranging over cultural, societal, trade and other interests. Polanyi’s vision of an embedded society arguably explains this pattern, at least in Europe. Third, the pattern then grows in the last 150 years or so to embrace things: namely labour per se, or land-related labour or product-related labour were turned into commodities. Finally, these commodities were recognized at some points as property in order to protect labourers (or creators) and to protect the capital invested that operated simultaneously in the manufacture and distribution of these commodities. From this, we should appreciate that: • The thing in IP law has varied from mental or intellectual labour to skills acquired, to confidential secrets, to know-how and techniques, and to personal dignity and reputation; • the motivations behind IP law vary from public interest to dignity to reward to securing investments. • From the nineteenth century, with the increased participation of corporations in the organizing, underwriting and financing of many aspects involved in generating and harnessing technologies and industries, IP law began to adopt a distinctively industrydriven character in terms of the nature and scope of rights. • The push to commoditize discrete things led to ingenious justifications, criteria and language which is betrayed in the discombobulated nature of IP law today. • Current discourses on whether to push the thing in IP law to outputs manifested without clear creative or human spirit are unsurprising given the landscape as sketched within this chapter. Existing criteria such as human authors and inventors, inventiveness, and personality can be viewed as fictional means to persuade society to harness innovation and creativity via property rights. Existing categories of sui generis rights, however, indicate another major consideration namely the protection of investments. At a fundamental level, Polanyi’s theory offers insights as to how to use motivations as a means of gauging laws in terms of market and society. One can appreciate the fact that all motivations which constitute the typology, lexicon and ethos of current IP law straddles several historical and justificatory bases. Polanyi’s explanations and warnings of an unregulated market is tempered by his invocations of the ‘double movement’ whereby we can stem all the ills that will arise from universal commoditization of every intangible output. How we should counter-reform will become more urgent and relevant as this new century of ours accelerates. Perhaps a simple Polanyian step towards such a consistent approach would be to ask purists and progressives to consider together the following statement as a universal rule of law: ‘The current pluralistic IP regime should declare human CREATORS to be recognized in line with the International Bill of Rights and be accordingly entitled to a perpetual right of recognition and remuneration.’

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

https://doi.org/10.1017/9781108989527.003 Published online by Cambridge University Press

part i

Science, Technology and Industry

https://doi.org/10.1017/9781108989527.004 Published online by Cambridge University Press

https://doi.org/10.1017/9781108989527.004 Published online by Cambridge University Press

II Sui Generis Protection of Non-creative Databases Caterina Sganga

II.I INTRODUCTION

One of the most controversial acts of EU harmonization in the field of copyright law, and one of the first copyright directives, was without doubt Directive 96/9/EC on the legal protection of databases.1 Many of its definitions, the narrow scope of its exceptions and the lack of coordination with general copyright law have been heavily criticized. Yet, the most challenged and discussed provision was and remains its Article 7, which introduced as a worldwide novelty the so-called sui generis right. The aim of the Directive was to create two paths for database protection, depending on their degree of originality. The first was conceived for databases that were original in structure and arrangement and used traditional copyright protection. The second was envisioned for databases that did not meet this originality threshold, but still required a substantial investment in obtaining, verifying and presenting their content, an investment which the EU legislator wanted to incentivize by protecting database makers from parasitic behaviors, unfair competition and freeriding. The latter entitlement was conceptualized as the right to prevent extraction and reutilization of the whole or substantial parts of the base, designed to be independent from copyright, and having a completely different rationale, requirements for protection, duration, exceptions. This new sui generis right was presented as the tool that would have bolstered the EU database industry against its fiercest competitors, and chiefly against the United States which, ironically, around the same period saw their judiciary increase the originality threshold for database protection under copyright, and their legislator cross out the idea of introducing any new form of exclusivity on collections of data and other materials.2 The sui generis right, which failed to spread internationally as much as the EU legislator desired, remained a European unicum, and was subject to strong critiques for its unbalanced nature and risky tendency to create informational and data monopolies, yet without performing an effective role in fostering industrial investments. Its introduction brought within the tangles of EU copyright law another entitlement which was meant to protect investments and not 1

2

Directive 96/9/EC of March 11, 1996 on the legal protection of databases, March 11, 1996, OJ L 77/20, March 27, 1996 [hereinafter Database Directive]. With the seminal case Feist Publications Inc v Rural Telephone Service Co 499 US 340 (1991), the US Supreme Court denied protection to a telephone directive, stating that a database that contained plain information without any minimum original creativity cannot be protected by copyright. On the US approach to database protection, also in comparison with the EU approach, see Estelle Derclaye, The Legal Protection of Databases: A Comparative Analysis (Edward Elgar 2008).

27 https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

28

Caterina Sganga

creativity, as most of the neighboring rights which copyright systems had already accommodated for decades. Yet, compared to traditional neighboring rights, Article 7 Database marked the debut of a very broad exclusivity akin to a property right, potentially perpetual, uncertain in its scope, subject to very narrow limitations and immune from copyright exceptions. This not only channeled into the copyright system a foreign element that was predestined to create systemic short-circuits and problems in the implementation by national legislators and even more by national courts, but it also opened the floodgate for an extension of exclusive rights to cover realms which traditionally belonged to the public domain, without adequate antibodies that could maintain the traditional balance set by copyright law. This chapter will provide an overview of the road that led to the Database Directive, analyzing and commenting on its most relevant provisions on the sui generis right, their rationale and interpretations (Section II.II). It will then offer a detailed overview of the evolution of key concepts and definitions related to Articles 7–11 Database in the case law of the CJEU (Section II.III), look at the assessment of the national implementations and overall impact of the sui generis right provided by the European Commission and by copyright scholars (Section II.IV), and conclude on outstanding challenges and the way forward (Section II.V). II.II THE ROAD TO THE DATABASE DIRECTIVE

Before the adoption of the Directive, and along the lines of Article 2(5) of the Berne Convention,3 most national copyright laws provided for the protection of collections of works that were original based on the selection and arrangement of content. Each Member State, however, presented a different approach, depending on the national requirements for copyright protection. In general, countries belonging to the droit d’auteur tradition protected only databases that were original enough to represent an intellectual creation, with different degrees of originality requested by courts, while countries from the common law tradition applied the skill and labor doctrine, thus protecting also nonoriginal databases that required “sweat of the brow” to be produced.4 In addition, some Member States provided special forms of protection for catalogues (e.g., Sweden, Denmark and Finland).5 Data from 1990 showed how 50 percent of European online database services were based in the UK, a country that protected a wide array of databases and offered more legal certainty than continental jurisdictions, where the situation was much more controversial and fragmented.6 This evidence of clear unbalance between Member States, coupled with the obstacles created for the internal market by the patchwork of legal solutions, called for a harmonizing intervention from the European Community, which had to find a midway between national approaches and, at the same time, cover with other entitlements those nonoriginal databases which would have fallen out from copyright protection in common law jurisdictions due to the increased originality threshold.7

3

4 5

6

7

Berne Convention for the Protection of Literary and Artistic Works, September 9, 1886, as revised at Stockholm on July 14, 1967 828 UNTS 221. On the protection of databases in international norms, see Daniel Gervais, ‘The Protection of Databases’ (2007) 82 Chicago Kent Law Review 1109, 1111–17. See the overview provided by Mark Davison, The Legal Protection of Databases (CUP 2003) 10–49. See Gunnar Karnell, ‘The Nordic Catalogue Rule’ in Egbert Dommering and Bernt Hugenholtz (eds) Protecting Works of Fact (Kluwer 1991) 67. As indicated in European Commission, Panorama of EC Industry 1990: over 165 sectors of manufacturing and service industries in focus, Brussels, 1990, at 30.17. ibid.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

Sui Generis Protection of Non-creative Databases

29

The first preparatory work mentioning the need to provide legal protection for databases for internal market needs was the Green Paper on Copyright and the Challenges of Technology (1988).8 The Green Paper asked stakeholders and the public whether the protection of compilations under copyright law should have been extended to databases containing materials not protected by copyright, and whether the preferred regulatory solution was copyright or a sui generis right.9 The responses were channeled in the follow-up to the Green Paper (1991),10 which indicated the intention to introduce as soon as possible blended solutions to cover both original and nonoriginal databases.11 A year later, the EC tabled the Initial Proposal, describing databases as a “vital tool in the development of an information market within the community,” particularly in light of the “exponential growth [. . .] in the amount of information generated and processed annually in all sectors of commerce and industry,” which “required investment [. . .] in advanced information management systems.”12 The new harmonized framework was deemed necessary for the development of a strong and competitive European database industry, still lagging behind compared to its main trading partners.13 This first draft covered only electronic databases and included both a protection by copyright and a “right to prevent unfair extraction from a database,” the second being very broad, albeit limited by compulsory licenses.14 The EC excluded the suitability of a copyright-only solution for two parallel reasons. On the one hand, copyright protection alone could have not covered all noncreative databases.15 On the other hand, harmonizing the originality standard to the level needed to protect as many databases as possible, without revolutionizing the continental model, would have excessively lowered the benchmark in droit d’auteur countries while still raising it in common law jurisdictions, thus reducing the protection the latter offered to a wide range of noncreative databases.16 On this basis, the proposal adopted a harmonized, higher threshold for copyright protection of original databases, based on the notion of originality and crossing out the lower “sweat of the brow” standard adopted in the common law environment. To cover nonoriginal databases that still required a 8

9 10

11

12

13 14

15

16

European Commission, Green Paper ‘Copyright and the Challenge of Technology: Copyright Issues Requiring Immediate Action’ COM(88) 172 final, ch. 6. ibid at para 6.7.1. Responses were more for copyright protection than for sui generis protection. See Michel M Walter and Silke von Lewinski, European Copyright Law: A Commentary (OUP 2010) 9.0.12; P Bernt Hugenholtz, ‘Something Completely Different: Europe’s Sui Generis Database Right’ in Susy Frankel and Daniel Gervais (eds) The Internet and the Emerging Importance of New Forms of Intellectual Property (Kluwer 2016) 207. European Commission, Working Programme of the Commission in the Field of Copyright and Neighbouring Rights. Follow-up to the Green Paper, COM (90) 584, 18 et seq. Proposal for a Council Directive on the Legal Protection of Databases, COM (92) 24 final, OJ C-156/4 (Initial Proposal), Recital 9. As explained in the Explanatory Memorandum of the Initial Proposal, Part 1, paras 2.2.11 and 5.1.1. Initial Proposal, Article 1. For an analysis of the drafting history of the directive, see P Bernt Hugenholtz, ‘Implementing the Database Directive’, in Jan JC Kabel and Gerard JHM Mom (eds) Intellectual Property and Information Law: Essays in Honour of Herman Cohen Jehoram (Kluwer 1998); Annemarie C Beunen, Protection for Databases: The European Database Directive and Its Effects in the Netherlands, France and the United Kingdom (Wolf 2007). Despite the UK proving to be a nonoriginal database-friendly jurisdiction thanks to its skill and labor standards (see, eg, Ladbroke Football Ltd v William Hill Football Ltd [1964] 1 WLR 273), and early case law from some continental jurisdictions admitted to protection of some nonoriginal databases as “information works” (see, eg, Cour de Cassation November 9, 1983, Droit de l’informatique 1984/1, 20; Cour de Cassation October 30, 1987, Droit de l’informatique 1988/1, 34), a mixed system such as the Dutch one shows much more reluctance, denying protection, for instance, to a telephone directory and a dictionary (Dale v Romme, Decision of January 4, 1991, [1991] NJ 2543 (no 608)). Explanatory Memorandum to the Initial Proposal (n 12) Part 1, para 5.1.1. See also Jens L Gaster, ‘The EU Council of Ministers’ Common Position Concerning the Legal Protection of Databases: A First Comment’ Entertainment Law Review (1995) 260.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

30

Caterina Sganga

qualitatively or quantitatively substantial investment for obtaining, verifying and presenting the materials, the proposal introduced a new sui generis right to prevent extraction and reutilization of the whole or a substantial part of their content.17 The discussion on the legal nature of this new right was expectedly heated. The Initial Proposal did not specify any definition, but for the introduction of the label sui generis right and its derivation “from regimes such as unfair competition law or the law repressing parasitic behavior.”18 However, since the EC had no intention to harmonize unfair competition law across the Union in light of the great differences among Member States,19 and of the incapability of unfair competition law to repress parasitic acts beyond those committed by competitors,20 the draft left to Member States the decision on how to implement the right.21 The proposal got the positive opinion of the Economic and Social Committee, while the European Parliament adopted several amendments,22 mostly intervening on the definition of the legal nature of the sui generis right, which became a fully-fledged intellectual property right to prevent “unauthorized” (and not “unfair,” as in the original proposal) extractions.23 The draft was then subject to a heated debate before the Council, where it took two full years to reach a Common Position, which substantially departed in terms of structure, degree of details and scope from the original proposal.24 The new text extended the protection to nonelectronic databases, and provided a much clearer distinction between provisions on copyright and provisions on sui generis right. The latter was described as a proprietary – and thus subjective, exclusive and transferable – right and not as a mere entitlement to protection based on unfair competition law and limited to relationships between competitors.25 This emerged clearly in the definition of its subject-matter, identified in a qualitatively/quantitatively substantial investment. With a careful balancing exercise, the proposal restricted the protection to uses involving the entire database or substantial parts thereof – once again based on the negative impact of the conduct on the investment – but it also eliminated compulsory licenses in favor of a much stricter list of exceptions. The hybrid nature of the new right made it possible also to provide its renewal every time the content of the database was subject to a substantial modification by virtue of a new qualitatively and/or quantitatively substantial investment.26 Both the Parliament and the Commission approved the revised proposal with few cosmetic amendments, enacting Directive 96/9/EEC as a “cornerstone of intellectual property protection in the new technological environment.”27 The sui generis right represented an absolute novelty in the IP (Intellectual Property) arena worldwide, and became so much of a flagship for the Community that the Commission tabled before the Committees of Governmental Experts of WIPO (World 17 18 19 20 21 22 23

24

25

26 27

ibid para 5.3.7. ibid para 5.3.6. ibid para 5.3.9. ibid para 5.3.10. ibid para 2.5. As in Walter and von Lewinski (n 10) para 9.0.4. See Explanatory Memorandum of the Amended Proposal for a Council Directive on the legal protection of databases, OJ C-308/1, p. 2. On the new proposal, see the comments of Simon Chalton, ’The Amended Database Directive Proposal: A Commentary and Synopsis’ (1994) 3 EIPR 94. Common position adopted by the Council on July 10, 1995, OJ C 288/14. The position, inter alia, streamlined the text, clarified the wording of several provisions and increased the number of Recitals from forty to sixty. The difference between the sui generis right and an entitlement deriving from unfair competition law is also that in the former case, the right is attributed a priori, while the latter sanctions behaviors a posteriori. For similar observations see Jerome H Reichman and Pamela Samuelson, ‘Intellectual Property Rights in Data?’ (1997) 50 Vanderbilt Law Review 51, 81; Hugenholtz (n 14) 187. As explained in the Common position (n 24) 11. European Commission, ‘Follow-up to the Green Paper on Copyright and Related Rights in the Information Society’ COM(96) 483 final, 8.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

Sui Generis Protection of Non-creative Databases

31

Intellectual Property Organization) a proposal for an international treaty on the sui generis protection of databases, to be conceived as a Berne Protocol or as a new instrument.28 Initially the idea was very well received by several delegations, including the USA, which responded with a partially different draft.29 The chairman of the Committee of Experts presented a merge of the EU and US proposals at the WIPO diplomatic conference in December 1996. The draft should have been discussed along the proposals for the two WIPO Internet Treaties,30 but the time available did not allow it. The only product of the discussion was a recommendation to follow up with the matter,31 but years of discussion did not lead to any consolidated product. The only explicit international references to databases remained, thus, Article 10(2) of the TRIPs Agreement32 (Trade-Related Aspects of Intellectual Property Rights) and Article 5 of the WCT (World Copyright Treaty), which use an almost identical language to require copyright protection for “compilation of data or other material,” whether in electronic or any other form, “which by reason of the selection or arrangement of their contents constitute intellectual creations.” Both provisions specify that the exclusivity should not extend to the data or material itself and should not prejudice any copyright subsisting on the database content. The two treaties, entered into force in 1995 and 2002, aimed at making sure that contracting States extended their copyright protection also to compilations of raw data and materials which are not subject to any exclusive right.33 Significantly, both the WCT and the TRIPs requested only a protection by means of traditional copyright, which entailed the need for the compilation to represent an “intellectual creation,” that is a product that meets the requirements set by national laws to be qualified as a protected “work” (originality and/or creativity, sweat of the brow, etc.). Against this background, the approach adopted by the EU legislator represented a worldwide unicum. II.III THE SUI GENERIS RIGHT IN THE EU DATABASE DIRECTIVE (96/9/EEC)

II.III.A The Directive in General In the original EC proposal, the new directive should have covered only those electronic databases that could not be eligible for copyright protection under national copyright law. This was usually the case of compilations of raw data and materials other than traditional 28

29

30

31 32

33

WIPO Committee of Experts, Basic Proposal for the Substantive Provisions of the Treaty on Intellectual Property in Respect of Databases to be Considered by the Diplomatic Conference, WIPO Doc CRNR/DC/6 (August 30, 1996). See Gervais (n 3) 1114–15, 1119. This proposal and the parallel legislative attempts within the USA are carefully examined by Philip J Cardinale, ‘Sui Generis Database Protection: Second Thoughts in the European Union and What It Means for the United States’ (2007) 6(2) Chicago-Kent Journal of Intellectual Property 157. The WCT and WPPT were indeed approved in that session. WIPO Performances and Phonograms Treaty, December 20, 1996, S. Treaty Doc No 105–17 (1997); 2186 UNTS 203; 36 ILM 76 (1997); WIPO Copyright Treaty, December 20, 1996; S. Treaty Doc No 105–17 (1997); 2186 UNTS 121; 36 ILM 65 (1997) WIPO doc CRNR/DC/88. Agreement on Trade-Related Aspects of Intellectual Property Rights, Apr. 15, 1994 – Marrakesh Agreement Establishing the World Trade Organization, Annex 1C, 1869 UNTS 3; 33 ILM 1197 (1994). For an overview of the drafting process and preparatory works see Mihaly Ficsor, The Law of Copyright and the Internet: The 1996 WIPO Treaties, Their Interpretation and Implementation (OUP 2002) ch. 2; Daniel Gervais, The TRIPS Agreement: Drafting History and Analysis (4th edn, Sweet & Maxwell 2012) 62 ff.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

32

Caterina Sganga

“works,” or where data were arranged by software.34 Despite their lack of originality or personal touch, such products still required substantial and risky investments to be developed, and had a relevant potential market, thus a top-down intervention from the European legislator was perceived as strongly needed to support the database industry.35 When the proposal reached the Council, the TRIPs Agreement and the WCT were either already approved or about to be approved, and with them the obligation for contracting parties to protect also nonelectronic databases via copyright. This led to the opening of the scope of the Directive to databases in any form, not only in the field of copyright but also in that of the sui generis right.36 The Directive devotes two separate chapters to the regulation of copyright and sui generis rights, mirroring also in its structure the legislative intention to maintain separated their respective justifications, general principles, requirements, rights and exceptions, with the aim of avoiding systematic confusion in the judicial evolution of the subject. According to Article 1 Database, the Directive covers collections37 of independent works, data and other materials, “arranged in a systematic or methodical way and individually accessible by electronic or other means,” with the exclusion of software programs used in its making or operation. Recital 17 specifies that a database may include any type of protected work or other material “such as texts, sound, images, numbers, facts and data,” with the exclusion of “the compilation of several recordings of musical performances on a CD,” both because this product does not meet the requirement for copyright protection and because it does not represent a substantial enough investment to be eligible under the sui generis right (Recital 19). On the contrary, the protection also covers materials necessary for the operation or consultation of certain databases, as in the case for indexation and thesaurus systems (Recital 20). The definition provided by the Directive is purposefully broad and comprehensive. The term “in any form” is conceived to be overarching and cover, exempli gratia, both online and offline databases (as in Recital 22), or static and dynamic databases.38 Similarly, Article 1 does not specify the amount of material necessary to have a “collection,” nor does it require the materials to come from one or multiple sources, or to be or not be created by the person or entity making the collection. Restrictions come, instead, from the requirements of independence of the materials, individual accessibility by any means39 and arrangement in a systematic or methodical way, which exclude from protection both creations that may rather amount to individual works or unitary nonprotectable creations, and unstructured or arbitrary accumulation of data.40 The definition of the subject-matter of the Directive is the product of a careful balancing between opposite considerations and features an attentive selection of terms and requirements. 34 35 36 37 38

39

40

As illustrated in the Explanatory Memorandum to the Initial Proposal (n 12) 3.1. ibid at 3.2. On this debate see Walter-von Lewinski (n 9) 9.1.1–9.1.4. The Preamble clarifies that “collection” should be understood as a synonym of “compilation” (Recital 13). On the broadness of the definition of databases see, eg, Simon Chalton, ‘The Effect of the EC Database Directive on United Kingdom Copyright Law in Relation to Databases: A Comparison of Features’ (1997) EIPR 278; Davison (n 4) 61 ff; Derclaye (n 2) 65–67; Mathias Leistner, ‘The Protection of Databases’ in Estelle Derclaye (ed) Research Handbook on the Future of EU Copyright (Edward Elgar 2009) 431. This requirement may exclude, for instance, materials of neural networks. See Thomas Dreier, ‘Die Harmonisierung des Rechtsschutzes von Datebanken in der EG’ (1992) GRUR International 739. See also FW Grosheide, ‘Database Protection: The European Way’ (2002) 8 Washington University Journal of Law and Policy 39. The concept of “systematic or methodical” is not interpreted strictly, although in some Member States, such as Germany and Austria, it is used to distinguish between collection works in general and database works, the latter being structured around logical criteria. On this see also Estelle Derclaye, ‘What Is a Database? A Critical Analysis of the Definition of a Database in the European Database Directive and Suggestions for an International Definition’ (2002) 5 Journal of World Intellectual Property 981.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

Sui Generis Protection of Non-creative Databases

33

Yet, the inevitable vagueness of some of the criteria and concepts used could not but give rise to heated academic debates and, expectedly, to a rich roster of judicial decisions, as we will see in more in Section II.III. The same happened, due to the novelty of most of its concepts and terms, to the sui generis right, introduced by Article 7. II.III.B The “Revolutionary” Sui generis Right Article 7(1) Database marks a clear distinction between the two sets of rights conferred by the Directive. The provision, in fact, describes the sui generis right as an entitlement conferred to the “maker” of a database “which shows that there has been qualitatively and/or qualitatively a substantial investment in either the obtaining, verification or preservation of the contents.” The use of the term “maker” and the connector “which” stands in clear opposition with the attribution of copyright to the “author” of the database, who is a natural or group of natural person(s), and a legal person only if admitted under national copyright law. At the same time, the requirement for protection is not the originality in the selection or arrangement of the materials, but the evidence of a substantial investment, measured in qualitative or quantitative terms, and directed to obtain, verify or preserve the content of the base.41 More specifications come from the Preamble. Recital 41 helps clarifying that the “maker” is a natural or legal person taking the initiative and bearing the risk of investing on the construction of the database. This clearly suggests that the Directive does not demand a direct engagement with development activities but only the assumption of initial steps – both organizational and financial – toward the endeavor. The key importance of the investment risks is also emphasized by the exclusion of subcontractors, and thus also employees, from the notion of “maker.”42 The key notion, however, remains that of “substantial investment.” “Investment” entails the use of “human, technical, and financial resources” (Recital 7), but also the use of time, effort and energy (Recital 2).43 Although Recital 39 refers to “financial and professional investment,” there is no other hint which would suggest the legislative intention to exclude from protection databases which are developed out of noncommercial efforts from private individuals who only later decide to exploit them.44 Most importantly, not all investments give rise to a sui generis protection, but only those which are necessary to obtain, verify or present the contents. This implies that an investment for the acquisition of an already-made database is not a sufficient basis to trigger the application of Article 7 Database while, on the contrary, an investment in technical devices necessary to obtain data, even if in an unstructured form, may

41

42 43

44

Along the same line see Grosheide (n 39) 55; Willliam Cornish, ‘1996 European Community Directive on Database Protection’ (1996) 21 Columbia Journal of Law and the Arts 1, 8; see also Gerald Dworkin, ‘Copyright, Patent or Protection for Computer Programs’ (1996) Fordham International Intellectual Property Law and Policy 183; for a broader overview Michal Koščík and Matˇej Myška, ‘Database Authorship and Ownership of Sui Generis Database Rights in Data-Driven Research’ (2017) 31 International Review of Law, Computers and Technology 43. As noted by Walter-von Lewinski (n 9) 9.1.6. See, eg, Paul Gaudrat, ‘Loi de transposition de la directive 96/9du 11 mars 1996 sur les bases de données: le champ de la protection par le droit sui generis’ (1999) 52 (1) RTD Com 100–02; Cornish (n 41) 9; Grosheide (n 39) 62 ff, analyzing also national case laws. On the debated treatment of so-called spin-off databases, see the detailed analysis of P Bernt Hugenholtz, ‘Program Schedules, Event Data and Telephone Subscriber Listings under the Database Directive: The “Spin-Off” Doctrine in the Netherlands and Elsewhere in Europe, Paper presented at Eleventh Annual Conference on International IP Law & Policy, Fordham University, New York, April 14–25, 2003.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

34

Caterina Sganga

be enough (Recital 39), as long as it was purposefully directed to the subsequent structuring of a database.45 The question of whether “obtaining” content could also include investments directed to create materials, instead, required, as we will see, an intervention of the CJEU (Court of Justice of the European Union).46 The exclusion of investments for the self-production of data is directed to avoid the privatization of data corpora which are generated by a single source, which would have strong anticompetitive effects.47 For this reason, Article 7 would still protect databases where the investment was directed to collect fully reelaborate raw data.48 The same distinction between generated and obtained data has to be made with regard to the investments made to verify the content, that is to make sure that information are updated and reliable, and to the presentation of the data, that is their structuring and arrangement, and their presentation via software or other means.49 To obtain protection, the investment should be substantial. This characterization introduces a benchmark that aims at excluding from the scope of Article 7 those databases the production of which did not require the assumption of an entrepreneurial risk that was high enough to justify a monopolization of their content.50 Substantiality has represented one of the most debated and controversial notions introduced by the Database Directive, for it triggered a plethora of different scholarly views and contrasting interpretations in national case laws.51 It generally translates into the notion of “considerable” from a qualitative or quantitative perspective, which Recital 7 connects to the cost necessary for copying or accessing the base independently. From this perspective, the legislative intent seems to qualify the substantiality of the investment on the basis of a comparison between the high costs needed to develop the base and the low cost of freeriding, with the exclusion of any other criteria and in line with the teleological justification of the sui generis right.52 Criteria used to establish whether the benchmark is met are both qualitative and quantitative, in order to take into account both the amount of effort, money and time invested in the endeavor, and the value of the specific skills, techniques and equipment used in obtaining, verifying and presenting the content.53 The justification underlying the introduction of the sui generis right also represents the reason of the careful definition of its subject matter. In fact, Article 7 extends the protection only to the whole of the content or to a substantial part thereof, evaluated qualitatively and/or qualitatively. This criterion clearly departs from the notion of originality which is used to determine the minimum 45

46 47

48 49 50

51 52

53

See the comprehensive analysis of Estelle Derclaye, ‘Database Sui Generis Right: What Is a Substantial Investment? A Tentative Definition’ (2005) 36 International Review of Intellectual Property 2, and the national case law account of Grosheide (n 39) 62 ff. See Section II.IV.B. See, eg Guido Westkamp, ‘Protecting Databases under US and European Law: Methodical Approaches to the Protection of Investments between Unfair Competition and Intellectual Property Concepts’ (2003) 34 International Review of Intellectual Property 772, 782. As in Leistner (n 38) 438. ibid at 439. Similarly see Westkamp (n 47) 781; and Michael Tappin et al., Laddie, Prescott and Vitoria on the Modern Law of Copyright (5th edn, Lexis 2018) 1076. Against the setting of a high threshold see Mathias Leistner, ‘Legal Protection for the Database Maker: Initial Experience from a German Point of View’ (2002) 33 International Review of Intellectual Property 439, 449–50; contra Derclaye (n 45) 10. See the account of the academic debate provided by Derclaye (n 45), 8–12. With the exclusion, for instance, of subjective criteria such as the financial situation of the investor or the size of the company, or secondary factors such as the potential market success of the database. See Walter-von Lewinski (n 9) 9.7.13–14. Broadly see Derclaye (n 2) 76 ff with ample references to national case law; similarly in Davison (n 4) 97 ff; Grosheide (n 39) 52 ff; Leistner (n 38) 439; Gaudrat (n 43) 101. On the amplitude of the qualitative criterion see Hugenholtz (n 14) 134–36.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

Sui Generis Protection of Non-creative Databases

35

excerpt still protected by copyright and thus amounting to a partial reproduction.54 The reason is quite simple and intuitive. Copyright protects creativity and the personal touch of the author, thus it also protects rightholders’ interest in remaining in control of any use of the work which, albeit minimal and thus not in competition with its normal exploitation, still involves fractions that carry its “spirit.”55 On the contrary, the protection offered by the sui generis right is directed only to secure the risky investments faced by the database maker against cheap acts of freeriding or unauthorized acts of exploitation that may endanger the market of the database.56 A confirmation of this reading comes from Recital 42, which specifies that the right to prohibit extraction and/or reutilization “relates not only to the manufacture of a parasitical competing product but also to any user who, through his acts, causes significant detriment, evaluated qualitatively or quantitatively, to the investment.” For this reason, the notion of “substantial” part has to be established on the same grounds of the “substantiality” of the investment, and consider any type of use, regardless of its commercial or noncommercial/private nature.57 A part may be substantial not only if the quantity of the materials extracted and reused is remarkable compared to the entire database, but also if the materials extracted, regardless of its quantity, required a qualitatively and/or quantitatively substantial investment to be collected, verified and presented.58 Along the same lines, the extraction and/or reutilization of insubstantial parts may still become substantial and impair the investment in specific circumstances, that is when the extraction and/or reutilization is repeated and systematic, insomuch as to conflict with a normal exploitation of the database, or to unreasonably prejudice the legitimate interests of the maker (Article 7(5) Database).59 The protection offered by the sui generis right is defined as the possibility to prevent the “extraction” and “reuse,” both terms which do not find any correspondence in copyright law nor in any related right. The notion of “extraction,” meaning “the permanent or temporary transfer of all or a substantial part of the contents of a database to another medium by any means or in any form,” is conceptually close to that of reproduction under Article 2 InfoSoc,60 or even broader, since also the transfer to another medium without a duplication may entail an extraction, as indirectly confirmed by the CJEU.61 As it is the case for every other exclusive right, the definition has to be read broadly and in light of the goal of Article 7, which is to guarantee to database maker an adequate return by protecting them against any unauthorized appropriation of the results of their investment.62 While it is uncontested that the transfer to “another” medium does not require a change in kind, but only that the origin and destination are different and independent from each other, it remains questionable whether some form of consultation or displaying on

54 55

56 57 58 59

60

61

62

As also in Grosheide (n 39) 54. The CJEU explained it in case C-5/08 Infopaq International A/S v Danske Dagblades Forening [2009] ECR I-6569. For a broader analysis see Caterina Sganga, ‘The Right of Reproduction’ in Eleonora Rosati (ed) Routledge Handbook on EU Copyright Law (Routledge 2021) ch. 14. As in Westkamp (n 47) 784. ibid. This interpretation will be later confirmed by the CJEU in a strain of decisions. See infra Section II.IV.C. Some countries classified these repeated and systematic extractions and reuse as extraction and reuse of substantial parts. In some national laws the two sets of acts are regulated adjacently. See, eg §87b (1), sentence 2, German Copyright Act; Art. 2 Dutch Database Act. See Walter-von Lewinski (n 9) 9.7.22. Directive 2001/29/EC of May 22, 2001 on the harmonization of certain aspects of copyright and related rights in the information society (2001) OJ L167/10. First, albeit not directly, in the BHB decision (infra n 72), and later in Directmedia and Apis-Hristovich (infra n 126 and 133). See infra Section II.IV.D. Directmedia and Apis-Hristovich (infra n 126 and 133)

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

36

Caterina Sganga

screen may fall under the prohibition of Article 7 Database.63 Recital 44 states that the right covers all instances where an extraction is needed to perform an onscreen display, which e contrario suggests that the displaying itself is outside the scope of the provision. The same can be said for consultations if no transfer is involved.64 The act of “reutilization” covers both acts of exploitation and acts performed without any commercial aim. Article 7(2)(b) refers to “any form of making available to the public” of all or a substantial part of the database, and specifies as examples the distribution of copies, and online and other forms of transmission. Despite the exclusion of rental and lending, this remains the most overarching and comprehensive part of Article 7, for it extends to any form of communication to the public and to the transfer of the database on a tangible support, the latter being subject to exhaustion.65 However, and differently than for the notion of “extraction,” the use of terms that are common in copyright law has facilitated the reference to general definitions and their judicial interpretation, easing the interpretative short-circuits caused by the introduction of new concepts within the tangles of national laws.66 To tackle the potential overlaps between different rights – sui generis over the content, copyright over the database structure, copyright and other exclusive rights on single materials – the Directive sets some basic rules to guide their interplay.67 Article 7(4) specifies that the sui generis right applies irrespective of the application of copyright on the structure and of the protection of the contents by copyright or other rights. Article 7 also clarifies that the sui generis right leaves unprejudiced any right existing on the database content, while Recital 46 reiterates that the sui generis right does not establish a new right in the works, data or materials, making a cumulation of protection on the very same piece of content highly unlikely to occur. At the same time, Recital 45 states that the sui generis right does not constitute an extension of copyright protection to items that are not eligible for it, such as mere facts or data, in this way highlighting the intention of the EU legislator to circumscribe the scope Article 7 Database and prevent information monopolies. Originally set in ten years and then prolonged to fifteen to ensure a full amortization of the investment,68 the term of protection starts from the first of January of the year following the date of full completion. The most controversial aspect remains, however, the possibility to extend the term in case of qualitatively or quantitatively substantial change of content (Article 10(3)), which may open the door for the conferral of perpetual exclusivity, and creates relevant legal uncertainty as to its scope, since the extent and type of change determine the subject matter of the term extension.69 When a static database is integrated by additional parts that required a substantial investment, in fact, the new term applies only to the new extensions. In the case of dynamic databases that require substantial investments to be updated, instead, the renewal concerns the entire base, which is thus subject to a potentially perpetual protection that goes

63 64 65 66 67

68 69

See, more extensively, Derclaye (n 2)104. ibid. Analogously see Hugenholtz (n 3) 213. This is the belief of Walter-von Lewinski (n 9) 9-7-36. In its Original Proposal, the Commission restricted its sui generis right to the parts of the database content that were not protected by any other right. When the Council modified Article 7 to exclude insubstantial part of the database from the sui generis protection, excluding compulsory licenses, the sui generis right was extended again to cover the entire base, from which it came the need to clarify the interpretation of potential cases of rights overlaps. See Statement of the Council Reasons, Common Position (n 24) no 14. See Explanatory Memorandum of the Amended Proposal (n 23) 6. See the attentive analysis of Derclaye (n 2)137 ff. On term perpetuity and the propertization of information see Reichman-Samuelson (n 25) 85–86. As for the date of completion, the burden of proof of the subsistence of the requirement for extension lies on the maker (Recital 54).

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

Sui Generis Protection of Non-creative Databases

37

much beyond the duration of the protection conferred by copyright.70 The reasons underlying this difference are said to lie on the rationale of the two rights: since the sui generis rights protect investments, it appears logical and consequential to grant additional protection every time a new investment is made to ameliorate the product. However, the potentially perpetual renewal is left without any stronger counterbalancing measure and, absent a system of registration of database rights, puts on competitors and users the cost of ascertaining duration and scope of the term extension, despite the imposition of the burden of proof of the changes on the database maker.71 II.III.C Lawful Uses and Exceptions Considering the implications of the sui generis right on access to information and the risk of informational monopolies, the EU legislator has introduced a set of provisions directed to strike a balance between rightholders’ and users’ interests. Article 8 Database is devoted to the rights and obligations of lawful users, “whose access to the contents of a database for the purpose of consultation results from the direct or indirect consent of the maker of the database.”72 The first paragraph excludes the possibility for rightholders to prevent a lawful user from extracting and/or reutilizing insubstantial parts of the database contents, evaluated qualitatively or quantitatively, for any purpose.73 In this sense, the provision adds little to Article 7, which already exclude from infringement extraction and reutilization of insubstantial parts, as long as they are not repeated and systematic and hurt the maker’s interests. The goal of Article 8, however, is probably different, and namely that of ensuring that the balance set by law and the scope of Article 7 are not modified by the market. Article 15, in fact, declares any contractual provision contrary to Article 8(1) null and void, using the same approach adopted in the Software Directive for the backup and interoperability exceptions.74 In order to avoid abuses, Article 7(5) Database provides an exception to lawful uses, ruling that a repeated and systematic extraction and reutilization of unsubstantial parts of a database amount to an infringement of the sui generis right if it conflicts with the normal exploitation of the base or which unreasonably prejudice the legitimate interests of the maker. Along the same lines, Articles 8(2) and (3) forbid lawful users to perform acts having similar effects, and to cause prejudice to the holder of copyright or related rights on works that are contained in the database. Exceptions are carefully tailored. Article 9 Database admits unauthorized extractions and reutilization of substantial parts of the database content for private purposes in case of nonelectronic databases, for illustration for teaching or scientific research, and for public security or administrative or judicial procedures. The provision, which is deemed exhaustive and of 70

71

72

73 74

The problems raised by the difference between static and dynamic databases are addressed also by Guido Westkamp, ‘EU Database Protection for Information Uses under an Intellectual Property Scheme: Has the Time Arrived for a Flexible Assessment of the European Database Directive?’ (2003) < https://ssrn.com/abstract=1115432> accessed 28 February 2021. As well pointed out by Jerome Reichman and Paul Uhlir, ‘Database Protection at the Crossroads: Recent Developments and Their Impact on Science and Technology’ (1999) 14 Berkeley Technology Law Journal 793, 801. The definition comes from case C-203/02 The British Horseracing Board Ltd and Others v William Hill Organization Ltd [2004] ECR I-10415, para 58. Recital 34 defines it indirectly by referring to the user to whom “the rightholder has chosen to make available a copy [. . .] whether by an on-line service or by other means of distribution.” On the notion see Vinciane Vanovermeire, ‘The Concept of the Lawful User in the Database directive’ (2002) 31 International Review of Intellectual Property and Competition Law 62. Accordingly, Derclaye (n 2) 127, suggests that the provision is “redundant and misleading.” The reference goes to Article 5(2) and Article 6 of Directive 91/250/EEC on the legal protection of computer programs (1991) OJ L122–42.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

38

Caterina Sganga

maximum harmonization, is limited to uses for noncommercial purposes and has to be interpreted narrowly.75 Differently than what is provided under Article 6 Database for copyright protection, there is no reference to the possibility for Member States to add further derogatory provisions from national copyright laws,76 and also the content of Article 9 exceptions is in some instances narrower if compared to the correspondent exceptions to copyright.77 Considering the exhaustive nature of the list of exceptions provided by Article 9 Database, subsequent directives have intervened to add new limitations to tackle emerging balancing needs that could not be addressed under the provision. An example comes from the Marrakesh Directive, which introduces a mandatory exception (also) to Articles 5 and 7 Databases in order to enable visually disabled individuals and authorized entities to make accessible copies of protected works and to communicate them to the public.78 Recently, and after years of debate,79 the Directive on Copyright in the Digital Single Market has tackled the problems that database protection raised for the operation of artificial intelligence (AI) agents and machine learning processes, introducing two mandatory textand-data-mining (TDM) exceptions.80 The first is dedicated to TDM activities for the purpose of scientific research, entailing reproductions and extractions made by research organizations and cultural heritage institutions (Article 3). The provision is not overridable by contract, allows the retention of copies of works, and requires the adoption of appropriate security measures and proportionality in the activities carried out. The second exception allows general TDM activities but subordinates the exception to the fact that rightholders have not expressly reserved such uses (Article 4). The TDM exceptions have been introduced to tackle the most evident criticalities that Article 7 Database have raised in the context of the AI and data economy, where information monopolies act as strong obstacles to data flows and sharing of data corpora, and thus to innovation and to the development of economies of scales.81 However, they represent only a circumscribed solution to some of the problems triggered by the sui generis right, offering no response to the challenges raised by the lack of flexibility of Article 9 Database and the nonalignment of general copyright exceptions and sui generis exceptions.

75

76

77

78

79

80

81

The exhaustiveness is derived from Recital 50. See, in this sense, Jean-Paul Triaille and Alain Strowel, Le droit d’auteur, du logiciel au multimédia: droit belge, droit européen, droit compare (Kluwer 1997) 287. Recital 52 only permits certain Member States to maintain their exceptions for rights comparable to the sui generis right. The reference mostly goes to the catalogue right and its exceptions (see supra n 5). Compare, eg, Article 6(2)(b) on the copyright exception for purposes of illustration for teaching and research, which covers also online transmission or transmission on a large screen, while the corresponding provision for the sui generis right, Article 9(b), is limited to extraction and does not extend to reutilization, which would cover the act of making available to the public. Directive (EU) 2017/ of September 13, 2017 on certain permitted uses of certain works and other subject matter protected by copyright and related rights for the benefit of persons who are blind, visually impaired or otherwise printdisabled (2017) OJ L242/6. See, inter alia, Christophe Geiger, Giancarlo Frosio and Oleksandr Bulayenko, ‘Text and Data Mining in the Proposed Copyright Reform: Making the EU Ready for an Age of Big Data?’ (2018) 49 International Review of Intellectual Property and Competition Law 814; Reto Hilty and Moritz Sutterer, ‘Position Statement of the Max Planck Institute for Innovation and Competition on the Proposed Modernisation of European Copyright Rules’ (March 4, 2017) accessed 28 February 2021; Thomas Margoni and Martin Kretschmer, ‘The Text and Data Mining Exception in the Proposal for a Directive on Copyright in the Digital Single Market: Why It Is Not What EU Copyright Law Needs’ (April 25, 2018) accessed 28 February 2021; Eleonora Rosati, ‘An EU Text and Data Mining Exception for the Few: Would It Make Sense?’ (2019) 13(6) Journal of Intellectual Property Law & Practice 429. Directive (EU) 2019/790 of April 17, 2019 on copyright and related rights in the Digital Single Market (2019) OJ L130/ 92. As in Geiger-Frosio-Bulayenko (n 79).

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

Sui Generis Protection of Non-creative Databases

39

II.IV THE SUI GENERIS RIGHT IN THE CASE LAW OF THE COURT OF JUSTICE OF THE EUROPEAN UNION

As foreseeable considering the many general definitions and broad terms used in the Directive and the relative novelty of the concepts it introduced, the CJEU had to intervene several times to untie important interpretative knots. Compared to the high number of cases in the field of general EU copyright law, the number of database decisions is relatively limited. Yet, their clarity and consistency are remarkable, as so has been their impact on the development of the discipline. For the sake of conciseness, this chapter analyzes only decisions concerning Article 7 Database and the definition of sui generis right, which are nevertheless the great majority of the cases issued by the CJEU on the Database Directive.82 In this context, the Court’s case law mostly intervened on four areas: (i) the definition of “database” and its content; (ii) the notion of “substantial investment” as a requirement for the sui generis right protection; (iii) the notion of “substantial part” and “insubstantial part” to define the subject matter and scope of the provision; and (iv) the scope of the exclusivity, that is the notions of “extraction” and “reutilization.” II.IV.A The Definition of “Database” and Its Content The CJEU’s first attempt to draw the boundaries of the notion of “database” under Articles 1(2) and 7 Database is marked by Fixtures Marketing Ltd v OPAP, one of the three Fixtures Marketing cases decided by the Grand Chamber in November 2004,83 all concerning the use of fixture lists of professional football matches, prepared yearly by a working group consisting of representatives of the clubs. Fixtures was the company retained by the organizers of English and Scottish leagues to manage the exploitation of fixture lists outside the UK. The defendant, OPAP – the Greek company having the national monopoly on gambling activities – used without authorization information from the lists and was thus sued for violation of Fixtures’ sui generis right. The interpretation provided by the CJEU is very useful to bring order in the assessment of the applicability of the Database Directive. The Court clearly excluded the need for the materials to come from external sources and for the database to be original, since the definition of a collection as “database” under Article 1 of the Directive is independent from the assessment of whether the same database qualifies for copyright protection under Article 3 and/or for the sui generis protection under Article 7.84 Then, it grounded its answer on a teleological reading of the Database Directive, finding several indications of the legislative intention to conceptualize the 82

83

84

This analysis will omit, for example, commenting on Case C-604/10 Football Dataco Ltd et al v Yahoo! UK Ltd et al [2012] EU:C:2012:115, through which the Court brought clarity on the requirement of originality to confer copyright protection to databases under Article 3 of the Directive, crossing out the application of any existing national requirement, such as the UK notion of skill and labor, and specified once again the autonomy of the protection conferred to database by copyright and by the sui generis right, underlining that they can subsist independently from each other and that the qualification of a collection as a database under Article 1 of the Directive does not require meeting the requirements of Articles 3 and/or 7 Database. Case C-444/02 Fixtures Marketing Ltd v Organismos prognostikon agonon podosfairou AE (OPAP) [2004] ECR I10549, which the Grand Chamber decided on the very same day (November 9, 2004) together with case C-46/02 Fixtures Marketing Ltd v Oy Veikkaus Ab [2004] ECR I-10365 and case C-338/02 Fixtures Marketing Ltd v Svenska Spel AB [2004] ECR I-10497. The Opinions were all delivered by Advocate General Stix-Hackl. See the comments of Mark J Davison and P Bernt Hugenholtz, ‘Football Fixtures, Horse Races and Spin-offs: The ECJ Domesticates the Database Right’ (2005) 27 European Intellectual Property Review 113. Fixture Marketing Ltd v OPAP para 26.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

40

Caterina Sganga

term “database” as having a wide scope “unencumbered by considerations of a formal, technical or material nature.”85 Such evidence ranges from the reference to “any form” (Article 1(1)) to the later inclusion of nonelectronic databases (Recital 14),86 the very broad exemplificative list of potential content in Recital 17,87 and the final version of Article 1(2), where the EU legislator eliminated the definition of a database as a collection of a “large number” of materials, thus revealing the willingness to offer protection to every database, regardless of its size.88 More specifically, the Court argued that the notion should be defined “in terms of its function”89 which, according to Recitals 10 and 12, is to store and process information.90 On this ground, it concluded that there should be “a collection of independent materials,” which are separable from one another without their informative, literary, artistic, musical or other value being affected,91 and “systematically or methodically arranged and individually accessible in one way or another.” The CJEU gave a broad reading of “autonomous information value,” including, for instance, geographical maps that are made up of independent data points. While the arrangement should not be physically apparent, the collection should be embedded in a fixed base and include means that make it possible to retrieve any independent material contained in it,92 the latter feature being the one that distinguishes a database from a plain collection.93 II.IV.B The Notions of “Substantial Investment” The very first clarification of the requirements of protection set by Article 7 Database, which have an inevitable impact on the definition of its content and scope, came from Fixtures Marketing Ltd v Oy Veikkaus Ab, the second decision of the Grand Chamber trio, followed by Fixture Marketing Ltd v OPAP, and Fixtures Marketing Ltd v Svenska Spel AB,94 which used an almost identical reasoning and language. In Fixtures Marketing v Oy Veikkaus, the key question was whether the notion of “obtaining” under Article 7 Database may cover also investments directed at the creation of the database content.95 To answer, the CJEU deemed necessary to define the extent of the protection conferred by the sui generis right.96 Using the teleological method of interpretation, it highlighted how Recitals 9, 10, 12 and 39 Database explain that the purpose of the sui generis right is to promote and protect investments in the storage and processing systems of existing information to foster the development of an information market,97 and “to safeguard the results of the financial and professional investments made in obtaining and collecting the contents of a

85 86 87

88 89 90 91 92

93 94

95 96 97

ibid para 20. ibid para 22. ibid para 23. According to Recital 17, “the term ‘database’ should be understood to include literary, artistic, musical or other collections of works or collections of other material such as texts, sound, images, numbers, facts, and data.” ibid para 24. ibid para 27. ibid para 28. ibid para 29. ibid para 30. This is derived from Recitals 21 and 13. Technical means can be, in this sense, “electronic, electromagnetic or electro-optical processes, in the terms of the 13th recital of the preamble to the directive, or other means, such as an index, a table of contents, or a particular plan or method of classification.” ibid para 31. As in the other cases, Svenska was a betting company sued by Fixtures after refusing to enter into a license agreement with the latter for the use of data from the fixture lists. See Fixtures/Svenska Spel, paras 11–13. As rephrased by the CJEU in Fixtures/Oy Veikkaus, para 29. ibid at para 31. ibid at para 33.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

Sui Generis Protection of Non-creative Databases

41

database,” thus excluding the creation of the materials from the definition of obtaining.98 Then, it underlined that Recital 19 excludes that the compilation of several recordings of musical performances on a CD may constitute “a substantial enough investment to be eligible under the sui generis right,” reading this statement as a confirmation of the fact that the resources used for the creation of works or materials “cannot be deemed equivalent to investment in the obtaining of the contents of that database and cannot, therefore, be taken into account in assessing whether the investment in the creation of the database was substantial.”99 On these bases, the Court concluded that the expression “investment in [. . .] the obtaining, verification or presentation of the contents” should be referred to the investment in the creation of the database as such,100 and thus to the resources “used to seek out existing independent materials and collect them in the database, and not to the resources used for the creation as such of independent materials.”101 Teleological arguments, and particularly the language used by Recitals 7, 39 and 40 also assisted the CJEU in defining “investment” as the deployment of human, financial or technical resources, where the quantitative assessment refers to quantifiable resources and the qualitative assessment to efforts which cannot be quantified, such as intellectual effort or energy.102 This led the CJEU to admit that even if the development of the database was connected to the exercise of a principal activity in which the database maker also created the material, the sui generis protection could still be claimed if the maker proved that obtaining, verifying or presenting the content required an additional, independent substantial investment.103 The CJEU used an almost identical reasoning in British Horseracing Board (BHB) Ltd v William Hill,104 another landmark case also decided on November 9, 2004. The case, however, is also important for the guidance it offered on the notion of “substantial part” and “insubstantial part” of the contents of a database, indispensable to define the subject matter and scope of the sui generis right. II.IV.C The Notion of “Substantial Part” and “Insubstantial Part” BHB concerned the use of a database containing a large amount of information on pedigrees of horses and prerace information from the UK. The defendant, William Hill, operated an online betting service that offered to its clients also information taken from the BHB’s feed. Despite the amount of data used represented only a very small proportion of the BHB’s database and it was arranged differently on William Hill’s website, BHB still sued the latter for infringement of their sui generis right. The BHB decision focused on the notion of “substantial part” since the referring court asked whether Article 7 Database could still apply where the systematic or methodical arrangement and the condition of individual accessibility of the materials extracted from the database had been altered by the person carrying out the extraction and/or reutilization.105 To answer, the CJEU used again the teleological approach. From Recital 32, which identifies among the main

98 99 100 101 102 103 104 105

ibid para 35. Similarly, see Fixture/OPAP, para 39, and Fixtures/Svenska Spel, para 25. Fixtures/Oy Veikkaus, para 39, and Fixtures/Svenska Spel, para 26. ibid para 35. ibid para 44. Similarly, see Fixture/OPAP, para 40, and Fixtures/Svenska Spel, para 27. Fixtures/Oy Veikkaus, para 38. Similarly, see Fixture/OPAP, paras 41–42, and Fixtures/Svenska Spel, para 28. Fixtures/Oy Veikkaus, paras 39–40. Similarly, see Fixture/OPAP, paras 45–46, and Fixtures/Svenska Spel, para 29. Supra n 72. BHB, para 68.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

42

Caterina Sganga

aims of the Directive that of preventing that a user, “through his acts, causes significant detriment, evaluated qualitatively or quantitatively, to the investment,” the Court derived the need to assess the substantiality of the part by referring to the investment required for its creation, and to the prejudice caused to the investment by extracting or reutilizing that part.106 In this sense, a quantitatively substantial part refers to a volume of data that is substantial compared to the entire database, and which required substantial resources to be deployed.107 A qualitatively substantial part is a part that required significant human, technical or financial investment for that material to be obtained, verified and presented, regardless of its size. In both cases, the intrinsic value of the materials affected, their importance for the database maker and the resources eventually used for their creation do not matter for the assessment of substantiality.108 Against this background, it appears evident that no change made by the person making the extraction and reutilization to arrangement and accessibility of the data may have any effect on the substantial or insubstantial nature of the part extracted and reutilized.109 Parallel to this, the CJEU clarified also the scope of the prohibition laid down by Article 7(5) Database. Looking at the goal of this safeguard clause,110 which is to prevent the circumvention of the sui generis right by acts which are not singularly relevant, but cumulatively may seriously prejudice the database maker’s investment,111 the CJEU deemed not relevant whether the acts were carried out to create of another database or in the exercise of other activities, since what mattered was only the impact on the maker’s economic interests.112 Similar arguments were raised in 2009 in Apis-Hristovich v Lacorda,113 where the contested act was the alleged extraction and reutilization by Lacorda of 82.5 percent of Apis-Hristovich’s database of legal materials, allegation denied by Lacorda, which argued that the content, albeit similar, was taken from other sources – mostly publicly accessible – and the database was differently and originally organized and structured. Remarkably, the CJEU admitted that the nonaccessibility to the public of the sources of certain materials may affect the assessment of whether there has been a substantial investment in obtaining them under Article 7(1) Database, but denied that this factor was enough to exclude the presence of an infringement.114 Similarly, the Court excluded that the nonprotectability under copyright of some of the database content exempted national courts from verifying the presence of the requirements for protection set by Article 7 Database and its eventual violation.115 However, the area where the CJEU was the most prolific – and understandably, considering the importance of such notions for the interpretation of the scope Article 7 Database, was the definition of the notions of “extraction” and “reutilization.” II.IV.D The Scope of “Extraction” and “Reutilization” The Court was called for the first time to rule on the matter in BHB, where it had to establish whether the protection offered by Article 7 Database also covered the use of data which, 106

ibid para 69. ibid para 70. 108 ibid paras 71–72, 78–79. 109 ibid para 81. 110 Common Position (n 24)), point 14. 111 BHB, para 86. 112 ibid para 87. 113 Case C-545/07 Apis-Hristovich EOOD v Lakorda AD (2009) ECR I-1627. 114 ibid paras 66–68. 115 ibid paras 69–70. 107

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

Sui Generis Protection of Non-creative Databases

43

although originally derived from a protected database, were obtained by the user from other sources. The question was dense of implications, since it aimed at understanding how far the exclusivity conferred by the sui generis right on data could go, particularly regarding indirect conducts. To provide a balanced interpretation, the CJEU referred once again to the objective of investment protection pursued by Article 7 Database,116 and highlighted that Recital 42 also specifies that the sui generis right “relates not only to the manufacture of a parasitical competing product but also to any user who, through his acts, causes significant detriment, evaluated qualitatively or quantitatively, to the investment.” This proves in the opinion of the Court, the irrelevance of the purpose underlying the extraction or reutilization of the database content.117 Reading these statements together with Article 7(2)(a) and (b), the CJEU concluded that the EU legislator intended to give to extraction and reutilization a very wide definition,118 as suggested by the use of expressions such as “by any means or in any form” and “any form of making available to the public,”119 which indicated that the two terms should be interpreted as referring “to any act of appropriating and making available to the public, without the consent of the maker of the database, the results of his investment, thus depriving him of revenue which should have enabled him to redeem the cost of the investment.”120 Against this background, the Court had no doubt in stating that Article 7(2) should also cover indirect conducts, since “acts of unauthorized extraction and/or reutilization by a third party from a source other than the database concerned are liable, just as much as such acts carried out directly from that database are, to prejudice the investment of the maker of the database.”121 The extension is not without limit, though. The CJEU excluded, in fact, that the sui generis protection could cover the consultation of a database,122 and underlined that where the database maker authorizes a third party to reutilize the content, they also consent that the database is made accessible to the public.123 This does not imply, however, that a lawful user who is authorized to consult the database may extract or reutilize its content, since the sui generis right does not get exhausted on the basis of the maker’s consent to consultation (see Recital 43).124 In other words, no matter if the maker has made the content of the database available to the public as a whole or in part, or authorized a third party to do so, any act of extraction (that is the transfer of content to another medium) and act of reutilization (that is the making available of the database to the public) require the authorization of the rightholder.125 Four years later, the Court returned to the same points in Directmedia Publishing GmbH v Albert-Ludwigs-Universitat Freiburg.126 Directmedia revolved around a collection of German poems (XVII–XIX centuries.) compiled by Mr. Knoop, arranged according to citation frequency and supplemented by bibliographic compilation, and allegedly copied by Directmedia, which marketed a CD-ROM containing 1,000 poems, 856 being also in Mr. Knoop’s list, which 116 117 118 119 120 121

122 123 124 125 126

BHB, paras 45–46, with reference to Recitals 42 and 48. ibid paras 47–48. ibid para 49. ibid para 51. ibid. ibid paras 53–54. For the CJEU, this conclusion is also backed by Article 7(2)(b), which excludes the application of exhaustion on the right to control extraction and reutilization of the contents. ibid para 54. ibid paras 56–57. ibid paras 58–59, also on the basis of Recital 44. ibid para 61. Case C-304/07 Directmedia Publishing GmbH v Albert-Ludwigs-Universitat Freiburg [2008] ECR I-7565.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

44

Caterina Sganga

Directmedia used as a guide although it took the texts from own digital resources. More explicitly than in BHB, the referring court asked whether the concept of extraction covered also the transfer or elements from one database to another upon visual consultation of the former and a selection based on the personal assessment of the person operating the move.127 At stake there was the need, according to Directmedia, to narrow down the interpretation of the sui generis right to cover only the physical transfer of all or of part of the database to another medium, but not its use as a source of consultation, information and critical inquiry.128 The CJEU grounded its answer on a literal, contextual and teleological interpretation of the Directive. It reiterated that Article 7 requires offering to the concept of extraction a wide definition and that, in order to protect the database maker from any act of freeriding by a user or a competitor,129 the concept of extraction could not be made dependent on the nature and form of the mode of operation used.130 This entails that the decisive criterion is the presence of a “transfer” of all or part of the database to another medium of whatever nature. It is immaterial, instead, the way how the transfer is made – manually or technically131 – whether the amount transferred is insubstantial,132 or whether the transfer leads to an arrangement that is different from the original database, since the Directive is clear in considering any unauthorized act of copying as a threat against the database maker’s interests.133 Similarly, no relevance should be given to the objective pursued by the act, its competitive or noncompetitive nature, or whether or not the act is part of an activity other than the creation of a database.134 Directmedia tried to oppose such a broad interpretation by arguing that a too-wide definition of the scope of the sui generis right would have established a pure ownership over information, promoting informational monopolies and infringing users’ right to free access to information.135 The CJEU, however, rejected the claim, arguing that users’ access rights were ensured by the fact that consultation fell outside the scope of Article 7,136 and that the Directive was sensitive to competition law concerns, as showed by Recital 46, which leaves unprejudiced EU and national competition rules,137 and by Article 16(3) Database, which calls the Commission to periodically report on the interferences between the sui generis right and free competition.138 Apis-Hristovich added to this framework the interpretation of the concepts of “permanent transfer” and “temporary transfer” under Article 7(2)(a),139 emphasizing that the distinction, lying in the duration of storage of the materials on another medium, is relevant only to assess the gravity of the infringement and thus the damages to be compensated.140

127 128 129 130 131

132 133 134 135 136 137 138 139 140

ibid para 22. ibid para 19. Directmedia, para 33, recalling Fixture/OPAP, para 35 and BHB, paras 32, 45, 46, 51, 67 Directmedia, para 35. ibid paras 36–38, backed also by the reference to Recital 14, which recalls the protection offered to nonelectronic databases, and Recital 21, which highlights that the protection afforded by the Directive does not require the materials contained in the database to “have been physically stored in an organised manner.” ibid para 43, in line with BHB, para 50. Directmedia, para 39, with reference to Recital 38. ibid para 47, in line with BHB, paras 47–48. ibid para 50. ibid para 52. ibid para 56. ibid para 57. ibid para 42. ibid paras 43–44.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

Sui Generis Protection of Non-creative Databases

45

One had to wait until 2013 to get the first answer of the CJEU on the much more challenging question of the applicability of Article 7 Database to the activities of metasearch engines. In Innoweb BV v Wegener ICT Media BV et al. 141 the plaintiff, Innoweb, ran a metasearch engine on car sales (GasPedaal), which used search engines from other websites to answer the queries of its users. The results were merged into one document with links to all the original sources. Every day, GasPedaal performed approximately 100,000 searches on Wegener’s AutoTrack website, which corresponded to approximately 80 percent of its collection. Each query, however, triggered the showing of only a small party of the AutoTrack’s content, always determined by the user through his search. On this basis, Wegener sued Innoweb for violation of its sui generis right and succeeded before the first instance court. Innoweb appealed, and the Regional Court of Appeal of the Hague decided to refer the case to the CJEU to receive clarification on whether the indirect activities of a metasearch engine met the requirements set by Article 7(1), (2) and (5) to have an infringement of the sui generis right.142 The question gave the opportunity to the Court to offer a more detailed interpretation of the concept of reutilization. First, the CJEU underlined the need to exclude from the definition the substantiality of the part reutilized, and to give to the concept a broad interpretation.143 Then, it recalled the objectives of the sui generis right to reiterate the broad definition of reutilization offered by Directmedia144 and, to determine the applicability of the provision to metasearch engines, it looked at the purpose of their activities and their effects. Metasearch engines provide access to the entire contents of other databases by means other than those intended by their makers. Since end users no long have the need to go to the original databases website, database makers are likely to lose revenues which they need to redeem the cost of the investment in building and operating the databases. This risk is not excluded if the result page hyperlinked to the original database webpage to have access to the database content, since the potential negative impact on the website traffic and advertisement inflow remains.145 And even if Article 7 Database does not cover consultations of freely accessible databases,146 the activity of a metasearch engine cannot be assimilated to a consultation, since the engine only provided indirect access to external databases to users who could have had access to and consult the same databased directly from the respective websites. In this sense, such activities “come close “to the manufacture of a parasitical competing product as referred to in Recital 42 [. . .] albeit without copying the information stored in the database concerned” and, “in view of the search options offered,” the metasearch engine “resembles a database, but without having any data itself.”147 For the CJEU, thus, metasearch engines perform acts of making available of the contents of other databased for the purpose of Article 7(2)(b), thus engaging in an unauthorized reutilization of the whole or substantial part of their contents, for they generally perform mirrored search on the entire databases they “scrap.”148

141 142 143

144 145 146 147 148

Case C-202/12 Innoweb BV v Wegener ICT Media BV et al. [2013] EU:C:2013:850. ibid paras 8–18. ibid paras 33–34, in line with BHB, para 51, and Directmedia, para 32, and as reiterated by case C-173/11 Football Dataco Ltd and others v Sportradar GmbH and others [2012] EU:C:2012:642, para 20. Innoweb, paras 37–38, as in BHB, para 67; Apis-Hristovich, para 49; Football Dataco, para 20. Innoweb, para 45. ibid para 46, as in BHB, para 53. ibid para 48. ibid paras 50 and 53.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

46

Caterina Sganga

II.IV.E Private Autonomy and the Balance of Countervailing Interests: The Ryanair Decision The most recent decision concerning the Database Directive, Ryanair v PR Aviation (2015),149 also revolved around the activities of a metasearch engine – PR Aviation – that operated a website on which consumers could search through the flight data of low-cost airlines, compare prices and book a flight on payment of a commission. The website gave access also to Ryanair dataset and booking system, despite access to the Ryanair website requested the application of its general terms and conditions, among which an exclusive distribution clause – which specified that only Ryanair.com was authorized to sell Ryanair flights – and a “permitted use” clause, which explicitly forbade using the website other than for limited private noncommercial purposes. Screen-scraping activities like the one performed by PR Aviation were prohibited unless the third party had directly concluded a written license agreement with Ryanair, and limitedly to the sole purpose of price comparison.150 The question to the Court was based on the premise that the Ryanair dataset could be classified as a database under Article 1(2) Database, but was not protected by copyright on the basis of Article 3 Database and/or of the sui generis right on the basis of Article 7 Database.151 It aimed at understanding whether the mandatory balance set by Articles 6(1), 8 and 15 of the Directive, winning over freedom of contract, should apply to any collection of works featuring the characteristics indicated by the general definition provided by the EU text, regardless of whether that the collection itself met the requirements to be actually protected by any of the two exclusive rights introduced by the Database Directive. The response of the CJEU was fully to the negative. The definition of Article 1(2) applies “for the purposes of this Directive,” which is “the legal protection of databases” by means of copyright and sui generis rights. In this sense, the fact that a database corresponds to the definition set out in Article 1(2) of Directive 96/9 does not justify the conclusion that it falls within the scope of the provisions of that directive governing copyright and/or the sui generis right if it fails to satisfy either the condition of application for protection by copyright laid down in Article 3(1) of that directive or the conditions of application for the protection by the sui generis right in Article 7(1) thereof.152

Article 6(1) Database could therefore apply only to databases protected by copyright; Article 8 only to databases protected by the sui generis right; and Article 15, which affirms the mandatory nature of certain provisions of the Directive by declaring null and void any contractual provision contrary to it, applies only when Articles 6 or 8 apply, thus it does not prevent the adoption of contractual clauses concerning the conditions of use of databases not protected by copyright or sui generis right under the Directive.153 In the opinion of the CJEU, this interpretation was in line with the general scheme of the Directive and the balance it sets out between the rights of database makers and the rights of lawful users. And it could not be argued that this would reduce the interest in claiming the protection instituted by EU law, since database makers would have more contractual freedom if operating outside its scope.154 In fact, the Database Directive offers automatic protection 149 150 151 152 153 154

Case C-30/14, Ryanair Ltd v PR Aviation BV [2015] EU:C:2015:10. ibid paras 15–16. ibid para 29. ibid para 35. ibid paras 36–39. ibid paras 40–41.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

Sui Generis Protection of Non-creative Databases

47

through two exclusive rights, with no administrative formalities nor any prior contractual arrangement needed,155 and limits the database maker’s freedom of contract only in circumscribed lawful uses to balance the broad exclusivity granted to rightholders – exclusivity which authors of nonprotected databases may claim only under national law or under the much weaker tool of contractual provisions.156 II.V THE IMPLEMENTATION OF THE DIRECTIVE AND ITS EFFECTS: THE 2005 AND 2018 EVALUATIONS AND THE WAY FORWARD

Already before its enactment, and widely after its adoption, the sui generis right was subject to strong critiques. Article 7 was challenged for being unclear in scope, not fit to stimulate innovation and growth, and triggering the risk of a data and information lockup, to the detriment of the scientific community and of competitors and other industries relying on the availability and free flow of data and information to conduct their business or research.157 To answer to such objections, the Commission conducted two assessment exercises. The first evaluation of the Directive was issued in 2005, with the aim to verify whether its policy goals had been achieved and whether the sui generis right had negative effects on competition.158 The Commission’s findings were remarkable, both from a legal and from economic perspective. The Evaluation provided some snapshots of the application of the new rules by national courts and authorities. Expectedly, while the notion of database found a uniform application, the definition of the sui generis right, due to its novelty, created a number of conflicting judgments, mostly on the definition of “substantial investment,” and on the treatment of spinoff databases and metasearch engines.159 Criticisms remained, particularly on the tilted balance between the interests of makers, lawful users and the general public, on the excessively broad and uncertain scope of the sui generis right, and on the too-narrow scope of exceptions.160 From an economic perspective, the evaluation concluded that the impact of the sui generis right on database production was unproven or rather negative, since the EU database production in 2004 fell back to preDirective levels,161 and the economic gap with the USA had not been reduced.162 The policy options opened at this stage were the repeal of the whole Directive, impossible considering internal market needs, and the withdrawal of the sui generis right to maintain only a copyright protection having a high originality threshold, following the successful US example.163

155 156 157

158

159 160 161

162

163

ibid para 42. ibid para 44. For a summary of the academic debate see Davison (n 4) 237 ff; and Derclaye (n 2) 271 ff; see also ReichmanSamuelson (n 25) 80–90. For an overview of the academic criticisms considered by the Commission in their last evaluation and of the complaints moved by stakeholders, see ‘Study in Support of the Evaluation of Directive 96/9/ EC on the Legal Protection of Databases. Annex 1 – Legal Analysis’ (2018) accessed 28 February 2021. European Commission, First Evaluation of Directive 96/9/EC on the legal protection of databases (DG Internal Market and Services Working Paper, Brussels, December 12, 2005). ibid 11–12, reporting the most significant case law from Member States. ibid 22–23, providing an overview of the complaints received by the Commission by stakeholders’ associations. ibid 5 and 20, reporting the results of the Gale Directory of Databases (GDD), and the responses of online surveys conducted by the Commission, which showed that 90 percent of interviewed database makers felt the sui generis right essential for their business. As reported in the GDD (ibid 22), showing a decrease in the share of global database production for Western Europe and an increase in the North American share. ibid 26.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

48

Caterina Sganga

Yet, since no significant administrative or other regulatory burdens on the database and other industries had been detected, no intervention was deemed necessary for the time being.164 The second Evaluation dates 2018. It was introduced by the Communication Building a European Data Economy in 2017 as necessary to determine the fitness of the regulatory framework with the needs of global data markets and industries working with machine-generated data and artificial intelligence agents.165 The Commission grounded the Evaluation on an external study which comprised online surveys, workshops with stakeholders and an in-depth comparative legal analysis of the state of the implementation across Member States, looking at data within the timeframe 2005–18.166 The Evaluation observed that the implementation and acceptance of the sui generis right had kept contentious, and the right itself remained a low-profile legal instrument generating limited interest among stakeholders.167 In addition, from 2005 the economic and technological use and value of data had witnessed an impressive shift, increasing the number of datasets that may be considered databases despite the narrowing down of the scope of Article 7 by the CJEU.168 The study highlighted a neat distinction between copyright and the sui generis right.169 National reports showed little practical interest and litigation on copyright database protection, with only a few national cases revolving around the definition of the concept of “author’s own intellectual creation” and “creative choice,” generally aligned to the CJEU guidelines in Football Dataco.170 The situation is largely different for the sui generis right.171National courts have been split on the interpretation of key concepts such as those of substantial investment, showing disagreement on the minimum threshold,172 with several Member States taking a permissive approach as long as the investment was not trivial.173 The same can be said for the notion of “substantial part,” with cases ranging from 20 percent to 50 percent of the database, and for the classification of indirect conducts such as those of metasearch engines (e.g., webscraping and the like).174 The legal analysis highlighted divergences also in the conceptualization of the database maker,175 while there seems to be a relative alignment to the CJEU case law on the general definition of extraction and reutilization.176

164 165 166

167

168 169 170

171 172

173

174 175 176

ibid. European Commission, Communication ‘Building a European Data Economy’ COM (2017) 9 final, 10.1.2017. See Study in support of the evaluation of Directive 96/9/EC on the legal protection of databases – Final report (2018) accessed 28 February 2021. The report is complemented by a legal analysis (n 157), an economic analysis and the synopsis of stakeholders online surveys, interviews and workshops accessed 28 February 2021. See the results of the public consultation ran by the Commission and the low level of turnout, as reported also in the Commission Staff Working Document ‘Evaluation of Directive 96/9/EC on the Legal Protection of Databases’ SWD (2018) 146 final, 12. ibid 13. ibid esp n 48. ibid 14, with reference to Bundsgerichtshof August 13, 2009 — I ZR 130/04 (on 1,000 poems that everyone should read) and Cour de Cassation (civ I) May 13, 2014, RIDA 2015, No 244. This is widely explained in Annex 1 – Legal Analysis (n 157) 45, esp para 6.1. ibid para 8.1, and the cases cited therein. All cases are also reported in Annex 6 – Country grids, available at accessed 28 February 2021. This has been called a ‘de minimis exclusion rule’ by Matthias Leistner, ‘Big Data and the EU Database Directive 96/ 9/EC: Current Law and Potential for Reform’ in Sebastian Lohsse, Reiner Schulze and Dirk Staudenmayer (eds) Trading Data in the Digital Economy: Legal Concepts and Tools (Nomos 2017) 30. Legal analysis (n 157) paras 8.2.2. and 12.2.2 with reference to infringement. Evaluation (n 167), para 6.1.5, with reference to French and Dutch cases. As confirmed by the limited number of cases cited in the Legal analysis (n 157) 68.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

Sui Generis Protection of Non-creative Databases

49

Despite such results, the Commission concluded that “engaging in a process of limited reform of the sui generis right would be, at this stage, largely disproportionate,” and would first “need to be substantial, and build a stronger case, considering the policy debates around the data economy.”177 This conclusion, however, did not close the debate on the shortcomings of the Database Directive. Aside from long-standing theoretical objections to the privatization of data and information, several academic and policy contributions have advanced very specific reform proposals to address the pitfalls in the Directive and in its national implementations. It has been maintained, for instance, that the concept of maker should be clarified, and identified in the person responsible for substantial investment, or in the “producer,” with a clearer split between “makership” and “authorship.”178 Some commentators have suggested to eliminate the notion of “substantiality” to evaluate the investment required for the sui generis protection, considering the uncertainties surrounding its definition and application. To counterbalance the move, they have proposed the introduction of a more stringent regime, featuring narrower rights, a significant substantiality threshold to assess infringements and more exceptions.179 Others have flagged the risk that this would go to the detriment of users, who are less deep pocket and more risk-adverse, but would be put in charge of keeping the floodgate of nonoriginal database protection shut. As an alternative, these contributions suggest removing the uncertainty associated with the substantiality criterion by standardization mechanisms such as a system of registration, annual meeting of EU and national judges, the creation of a database of national decisions and so forth.180 Similar criticisms on its uncertainty and proposals of reform concern the definition of the subject matter of the investment necessary to obtain sui generis protection, and the notion of substantial part and of repeated and systematic use of insubstantial parts to draw the boundaries of infringement. As to the former, both commentators and stakeholders have underlined the difficulties in distinguishing between obtaining and creating data and other content, especially in the context of the new (collaborative) data economy and of the Internet of Things.181 To tackle such problems, they have proposed either to extend the sui generis right to cover also investments in the creation of database materials, or to introduce other exclusive rights over data if needed, separating them more clearly from the sui generis right. In order to counterbalance the increased risks of informational monopolies, they also envisioned the provision of ad hoc balancing tools, such as compulsory licensing schemes or tailored exclusions of specific categories of data.182 As to the latter, critics refer particularly to the conflicting output of national case laws and to the doubts surrounding the treatment of indirect acts of extraction and reutilization, both having substantial chilling effects on the activities of lawful users and competitors, especially after the Innoweb decision. While license agreements remain a valid alternative to more paternalistic regulatory solutions, they may prove ineffective in case of unbalances in bargaining power and polarization and centralization of data ownership – situations where compulsory licensing schemes would still be the most effective solution to solve market failures.183 177 178 179 180 181 182 183

Evaluation (n 167) 47. See, eg Koščík and Myška (n 41) 53; Beunen (n 14) 154. See the literature overview provided by Derclaye (n 45) 62 ff. As by Lionel Bently and Estelle Derclaye, main authors of the Legal Analysis (n 157) 57. The reference goes, for example, to sensor-generated data, meteorological data, etc. See Evaluation (n 167) 35–37. See, inter alia, Westkamp (n 70) 7; Hugenholtz (n 14) 89. As this is generally the case in intellectual property law. In copyright, see Peter Menell, ‘A Remix Compulsory Licensing Regime for Music Mashups’ in Michelle Bogre and Nancy Wolff (eds) The Routledge Companion to Copyright and Creativity in the 21st Century (Routledge 2020).

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

50

Caterina Sganga

The most problematic aspect of the sui generis protection, however, remains the regulation of exceptions. The list provided by Article 9 is criticized for its exhaustiveness, optional nature, scarce or no coordination with general copyright exceptions under the InfoSoc Directive, and incapability of accommodating basic countervailing interests, especially with regard to the reuse of data.184 While the introduction of the TDM exceptions have partially tackled this issue, the very limited room left to limitations in the field of sui generis right makes the latter a much more absolute entitlement than copyright is, leading rightholders to privilege Article 7 as the main tool to protect their interests. Against this background, the EU legislator had plenty of hints to assess reflect on the impact of the Database Directive on the EU policies on the data economy, and how such hints were univocally pointing to specific directions. Still, the most recent preparatory works and reforms acts and proposals limited their interventions on Directive 96/9/EC on very sectorial aspects, without really taking care of the coordination between the EU Data Package and EU database law. II.VI THE DATABASE DIRECTIVE IN THE CONTEXT OF THE EUROPEAN STRATEGY FOR DATA: A MISSED REVOLUTION

Early in 2017, in the Communication “Building a European data economy”185 the Commission highlighted the need to intervene on the EU legislation in order to develop an ecosystem that could facilitate the cooperation between market actors, users and public entity in making data accessible and reusable, and thus in allowing the extraction of their value and the development of applications having great economic, technological and social potential. To this end, and in line with the GDPR,186 the Communication underlined the need to create a clear policy and legal framework for the data economy, eliminating obstacles to the circulation of data and facing legal uncertainties created by new technologies. At the same time, it moved upfront the goal of launching consultations, studies, and impact assessments to achieve the free circulation of data, the portability of non-personal data, the interoperability of data and infrastructures, and a clearer regulation of ownership, access and transfer of machine-generated data, also by rebalancing the contractual power of SMEs, start-ups and big companies and reducing lock-in effects for consumers.187 Regulatory solutions envisaged by the text ranged from the full openness of data to the creation of a data producers’ rights, flanked by compromise options such as the introduction of compulsory licenses, contractual standards, boilerplate fairness and transparency clauses and exceptions subordinated to the payment of fair compensation.188 The 2017 Communication represented the foreground for the introduction of the Regulation on the free flow of non-personal data,189 and the 2019 Open Data Directive, which modified the 184 185

186

187 188 189

See Evaluation (n 167) 30. European Commission, Communication “Building a European data economy”, COM(2017) 09 final, preceded by the Communication “Towards a thriving data-driven economy”, COM(2014) 442 final, which already recognized the need to introduce modern and coherent EU-wide norms to allow the free flow of data across borders, since “the complexity of the current legal environment together with the insufficient access to large datasets and enabling infrastructure create entry barriers to SMEs and stifle innovation” (ibid at 3). Regulation (EU) 2016/679 of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation), OJ L-119/1 of 4 May 2016. Building a European data economy, p.11. See Herbert Zech, ‘Building a European data economy’, in IIC, 48(5) (2017) pp.501 ss. Regulation (EU) 2018/1807 of 14 November 2018 on a framework for the free flow of non-personal data in the European Union, OJ L-303/59 of 29 November 2018.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

Sui Generis Protection of Non-creative Databases

51

2013 PSI Directive and broadened the range of data from public bodies that are subject to the obligation of being made available for commercial and non-commercial reuse. Together with other important step forwards in the definition of standard procedures, contracts and technical features, the ODD made it practically impossible for public entities to exercise any exclusive right on data they detain, including rights under Article 7 Database, yet without excluding their ownership.190 In 2020, another Communication (“A European strategy for data”)191 set the goal to “The aim is to create a single European data space – a genuine single market for data, open to data from across the world – where personal as well as non-personal data, including sensitive business data, are secure and businesses also have easy access to an almost infinite amount of high-quality industrial data, boosting growth and creating value, while minimising the human carbon and environmental footprint.”192 To this end, the Commission planned four lines of action, the first and key one being a set of regulatory interventions to create a data governance framework that (i) ensures the functioning of “common European data spaces” based on principles of findability, accessibility, interoperability and reuse (FAIR) and on data altruism; (ii) identifies high value datasets to be made publicly available for free, in machine-readable form via standardized APIs, in light of their potential for innovation and PMIs; (iii) incentivized B2B and B2G data sharing, by intervening on datra ownership and existing IP rights – particularly database rights – in order to increase data adccdess and reuse.193 The first intervention, the Data Governance Act (DGA), extended the principles of the ODD to a broader range of public body information, covered by IP rights, commercial or statistic confidentiality or other limitations due to personal data protection.194 Yet, the Commission did not intervene on the Database Directive, which still theoretically admits the possibility for public sector bodies to hold rights over data obtained, verified and organized by public investments, albeit without being able to exercise them.195 A few months later, the Parliament Resolution on a European strategy for data196 requested the Commission to introduce as early as possible a draft Data Act to allow a broader and more equitable flow of data in every sector, creating EU data spaces that facilitate cross-sectorial and cross-border data exchange between industry, academia, relevant stakeholders and the public sector.197 To this end, the Parliament suggested to clarify data ownership and access regimes, thus tackling market unbalances caused by the concentration of data control in the hands of a few players, to the detriment of SMEs.198 The proposed Data Act (DA), issued in February 2022,199 has been conceived as a horizontal instrument having five goals, which are (i) to facilitate access and reuse of data by consumers and market players, (ii) to allow public bodies 190

191

192 193 194

195 196

197 198 199

Similarly in Mireille Van Eechoud, ‘A Serpent Eating Its Tail: the Database Directive Meets the Open Data Directive’ in IIC, 52 (2021) pp.375 ss. European Commission, Communication “A European strategy for data”, COM (2020) 66 final. On the change of paradigm on the conceptualization of data ownership, see Maria Lilla Montagnani, ‘Dati e proprietà intellettuale in Europa: dalla “proprietà” all’”accesso”’, in Diritto dell’economia 66(1) (2020) pp.539 ss. Ibid at 5. Ibid at 17. Regulation EU 2022/868 of 30 May 2022 on European Data Goverrnance and amending Regulation EU 2018/1724 (Data Governance Act), L 152/1 of 3 June 2022. Also in Van Eechoud (n . . .) at 377. Resolution of the European Parliament of 25 March 2021 of a European strategy for data (2020/2217(INI)) (2021/C 494/ 04). Ibid at 3. Ibid at 4. Proposal of a Regulation on harmonised rules on fair access to and use of data (Data Act), COM (2022) 68 final (DA).

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

52

Caterina Sganga

to use, in exceptional circumsntaces, data detained by companies and platforms; (iii) to make it easier to switch between cloud and edge services; (iv) to provide safeguards against illegal data transfers towards third countries’ governments; and (v) to impose interoperability standards for data. In this context, and among several other measures, the Commission decided to intervene on the Database Directive. However, the reform it proposed has a much more limited scope than what the two impact assessments suggested as necessary to correct existing flaws.200 Article 35 of the proposed Data Act, in fact, introduces only a mere clarification of the subject matter covered by Article 7 Database, excluding databases whose data have been generated or obtained from the use of a product or service. The goal is to allow users to use and share their data with third parties, that is to exercise the rights granted to them by Articles 4-5 DA. Recital 84 only specifies that the provision aims at avoiding the risk that the holder of data obtained or generated from the physical components of an IoT product or service claims a sui generis right under Article 7 Database, thus frustrating users’ prerogatives under the Data Act. In this sense – the Preamble continues – Article 35 DA does not introduce any new rule, but only clarifies that Article 7 Database does not apply in such cases “since the requirements for protection are not fulfilled”. Although the goal pursued by the EU legislator was limited to excluding producers’ control over data generated by IoT devices and allowing the application of Articles 4-5 DA on data stemming from the use of such products/services, the regulatory options available were still many. Some of them would have allowed intervening on general rules and definitions, such as the notion of substantial investment and its subject matter, or the notion of substantial part in assessing the presence of a violation.201 The EU legislator, however, decided to adopt the simplest solution, de facto and indirectly modifying the scope of Article 7 Database without taking the opportunity for a conceptual reordering of EU database law,202 and without excluding that Member States can regulate differently matters outside the scope of the Database Directive (as data stemming from the use IoT devices are now).203 At the same time, the proposed DA does not introduce any remedy against the contractual or TMPs overridding of the protection it provides in its Article 4,204 and does not put forward key reforms advocated for by the 2018 Impact Assessment, such as the exclusion of public bodies from the range of potential rightholders, the introduction of a research exception going beyond TDM purposes, and the streamlining of copyright and sui generis right exceptions.205 If approved as it stands today, the Data Act will leave unsolved most of the flaws of the Database Directive highlighted in the past years by scholars and stakeholders, and possibly create further problems, despite all the good intentions showed by the European Strategy for Data.

200

201 202

203

204 205

Study to Support an Impact Assessment for the Review of the Database Directive – Final Report, January 2022, available at https://ec.europa.eu/newsroom/dae/redirection/document/83514 (last accessed 23 December 2022). Ibid, at 3-4. In fact, Article 35 DA is based on the wrong assumption that data produced by IoT devices do not meet the requirements of protection of Article 7 Database, while this is not the case, for instance, for investments directed to install sensors that are able to collect data on the use of a specific device. At the same time, the definition offered by Article 35 DA to “data generated by the use of a product connected to a service” is much broader than the subject matter of IoT devices, for it may easily cover also data transmitted by a physical object through a communication service, such as in the case of use of a fidelity card in a shop. As in Husovec-Derclaye, p.11. The Impact Assessment explicitly suggested to intervene on the matter (Impact Assessment, supra, note XXX, pp.56 ss.). Ibid at 57. Ibid at 52 ff et seq.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

Sui Generis Protection of Non-creative Databases

53

II.VII CONCLUSIONS

At the time of the discussions leading to its introduction in 1996, it was already clear that the sui generis right would have brought within the tangles of EU copyright law a “foreign” entitlement which, despite being investment-protection-driven as many other neighboring rights, had little to share with this category, and much more pervasive effects. From the very limited range (and closed list of) exceptions to the potentially perpetual term of protection and the very wide – and similarly very uncertain – scope of exclusivity, the sui generis right introduced under the umbrella of copyright law and exclusive right that “propertized” not only materials already protected by copyright or other rights, but also data and information traditionally belonging to the public domain for conscious and consolidate policy choices. A quarter of century after its enactment, the Database Directive has been subject to two rounds of evaluation by the European Commission, and its sui generis right has been articulated by the CJEU and national courts with a wide array of doctrines and specifications. Some of these interventions have brought more clarity, tempering the chilling effect that the legal uncertainty surrounding Article 7 Database had on the activities of users, researchers and market actors. Some others have only contributed, with their pitfalls, to highlight the shortcomings of the Directive, still without any impact on the legislative agenda of EU policymakers. With the increasingly more important role played by data flows and data sharing in the new data economy and for the development of artificial intelligence agents and machine-learning processes, the obstacles posed by Article 7 Database to EU data research, industry and market have become so prominent to force the legislator to turn back its attention to the sui generis right and correct its most evident distortions. Two new exceptions for text and data mining have been introduced by the Copyright Directive In 2019, to balance between exclusivity and access to database, but with an effective impact only on non-profit research activities. There were discussion on the opportunity to introduce a data producer’s right on machine-generated data,206 while the ODD has excluded the application of the sui generis right if owned by public entities, yet without intervening on its attribution. Later on, the European Strategy for Data and its first product, the DGA, have extended the ODD principles to a broader range of information held by public bodies but subject to third-party rights, confidentiality obligations or other limitations for data protection, introducing specific conditions for their reuse, but without coordinating the new provisions with the Database Directive, which is only blocked in its application without discussing its ultimate applicability to data obtained, verified and structured by public entities or funding.207 According to the Commission’s plan and on the basis of the last Impact Assessment, the proposed Data Act should have intervened more incisively on EU database law to correct its flaws and align it to the EU policy goals in the field of data. Unfortunately, the mountain roared and brought forth a mouse, for it missed to intervene on the most pressing shortcomings of the Database Directive and introduced, to a certain extent, further unclear elements that are prone to create additional interpretative problems in the future.

206 207

Building a European Data Economy, at 13. Cfr. M.van Eechoud, cit., p.377.

https://doi.org/10.1017/9781108989527.005 Published online by Cambridge University Press

III Test Data Exclusivity An Elusive Pursuit to Strike a Balance between Affordable Drugs and Investment Incentives Daria Kim

III.I INTRODUCTION

Test data exclusivity (TDE) is a regulatory mechanism of suspending price competition in the drug market rationalised on the grounds of protecting innovation incentives of originator companies. Under international trade law, it is considered a subset of intellectual property (IP).1 While its impact on innovation remains ambiguous, TDE has been criticised at length as an impediment to affordable medicine and the right to health,2 an obstacle to transparency in medical research,3 a case of ‘ratcheting up’ of international IP standards,4 and a threat to TRIPS flexibilities.5

Senior research fellow, the Max Planck Institute for Innovation and Competition (LLM, MA, Dr iur), daria.kim@ip. mpg.de. 1 Article 39 of the Agreement on Trade-Related Aspects of Intellectual Property Rights (hereinafter the TRIPS Agreement) mandates that WTO members shall protect test data submitted for the authorisation of the pharmaceutical or agricultural chemical products, ‘the origination of which involves a considerable effort’ against unfair commercial use and, in addition, against disclosure. 2 Srividhya Ragavan, ‘The (Re)newed Barrier to Access to Medication: Data exclusivity (2018) 51 Akron Law Review 1163; Srividhya Ragavan, ‘The Drug Debate: Data Exclusivity Is the New Way to Delay Generics’ (2018) 50(2) Connecticut Law Review Online 3; Carlos Correa, ‘Designing Patent Policies Suited to Developing Countries Needs’ (2008) 10 Econômica, Rio de Janeiro 82; Sean Baird, ‘Magic and Hope: Relaxing Trips-Plus Provisions to Promote Access to Affordable Pharmaceuticals’ (2013) 33(1) Boston College Journal of Law & Social Justice 107. 3 Transparency-based arguments concern not only TDE but trial sponsors’ control over test data more broadly. See Trudo Lemmens, ‘Leopards in the Temple: Restoring Scientific Integrity to the Commercialized Research Scene’ (2004) 32 Journal of Law, Medicine & Ethics 641; Trudo Lemmens and Candice Telfer, ‘Access to Information and the Right to Health: The Human Rights Case for Clinical Trials Transparency’ (2012) 38 American Journal of Law & Medicine 63; Trudo Lemmens, ‘Pharmaceutical Knowledge Governance: A Human Rights Perspective’ (2013) 41 Journal of Law, Medicine & Ethics 163. 4 See, eg, Susan Scafidi, ‘The “Good Old Days” of TRIPS: The U.S. Trade Agenda and the Extension of Pharmaceutical Test Data Protection’ (2004) 4 Yale Journal of Health Policy, Law & Ethics 341; Michael Handler and Bryan C Mercurio, ‘Intellectual Property’ in Simon Lester and Bryan C Mercurio (eds) Bilateral and Regional Trade Agreements: Commentary and Analysis (CUP 2009) 308, 341; Annette Kur and Henning Grosse Ruse-Khan, ‘Enough Is Enough: The Notion of Binding Ceilings in International Intellectual Property Protection’ (2009) Max Planck Institute for Intellectual Property, Competition & Tax Law Research Paper Series No 09-01, 13; footnote 39. 5 See, eg, Ellen FM ‘t Hoen, Pascale Boulet and Brook K Baker, ‘Data Exclusivity Exceptions and Compulsory Licensing to Promote Generic Medicines in the European Union: A Proposal for Greater Coherence in European Pharmaceutical Legislation’ (2017) 10 Journal of Pharmaceutical Policy and Practice 1. For a comprehensive analysis, see Owais H Shaikh, Access to Medicine Versus Test Data Exclusivity: Safeguarding Flexibilities Under International Law (Springer 2016).

54 https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

Test Data Exclusivity

55

In line with the overarching theme of this collection, this chapter examines to what extent TDE is aligned with the normative theory of IP. The answers to this question can vary depending on which normative theory of IP is taken as a baseline. Conventionally, TDE has been justified based on the need to compensate the originator companies for the high costs of drug development. As the industry never ceases beating the drum of high costs of research and development (R&D),6 it has become almost commonplace to equate ex post compensation with ex ante innovation incentives. When introducing TDE, the legislators in the USA7 and the EU8 sought to reach a compromise between the policy goal of protecting incentives of drug innovators and that of facilitating generic entry. Where exclusivity becomes a policy instrument of choice, a trade-off between social benefits is inevitable. In the case of TDE, a trade-off occurs precisely between two policy objectives at stake: the protection of returns on investment – conventionally associated with innovation incentives – comes at the expense of the affordability of medicinal products. Furthermore, when complying with the TRIPS standard or implementing obligations under the free trade agreements, lawmakers need to decide on the modalities9 of TDE that jointly determine the strength, length, and scope of protection. The configuration of such modalities might either mitigate or exacerbate the policy trade-off. The vision that ‘legislation should provide for an optimal balance between innovative medicinal products and generic medicines’10 is sensible and well intended. However, the questions of how exactly such balance should be moulded into the legal provisions, to what extent affordability should be ‘sacrificed’, and how one might know that certain rules provide for an optimal cost–benefit ratio remain open-ended. Which term of protection is more optimal: five years as in the USA,11 or ‘8+2+1’ years as in the EU?12 Do residents of the EU enjoy greater benefits of drug innovation? This chapter approaches the question of justification of TDE from a welfare perspective and examines analytical issues of translating the idea of balancing innovators’ returns and generic entry into the legal norms. A mere statement that a legal rule ‘pursues’ or ‘reflects’ a socially valuable objective does not automatically justify it – a mere reference to policy goals might simply ‘conceal the legitimation of a rule behind a façade of “explanation”’.13 To state that TDE protects economic 6

7 8

9

10 11

12 13

International Federation of Pharmaceutical Manufacturers & Associations (IFPMA), ‘Data Exclusivity: Encouraging Development of New Medicines’ (2011) ; European Federation of Pharmaceutical Industries and Associations (EFPIA), ‘The Pharmaceutical Industry in Figures Key Data’ (2020) ; Pharmaceutical Research and Manufacturers of America (PhRMA), ‘2020 PhRMA Annual Membership Survey’ (accessed 21 Feb. 2021). Drug Price Competition and Patent Term Restoration Act of 1984, Pub L No 98-417 §1538, 98 Stat 1585 (1984). Recital 9 of Directive 2001/83/EC (OJ L 311, 28.11.2001) states: ‘Experience has shown that it is advisable to stipulate more precisely the cases in which the results of toxicological and pharmacological tests or clinical trials do not have to be provided with a view to obtaining authorization for a medicinal product which is essentially similar to an authorized product, while ensuring that innovative firms are not placed at a disadvantage.’ See also European Commission, ‘Report on the Basis of Article 71 of Regulation (EEC) No 2309/93’ (23.10.2001) COM(2001) 606 final 21. Such modalities include the duration of protection, whether protection is limited to new chemical entities, whether it is susceptible to exceptions and limitations, etc. European Commission (n 8) 21 (emphasised in the original). The USA provides for a five-year protection term only for pharmaceutical products that contain a new chemical entity. 21 CFR § 314.108(b)(2). Directive 2001/83/EC (n 8) art 10. Mark van Hoecke, ‘Legal Doctrine: Which Method(s) for What Kind of Discipline?’ in Mark van Hoecke (ed) Methodologies of Legal Research: Which Kind of Method for What Kind of Discipline? (Hart 2011) 8 (quoting Aleksander Peczenik, ‘Scientia Juris: Legal Doctrine as Knowledge of Law and as a Source of Law’ in Enrico Pattaro (ed) A Treatise of Legal Philosophy and General Jurisprudence, 4 (Springer 2005) 4).

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

56

Daria Kim

investments and, in so doing, supports drug innovation does not suffice. First, one needs to be confident in the underlying causal mechanism. Second, from a welfare-based perspective, not only should the benefits of protection outweigh its costs, but the ratio of social costs and benefits should be more favourable relative to alternative policy options. The analysis is structured as follows. Part III.II starts by outlining the standard justification for TDE that, on the surface, appears to be in line with the mainstream economic theory of exclusive IP rights and introduces general caveats that fall under the ‘calculability’ critique of IP14 (Section III.II. A). Sections III.II.B and III.II.C present a more nuanced view on how the incentive theory applies to clinical trials, which market ‘fails’ and which respective roles TDE and patents are meant to play in this regard. Evidence on welfare costs and benefits of TDE is discussed in Sections III.II.D and III.II. E. Section III.II.F highlights some conceptual pitfalls of trying to square up welfare costs and benefits of TDE and translating the idea of balancing into the legal modalities of TDE. Societal concerns regarding TDE beyond deadweight loss are discussed in Section III.II.G. Part III.III concludes by reinforcing the argument that the potency of TDE to effectively delay generic entry as such does not justify TDE as an optimal innovation instrument and that private returns on drug R&D should not be assumed to convert into social benefits in the form of drug innovation. This chapter does not discuss in detail the legal standard of test data protection under the TRIPS Agreement;15 nor does it analyse national provisions implementing it.16 The term ‘data exclusivity’ is used in a jurisdiction-neutral sense to refer to a regulatory instrument precluding the marketing approval of a generic product by way of reference to the originator’s test data. The analysis focuses on non-summary clinical trial data17 – data beyond the information that can be disclosed in scientific publications by trial investigators or trial registers upon the trial completion or published by drug authorities upon the drug approval. III.II TDE THROUGH A LENS OF INNOVATION INCENTIVES THEORY

III.II.A The Conventional Rationale and General Disclaimers From a law-and-economics perspective, TDE presents a means to encourage innovation by temporarily restraining generic competition.18 The delay of price competition is generally assumed to remedy a market failure of insufficient economic incentives arising due to the public-good nature of knowledge and uncertainty as a factor of innovative activity.19 The lack of 14

15

16 17

18

19

Robert P Merges, ‘Philosophical Foundations of IP Law: The Law and Economic Paradigm’ in Peter Menell and Ben Depoorter (eds) Research Handbook on the Economics of Intellectual Property Law, vol 1 (Edgar Elgar 2019) 81. This issue has been analysed extensively in the literature. See, eg, Carlos M Correa, ‘Unfair Competition under the TRIPs Agreement: Protection of Data Submitted for the Registration of Pharmaceuticals’ (2002) 3 Chicago Journal of International Law 69; Aaron Xavier Fellmeth, ‘Secrecy, Monopoly, and Access to Pharmaceuticals in International Trade Law: Protection of Marketing Approval Data Under the TRIPs Agreement’ (2004) 45(2) Harvard International Law Journal 443; Peter K Yu, ‘The Political Economy of Data Protection’ (2010) 84 Chicago-Kent Law Review 777; Nuno Pires de Carvalho, The TRIPS Regime of Patents and Test Data (5th edn, Kluwer Law International 2018). For an overview of national provisions on TDE, see Shaikh (n 5); IFPMA (n 6). In a broad sense, the term ‘clinical trial data’ comprises not only data gathered in a clinical trial but also documents that lay out how a trial is designed and how data is collected, including the trial protocol, case report forms, clinical study reports, etc. Individual patient-level data presents the most detailed and comprehensive records. In the EU, TDE was introduced under Directive 87/21/EEC (OJ L 15, 17.1.1987). The currently applicable framework is comprised of the provisions contained in Directive 2001/83/EC (OJ L 311, 28.11.2001) and Regulation (EC) No 726/ 2004 (OJ L 136, 30.4.2004). Kenneth J Arrow, ‘Economic Welfare and the Allocation of Resources for Invention’ in Universities-National Bureau Committee for Economic Research & Committee on Economic Growth of the Social Science Research Council (ed) The Rate and Direction of Inventive Activity: Economic and Social Factors (Princeton University Press 1962) 616.

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

Test Data Exclusivity

57

protection against imitation is posited to cause a trade-off between the benefits of price competition (static efficiency) and innovation incentives (dynamic efficiency).20 In this sense, TDE replicates the economic logic of exclusive IP rights of allowing innovators to internalise external benefits – put simply, to prevent ‘free-riding’. Yet, the mere statement that TDE is an effective measure of delaying generic competition is insufficient to draw a conclusion on its normative validity. First, a regulatory intervention should be based on a plausible causal relationship between a behavioural change of a target group induced by such intervention and the achievement of the policy goal.21 To emphasise, the ultimate normative goal of IP is not to protect the innovator’s returns on R&D as such, but to promote the social benefits of innovation. While the premise is that supra-competitive profits can facilitate innovation incentives and dynamic efficiency,22 it is important to distinguish several linkages within the logical chain connecting private returns on R&D and the ultimate social benefits of IP protection. The realisation of social benefits would depend on whether the legal protection can yield adequate returns on R&D; whether the economic reward will encourage inventive activity; whether an inventive activity will lead to successful R&D outcomes; whether such outcomes will contribute to social welfare.23 Given that multiple factors can bear on the functional relationship underlying each linkage,24 no linear association between innovation and IP protection can be presumed. In other words, the social benefits of innovation cannot be assumed to increase monotonically with greater – broader, longer, stronger – exclusivity.25 Second, it should be noted that exclusive IP rights do not eliminate the trade-off between static and dynamic efficiency but swop costs and benefits: under exclusivity, the prospective gain in dynamic efficiency is offset by the loss in static efficiency due to the delayed price competition. Given that a trade-off between efficiencies is inherent in the mechanism of exclusive IP rights,26 the justification inquiry is a matter of establishing not only whether welfare benefits outweigh costs, but also whether the cost–benefit ratio of a particular form of IP protection is more favourable relative to alternative policy options.27 In this view, at issue is not whether IP in 20

21

22

23 24

25 26

27

See William S Comanor, ‘The Political Economy of the Pharmaceutical Industry’ (1986) Journal of Economic Literature 24 (3) 1178, 1213. See generally David W Barnes, ‘The Incentives/Access Tradeoff’ (2010) 9 (3) Northwestern Journal of Technology and Intellectual Property 96. Before the adoption of the Hatch–Waxman Act, some US economists regarded confidentiality of test data submitted for marketing authorisation as ‘an extremely important issue to the sponsor’s return on investment in nonpatentable know-how’ and the protection against disclosure as ‘the secretive imperative [. . .] of the innovation process’. William L Casey, John E Marthinsen and Laurence S Moss, Entrepreneurship, Productivity, and the Freedom of Information Act: Protecting Circumstantially Relevant Business Information (Lexington Books 1983) 201. Paul Sabatier and Daniel Mazmanian, ‘The Implementation of Public Policy: A Framework Analysis’ (1980) 8 Policy Studies Journal 538. Appropriability of returns on innovative activity is regarded as one of the fundamental factors of a technological regime. See Stefano Breschi and Franco Malerba, ‘Sectoral Innovation Systems: Technological Regimes, Schumpeterian Dynamics, and Spatial Boundaries’ in Charles Edquist (ed) Systems of Innovation: Technologies, Institutions, and Organizations (Routledge 1997) 135. Louis Kaplow, ‘The Patent-Antitrust Intersection: A Reappraisal’ (1984) 97 Harvard Law Review 1813, 1823 ff. ibid 1824 (noting that the understanding of such functional relationships is ‘quite limited’). See also Giovanni Dosi, Luigi Marengo and Corrado Pasquali, ‘How Much Should Society Fuel the Greed of Innovators? On the Relations between Appropriability, Opportunities and Rates of Innovation’ (2006) 35 Research Policy 1110, 1113 (noting that ‘no model of invention and innovation [. . .] is possible without a reasonable account of inventive and innovative opportunities and their nature’). See generally Dosi, Marengo, and Pasquali (n 24). Gideon Parchomovsky and Peter Siegelman, ‘Towards an Integrated Theory of Intellectual Property’ (2002) 88 Virginia Law Review 1455, 1458. Carl V Patton, David S Sawicki and Jennifer Clark, Basic Methods of Policy Analysis and Planning (3rd edn, Routledge 2013) 243 ff. On effectiveness and efficiency as criteria of good governance, see Marie-Louise

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

58

Daria Kim

an abstract sense can be justified, but rather whether the way IP protection is designed and implemented can justify its existence.28 A general disclaimer regarding the welfare analysis of IP is worth highlighting: the feasibility of carrying out welfare evaluation is contingent upon the condition that all relevant effects can be identified and evaluated. While the concepts of welfare function, dynamic efficiency, and a static–dynamic efficiency trade-off can serve as the guiding principles, they hardly provide a measure of ‘how much’ exclusivity is necessary to optimise costs and benefits when the modalities of IP protection are configured. In this regard, the economic analysis of IP rights has produced puzzling conclusions29 and led some to believe that IP rights are ‘faith-based’ rather than empirically defensible.30 Even though the notion of ‘an optimal level of exclusivity’ is often referred to as a normative benchmark, the question of where precisely it is located has been subject to extensive inquiry.31 In this regard, welfare analysis of IP can illustrate a general caveat to the application of welfare economics in policymaking: while a social welfare function is regarded as an apt and powerful concept, its practical usefulness might be illusory, as an allencompassing welfare analysis of a regulatory intervention is rather unrealistic.32 This, however, does not mean that any attempt to contextualise the costs and benefits of IP protection and to make a value judgement is not worth pursuing. In what follows, let us consider how the incentive-based theory of IP rights applies in the case of pharmaceutical test data.

28

29

30

31

32

Bemelmans-Videc, Ray C Rist and Evert Vedung, Carrots, Sticks & Sermons: Policy Instruments & Their Evaluation (Transaction Publishers 2010) 7–8. This is in line with a consequentialist perspective, according to which ‘the aggregate positive social welfare contribution of IP law’ results from the sum of welfare-maximising ‘micro cases’. Merges (n 14) 77. Richard A. Posner, ‘Intellectual Property: The Law and Economics Approach’ (2005) Journal of Economic Perspectives 19(2) 57, 59 (observing that, ‘[u]nfortunately, economists do not know whether the existing system of intellectual property rights is, or for that matter whether any other system of intellectual property rights would be, a source of net social utility, given the costs of the system and the existence of alternative sources of incentives to create such property’ (emphasis added)); William M Landes and Richard A Posner, Economic Structure of Intellectual Property Law (Harvard University Press 2003) 9 (noting that ‘a belief that without legal protection the incentives to create [intellectual] property would be inadequate [. . .] cannot be defended on the basis of current knowledge’). See also Daniel R Cahoy, ‘An Incrementalist Approach to Patent Reform Policy’ (2006) 9 Legislation and Public Policy 587, 589 (recapping that ‘the degree to which current patent systems promote innovative behaviour remains surprisingly unclear’); George L Priest, ‘What Economists Can Tell Lawyers about Intellectual Property: Comment on Cheung’ (1986) 8 Res Law & Economics 19, 21 (concluding that ‘economists have virtually nothing to say to lawyers’); Fritz Machlup, The Production and Distribution of Knowledge in the United States (Princeton University Press 1962) 176 (stating that ‘[t]he absence of any empirical evidence for either the claim or its denial that the patent system is an effective promoter of inventive research [. . .] is most frustrating’). Mark A Lemley, ‘Faith-Based Intellectual Property’ (2015) 62 UCLA Law Review 1328, 1332 (summarising that ‘[i]n short, we just didn’t know very much about whether patents survived cost–benefit analysis’). Merges (n 14) 77 ff. There is a large body of research on the optimal design of patent rights: see, eg, William D Nordhaus, Invention, Growth and Welfare (1969); Fredriek M Scherer, ‘Nordhaus’ Theory of Optimal Patent Life: A Geometrical Reinterpretation’ (1972) 62 American. Economic Review 422; Paul Klemperer, ‘How Broad Should the Scope of Patent Protection Be?’ (1990) 21 RAND Journal of Economics 113; Nancy T Gallini, ‘Patent Policy and Costly Imitation’ (1992) 23 RAND Journal of Economics 52; Vincenzo Denicolò, ‘Patent Races and Optimal Patent Breadth and Length’ (1996) 44 Journal of Industrial Economics 249; Ted O’Donoghue et al., ‘Patent Breadth, Patent Life, and the Pace of Technological Progress’ (1998) 7 Journal of Economics & Management Strategy 1; Ian Ayers and Paul Klemperer, ‘Limiting Patentees’ Market Power Without Reducing Innovation Incentives: The Perverse Benefits of Uncertainty and Non-Injunctive Remedies’ (1999) 97 Michigan Law Review 985; Daniel J Gifford, ‘How Do the Social Benefits and Costs of the Patent System Stack up in Pharmaceuticals?’ (2004) 12 Journal of Intellectual Property 75; Margaret K Kyle, ‘The Alignment of Innovation Policy and Social Welfare: Evidence from Pharmaceuticals’ (2020) 20 Innovation Policy and the Economy 95, 103. Richard E Just, Darrell L Hueth and Andrew Schmitz, The Welfare Economics of Public Policy: A Practical Approach to Project and Policy Evaluation (Edward Elgar 2004) 41; Mark D White, The Oxford Handbook of Ethics and Economics (OUP 2019) 66 (noting that a social welfare function is a ‘theoretical construct’).

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

Test Data Exclusivity

59

III.II.B Which Market ‘Fails’ and Which Activity Is Incentivised? In view of a market failure-based justification of exclusive IP rights, one should start by clarifying which market, why, and in what way, fails in the case of test data.33 At the outset, it should be noted that in contrast to technological ‘reverse-engineerable’ information, primary (non-summary) clinical trial data constitutes an ‘impure’ public good – a good that is nonrivalrous in use, but partially – in the case of clinical trial data, to a large extent – excludable.34 First, once primary (individual-level) test data is generated in clinical studies, trial sponsors become data ‘custodians’35 capable of exercising exclusive control over third-party access, even if the law does not allocate property rights in individual patient-level test data to trial sponsors.36 Second, non-summary test data cannot be ‘reverse-engineered’ from scientific publications reporting trial results or the summary-level information published in trial registers. Nor can the primary data be ‘reconstructed’ from a drug, in contrast to the technical information – such as chemical composition and formulation – underlying a drug. Third, once trial data is submitted to a regulatory authority in support of an application for drug marketing authorisation, the professional duty of confidentiality applies.37 Furthermore, generic companies do not obtain actual access to the originator’s data during the abbreviated marketing authorisation, let alone independent researchers or the public at large.38 Neither do third parties gain access to primary test data upon the expiration of TDE. In sum, nonsummary data does not ‘spill over’ into the public domain and can be characterised as a good to a large extent excludable even in the absence of exclusive rights vested in trial sponsors. Thus, no externality-induced market failure due to non-excludability of test data as such can be identified, to begin with, which IP protection would need to remedy. Economic externalities – that is, uncompensated benefits external to the originator companies – can arise where a generic product is approved based on the evidence underlying the originator drug, the so-called referential use of the originator’s test data.39 TDE is tasked to internalise such benefits otherwise accruing to generic competitors in the form of saved costs of conducting clinical trials.40 The peculiarity of TDE subsists in that it does not intend to remedy 33

34 35

36

37

38 39

40

A market failure is defined as a failure ‘of a more or less idealized system of price-market institutions to sustain “desirable” activities or to estop “undesirable’ activities”‘, whereby the desirability of activity is evaluated relative to an ‘explicit or implied maximum-welfare problem’. Francis M Bator, ‘The Anatomy of Market Failure’ (1958) 72(3) Quarterly Journal of Economics 351, at 351. Richard Cornes and Todd Sandler, The Theory of Externalities, Public Goods, and Club Goods (CUP 1996) 4. Under European law, exclusive control of clinical trial sponsors over primary non-summary trial data stems from the obligation to protect all information and data from clinical trials inter alia against unauthorised or unlawful access, disclosure, and dissemination. Regulation 536/2014/EU (OJ L 158, 27.5.2014), art 56. Besides, drug companies – trial sponsors – are usually in the position to exclude third-party access to primary trial data through complex contractual arrangements. Furthermore, factual secrecy as well as trade secret law, where applicable, can restrict third-party access to data. On the legal determinants of trial sponsors’ control over clinical trial data in the EU context, see Daria Kim, Access to Non-Summary Clinical Trial Data for Research Purposes Under EU Law (Springer 2021) 79 ff. Test data exclusivity should be distinguished from the duty of confidentiality; the latter is guaranteed in the EU under Article 339 of the Treaty on the Functioning of the European Union and applies to EU institutions, including the European Medicines Agency. As discussed in III.II.G. For that, a generic applicant needs to demonstrate bioequivalence with the originator product. For details in the EU context, see European Medicines Agency, ‘Investigation of Bioequivalence’ (accessed 28 Feb. 2021). Generic companies are exempted from conducting clinical trials based on the ethical and efficiency considerations, namely, to prevent unnecessary exposure of trial participants to health risks and to avoid wasteful duplication of research.

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

60

Daria Kim

a distinct public good problem in a ‘market for data’. Even though clinical trial data-sharing occasionally takes place,41 returns on drug R&D are earned predominantly in a drug market. Thus, rather than test data as such, TDE aims to protect the originator’s competitive advantage in a drug market where absent the protection generic competition would not allow the originator to earn returns on R&D investment. The problem of insufficient incentives to generate test data is rooted in a different market failure than the one tackled by patents. In the case of technological inventions, the lack of incentives is attributed to the market failure caused by the public-good nature of information (Arrow’s ‘information paradox’42), whereas the lack of incentives to generate test data is caused by the market failures of information asymmetries and insufficient testing43 (‘the market for lemons’44). Thus, the regulation of drug marketing authorisation involves both ‘stick’ and ‘carrot’ components corresponding to two distinct policy objectives: the ‘stick’ refers to the requirements to conduct clinical trials and gather and submit sufficient evidence on drug safety and efficacy to mitigate the risks to public health;45 the ‘carrot’ refers to TDE aiming to protect economic incentives for innovation. The former prompts the latter as high costs of generating evidence necessitate some form of guaranteed compensation. However, it is worth highlighting that TDE as such is not essential for achieving the policy objective of mitigating risks related to drug safety and quality and, thus, cannot be credited with incentivising the optimal testing and generation of sufficient evidence.46 Drug companies simply have no leeway to decide whether to conduct trials or not if they intend to commercialise medicinal products. In view of the foregoing, a more nuanced understanding of TDE, as an innovation incentive, can be narrowed down to ex post compensation for drug development costs. In this sense, it performs the same function as patents, namely, to protect the originator’s competitive advantage in a market. III.II.C A Partial Overlap between Patents and TDE from a Normative Perspective The common denominator between patents and TDE is their potential to delay competition by imitation. In the case of patents, it results from the right to exclude third parties from using the patented subject matter; in the case of test data, it stems from the exception to the drug authorities’ practice of approving generic products based on the originators’ evidence. While both patent protection and TDE are rationalised as innovation incentives, several divergences in the modes of protection have been pointed out by scholars: the grant of TDE does not stipulate non-summary data disclosure as a quid pro quo for protection;47 public interest 41

42 43

44

45

46

47

For instance, where a generic product is authorised by the originator (the so-called branded generics), or where data is shared for research purposes. On the latter issue, see III.II.G. Arrow (n 19). For definitions, see John O Ledyard, ‘Market Failure’ in Steven N Durlauf and Lawrence E Blume (eds) The New Palgrave Dictionary of Economics (2nd edn, Palgrave Macmillan 2008) 300, 301. George A Akerlof, ‘The Markets for “Lemons”: Qualitative Uncertainty and the Market Mechanism’ (1970) 84 Quarterly Journal of Economics 488; Ariel Katz, ‘Pharmaceutical Lemons: Innovation and Regulation in the Drug Industry’ (2007) 14 Michigan Telecommunications and Technology Law Review 1. The ‘stick’-type regulation is characteristic of policies intending to mitigate public health and environmental risks, in particular, through safety and quality standards. For definitions, see Bemelmans-Videc, Rist, and Vedung (n 27) 10. See Jerome H Reichman, ‘Rethinking the Role of Clinical Trial Data in International Intellectual Property Law: The Case for a Public Goods Approach’ (2009) 13 Marquette Intellectual Property Law Review 1, 40 (criticising the proposition that ‘strengthening the incentives that flow from data exclusivity, as distinct from patents, [can] improve the quality of clinical trial’). Rebecca S Eisenberg, ‘Data Secrecy in the Age of Regulatory Exclusivity’ in Rochelle C Dreyfuss and Katherine J Strandburg (eds) The Law and Theory of Trade Secrecy: A Handbook of Contemporary Research (2011) 468–69.

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

Test Data Exclusivity

61

safeguards are often missing;48 nor does TDE hinge on ‘any of the usual patentability requirements’.49 The argument that TDE does not provide for data disclosure in lieu of protection is valid in so far as individual patient-level data is concerned. Clinical trials constitute translational research that adapts the findings from basic science for the clinical setting.50 A distinct benefit of trial data relative to the technical teaching disclosed in patent applications subsists in that it provides extensive evidence on safety and efficacy. Besides the summary-level results published in trial registers, more detailed information can be disclosed by trial investigators in scientific publications and by drug authorities in the assessment reports.51 Yet, without access to the primary data,52 the robustness of the original conclusions regarding the benefit–risk balance of medicinal products cannot be validated, which contradicts the very principle of open science.53 Given the public interest at stake, it appears absurd that researchers and public health professionals need to petition for access to non-summary data to validate the trial design and outcomes – SARS-Cov-19 vaccines is an infamous case in point.54 Furthermore, the argument that TDE – by analogy with patents – should promote disclosure might appear less convincing if one takes a closer look at how patents have been coping with this function.55 As for public interest safeguards, concerns have been raised that TDE can interfere with the effective use of compulsory licences related to patented medicinal products.56 However, it appears questionable that TDE presents an insurmountable obstacle in this regard, given that such licences have been granted by the WTO (World Trade Organisation) members in the past.57 It should be noted that the TRIPS Agreement does not prohibit exceptions to and limitations of TDE.58 A temporary suspension of TDE that might be needed to exercise the rights under a compulsory licence would also be in line with the Doha Declaration that affirms that the TRIPS Agreement should not ‘prevent Members from taking measures to protect public health’59 and that it ‘can and should be interpreted and implemented in a manner supportive of 48

49

50

51

52

53 54

55

56 57

58

59

See Shaikh (n 5) 197 (concluding that many jurisdictions ‘have failed to include appropriate exceptions and limitations in their laws’). John R Thomas, ‘Toward a Theory of Regulatory Exclusivities’ in Ruth L Okediji and Margo A Bagley (eds) Patent Law in Global Perspective (OUP 2014) 360–61. Doris McGartland Rubio et al., ‘Defining Translational Research: Implications for Training’ (2010) 85(3) Academic Medicine 470, 471–72. In the case of EMA, an assessment report includes documents related to the evaluation of medicines, such as product information and assessment history, authorised via the centralised procedure. See European Medicines Agency, ‘European Public Assessment Reports: Background and Context’ (accessed 28 Feb. 2021). This concerns not only individual patient-level data but also meta-data, such as trial protocols and related documentation. See also III.II.G. Peter Doshi et al., ‘Open Letter to the Department of Health and Human Services’ (20 Oct 2020) (accessed 11 Mar. 2021). See, eg, ‘The Disclosure Function of the Patent System (Or Lack Thereof)’ 118 (6) Harvard Law Review (2005) 2007; Bronwyn H Hall and Dietmar Harhoff, ‘Recent Research on the Economics of Patents’ 4 Annual Review of Economics (2012) 541, 549 (observing that economic research finds ‘little empirical evidence as to the extent of [patent] disclosure and its economic impact’). ‘t Hoen, Boulet, and Baker (n 5) 1; Shaikh (n 5) 12, footnote 44 (with further references). See ‘Examples of Health-Related Compulsory Licenses’ (accessed 28 Feb. 2021). In particular, the TRIPS Agreement does not stipulate any minimum duration of protection, neither does it outlaw an opposition procedure, reliance on foreign approvals, exceptions to TDE, if a patent protecting a drug at issue is subject to a compulsory licence. See Shaikh (n 5) 14. World Trade Organization, ‘Declaration on the TRIPS Agreement and Public Health’ (14 Nov 2001) WT/MIN(01)/ DEC/2 para 4.

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

62

Daria Kim

WTO members’ right to protect public health and, in particular, to promote access to medicines for all’.60 That TDE is not contingent on the patentability criteria is not necessarily in conflict with the innovation rationale of protection, if one takes a view that drug R&D can result in unpatentable but safe and efficacious medicinal products. In this regard, TDE can offer ‘a gap-filling alternative to patents’61 providing for a ‘safety net’ for drugs that are either unpatentable or where patent protection might lapse.62 The need for an additional layer of protection might not indicate a deficiency63 of patents in the pharmaceutical industry but rather can be attributed to objectively higher R&D costs64 compared to other industries.65 At the same time, the fact that TDE is not conditioned on the patentability assessment, and yet it provides for in a way a functional equivalent66 of patent protection, raises a curious issue: What actual role do patentability requirements play in drug innovation? If patentability criteria – especially the nonobviousness requirement regarded as the ‘ultimate condition of patentability’67 – are only weakly associated with the therapeutic value of drugs,68 which is arguably a true measure of the social value of drug innovation, are patents at all a relevant instrument for discerning which advances ought to be rewarded and incentivised in this context? Notably, empirical research shows that neither TDE nor patents correlate with the therapeutic value of drugs.69 Patents are known to promote the rate and, to some extent, the direction of innovation,70 but not

60 61 62

63

64

65

66

67 68

69

70

ibid. Reichman (n 46) 38. ibid 39–42 (arguing that TDE should not be ‘entirely dismissed as an irrelevant appendage to the patent system’ for the incentive rationale can have some ‘legitimacy in at least a narrow range of cases’, such as where drugs not meeting the patentability requirement of non-obviousness can yield ‘surprising therapeutic benefits’). At the same time, it is admitted that the innovation incentive rationale for TDE is ‘usually overstated’. ibid 40. See also Benjamin N Roin, ‘Unpatentable Drugs and the Standards of Patentability’ (2009) 87 Texas Law Review 503, 506 (noting that ‘there is little reason to believe that the drugs denied protection are any less valuable than patentable drugs’). Thomas (n 49) 360 (pointing out that the view of regulatory exclusivities, such as TDE, as innovation incentives ‘implies that the patent system necessarily possesses failings that are unique to pharmaceutical research and development [which] contrasts starkly with the usual account that patents have provided more value to the pharmaceutical industry than to other industries’). Joseph A DiMasi, Henry G Grabowski and Ronald W Hansen, ‘Innovation in the Pharmaceutical Industry: New Estimates of R&D Costs’ (2016) 47 Journal of Health Economics 20. See also Pharmaceutical Research and Manufacturers of America (PhRMA), 2020 PhRMA Annual Membership Survey (2020) (accessed 28 Feb. 2021). Richard C Levin et al., ‘Appropriating the Returns from Industrial Research and Development’ (1987) Brookings Papers on Economic Activity 783. While patents and TDE can be viewed as functional equivalents in a sense of delaying generic entry, protection under TDE is stronger in view of its incontestability by competitors. See also Fabian Gaessler and Stefan Wagner, ‘Patents, Data Exclusivity, and the Development of New Drugs’ (2020) The Review of Economics and Statistics, https://doi.org/ 10.1162/rest_a_00987 (showing that TDE can be an effective policy instrument to motivate drug development and introduction where patent protection is not available or uncertain). John F Witherspoon (ed) Nonobviousness: The Ultimate Condition of Patentability (Bureau of National Affairs 1980). Roin (n 62) 503 (pointing out that the view that patents are successful in promoting pharmaceutical innovation ‘overlooks a serious shortcoming in the drug patent system: the standards by which drugs are deemed unpatentable under the novelty and nonobviousness requirements bear little relationship to the social value of those drugs or the need for a patent to motivate their development’). Margaret K Kyle, ‘Are Important Innovations Rewarded? Evidence from Pharmaceutical Markets’ (2018) 53 Review of Industrial Organization 211, 217 (noting that ‘data exclusivity policies are not tied to therapeutic value’; neither do the patent term, the number of patents or market exclusivity positively correlate with therapeutic importance). See, eg, Joshua Krieger, Danielle Li and Dimitris Papanikolaou, ‘Missing Novelty in Drug Development’ (2018) NBER Working Paper 24595 (accessed 28 Feb. 2021) (showing that ‘risk aversion leads pharmaceutical firms to underinvest in radical innovation’).

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

Test Data Exclusivity

63

innovation quality.71 While compliance with the safety and efficacy requirements safeguards the therapeutic characteristics of drugs, the marketing authorisation regulation does not provide for an incentive to enhance the therapeutic value.72 According to one view, patents and TDE play distinct roles in incentivising drug innovation, in that patents are tasked to induce an invention, while TDE ‘responds more directly to the economics of funding non-inventive but resource-intensive regulatory trials’.73 However, such reading renders irrelevant the inducement theory of patents,74 according to which patents encourage not only an invention but also its development and commercialisation.75 Even if the costs of drug R&D can be higher relative to other industries,76 the tendency to scale-up exclusivity does raise the question of when enough is enough,77 especially, given other instruments, such as patent term extension/ supplementary protection certificates justified on similar grounds, namely, to compensate for the lengthy drug development process.78 Moreover, research on the relationship between IP protection and innovation does not allow one to assume that an extension of exclusivity would yield a proportionate increase in innovation.79 III.II.D Evidence on the Social Benefits of TDE Empirical studies examining the relationship between TDE and the level of R&D investment and innovation activity are scarce.80 A 2013 study compared 45 jurisdictions with and without TDE and did not establish any statistically significant correlation between the amount of investment made by the pharmaceutical industry and the existence of TDE in a particular jurisdiction.81 More recent research examined whether drug R&D projects tend to continue where the invalidation of at least one patent covering a drug candidate undergoing clinical trials 71

72

73

74

75

76 77 78 79

80

81

Product quality is understood in a sense of product functionality and performance, reliability, conformance with the established standards of safety standards, durability, etc. The requirement of industrial application sets a too low standard to perform such functions. Under Article 57 of the European Patent Convention, an invention ‘shall be considered as susceptible of industrial application if it can be made or used in any kind of industry’. The threshold of industrial applicability is relatively low: an invention that does not comply with the generally accepted laws of physics would not fulfil the requirements of Article 57. Case T 0541/96, the decision of 7 Mar. 2001, point 6.1 of the Reasons, ECLI:EP:BA:2001:T054196.20010307. This issue relates to the debate regarding comparative effectiveness trials. See, eg, Laura L Hester et al., ‘Publication of Comparative Effectiveness Research Has Not Increased in High-impact Medical Journals, 2004–2013’ (2017) 84 Journal of Clinical Epidemiology 185. Antony Taubman, ‘Unfair Competition and the Financing of Public-Knowledge Goods: The Problem of Test Data Protection’ (2008) 3 Journal of Intellectual Property Law & Practice 591, 595. Reichman (n 46) 41 (pointing out that ‘federal funding policy often leaves only the downstream research burdens to the pharmaceutical companies, [hence,] it is largely the costs of clinical trials that justifies strong pharmaceutical patents – and correspondingly high pharmaceutical prices – in the first place’). See also Yu (n 15) 784 ff. Roberto Mazzoleni and Richard R Nelson, ‘The Benefits and Costs of Strong Patent Protection: A Contribution to the Current Debate’ (1998) 27 Research Policy 273, 276-277. Above (n 65). Dosi, Marengo, and Pasquali (n 24). For further arguments casting doubt on the economic justification for TDE, see Yu (n 15) 784 ff. See Kaplow (n 23) 1824 (pointing out that ‘our knowledge of the functional relationships between the separate links in the chain connecting patent life to social benefits remains quite limited’); see also below (nn 90–94) and the accompanying text. See, eg, Gaessler and Wagner (n 66) 3 (acknowledging that ‘there is little empirical evidence on how the overall duration of market exclusivity relates to originators’ innovation efforts’). On the methodological challenges of conducting empirical studies on the effects of IP rights on innovation, see Heidi L Williams, ‘Intellectual Property Rights and Innovation: Evidence from Health Care Markets’ (2016) 16 Innovation Policy and the Economy 53. Mike Palmedo, ‘Do Pharmaceutical Firms Invest More Heavily in Countries with Data Exclusivity?’ (2013) 21 Currents: International Trade Law Journal 38, 44.

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

64

Daria Kim

reduces the expected market exclusivity and where drug sponsors would need to rely on TDE as the source of protection.82 Having demonstrated the sensitivity of the decision making regarding the continuation of an R&D project to the anticipated protection against generic competition, the authors conclude that their ‘findings have important implications: data exclusivity emerges as an effective policy instrument to provide market exclusivity in cases where the remaining patent term is short relative to the lengths of needed clinical trials, or where patent protection is uncertain’.83 Such outcome confirms a wide-held consensus that protection against price competition plays a role in facilitating innovation activity in the research-based pharmaceutical sector.84,85 However, one would assume that sensitivity of the decision making to the availability of protection in the drug market would not change substantially if the forms of protection were reversed – if, for some reason, TDE would not be available and a drug sponsor could rely only on patent protection. A study by Goldman et al.86 estimated how a hypothetical extension of the term of exclusivity for new chemical entities from five years, as currently provided in the USA, to twelve years might affect the revenues of pharmaceutical companies, the rate of new drug approval and consumers’ longevity. Based on a microsimulation model,87 the study projects that the short-term effect of such extension would be, on average, a 5 per cent increase in companies’ revenues, while the long-term effect would amount to 228 additional drugs approved in the USA between 2020 and 2060 and the life expectancy would rise by 1.7 months corresponding to $10,400 in net benefits.88 These estimations are based on the study by Acemoglu and Linn,89 who found that an increase in the total therapeutic class-level revenue by 1 per cent yields up to a 3–4 per cent increase in the annual drug approval. The relationship between profitability and R&D productivity in the pharmaceutical sector is a much more complex subject than can be covered in this chapter. Suffice it to say that no linear relationship between the rate of the private returns on R&D and the rate of innovation can be presumed, first and foremost, due to the complexity and uncertainty of drug discovery and development. Notwithstanding the increase in R&D expenditures publicised by drug companies,90 there has been a widespread sentiment regarding the ‘productivity crisis’ in the

82

83 84

85

86

87

88 89

90

Gaessler and Wagner (n 66). The study examined the dataset of 935 drug candidates undergoing clinical trials, for which the decision on patent opposition was published before the completion of trials. ibid 6. ibid 33 (concluding that the study confirms that, ‘as posited by the theoretical literature on the incentives for innovation, R&D efforts of firms are muted in case of reduced periods of market exclusivity’). The convergence of several factors, such as high costs of innovation and low costs of imitation, the ease of reverseengineering and the lack of alternative means to protect competitive advantage against imitation, explains greater reliance of the research-based pharmaceutical industry on exclusive IP rights compared to other industrial sectors. See Luigi Orsenigo and Valerio Sterzi, ‘Comparative Study of the Use of Patents in Different Industries’ (2010) KITeS Knowledge, Internationalization and Technology Studies Working Paper 33/2010 7. A number of empirical studies support this proposition. See James Bessen and Michael J Meurer, ‘Lessons for Patent Policy from Empirical Research on Patent Litigation’ (2005) Boston University School of Law Working Paper No 05-22; Levin et al. (n 65); Edwin Mansfield, ‘Patents and Innovation: An Empirical Study’ (1986) 32 Management Science 173, 175. Dana P Goldman et al., ‘The Benefits from Giving Makers of Conventional “Small Molecule” Drugs Longer Exclusivity Over Clinical Trial Data’ (2011) 30 Health Affairs 84. Darius N Lakdawalla et al., ‘U.S. Pharmaceutical Policy in a Global Marketplace’ (2008) 27(1) Health Affairs, https:// doi.org/10.1377/hlthaff.28.1.w138. Goldman et al., (n 86) 88. Daron Acemoglu and Joshua Linn, ‘Market Size in Innovation: Theory and Evidence from the Pharmaceutical Industry’ (2003) 119 Quarterly Journal of Economics 1049. PhRMA (n 6).

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

Test Data Exclusivity

65

pharmaceutical sector.91 Even though recent research finds signs of productivity improvement,92 they have been associated with the advent of more targeted and precise research methods.93 Meanwhile, the rate of new drug approval has been fluctuating94 and any attempt to explain the underlying dynamics would require a more comprehensive understanding of the relevant determining factors. Such factors are likely to reflect the specifics of the national innovation system and healthcare; therefore, the projections based on the historical data from the pharmaceutical industry in the USA should not be extrapolated to inform the decision making in other jurisdictions. Important is that, instead of maximising private returns, IP law should aim at balancing the costs and benefits of IP protection. As acknowledged by Goldman et al., the extension of data exclusivity protection would require a trade-off between the current and future generations.95 Along the same lines, Gaessler and Wagner point out that, even though TDE can be ‘an effective policy lever to restore incentives in case of long drug development periods’ – especially where patent protection is not available – extending the term of exclusivity as a policy measure ‘is not uncontested’, given its effects on the affordability of drugs.96 III.II.E Deadweight Loss as a Social Cost of TDE The social cost of exclusive IP rights is defined as a loss in static efficiency due to restrained price competition leading to prices above the competitive level (the deadweight loss).97 Economic studies produce divergent estimates in the case of patented drugs.98 A universal calculus is hardly possible as the magnitude of loss in static efficiency – as well as the prospective gain in dynamic efficiency – can vary considerably from one country to another99 depending on factors such as

91

92

93

94

95 96 97

98

99

Dennis Lendrem et al., ‘R&D Productivity Rides Again?’ (2015) 14 Pharmaceutical Statistics 1; Jack W Scannell et al., ‘Diagnosing the Decline in Pharmaceutical R&D Efficiency’ (2012) 11 Nature Reviews Drug Discovery 191; Fabio Pammolli, Laura Magazzini and Massimo Riccaboni, ‘The Productivity Crisis in Pharmaceutical R&D’ (2011) 10 Nature Reviews Drug Discovery 428. Fabio Pammolli et al., ‘The Endless Frontier? The Recent Increase of R&D Productivity in Pharmaceuticals’ (2020) 18 Journal of Translational Medicine 162. ibid 12–13. See also Steven M Paul et al., ‘How to Improve R&D Productivity: The Pharmaceutical Industry’s Grand Challenge’ (2010) 9 Nature Reviews Drug Discovery 203; Jörg Mahlich, Arne Bartol and Srirangan Dheban, ‘Can Adaptive Clinical Trials Help to Solve the Productivity Crisis of the Pharmaceutical Industry? A Scenario Analysis’ (2021) 11 Health Economics Review 4. FDA’s Center for Drug Evaluation and Research, ‘Advancing Health through Innovation: New Drug Therapy Approvals 2019’ 6 (accessed 28 Feb. 2021). Goldman et al. (n 86) 85. Gaessler and Wagner (n 66) 34. Richard A. Posner, ‘The Social Costs of Monopoly and Regulation’ (1975) 83 (4) Journal of Political Economy 807, at 807. But see Gifford (n 31) 122–23 (arguing that, ‘[t]o the extent that the exclusivity conferred on a patentee is necessary to generate an invention, the resulting deadweight loss is not a social loss at all’; only ‘to the extent that the exclusivity is unnecessary [. . .] it can become a social cost’). See, eg, Shubham Chaudhuri, Pinelopi K Goldberg and Panle Jia, ‘Estimating the Effects of Global Patent Protection in Pharmaceuticals: A Case Study of Quinolones in India’ (2006) 96(5) American Economic Review 1477 (projecting ‘substantial welfare losses to the Indian economy, even in the presence of price regulation [whereby the] overwhelming portion of this welfare loss derives from the loss of consumer welfare’). But see Mark Duggan, Craig Garthwaite and Aparajita Goyal, ‘The Market Impacts of Pharmaceutical Product Patents in Developing Countries: Evidence from India’ (2016) 106 American Economic Review 99 (providing more moderate estimates); Margaret Kyle and Yi Qian, ‘Intellectual Property Rights and Access to Innovation: Evidence from TRIPS’ (2013) NBER Working Paper 20779 23 (concluding that ‘the existence of patents on a molecule in a country does not always block generic imitation’). Kyle (n 31) 103.

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

66

Daria Kim

the competitiveness of both the originator and generic branches of the domestic pharmaceutical industry, the market conditions, as well as the specifics of the national healthcare system. Brought to the fore under the heading of access to affordable drugs, the deadweight loss due to the delayed generic entry has been the major point of criticism of both patents and TDE.100 Studies show that, on average, TDE expires several years earlier101 than patents protecting a new chemical entity.102 However, even where protection terms might overlap, TDE cannot be considered redundant due to its incontestability.103 Theoretically, cumulative protection under patents and TDE can yield higher returns on investment and, hence, greater incentive to innovate. Yet, even though this argument has been conventionally put forth as a justification for TDE,104 the welfare perspective advocates for balancing social welfare costs and benefits of IP, rather than maximising private returns. III.II.F Considerations Regarding Balancing Welfare Costs and Benefits of Exclusivity TDE poses a similar policy dilemma as other exclusive IP rights, namely, to what extent price competition should be restrained to promote innovation in an optimal way.105 While the idea of balancing static and dynamic efficiency might be appealing, how exactly it should be implemented at the operative level of law is unclear, in particular, due to the difficulty in assessing and projecting effects on dynamic efficiency106 and the uncertainty about the functional relationship between the appropriability conditions and R&D productivity.107 Neither should one downplay the importance of estimating the deadweight loss of exclusivity: an extension of the protection term cannot be justified by projecting potential benefits based on the historical data without measuring the corresponding loss. In this view, the shortcoming of the policy recommendation by Goldman et al.108 can be seen in that it does not factor into the analysis the deadweight loss resulting from the proposed extension of the TDE term. The expected increase in drug companies’ revenues by 5 per cent under the restrained price competition constitutes the transfer from consumer to producer.109 However, the increase in revenues alone does not tell us anything about the magnitude of the induced deadweight loss, which would depend on the particular market conditions, as well as the specifics of the national policies for drug procurement and health insurance. Theoretically, it is possible that the total dollar value of the increased longevity projected by Goldman et al., in discounted terms, can be 100 101

102

103 104 105

106

107 108 109

Above (n 2). See Erika Lietzan, ‘The “Evergreening” Metaphor in Intellectual Property Scholarship’ (2019) 53 Akron Law Review 805, 854 (and studies referenced therein). While the protection obligations under Article 39 of the TRIPS Agreement concern pharmaceutical and agricultural chemical products utilising new chemical entities, the extension of protection to known chemicals has been stipulated under the free trade agreements. Thomas (n 49) 357. IFPMA (n 6) 5–6. Meir Perez Pugatch, ‘Intellectual Property, Data Exclusivity, Innovation and Market Access’ in Pedro Roffe, Geoff Tansey and David Vivas Eugui (eds) Negotiating Health: Intellectual Property and Access to Medicines (Earthscan 2006) 98; Daria Kim, ‘Rationalising Innovation Incentives under Drug Approval Regulation from the National Policymaking Perspective’ in Bryan Mercurio and Daria Kim (eds) Contemporary Issues in Pharmaceutical Patent Law: Setting the Framework and Exploring Policy Options (Routledge 2017) 204. Wolfgang Kerber, ‘Should Competition Law Promote Efficiency? Some Reflections of an Economist on the Normative Foundations of Competition Law’ in Josef Drexl, Laurence Idot and Joel Moneger (eds) Economic Theory and Competition Law (Edward Elgar 2009) 93, 98. Kaplow (n 23) 1824. Goldman et al (n 86). Richard A Posner, ‘The Social Costs of Monopoly and Regulation’ 83 (4) Journal of Political Economy (1975) 807.

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

Test Data Exclusivity

67

less than the dollar value of the corresponding deadweight loss. If so, the term extension would not be an effective110 measure for promoting drug innovation. Furthermore, one would need to compare the cost–benefit ratio of the alternative policy measures, including the non-extension baseline, to identify the more optimal instrument that would minimise costs and maximise benefits. All things being equal, it is possible that other regulatory instruments can accomplish the policy objective more efficiently by achieving a matching rate of new drug introduction with lower deadweight loss. Another objection to the reasoning by Goldman et al.111 is that it is based on the concept of Kaldor–Hicks efficiency, according to which the allocation of resources is efficient if the gains to the beneficiaries, due to the change in resources allocation, outweigh the losses to those who are put at a disadvantage. It is worth highlighting that the application of Kaldor–Hicks efficiency as a normative benchmark in public health policy has been criticised as normatively deficient and unacceptable.112 In situations where exclusive rights cover medicinal products, the deadweight loss due to the delayed price competition implies limited affordability of drugs. The static–dynamic efficiency trade-off emanating from exclusivity means that lives and health that could be saved or improved today, if drug prices were at a competitive level, are ‘traded-off’ for the lives and health improved in the future due to the sustained incentives of innovators. The ethical skewness of this logic is self-evident and raises the question: have we, as a society, exhausted all thinkable alternatives of dealing with the appropriability issue in drug innovation? If the goal is not to maximise appropriability but to minimise the trade-off between social benefits, the cost-sharing approach long advocated by scholars113 and health activists114 manifestly provides for a more appropriate instrument. Notably, in the EU, it has been implemented in relation to chemical test data under the REACH Regulation (Registration, Evaluation, Authorisation and Restriction of Chemicals),115 which protects the commercial interests of the first registrant of a chemical substance by way of requiring subsequent registrants to share the cost of generating test data ‘in a fair, transparent and non-discriminatory way’.116 On balance, TDE promises potential future benefits in terms of innovative medicines while imposing critical social costs in terms of the affordability of drugs. How can and should these parts of the equation be squared up? To emphasise, the idea of balancing static and dynamic efficiency is not a unique challenge to TDE and can be viewed as an unfeasible-to-operationalise normative standard across IP rights if by ‘to operationalise’ one means to define the length

110

111

112 113

114

115 116

‘Cost-effectiveness’ refers to the ‘goal-realisation’ achieved by a policy instrument, taking into account both its positive and negative impact. Bemelmans-Videc, Rist, and Vedung (n 27) 7. Goldman et al. (n 86) 89 (concluding that ‘Americans would benefit in the long term from a longer period of data exclusivity’). White (n 32) 480. With regard to IP rights in general, see Cristiano Antonelli, ‘Towards Non-Exclusive Intellectual Property Rights’ in Cristiano Antonelli and Albert N Link (eds) Routledge Handbook of the Economics of Knowledge (Routledge 2015); Cristiano Antonelli, ‘Technological Knowledge as an Essential Facility’ (2007) 17 Journal of Evolutionary Economics 451. With regard to test data in particular, see Fellmeth (n 15); Robert Weissman, ‘Public Health-Friendly Options for Protecting Pharmaceutical Registration Data’ (2006) International Journal of Intellectual Property Management 1(1/2) 113; Reichman (n 46) 42 ff. Judit Rius Sanjuan, James Love, and Robert Weissman, ‘Protection of Pharmaceutical Test Data: A Policy Proposal’ (2006) KEI Research Paper 1 (accessed 28 Feb. 2021). Regulation (EC) No 1907/2006 (OJ L 396, 30.12.2006). ibid arts 27, 30.

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

68

Daria Kim

and breadth of exclusive rights in a way to provide for a certain amount of exclusivity as a universally optimal cost–benefit ratio.117 III.II.G Losses Beyond Static Inefficiency In recent years, a parallel discourse around access to clinical trial data has emerged prompted by the initiatives of the drug authorities including the European Medicines Agency,118 the US Food and Drug Administration,119 and Health Canada.120 The medical community has argued relentlessly that access to primary evidence from clinical trials is a paramount principle of scientific research, and that such data provides ample opportunities to conduct medical research and promote public health. In broad terms, the benefits of post-trial access to individual patientlevel data fall into two categories in line with two types of secondary data analysis: confirmatory analysis can ensure greater transparency and confidence in drug safety and efficacy,121 while exploratory data analysis can generate additional insights into disease pathways, identify new promising endpoints,122 or otherwise contribute to medical knowledge beyond the questions and hypotheses investigated in the original trial.123 To the extent to which TDE might restrict access, such benefits remain unrealised and, thus, impose distinct social costs. Notably, such costs of TDE were anticipated before the Hatch–Waxman Act – the world’s first law introducing TDE – was enacted in the USA in 1984.124 In their 1980 article, McGarity and Shapiro125 point out that confidential treatment of test data can ‘hamper scientific progress, deny consumers the opportunity to make fully informed product use decisions, increase the risks that agency decisions based on faulty data or analysis will remain undiscovered, and encourage potentially hazardous duplicative human testing’.126 117 118

119

120

121

122

123 124 125

126

On the calculability critique of the welfare theory of IP rights, see Merges (n 14) 77 ff. European Medicines Agency, ‘European Medicines Agency Policy on Publication of Clinical Data for Medicinal Products for Human Use. Policy 0070’ (21 Mar. 2019) (accessed 28 Feb. 2021). In 2013, the US Food and Drug Administration (US FDA) issued a request for comments proposing a system of proactive release of raw, de-identified and masked individual patient-level data and acknowledging that safety and effectiveness data that are submitted for drug MA has ‘a tremendous potential to help address critical challenges and provide new opportunities for innovation in medical product development, including for human drugs, medical devices, and biological products’. US FDA, ‘Availability of Masked and De-Identified Non-Summary Safety and Efficacy Data: Request for Comments’ (4 June 2013) 78(107) Fed. Reg. 33422. By ‘masked data’, the US FDA refers to ‘data with information removed that could link it to a specific product or application’. In 2018, Health Canada conducted a public consultation on the public release of clinical trial data. See Health Canada, ‘Draft Guidance Document, Public Release of Clinical Information’ (10 April 2018) (accessed 28 Feb. 2021). In this regard, it is important to distinguish between the secondary analysis by drug authorities in the course of the regulatory approval of a drug and the secondary analysis by independent researchers. Primary analysis is usually conducted by trial investigators. See, eg, Institute of Medicine, Committee on Strategies for Responsible Sharing of Clinical Trial Data, Sharing Clinical Trial Data: Maximizing Benefits, Minimizing Risk (The National Academies Press 2015); US FDA (n 119) 33423; FDA, ‘Advancing Regulatory Science at FDA: A Strategic Plan’ (2011) (accessed 28 Feb. 2021) 12; Finn Gustafsson et al., ‘Maximizing Scientific Knowledge from Randomized Clinical Trials’ (2010) 159(6) American Heart Journal 937– 43; Lesley A Stewart and Jayne F Tierney, ‘To IPD or Not To IPD? Advantages and Disadvantages of Systematic Reviews Using Individual Patient Data’ (2002) 25(1) Evaluation & The Health Professions 76. Institute of Medicine (n 122). Above (n 7). Thomas O McGarity and Sidney A Shapiro, ‘The Trade Secret Status of Health and Safety Testing Information: Reforming Agency Disclosure Policies’ (1980) 93 Harvard Law Review 837. ibid 838.

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

Test Data Exclusivity

69

It is worth highlighting that, while TDE provides for protection against the referential use by drug authorities,127 the problem of access is much broader as it concerns trial sponsors’ control over all data gathered in trials, most critically, data from failed trials.128 Primary data collected in trials usually stays with trial sponsors, whether or not it is submitted to drug authorities,129 and it remains at their discretion whether or not to grant access to non-summary data to third parties for secondary data analysis. While a detailed analysis of this long-standing contentious issue goes beyond the scope of this chapter,130 the remainder of the discussion focuses on two questions: first, whether third-party access to non-summary clinical trial data for secondary data analysis – and policy measures enabling such access – can render TDE protection void; second, whether TDE might, in fact, facilitate data sharing and secondary data analysis. Each is addressed in turn. III.II.G.I Concerns over the Potential Misuse of the Accessed Data Whether the release of non-summary data by drug authorities can render TDE void is important to clarify considering the arguments advanced by the pharmaceutical industry that such measure would ‘impede’ their innovation incentives. In particular, originator companies claimed that generic competitors would be able to ‘re-submit’ the accessed data to drug authorities in other jurisdictions, which would hinder protection against price competition.131 Such arguments suggest that pro-access regulatory measures can give rise to a policy trade-off different from the one discussed earlier, namely, a trade-off between social benefits of multiple secondary data analyses, on the one hand, and those arising due to the protection of innovation incentives, on the other hand. However, it is doubtful whether concerns expressed by drug companies can be substantiated. In jurisdictions that provide for TDE – including the USA,132 the 127

128

129

130 131

132

Protection under TDE is based on the idea of excludability: while subsequent applicants are excluded from relying on the originator’s test data, the data holder can still authorise the ‘referential’ use of safety and efficacy data. Prompt access to data from failed trials can preclude duplication of mistakes and wasteful research. See Harvard University’s Multi-Regional Clinical Trials Center (MRCT), ‘Comments to Docket No FDA–2013–N–0271: Availability of Masked and De-identified Non-Summary Safety and Efficacy Data’ (5 August 2013) (accessed 28 Feb. 2021). The requirements and assessment practices of drug authorities with regard to test data vary among jurisdictions. The US FDA is known to be unique in that it re-analyses individual patient-level data. The European Medicines Agency has not been routinely requesting the submission of individual patient-level data, at least, not until recently. See European Medicines Agency, ‘Note for Guidance on the Inclusion of Appendices to Clinical Study Reports in Marketing Authorisation Applications’ (23 June 2004) CHMP/EWP/2998/03/Final. See also Franz König et al., ‘Sharing Clinical Trial Data on Patient Level: Opportunities and Challenges’ (2015) 57 Biometrical Journal 8. For a detailed analysis, see Kim (n 36). For an overview of arguments submitted by the pharmaceutical industry during the public consultation conducted by the European Medicines Agency, see Daria Kim, ‘Knowledge Sharing as a Social Dilemma in Pharmaceutical Innovation’ (2016) 71 Food and Drug Law Journal 673, 681 ff. Similar arguments were advanced during the public consultation by the US FDA on the availability of masked and de-identified non-summary safety and efficacy data. See ‘Availability of Anonymized Non-Summary Safety and Efficacy Data. Request for Comments’ (accessed 28 Feb. 2021) (in particular, the submissions by Roche, Novartis, and PhRMA). The US Federal Food, Drug, and Cosmetic Act does not state explicitly whether, during the term of test data protection, a generic drug can be approved based on the submission of a new drug application. However, according to the explanation provided by the US FDA, the new drug product exclusivity does not provide ‘any protection from the marketing of a duplicate version of the same drug product if the duplicate version is the subject of a full new drug application submitted under 505(b)(1) of the [Federal Food, Drug, and Cosmetic] Act’. See U.S. Department of Health and Human Services, US Food and Drug Administration, ‘Frequently Asked Questions for New Drug Product Exclusivity’ (accessed 28 Feb. 2021) (emphasis added).

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

70

Daria Kim

EU,133 Russia,134 and China135 – a generic product can be de jure approved, provided that the applicant submits ‘own’ efficacy and safety data, that is, based on a complete application dossier. Conducting clinical trials is not a secretive activity – trials are subject to approval and registration and, even though this requirement is not always observed, applicants for the drug marketing authorisation are obliged to provide clinical trial information, including unique trial identifiers.136 In other words, one cannot simply re-submit any data to a drug authority in other jurisdictions, without being verified where the data originates from and whether the application submission is legitimate. In some jurisdictions, the submission of test data is not required to begin with, as a drug candidate can be approved based on the marketing authorisation issued by other drug regulators, such as the US Food and Drug Administration or the European Medicines Agency. All this can explain why the argument that competitors can misuse data released by the European Medicines Agency under its transparency policy has not succeeded at the Court of Justice of the European Union,137 where several drug companies objected to the disclosure of clinical study reports submitted for marketing authorisation. Accordingly, access to data submitted to drug authorities would not cause a trade-off between greater transparency and validity of safety and efficacy claims, on the one hand, and the protection of incentives by way of TDE, on the other hand.138 In this view, the legislator could enable access to non-summary test data without rendering TDE void and without ‘fortifying’139 TDE by market exclusivity,140 given that the requirement of submitting ‘own’ data can serve as a market entry barrier.141 It is worth highlighting in this regard that, even if generating ‘own’ data by generic companies remains a purely hypothetical option,142 providing for such possibility de

133 134 135 136 137

138

139

140

141 142

Kim (n 36) 148 ff. Federal Law of the Russian Federation No 61-FZ of 4 April, 2010 ‘On Marketing of Medicinal Products’, art 18. Kim (n 36) 151, footnote 155. Under EU law, this requirement is laid down under Regulation 536/2014/EU (OJ L 158, 27.5.2014), Annex IV (A). In PTC Therapeutics International v EMA, the CJEU noted that the claim that clinical study reports could be misused by competitors for marketing authorisation purposes was ‘unfounded in law’, since a competitor ‘would still have to carry out its own relevant studies and trials’. Case T-718/15, PTC Therapeutics International v EMA [2018] ECLI:EU:T:2018:66, para 91 (emphasis added). In Amicus Therapeutics v EMA, the CJEU held that the arguments concerning potential misuse of the clinical study reports are ‘vague’, and that the applicants ‘have not put forward any specific argument to show that the alleged danger in certain third countries is real’. Case T-33/17, Amicus Therapeutics v EMA [2018] ECLI:EU:T:2018:595, para 84. See also Case C‑576/19 P, Intercept Pharma Ltd. v EMA, [2020] ECLI:EU:C:2020:873, para 53 (pointing out that ‘a mere unsubstantiated claim relating to a general risk of [data] misuse’ does not enable ‘the Courts of the European Union to understand how disclosure of those data would be likely concretely and reasonably foreseeably to undermine the commercial interests of the persons concerned thereby’). The issue of whether disclosure of non-summary test data can have implications for innovation incentives beyond TDE and can impact other determinants of competitive advantage of originator companies, including patent protection for drugs that can be developed based on secondary data analysis, goes beyond the scope of this chapter. For an in-depth analysis, see Kim (n 36). Eisenberg (n 47) 488 (suggesting that, in order to mitigate competitive harms, the ‘gap in protection could be fixed by amending the statute to specify that during the period of regulatory exclusivity, nobody could submit an NDA for the same chemical entity using data previously submitted in support of the NDA for the reference product without the permission of the holder of the NDA for the reference product’). As noted earlier, this mechanism is already in place in the USA (above (n 132)). The difference between TDE and market exclusivity lies in that under TDE a generic product can, theoretically, be authorised based on ‘own’ data, while under market exclusivity a competing product cannot be authorised even if the applicant submits evidence from new trials. For generic companies, the costs of conducting trials are considered to be prohibitive. Gaessler and Wagner (n 66) 10, footnote 3 (reporting that their ‘interviews with executives from pharmaceutical companies did not reveal any cases where firms duplicated clinical trials’).

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

Test Data Exclusivity

71

jure is inconsistent with the legislator’s intention to exempt generic applicants from conducting trials in the first place, given the ethical considerations.143 In principle, the subsequent applicant should not be allowed to submit ‘own’ data. Notably, the regulation of chemical data is more consistent in this regard, as it explicitly prohibits duplicative studies involving vertebrate animals.144 III.II.G.II Can Protection under TDE Facilitate Data Sharing and Exploratory Data Analysis? Post-trial exploratory analysis of non-summary trial data can open new avenues in medical research and drug development not directly related to the research questions explored in the original trial in which data was collected.145 Given that information and data are goods non-rival in use, benefits in terms of new knowledge can be multiplied if the use of data as a research input is maximised.146 Furthermore, research opportunities are enhanced if datasets from multiple trials can be aggregated for meta-analysis.147 As such, TDE does not stand in the way of the realisation of such benefits, as trial sponsors (data holders) can contractually share non-summary trial data irrespective of whether it has been submitted to the drug authorities or not, provided that necessary safeguards for personal data protection and other legal obligations are observed. To some extent, such practice has been taking place under the data-sharing policies of drug companies that provide for access to anonymised patient-level data for non-commercial research purposes, usually subject to review of research proposals.148 The question is rather whether under such conditions the full value of data as a research input can be realised. Can TDE provide an incentive to share data for secondary data analysis? Such assumption would be in line with the market-opening theory of patent rights, according to which exclusive rights can enable licensing of information. The difference between technical teaching underlying an invention and test data is that the ‘information paradox’149 does not arise in the case of non-summary data.150 In the absence of pervasive externalities, trial sponsors can internalise the benefits of data sharing through contractual means. However, their motivation to grant thirdparty access for exploratory data analysis depends on multiple factors and can be inhibited by competitive concerns. Even if the accessed data cannot be de jure re-used for drug marketing 143

144

145

146 147

148

149 150

The abbreviated marketing authorisation by way of referencing the originator’s test data was introduced to avoid duplicative testing and prevent unnecessary exposure of trial participants to health risks. Directive 2001/83/EC, rec 10; Directive 87/21/EEC, rec 4; Case C-368/96, The Queen v The Licensing Authority [1998] ECLI:EU:C:1998:583, paras 69–71. Regulation (EC) 1907/2006, art 26(3). See also Regulation (EC) 1107/2009 (OJ L 309, 24.11.2009), rec 40 (stating that ‘duplication of tests and studies on vertebrates should be prohibited’, and that ‘[f]or the purpose of developing new plant protection products, there should be an obligation to allow access to studies on vertebrates on reasonable terms and the results and the costs of tests and studies on animals should be shared’). See, eg, Stewart and Tierney (n 122). For a detailed overview of various types of secondary data analysis, see Kim (n 36) 45 ff. OECD, Data-Driven Innovation: Big Data for Growth and Well-Being (OECD Publishing 2015) 196. See, eg, Jayne F Tierney et al., ‘How Individual Participant Data Meta-Analyses Have Influenced Trial Design, Conduct, and Analysis’ (2015) 68 Journal of Clinical Epidemiology 1325; Mark Simmonds, Gavin Stewart and Lesley Stewart, ‘A Decade of Individual Participant Data Meta-Analyses: A Review of Current Practice’ (2015) 45 Contemporary Clinical Trials 76–83. See, eg, statistics on data requests under the Clinical Study Data Request Consortium (accessed 28 Feb. 2021). Arrow (n 19) 615. On clinical trial data as an excludable good, see (nn 35–37) and the accompanying text.

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

72

Daria Kim

authorisation,151 research-based drug companies argued that third-party analysis can facilitate the development of competing products and in that way hinder innovation incentives.152 The meaning of ‘innovation incentive’ in this context is distinct from protection against generic competition and is associated with the protection of competitive advantage in competition in innovation. While the extent to which competitive concerns of trial sponsors can be justified would depend on the individual circumstances of a case, a closer look at the possible constellations suggests that such concerns are likely to be overstated.153 Enabling access to (anonymised) individual patient-level data for secondary research purposes could be rationalised by analogy with the experimental use exception under patent law. By allowing the information disclosed in a patent to serve as research input, the exception enables the ‘standing-on-shoulders’ effect154 and promotes dynamic competition. By way of this analogy, TDE – if viewed as an IP form of protection – should not only protect the originator’s returns on R&D investment but also enable the use of data as an input for subsequent research and knowledge creation.155 However, the analogy is not perfect as there is no public access to trial data, in contrast to technical information disclosed in a patent. In other words, the experimentaluse defence would be of no use as long as access to data remains subject to the authorisation of data holders. Considering the foregoing, the inconsistency of TDE with the economic rationale of exclusive IP rights can be viewed in that TDE is not designed to promote dynamic competition by enabling the use of data as an input for follow-on research and innovation. To the extent originator companies’ control over non-summary data – enabled through TDE or other legal and factual means – can preclude the realisation of the research value of data, it can impose a distinct social cost in terms of missed opportunities for drug innovation. III.III CONCLUDING REMARKS

In line with the leitmotif of the edited collection, this chapter intended to examine to what extent TDE is aligned with the economic logic of IP protection. The similarity has been observed in that TDE is meant to resolve a ‘static–dynamic efficiency trade-off’ in drug innovation. However, the fact that TDE provides for an effective mechanism of delaying generic competition and securing private companies’ returns on investment in drug R&D does not automatically justify it under a welfarebased theory of IP law. By identifying challenges with evaluating both social costs and benefits of TDE, the analysis highlights that the idea of balancing static and dynamic efficiency might be too unrefined to serve as an operative normative standard when designing the modalities of legal protection that could fashion an ‘optimal measure’ of exclusivity. Overall, the chapter reinforces the argument that protection of (private) returns on R&D investment should not be equated with social benefits in terms of drug innovation, and that policymakers should not strive to maximise exclusivity in pursuit of greater innovation. 151 152 153 154

155

As argued at III.II.G.I. For an overview of arguments, see Kim (n 131) 686 ff. For an analysis, see Kim (n 36) 222 ff. It should be noted, however, that the experimental use exception under patent law is usually tailored rather narrowly, and it is questionable whether it provides for an optimal instrument to realise the full value of the protected technology as a knowledge input. Rather, such role is ascribed to licensing. See Suzanne Scotchmer, Innovation and Incentives (MIT 2004) 127 ff. On the importance of enabling the use of IP-protected subject matter as an innovation input, see Cristiano Antonelli, ‘A Reappraisal of the Arrovian Postulate and the Intellectual Property Regime: User-Specific Patents’ (2019) 47 European Journal of Law and Economics 377; Antonelli, ‘Technological Knowledge as an Essential Facility’ (n 113).

https://doi.org/10.1017/9781108989527.006 Published online by Cambridge University Press

IV Copyright in Works Created by Artificial Intelligence Between Creativity and Investments Enrico Bonadio, Luke McDonagh, and Plamen Dinev

IV.I INTRODUCTION

Artificial intelligence (AI) has captured the attention of copyright lawyers fascinated by the thought of machines creating works of art, music and literature. There is no doubt that, as has often happened in the past during previous waves of technological advances, AI platforms – and especially, machine learning – have brought with them new opportunities as well as challenges. Machine learning is an AI application enabling programs to learn and progress automatically from experience. Its main feature is accessing data and often using it for the purpose of creating outputs, including music, literature, movies and art. Amounts of data are observed and analysed by the machine, which enables the latter to learn and then make creative decisions leading to final outputs that, as precise works of art, are often not foreseeable by the people who developed and started the initial program. Such a process is characterised by the absence of substantial human intervention or assistance after the program is operated, and by using algorithms – namely a sequence of instructions aimed at solving a problem or performing a computation. This can be labelled ‘algorithmic creativity’, that is, the way by which AI creates new works. There is no doubt, therefore, that AI technologies can create tangible and intangible outputs. If such musical, literary and artistic expressions were created by humans, no one would object to them being considered as copyright works. AI-created works are certainly capable of captivating an audience and stimulating emotions in a similar fashion to works of music, literature or art produced by human beings;1 some AI-created works have even been accorded an economic value in the art market. Yet, as machines have learned to mimic human creativity, the copyright world has accordingly entered an AI-driven uncharted territory.2 Whether AI-generated works can be protected This chapter further develops some parts of a previous article published by Enrico Bonadio and Luke McDonagh, ‘Artificial Intelligence as Producer and Consumer of Copyright Works: Evaluating the Consequences of Algorithmic Creativity’ (2020) 2 Intellectual Property Quarterly 112–37. 1 J Grubow, ‘O.K. Computer: The Devolution of Human Creativity and Granting Musical Copyrights to Artificially Intelligent Joint Authors’ (2018) 40 Cardozo Law Review 409 (noting, with specific reference to music, that ‘Al learns the notes, rhythms, and other musical elements of each work, it assigns weights to them until it can accurately predict subsequent notes and rhythms within a genre. Each note output is a subsequent input for generating a musical phrase. The weights, linked to specific neurons and layers of the neural network, resemble human emotions when we hear music we like – chemical interaction between two neurons fire, triggering the release of pleasant-feeling hormones.’ 2 It is also worth noting that many algorithms are trained using human labour; P Tubaro, A Casilli and M Coville, ‘The Trainer, the Verifier, the Imitator: Three Ways in Which Human Platform Workers Support Artificial Intelligence’ (2020) 7 Big Data & Society 1.

73 https://doi.org/10.1017/9781108989527.007 Published online by Cambridge University Press

74

Enrico Bonadio, Luke McDonagh, Plamen Dinev

by copyright is doubtful in several jurisdictions, largely due to the existence of core substantive criteria (such as originality and authorship) which have traditionally been associated with human creativity. Indeed, there is no scholarly consensus on whether copyright – still anchored to romantic views focused on the centrality of the human author – is an appropriate framework to regulate machine-produced output in the first place. In this chapter we consider the way copyright laws in key jurisdictions (including the USA and EU) deal with AI-generated works at present. We also examine the UK’s legislation on point – the world’s first copyright provision specifically designed to address the advent of (and encourage investment in) computer-generated works3 – before considering alternative solutions, such as leaving AI-generated output in the public domain; or introducing a new, narrow and time-limited sui generis right which aims to strike a fair balance between incentivising investment in AI and the need to safeguard human creativity. IV.II AUTHORSHIP, ORIGINALITY AND THE NEED FOR HUMAN INVOLVEMENT

The question of ‘who is the author’ of the final machine-generated output(s) is crucial because under most copyright regimes around the world the author of the work is also the first owner of the copyright.4 Traditionally, authors have (inevitably) been human, and copyright’s rationales for recognising, rewarding and incentivising creativity through intellectual property have human authors in mind. By contrast, AI enables programmers and/or users (who might turn into authors and artists in their own right) to use a system that is self-contained enough in key decision-making abilities to operate autonomously.5 Therefore, one could argue that automation in this field – in the form of a machine takeover of the creative process from humans – is antithetical to the concept of authorship underpinned by copyright.

3

4 5

House of Lords Debate, Hansard, ‘Copyright, Designs and Patents Bill Hl’ (Volume 489: debated on Thursday 12 November 1987), available at accessed 15 September 2021. Ownership gives the author exclusive rights, e.g., in relation to sales, licences or other forms of control of the work. A Bridy, ‘The Evolution of Authorship’ (2016) 39 Columbia Journal of Law & the Arts 395, 395. That machines can be truly creative and autonomous is not unanimously accepted. Several commentators stress that human beings must still be considered in control of the whole AI-generative process; see, for example, Ginsburg and Budiardjo, ‘Authors and Machines’ (2018) 34 Berkeley Technology Law Journal 343, 344, 401–05 (noting that ‘[n]o machine is itself a source of creativity’ and that every action of a machine is ‘the product of the precise articulation of commands by a human programmer or machine operator’. These authors believe that today’s generative machines are at best ‘faithful’ and ‘obedient’ agents of the humans who interact with them, and that execute specific instructions from them. A similar view is voiced by Hedrick, ‘I Think, Therefore I Create’ (2019) 8 NYU Journal of Intellectual Property & Entertainment Law 324, 365, arguing that ‘there are human programmers and users who write the algorithm’s code, set the objective functions and other parameters of the algorithm, and decide whether the algorithm is creating the desired outputs or whether it ought to be tweaked. These humans are masterminding the creative process; even complex Al models are simply following the humans’ commands (or at least creative guidelines, criteria, and rules)’ (332). Hedrick also argues that ‘[e]ven when an algorithm generates something H-creative (‘historically creative’, i.e., never before created by humans), such creativity is the result of the instructions and capabilities programmed by its creator and is therefore dictated by the (creative) choices of the programmer or user’ (339). She further notes that ‘[the] programmer or user therefore “superintends” and “masterminds” the work of the algorithm, providing it with parameters that guide its functionality and data that determines its trajectory’ (353); and that ‘algorithms can be programmed to exhibit apparent creativity as the result of built-in randomness and other rules, including commands to break certain rules in order to create more unique works. However, that creativity is still the result of those rules and of the creative choices made by the programmer and the user’ (359).

https://doi.org/10.1017/9781108989527.007 Published online by Cambridge University Press

Copyright in Works Created by Artificial Intelligence

75

The requirement for authors/artists to be human is an assumption common in many copyright regimes.6 This principle is enshrined in international treaties such as the Universal Declaration of Human Rights and the International Covenant on Economic, Social and Cultural Rights. The former states that ‘everyone’ has the right to benefit from the protection of the moral and material interests resulting from their works (Art. 27 UDHR; Art. 15 ICESCR). The word ‘everyone’ clearly suggests a human element. Furthermore, several national laws in Europe limit authorship to natural persons. Spanish law, for example, provides that the author is the natural person creating the work;7 French law suggests that only a natural person can be the author;8 and German law provides that copyright protects the author in his or her intellectual/ personal relationship to the work.9 Likewise, although there is no specific human authorship requirement in the US Copyright Act,10 in the Compendium of its practices the US Copyright Office emphasises the importance of the human element in the creative process; indeed, it only registers an original work of authorship ‘provided that the work was created by a human being’.11 The US Copyright Office is even clearer when it comes to works generated by new technologies: ‘the Office will not register works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from a human author’.12 Although these provisions do not have a binding effect,13 they nonetheless demonstrate that the US copyright system takes an unfavourable view towards protecting AI- and machine-created works via copyright. This attitude has been visible since the early days of computer technological advance. When in 1956 the US mathematicians Martin Klein and Douglas Bolitho tried to register the computergenerated song ‘Push Button Bertha’, the US Copyright Office refused the registration, adding that no one had ever registered music created by a machine;14 and in 1964 the same office refused to register a design for a tile floor because it had been produced by a machine using random geometric patterns, asserting that the design did not constitute the ‘writing of an author’.15 A similar outcome was reached more recently by the Australian courts. In Acohs Pty Ltd v Ucorp Pty Ltd, the Federal Court of Australia considered that the underlying HTML code for information sheets generated by a computer program did not have any author, and therefore could not be protected by copyright.16

6

7

8

9 10

11 12 13

14 15

16

S Ricketson, ‘People or Machines: The Berne Convention and the Changing Concept of Authorship’ (1991–1992) 16 Columbia Journal of Law & the Arts 1, 8 and 11 (adding that ‘[t]here must still be evident some human contribution to the form of the work for which protection is claimed. To put it crudely, the work must not be generated by a machine or be the result of some organized industrial undertaking wherein it is impossible to identify an individual human creator or creators’). See Ley 22/11 sobre la Propiedad Intelectual de 1987, Preambulo: ‘los derechos que corresponden al autor, que es quien realiza la tarea puramente humana y personal de creacion de la obra y que, por lo mismo, constituyen el nucleo esencial del objeto de la presente Ley’. French Code de la Propriété Intellectuelle art L112–1. The French Code of intellectual property defines copyrightable subject-matter as ‘oeuvres de l’esprit’. German Copyright Act (Urheberrechtsgesetz, UrhG) art11. This was confirmed by a dictum in Urantia Foundation v Maaherra, 895 F Supp 1337 (D. Ariz. 1995) at [957] (dealing with copyright in a work supposedly created by celestial voices). Compendium of US Copyright Office Practices §101 3rd edn (2017) 306. Compendium of US Copyright Office Practices §101 (2017) 313. The Compendium of US Copyright Office Practices is a manual published by the Office, used by staff as a guide to the Office’s policies and procedures. See A Bridy, ‘The Evolution of Authorship’ (2016) 39 Columbia Journal of Law & the Arts 395, 395–97. US Copyright Office, Register of Copyright, 67th Annual Report of the Register of Copyrights (1964)7–8; see also Hedrick, ‘I Think, Therefore I Create’ (2019) 8 NYU Journal of Intellectual Property & Entertainment Law. 324, 365. Acohs Pty Ltd v Ucorp Pty Ltd [2012] FCAFC 16.

https://doi.org/10.1017/9781108989527.007 Published online by Cambridge University Press

76

Enrico Bonadio, Luke McDonagh, Plamen Dinev

What about Europe? Under European Union (EU) law there is no explicit authorship requirement; the concept of authorship in EU copyright law is linked to the originality requirement. In Infopaq, the latter has been interpreted by the Court of Justice of the European Union (CJEU) as requiring the work to be the ‘author’s own intellectual creation’.17 And in Painer, the CJEU clarified that an intellectual creation is an author’s own if it reflects their personality.18 This would be the case, the court added, if the author were able to express free and creative choices, that is, a ‘personal touch’. This suggests that the originality requirement involves some degree of human authorship. This point is reinforced by a remark by Advocate General Trstenjak in his Opinion in Painer: ‘only human creations are . . . protected’.19 This statement does not bode well for the proposition that works created by a machine should be considered original and therefore copyright; rather, the view of AG Trstenjak confirms the personal and anthropocentric nature of the EU copyright regime – a characteristic embedded within the civil law countries of continental Europe that have historically focused on the droit d’auteur approach to authorship as emblematic of human creativity and personality. IV.II.A Adopting an Objective Originality Standard: A Futureproof Solution? Generally, the originality requirement may be framed either subjectively or objectively.20 Under the former approach, the emphasis is on the author’s intent to create an original work. This assessment focuses on their personal characteristics, choices and subjective feelings. In contrast, objective originality is more concerned with the audience’s perception of the result. This latter approach is based on an external, general (common) standard of human creativity and intellectual potential (as opposed to the process itself).21 The CJEU’s originality standard, for instance, and its emphasis on the creativity involved in producing the work (the author must stamp the work with a ‘personal touch’ which reflects their free and creative choices) is evidently centred on a subjective interpretation. That there is a basic assumption under EU copyright law that the author of any copyright work must be human raises the question of whether having such a narrow and anthropocentric standard is desirable and, crucially, futureproof.22 Drawing on (inter alia) philosophy literature, de Cock Buning argues that in traditional philosophy, creativity is defined as the ability to create something that is original, valuable and surprising.23 This typically requires intent, desire or belief – attributes that would be absent in, for example, machines. This definition, the argument goes, is largely based 17 18 19

20

21 22

23

Infopaq International A/S v Danske Dagblades Forening (C-5/08) EU:C:2009:465; [2012] Bus LR 102 at [35]. Painer v Standard Verlags GmbH (C-145/10) EU:C:2011:798; [2012 ECDR 6] at [88]. Painer v Standard Verlags GmbH, Opinion of AG Trsteniak (12 April 2011) (C-5/08) EU:C:2009:465; [2011] ECDR 13 at [121]. See generally M Jovanovic, ‘The Originality Requirement in EU and US: Different Approaches and Implementation in Practice’ (2020) available at accessed 15 September 2021. ibid.. M De Cock Buning ‘Autonomous Intelligent Systems as Creative Agents under the EU Framework for Intellectual Property’ (2016) 7 European Journal of Risk Regulation 321; see also R Denicola, ‘Ex Machina: Copyright Protection for Computer-Generated Works’ (2016) 69 Rutgers Law Review 251, 286 (arguing, from a different (authorship) angle and the perspective of US law, that ‘if a user’s interaction with a computer prompts it to generate its own expression, the result is excluded from copyright. This is a tenuous and ultimately unnecessary and counter-productive distinction. It denies the incentive of copyright to an increasingly large group of works that are indistinguishable in substance and public value from works created by human beings’). B Gaut, ‘The Philosophy of Creativity’ (2010) 5 Philosophy Compass 1034; P Carruthers, The Architecture of the Mind (OUP 2006) as cited in M De Cock Buning ‘Autonomous Intelligent Systems as Creative Agents under the EU framework for Intellectual Property’ (2016) 7 European Journal of Risk Regulation.

https://doi.org/10.1017/9781108989527.007 Published online by Cambridge University Press

Copyright in Works Created by Artificial Intelligence

77

on a dualistic view of the ‘mind–body’ problem whereby the (non-physical) mind is clearly separated from the (physical) body, a notion that also appears to be reflected in the CJEU’s originality standard (which emphasises the non-physical).24 However, modern philosophers tend to adopt a monistic approach where the mind and body are viewed as a whole, whereby all behaviour, including creativity, belongs to the brain, an interpretation which is also supported by modern neuroscience and computer science.25 Viewing creativity through this monistic angle, the emphasis is instead on the final output of the creative process, as opposed to the personal and conscious touch of the author. De Cock Buning thus argues that if judges and juries are to follow a modern view of creativity and no longer focus on the author’s personal choices and feelings – but rather on the work they have created irrespective of the actual process – ‘the result of machine creativity could be compared to the result of human creativity objectively, without ‘prejudice’’’.26 In other words, what would be required to assess computer-produced works as ‘original’ is an objective interpretation of originality which focuses on the final output per se – regardless of whether there has been a human being involved in its generation. This argument has also been raised in the context of US law by Yaniski-Ravid and Velez-Hernandez who argue that adopting an objective standard would allow judges to assess works made not only by humans (who act with intention) but also those by creative robots ‘for which it remains difficult to understand the concept of consciousness and intention’.27 In a case in Shenzhen, for instance, a Chinese court that considered a series of AI-produced news articles to be copyright appears to have taken this very approach.28 This type of assessment effectively focuses on the external relationship between the expression, the way the audience sees the work and the resemblance of the expression to other works.29 The interpretation undertaken by a judge or jury, considering the field of art, music or literature and the public to which it is addressed, as well as its closeness to pre-existing output, would be the crucial factors for the purpose of determining originality. This would evidently shift the focus from a subjective intention of the human author (which is obviously absent in circumstances where robots create independently or with a considerable degree of autonomy) towards the objective opinion of viewers or listeners of the work. This approach could pave the way for the protection of modern AI-generated artworks such as Jason Allen’s ‘Théâtre D’opéra Spatial’ – created through Midjourney (AI that produces images based on textual description, similar to OpenAI’s DALL-E 2) – which won the top prize at the Colorado State Fair’s fine art competition in September 2022.30 Yet, as we have seen, the current EU originality requirement requires a human touch, which inevitably makes any objective interpretation difficult, if not impossible; and clearly does not fit

24

25 26

27

28

29

30

M De Cock Buning ‘Autonomous Intelligent Systems as Creative Agents under the EU Framework for Intellectual Property’ (2016) 7 European Journal of Risk Regulation 315, 321. B Turner, The Body and Society: Explorations in Social Theory (Sage 2008) 78. M De Cock Buning ‘Autonomous Intelligent Systems as Creative Agents under the EU Framework for Intellectual Property’ (2016) 7 European Journal of Risk Regulation 315 (points out that this approach is also supported by leading computer scientists, including the creator of the Painting Fool); see G Ritchie, ‘Some Empirical Criteria for Attributing Creativity to a Computer Program’ (2007) 17 Minds & Machines 67–99. S Yaniski-Ravid and L Velez-Hernandez, ‘Copyrightability of Artworks Produced by Creative Robots and Originality: The Formality-Objective Model’ (2018) 19 Minnesota Journal of Law, Science & Technology 1, 33–35, 41–42. See Z Yangfei, ‘Court Rules AI-Written Article Has Copyright’ (9 January 2020), China Daily, accessed 8 September 2022. S Yaniski-Ravid and L Velez-Hernandez, ‘Copyrightability of Artworks Produced by Creative Robots and Originality: The Formality-Objective Model’ (2018) 19 Minnesota Journal of Law, Science & Technology 1. B Fernando, ‘How an AI-Generated Artwork Won the Top Prize in a US Art Competition’ (4 September 2022) accessed 8 September 2022 (while the competition judges were reportedly unaware that Midjourney is an AI program, they stated that the winner would have been the same even if they had known at the time).

https://doi.org/10.1017/9781108989527.007 Published online by Cambridge University Press

78

Enrico Bonadio, Luke McDonagh, Plamen Dinev

into the current EU acquis on the test. In other words, what would be required is a major paradigm shift in EU legislation and case law. Alternatively, this could be accommodated through the creation of a new sui generis right – specifically designed to deal with machinegenerated output and having its own originality standard – a point to which we return later. IV.II.B Protecting Investment in AI: The UK’s ‘Pragmatic’ Approach A pragmatic approach to dealing with works created by machines is visible in the UK Copyright, Designs and Patents Act 1988 (CDPA), s. 9(3) of which provides that: ‘[i]n the case of a literary, dramatic, musical or artistic work which is computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken’.31 This provision was first proposed in the late 1980s. The House of Lords recognised that, following technological advances since 1956 (when the former Copyright Act was enacted), there may now be instances where computers produce works to which it is not possible to ascribe human authorship.32 The provision was specifically designed to go beyond merely protecting works generated by using a computer as a tool or ‘clever pencil’ – it is clear that the House was cautious of advances in artificial intelligence in particular.33 In terms of the rationale behind the provision, investment in AI was a key priority: ‘the far-sighted incorporation of computergenerated works in our copyright system will allow investment in artificial intelligence systems, in the future, to be made with confidence’ (per the Earl of Stockton).34 Indeed, while copyright law would have little meaning as an incentive mechanism from the perspective of a machine – which would presumably lack consciousness (as of now) and thus produce works irrespective of the existence of exclusive rights or other rewards – having an explicit provision which covers computer-generated output will almost certainly encourage humans to invest in AI and thus produce socially valuable works.35 In terms of specific provisions, CPDA defines a computer-generated work as one created by a computer in circumstances where there is no human author.36 S.9(3) CDPA essentially introduces a legal fiction. It considers the author a person who has not directly created the work but has merely made the necessary arrangements for such production to take place. In this sense the provision expands the notion of author,37 considering the objective creation of the 31

32

33

34

35

36 37

This provision just applies in cases of literary, dramatic, musical or artistic works generated by computers. It seems thus to suggest that other types of works that may be produced by machines such as movies, software or databases, are not copyrightable instead: see A Chiabotto, ‘Intellectual Property Rights over Non-human Generated Creations’ (2017) available at . House of Lords Debate, Hansard, ‘Copyright, Designs and Patents Bill Hl’ (Volume 489: debated on Thursday 12 November 1987) available at accessed 15 September 2021. See also UK Intellectual Property Office, ‘Artificial Intelligence Call for Views: Copyright and Related Rights’ (23 March 2021) available at accessed 15 September 2021. House of Lords Debate, Hansard, ‘Copyright, Designs and Patents Bill Hl’ (Volume 489: debated on Thursday 12 November 1987), available at accessed 15 September 2021. See also UK Intellectual Property Office, ‘Artificial Intelligence Call for Views: Copyright and Related Rights’ (23 March 2021) available at accessed 15 September 2021. See Copyright, Designs and Patents Act 1988 s.178. The need for an expansion of the concept of author has been acknowledged for long time in legal scholarship, including in the USA since the 1980s: see T Butler, ‘Can a Computer be an Author?’ (1982) 4 Hastings Communication & Entertainment Law Journal 707, 744–45 (noting that ‘[w]hen courts find that a given product

https://doi.org/10.1017/9781108989527.007 Published online by Cambridge University Press

Copyright in Works Created by Artificial Intelligence

79

output, and then locating the most plausible nearby ‘author’ (and owner), which could also be a company.38 This provision could plausibly be extended by courts to cover AI-produced output. One criticism of doing so is that it would set the stage for an expansion of corporate ownership of copyright works, something evident in the decision of the Chinese court in Shenzhen, referenced above.39 S.9(3) CDPA has been criticised for not clearly addressing the issue of originality (and for not being compliant with the EU acquis).40 Indeed, this provision could be viewed as an exception to the originality requirement as it has been historically interpreted by UK courts.41 Traditionally under UK law a work was considered original if it was the result of its author’s own ‘skill, labour and judgement’ – though post-Infopaq the ‘intellectual creation’ test has been used by UK courts. In any event, in AI scenarios the works produced (by a machine) do not directly originate from any author employing ‘skill and labour’ – or indeed ‘intellectual creativity’ (unless the machine itself is considered an author).42 The legal fiction created by s.9(3) CDPA, however, is not an isolated phenomenon in UK copyright law. S. 9(2) CDPA also considers as ‘author’ the producer of a sound recording, the producer of a film (together with the principal director), the person making the broadcast and the publisher of a typographical arrangement of a published edition.43 And such ‘authors’ (e.g., producers) are often tied via contract to corporate entities (e.g., record companies).44 There is indeed continued heated debate on the issue of originality and computer-generated works under s.9(3) CDPA. Goold, for instance, argues that while the legislative history behind the provision suggests that such works are still subject to originality, this interpretation is inherently contradictory.45 If an AI-generated work is to pass the requirement, the argument goes, there must presumably be a human supplying the necessary originality. If this is the case,

38

39

40

41

42

43

44

45

of AI software is authored by machine rather than a person, the court should presume the existence of a fictional human author’). Also, the CDPA has opted to use the term ‘author’ when referring to the creator of a computer-generated work rather than, for example, the more neutral word ‘maker’, which is used both in the EU Database Directive to define the person that comes up with a database (art 7.1). One may thus wonder whether the UK CDPA has tried to ‘humanise’ computer-generated works by avoiding using impersonal and neutral terminology; see also Hart, ‘Author’s Own Intellectual Creation’ (1993) 9 Computer Law and Security Review 164, 165. Z Yangfei, ‘Court Rules AI-Written Article Has Copyright’ (9 January 2020), China Daily, accessed 18 March 2020. See M Iglesias et al., ‘Intellectual Property and Artificial Intelligence: A Literature Review’ (Publications Office of the EU 2019) at 15, and the sources cited therein, accessed March 2020. A Rahmatian, ‘Originality in UK Copyright Law: The Old “Skill and Labour” Doctrine under Pressure’ (2013) 44 International Review of Intellectual Property and Competition Law 4. However, there is some disagreement regarding this issue. See also A Guadamuz, ‘Do Androids Dream of Electric Copyright?’ [2017] Intellectual Property Quarterly 169, 176; Lionel Bently, mentioned by Begoña González Otero and Joao Pedro Quintais, ‘Before the Singularity: Copyright and the Challenges of Artificial Intelligence’ Kluwer Copyright Blog (2018) accessed 18 March 2020 (stressing that the CDPA provisions on computer-generated works do not offer a useful model for protecting AI outputs, because of their incompatibility with the EU copyright acquis and failure to address the issue of originality). See also D Vaver, ‘Translation and Copyright: A Canadian Focus’ (1994) 4 European Intellectual Property Review 159 (noting, in relation to the expansion of the concept of author, that ‘computer-generated works join the list of other works for which the UK Act has created a fictitious author: the producer of a film or sound record, the maker of a broadcast, the provider of a cable service programme – almost all of which are equally fictitious persons, that is, corporations rather than humans’). S Ricketson, ‘People or Machines: The Berne Convention and the Changing Concept of Authorship’ (1991–92) 16 Columbia Journal of Law & the Arts 1, 16. See P Goold, ‘The Curious Case of Computer-generated Works under the Copyright, Designs and Patents Act 1988’ (2021) 2 Intellectual Property Quarterly 120–30, 124.

https://doi.org/10.1017/9781108989527.007 Published online by Cambridge University Press

80

Enrico Bonadio, Luke McDonagh, Plamen Dinev

however, the work will clearly not be ‘computer-generated’ as required by the legislation.46 As of 2021, there has only been one case to specifically interpret s.9(3) CDPA47 which unfortunately has failed to provide an answer to this crucial question.48 The debate, therefore, continues. Finally, while s.9(3) CDPA was the world’s first copyright legislation attempting to deal with computer-generated works,49 it is notable that – more than three decades later – only a handful of other (common law) jurisdictions have followed this approach.50 As discussed above, the original rationale behind the provision was to encourage investment in AI. If there is evidence that the provision is effective and achieves its aims, it would indeed be reasonable to maintain the current law and even duplicate it in other jurisdictions. Guadamuz, for instance, describes the UK’s model as ‘the most efficient’, arguing it could help to ensure that companies keep investing in AI, safe in the knowledge that they will get a return on their investment.51 He also argues that it should be adopted more widely.52 It is nevertheless notable that significant investment in AI (and AI-generated works) has taken place in key jurisdictions such as the USA, the world leader in the field, despite the clear lack of any such provisions in US copyright law.53 It is also not difficult to see why others have taken the opposite view – arguing that not only should other jurisdictions think twice before duplicating this provision, but that the UK should actually consider abolishing it – given that key questions are yet to be answered.54 IV.III EVALUATING THE PUBLIC DOMAIN AND SUI GENERIS RIGHT AS SOLUTIONS

Much of the above discussion has focused on how copyright deals with AI-generated works at present. However, not everyone agrees that the copyright system is an appropriate tool to regulate such output in the first place. Mezei, for instance, argues that copyright’s traditional humancentric approach to authorship should be preserved, given that its core tenets are inextricably and inherently linked to human creativity.55 Until we are certain – the argument goes – that society, as opposed to a few major stakeholders, would benefit from an AI-copyright regime (and

46 47

48

49

50

51 52 53

54

55

ibid. Nova Productions Ltd v Mazooma Games Ltd & Ors [2006] EWHC 24 (Ch) (holding that a computer game player was not the author of screenshots taken while playing the game, as he had not undertaken the relevant arrangements necessary for their creation). Stephens K, ‘Who Owns an AI-generated Invention?’ (December 2019), available at accessed 15 September 2021. House of Lords Debate, Hansard, ‘Copyright, Designs and Patents Bill Hl’ (Volume 489: debated on Thursday 12 November 1987), per Lord Young of Graffham, available at accessed 15 September 2021. Namely, Ireland, New Zealand, India, South Africa and Hong Kong. This provision also inspired the 1990 WIPO Draft Model Copyright Law, which provides that the original owner of the economic rights in a computer-generated work may be either the person or entity ‘by whom or by which the arrangements necessary for the creation of the work are undertaken’ or the person or entity ‘at the initiative and under the responsibility of whom or of which the work is created and disclosed’: see International Bureau of WIPO, Preparatory Document, Draft Model Law on Copyright, 258–59 (No CD/MPC/III2, 30 March 1990).

A Guadamuz, ‘Do Androids Dream of Electric Copyright?’ [2017] Intellectual Property Quarterly 169, 186. UK Intellectual Property Office, ‘Artificial Intelligence Call for Views: Copyright and Related Rights’ (23 March 2021) available at accessed 15 September 2021. P Goold, ‘The Curious Case of Computer-generated Works under the Copyright, Designs and Patents Act 1988’ Intellectual Property Quarterly (2021) 2 120–30, 124 (arguing that it is unclear how the provision fits with originality). P Mezei, ‘From Leonardo to the Next Rembrandt: The Need for AI-Pessimism in the Age of Algorithms’ (2020) accessed 30 May 2020.

https://doi.org/10.1017/9781108989527.007 Published online by Cambridge University Press

Copyright in Works Created by Artificial Intelligence

81

there is convincing economic evidence and policy argument), it would be best not to regulate algorithmic creativity at all.56 Given the existing problems with copyright law in this context (EU law relies on a human-centric subjective approach to originality and even UK law, which has adopted a ‘pragmatic’ approach to deal with this specific issue, has failed to provide concrete answers), we now consider various alternative approaches to regulation, including leaving AI-generated works in the public domain or adopting a new sui generis right. IV.III.A The Public Domain Solution A tempting option could be to deem the works created by machines as automatically entering the public domain. As there is no human author directly involved in the creative production of the output, we could argue that no one should be able to claim exclusive rights over it.57 Works produced by robots would thus be comparable to things found in nature, such as music that the wind generates when it moves through wind chimes or the sounds of a waterfall, or birds singing at dawn58 – outputs which cannot be monopolised by anyone.59 There is also an (obvious) incentiverelated argument that supports this position. Computers and machines are not able to respond to the incentives offered by copyright – and therefore their works should remain in the public domain (at least until technology evolves deeply enough to give machines some sort of human-like consciousness).60 This solution would also neutralise the anti-competitive risks that an over-proliferation of strong and long monopolistic rights, owned by corporate entities and protecting AI produced output, may bring. The scenario would turn even more anti-competitive if a ‘commissioned work’ approach was used in relation to these works – with the programmer/employer considered the author and thus copyright owner. Under these circumstances AI companies may be attracted by the idea of securing ownership of exclusive rights at an unprecedented rate and thus decide to hoard access to their AI technologies, to always remain the ‘employers’ and therefore the rightholders. This would consolidate the dominant position of a few tech companies. The public domain option is not just a theoretical proposal. It is already operative in jurisdictions such as the USA and Australia (where this appears to be the default position for AI-generated works).61 Several scholars support this approach as well.62 But what are the drawbacks of this? Would 56 57

58

59

60

61

62

ibid. D Gervais, ‘The Machine As Author’ 105 Iowa Law Review (2019) 2053 2062 (noting that ‘machine productions are not protectible by copyright once the machine has crossed what the Article calls the autonomy threshold and is no longer a tool in the user’s hands or a reflection of its (human-made) program’). A Khoury, ‘Intellectual Property Rights for Hubots: On the Legal Implications of Human-like Robots as Innovators and Creators’ (2016–2017) 35 Cardozo Arts & Entertainment Law Journal 668. D Lim, ‘AI & IP: Innovation & Creativity in an Age of Accelerated Change’ (2019) 52 Akron Law Review 813 840 (noting how according to several commentators the fall into the public domain of these works will provide fertiliser ‘to give birth to new artistic genres and whole new areas of innovation, where humans could build freely upon initial machine-output’). R Clifford, ‘Intellectual Property in the Era of the Creative Computer Program’ (1997) 71 Tulane Law Review 1675 702–03. In February 2022, the US Copyright Office rejected Stephen Thaler’s second request to register a copyright claim in a work reportedly created by artificial intelligence, named ‘A Recent Entrance to Paradise’. See A Robertson, ‘The US Copyright Office Says an AI Can’t Copyright Its Art’ (21 February 2022) accessed 8 September 2022. See for example Bently L, mentioned by B Otero and J Quintais, ‘Before the Singularity: Copyright and the Challenges of Artificial Intelligence’ Kluwer Copyright Blog (2018) accessed 18 March 2020; M Perry and T Margoni, ‘From

https://doi.org/10.1017/9781108989527.007 Published online by Cambridge University Press

82

Enrico Bonadio, Luke McDonagh, Plamen Dinev

the refusal to protect algorithmic creativity discourage investments in, and dissemination of, these creative technologies? AI is a resource-intensive field, which raises the need to secure some sort of exclusive rights in the products of this investment. In this manner, some scholars argue that denying legal protection of machine-generated outputs may not have the effect of increasing the public domain in the long run; in this view, this would have the effect of reducing the incentives to create new AI works, and thus may ultimately lead to a lower number of these outputs, and accordingly a decrease in works that would eventually enter the public domain.63 Similarly, some scholars raise the speculative prospect that the arts, education, medicine, technology, among others, could suffer, resulting in loss of investments into valuable research and future AI applications.64 On the other hand, given that other forms of protection – for example, copyright and, in some cases, patents – are available for the programs themselves, even if not the outputs, one may argue that a sufficient incentive already exists. Regardless of these speculative arguments, there is little doubt that the debate will soon reach the legislative field, with all the attendant lobbying that takes place. IV.III.B A Sui Generis Right We have seen that at present copyright regimes may not be fit to accommodate what is produced by algorithmic creativity; and indeed, copyright laws in several jurisdictions do not explicitly protect machine-created works.65 Yet, whether we agree or not, corporate entities may soon lobby for a form of exclusive rights to protect the final outputs of machine-driven processes. Could an acceptable compromise be offered by a sui generis system – a kind of protection that could incentivise the development of and use of AI creative platforms while at the same time safeguarding human ingenuity?66 The benefit of a sui generis regime (as opposed to using the strong conventional copyright to protect such works) would be that rightholders could be given only a thin scope of

63 64 65 66

Music Tracks to Google Maps: Who Owns Computer Generated Works?’ (2010) 26 Computer Law and Security Review 621; D Gervais, ‘The Machine As Author’ 105 Iowa Law Review (2019) 2053; G Gabison, ‘Who Holds the Right to Exclude for Machine Work Products? (2020) 1 Intellectual Property Quarterly. Available at SSRN accessed 28 June 2020; M Kop, ‘AI & Intellectual Property: Towards an Articulated Public Domain’ (2019) available at accessed 28 June 2020. K Hristov, ‘Artificial Intelligence and the Copyright Dilemma’ (2016) 57 IDEA: IP Law Review 431 439. K Hristov, ‘Artificial Intelligence and the Copyright Dilemma’ (2016) 57 IDEA: IP Law Review 431 439. The lack of human authorship/originality is a key factor in this regard. A Lauber-Rönsberg and S Hetmank, ‘The Concept of Authorship under Pressure: Does Artificial Intelligence Shift Paradigms?’ (2019) 14 Journal of Intellectual Property Law & Practice 509, 577. See also S García, ‘Las obras creadas por sistemas de inteligencia artificial y su protección por derecho de autor’ [2019] InDret, Revista para el análisis del derecho; Ramalho, ‘Will Robots Rule the (Artistic) World?’ (2017) 21 Journal of Internet Law 12, 16–20 (proposing to adopt a (i) regime similar to the EU database right under the EU Database Directive (Directive 1996/9), which notoriously aims at protecting investments made in producing compilation of data; or (ii) a sort of ‘disseminator’s right’ comparable to the publisher’s right in the publication of previously unpublished works provided by the EU Term of Protection Directive (Directive 93/98), replaced by Directive 2006/116). The second proposal is more concerned with enhancing the accessibility of AI-produced content: while such a right would be inherently economic, it would allow the rightholder to extract value out of the creations and at the same time aim at encouraging the dissemination of algorithmic output. For academic opinions against sui generis systems of protection of AI-generated outputs, see P Goold, ‘The Curious Case of Computer-generated Works under the Copyright, Designs and Patents Act 1988’ (2021) Intellectual Property Quarterly; see also P Mezei, ‘Intervention to WIPO’s AI and IP Conversation’ (23 June 2020) accessed 28 June 2020 (expressing a ‘pessimistic view’ on copyright protection of AI-generated output and arguing that the maximum protection such output should receive would be a new sui generis regime, ‘the scope of which can only be set following a deep empirical and multi-disciplinary analysis of the marketability of outputs (economic analysis) as well as the individual and cultural/ social consequences (cultural analysis) of an AI-centred protection upon an human-centred regime, where the ultimate goal shall remain that copyright law aims the development of individual humans and the society/humankind in general’).

https://doi.org/10.1017/9781108989527.007 Published online by Cambridge University Press

Copyright in Works Created by Artificial Intelligence

83

protection, allowing them to prevent others from exploiting exact copies of the machine-generated work. In this view, such a right would essentially give protection against literal copying only.67 As to the length of protection, unlike the typical copyright duration, a very short term could be applied in the AI context, for example, three years from the date of publication.68 Why should a thinner and shorter right be preferred?69 The usual justifications for the long duration in the context of human authors would not apply. In fact, providing AI developers and companies incentives for AI-created works by offering the strong traditional copyright protection (and long duration) may even lead to fewer human-generated works being created in the long run.70 AI creative capacity is potentially both more vast and speedier than human capacity.71 The risk is a devaluation of human intellectual ingenuity and a marginalisation of the human creative potential72 Just as automation threatens (and, over time, eliminates) jobs in manufacturing, AI creativity could threaten the value of human authorship. The provision of a limited sui generis right might neutralise this risk by providing an incentive to recognise the creativity of AI works without giving machines an equal level of protection to humans.73 The fact that only literal copying would be prohibited would leave human creators free to adapt, transform and reinterpret AI-generated works and thus to use them for creative purposes. In fact, such a feature of the proposed sui generis right may fit with the characteristics of many AI machines such as Amper, an AI music composition system.74 Overall, the introduction of a thin and time-limited right could achieve an appropriate balance: while some incentives would still be given to the developers of AI creative technologies via the offer of exclusive rights aimed at preventing the exploitation of the final output by third parties, the reduced scope and duration of the protection will leave human creators with enough freedom and motivation to create. Such a balance would preserve value in human ingenuity (fully protectable by copyright) and at the same time protect machine-produced outputs sufficiently (and thus encourage investments in and use of AI technologies).75 What would the requirement for attracting the sui generis protection be? An originality test as assessed and interpreted objectively and contextually would be appropriate (see the detailed 67

68

69

70

71

72

73

74 75

For a proposal attributing a thin scope of protection to programmers, see also R Yu, ‘The Machine Author: What Level of Copyright Protection Is Appropriate for Fully Independent Computer-Generated Works?’ (2017) 165 University of Pennsylvania Law Review 1245 1268–69. The length of term suggested in a recent AIPPI ‘Study Question’ by Dutch delegates; see AIPPI Summary Report: 2019 Study Question on Copyright/Data Copyright in Artificially Generated Works 7, accessed 19 March 2020; E Bonadio and L McDonagh, ‘Artificial Intelligence as Producer and Consumer of Copyright Works: Evaluating the Consequences of Algorithmic Creativity’ (2020) 2 Intellectual Property Quarterly. S. 12(7) of the CDPA states that copyright in computer-generated works ‘expires at the end of the period of 50 years from the end of the calendar year in which the work was made’. D Gervais, ‘The Machine As Author’ 105 Iowa Law Review (2019) 2053. On intrinsic motivations as factors which stimulate creativity, see C Buccafusco and C Sprigman, ‘Experiments in Intellectual Property’ in P Menell and D Schwartz (eds) Research Handbook on the Economics of Intellectual Property Law (Edward Elgar 2016). AIPPI Summary Report 7–8 accessed 19 March 2020 (French Group position). AIPPI Summary Report 13 accessed 19 March 2020 (UK Group position). See again AIPPI Summary Report 17 (German Group) (noting that a shorter term of protection is justified in light of ‘the reduction in costs by using the AI in generating the works . . . to safeguard the rights of traditional authors from being replaced by the cheaper labour of AI’). J Grubow, ‘O.K. Computer’ (2018) 40 Cardozo Law Review 409, 416. AIPPI Summary Report: 2019 Study Question on Copyright/Data Copyright in Artificially Generated Works 13 (UK Group position).

https://doi.org/10.1017/9781108989527.007 Published online by Cambridge University Press

84

Enrico Bonadio, Luke McDonagh, Plamen Dinev

discussion on originality above).76 Judges could consider the work’s aesthetic, literary or artistic similarity to existing works when considering originality, in the sense of being sufficiently distinguishable from prior works. While this would necessarily differentiate the originality standards between human-created works and AI works, it is worth noting that the requirements for attracting protection can vary even depending on the kind of output. Under UK law, for example, while graphic works, photographs, sculptures and collages attract copyright based on originality ‘irrespective of artistic quality’, other fruits of human creativity like works of artistic craftsmanship are protected only if they reach a certain threshold.77 There is no doubt that protecting AI-generated works will shift the focus from the subjective element of traditional creative processes (the centrality of the human author) to the objective outputs produced by machines, thus changing the emphasis from authors to works.78 Yet, as mentioned, the negative effects of such a shift (in terms of a progressive marginalisation of human ingenuity) could be contained, if not almost fully neutralised, if a proper balance between the above needs is found. Who would the owner of such a sui generis right be? The UK AIPPI Group identifies two alternative approaches, that is, the proximity and the investment approaches.79 By using the former criterion, the owner could be: (i) the natural or legal person that is most closely associated with the creative output – for example the person who comes up with the code (coder); or (ii) the person who identifies the objective to be reached (goal selector); or (iii) the person who chooses the input data (data selector); or (iv) the person who trains the AI (trainer); or (v) the person who carries out a qualitative or aesthetic selection of a work from a number of new artificially generated works (output selector).80 Unsurprisingly, the UK Group of AIPPI favours the investment approach on the basis of legal certainty, thus clearly arguing for corporate ownership, that is, that the natural or legal person who invests in the project should be considered the owner of the sui generis right.81 EU law seems to support this approach. The EU Info-Society Directive in Recital 5 recognises that: ‘[t]echnological development has multiplied and diversified the vectors for creation, production and exploitation. While no new concepts for the protection of intellectual property are needed, the current law on copyright and related rights should be adapted and supplemented to respond adequately to economic realities’ (emphasis added). Recital 5 mentions ‘related rights’ as referring to rights protecting outputs such as cinematographic works, sound recordings and broadcasts. And a sui generis right protecting AI-created works might fit well into the ‘related right’ category that aims at incentivising investments in crucially relevant technological fields.82 76 77 78

79 80 81

82

AIPPI Summary Report 16 (UK and Singapore Groups position). CDPA s4(1) s4(1). R Denicola, ‘Ex Machina: Copyright Protection for Computer-Generated Works’ (2016) 69 Rutgers Law Review 251, 270. See also de Cock Buning, ‘Autonomous Intelligent Systems as Creative Agents under the EU framework for Intellectual Property’ (2016) 7 European Journal of Risk Regulation 310 (noting that ‘if courts no longer assess the author, but rather the work he created, regardless of the process by which he came by it, the result of machine creativity could be compared to the result of human creativity objectively, without ‘prejudice’). AIPPI Summary Report 16–17 (UK Group position). ibid. ibid. Finally, no moral rights should follow the introduction of this sui generis form of protection. After all, UK copyright law currently excludes the applicability of moral rights to computer-generated works. On the inappropriateness of the proximity approach when it comes to attributing copyright in AI-generated artworks, see also Megan Svedman ‘Artificial Creativity: A Case Against Copyright for AI-Created Visual Artwork’ (2020) 1 Intellectual Property Theory 5. See also T Bond and S Blair, ‘Artificial Intelligence & Copyright: Section 9(3) or Authorship without an Author’ (2019) 14 Journal of Intellectual Property Law & Practice 423 (noting that the solution may lie in recognising computer-created works as deserving of only economic rights similar to those offered to movies, broadcasts, sound recordings and typographical arrangements). After all, a sui generis approach had been advocated by Karl Milde back in 1969: see Milde, ‘Can a Computer Be an “Author” or an “Inventor”?’ (1969) 51 Journal of the Patent and Trademark Office Society 378, 402 (stressing ‘the necessity of a new type of protection, sui generis to the problem of computer talk, which would also make practical the enforcement of any right granted’).

https://doi.org/10.1017/9781108989527.007 Published online by Cambridge University Press

Copyright in Works Created by Artificial Intelligence

85

There are also arguments that caution against the introduction of a sui generis system. Creating sui generis laws to accommodate the needs of a certain sector or industry fails to keep copyright regimes technology-neutral.83 Yet, it could be counter-argued that copyright laws bring with them the seed of ‘differentiation’. Across the decades, these laws have produced rules in relation to different works. For example, the range of copyright works are quite distinct from each other, from literary, artistic and musical works to more entrepreneurial (and investmentdriven) kinds of subject matters including typographical arrangements of published editions, original compilations of data and sound recordings, broadcasts and movies. These are exactly the kinds of works highlighted in other chapters of the present volume. Perhaps the greatest concern with the enactment of a new right would be the danger of increased corporate ownership. This is undoubtedly a worry. However, the above-mentioned Shenzhen case demonstrates that in the absence of a thin sui generis right, corporate entities will attempt to claim the full rights of copyright over AI-created works. Thus, even if there are risks attached to a sui generis system, if it is enacted in a balanced way the risk of ending up consolidating and overprotecting monopolistic rent-seeking power may be significantly reduced, if not completely ruled out. IV.IV CONCLUSION

The debate on whether and how AI-produced works could and should be protected by copyright fits well into the scope of analysis of this book. Indeed, the present volume aims to spark an academic discussion on whether and to what extent modern intellectual property rights should reward entrepreneurial investments over human creativity and ingenuity. It is clear that provisions on copyright in computer-generated works (which also accommodates output created by AI) go in that direction. While these rules cannot incentivise and reward machines, they can encourage and recompense companies which invest time and money to develop AI/robots capable of generating art, music and literature. In this view, what S.9(3) CDPA achieves is to promote and protect investments in new technologies. Should a sui generis right of the type recommended in this article be finally adopted by some countries, such a regime would follow a similar line of thinking. It would be a right aimed at incentivising and rewarding investors in AIcreative technologies. It would however provide a guarantee which provisions like S.9(3) CDPA cannot give: that is, a much-needed balance between encouraging investments in algorithmic creativity and safeguarding human ingenuity. The UK approach of granting machine-generated works copyright protection (which lasts for 50 years after the death of the fictional author) goes too far,84 offering too high a level of protection. Indeed, a more limited protection for this type of work would significantly reduce (if not eliminate) risks of a futuristic (and scary) scenario where robotic creativity – even that which relies on human labour for training of AI – might gradually displace human ingenuity.85

83

84

85

F De Rouck, ‘Moral Rights & AI Environments: The Unique Bond between Intelligent Agents and Their Creations’ (2019) 14 Journal of Intellectual Property Law & Practice 203. Section 12(7) of the CDPA provides that copyright in computer-generated works ‘expires at the end of the period of 50 years from the end of the calendar year in which the work was made’. M Gahntz, ‘The Invisible Workers of the AI era’ towards Data Science (December 2018)

https://doi.org/10.1017/9781108989527.007 Published online by Cambridge University Press

V Plant Variety Protection and Investment Viola Prifti Independent researcher in intellectual property law

V.I INTRODUCTION

The terms “plant variety” and “investment” may at first appear unrelated. While “plant variety” may bring concepts such as “food,” “nature,” and “farming” to our mind, the word “investment” echoes money and profit. In a market economy, all these concepts are interrelated because money is the medium of exchange that measures and gives value to good and services. While plants and their varieties existing in nature are a result of nature’s “creativity,” any kind of improvement made upon these products of nature is due to intellectual and financial investment. Since the dawn of time, farmers have dedicated considerable time and effort to improving seed and domesticating wild varieties. The importance of plant breeding for agriculture and for other industrial sectors led countries to invest in the seed industry and further develop the sector. For example, the Italian agronomist Nazareno Strampelli helped Italy become self-sufficient in wheat production, which was thereafter exported.1 His work was followed by the “Green Revolution,” for which Normal Borlaug was awarded the Nobel Peace Prize for his valuable contributions in the field. Today the value of plant breeding for the economy is apparent. In 2021, the global seed market was valued at $63.0 billion and it is estimated to grow to $86.8 billion by 2026.2 Increased investments seem to be paralleled by market concentration3 and stronger intellectual property (IP) rights, especially for new plant breeding techniques.4 This chapter will explore rights on plant varieties as a means of recouping investments. The following sections will firstly clarify the legal definition of “plant variety” and related rights. Second, it will scrutinize the wording of different laws to show the “investment” intent of the legislator in drafting related IP laws as well as offer evidence of the increased investments in the field. Lastly, it will conclude with a few considerations on laws that can help society advance in view of ethical principles. 1

2

3

4

Lorenzetti, R (2000) La scienza del grano. Nazareno Strampelli e la granicoltura italiana tra le due guerre, Ministero per i Beni e le Attività Culturali, Saggi 58. “Seeds Market by Type (Genetically Modified & Conventional), Trait (Herbicide Tolerance, Insect Resistance), Crop Type (Cereals & Grains, Oilseeds & Pulses, Fruits & Vegetables), and Region: Global Forecast to 2026” by MarketsandMarkets, available at . Accessed 14 July 2021. See also the global seed exports by the International Seed Federation, available at . Accessed 14 July 2021. “Concentration in Seed Markets: Potential Effects and Policy Responses” available at . Accessed 14 July 2021. MA Kock, “Open Intellectual Property Models for Plant Innovations in the Context of New Breeding Technologies” (2021) 11 Agronomy 1218. .

86 https://doi.org/10.1017/9781108989527.008 Published online by Cambridge University Press

Plant Variety Protection and Investment

87

V.II WHAT IS PLANT VARIETY PROTECTION?

Art. 27.3(b) of the Trade-Related Intellectual Property Rights Agreement (TRIPS Agreement) requires its members to “provide for the protection of plant varieties either by patents or by an effective sui generis system or by any combination thereof.” The United States and Australia have in place a patent system for plant varieties and parts thereof, while most countries have mostly adopted the provisions of the International Convention for the Protection of Plants (UPOV Convention)5 on plant varieties in combination with patent rights on plant parts. Some Asian countries, such as India, have adapted the UPOV Convention to serve their national interests by introducing specific provisions in favor of farmers.6 Hence, different systems of protection are available for protecting plant varieties. This chapter will use the terms “plant variety protection” to indicate both these systems. V.II.A The Concept of “Invention” for Plant Varieties Before analyzing the different IP laws on plant varieties, it is worth shedding light on the concept of “plant variety” as used in legal jargon. The TRIPS Agreement does not require countries to provide for rights on plants occurring in nature. These types of plants are heterogeneous and do not reproduce true to type. Plant varieties bred with human assistance, on the other hand, have uniform characteristics that can be transmitted over generations. Plant breeding, indeed, aims to create improved plant varieties that can resist climate change and give higher yields as well as satisfy consumers’ organoleptic and nutritional needs. When homogenous plant varieties retain their properties throughout the years, breeders can cater for a large market and commercialize their seeds. Thus, the TRIPS Agreement intends to grant intellectual property rights to those engaging in plant breeding. To better understand the object of protection, see Table 5.1, which offers an illustration of the taxonomic ranking of potatoes. The UPOV Convention grant rights on the variety, not the plant species or genus. Breeders, indeed, cannot create new plant species, but instead can modify and enrich them to create new varieties of plants with enhanced characteristics. But patent protection can extend to plant species and plant varieties given that patent rights are often granted on DNA sequences. One can certainly table 5.1. Classification of the potato plant/Solanum tuberosum Kingdom Division Class Order Family Genus Species Variety

Plantae Magnoliophyta Magnoliopsida Solanales Solanaceae – potato family Solanum L. – nightshade Solanum americanum Mill. var. americanum and var. nodiflorum

Source: USDA Natural Resources Conservation Service.7

5

6

7

Please, note that UPOV is the acronym for the French version of the Convention. The UPOV text is available at . Accessed 14 July 2021. “The Protection of Plant Varieties and Farmers Rights Act, 2001” available at . Plant taxonomy and their characteristics can be searched via the USDA website available at . Accessed 16 July 2021.

https://doi.org/10.1017/9781108989527.008 Published online by Cambridge University Press

88

Viola Prifti

raise a few questions regarding IP protection of plant varieties. While the legal issues will be clarified in the following paragraphs, here will follow a few brief considerations on moral concerns.8 Morals may stem from various sources. Religious beliefs, for example, consider any attempt to own life forms as immoral. Since God is the only creator of life, it is difficult to envisage humans as creators. Breeders can improve the genetic material embedded in plants, but they can’t create a new plant from scratch. Humans can work only with the elements provided by nature, and in doing so, they need to obey precise laws and energy cycles.9 It is, indeed, the interaction of these energy cycles between the various elements that creates and controls life on Earth. In this view, humans cannot be superior to nature and cannot invent, only innovate. Morals based on law, on the other hand, accept legal fiction when it serves legal purposes and realizes social goals. This means that humans’ efforts to improve plants can be deemed an “invention.” While ancient breeders have undoubtedly applied considerable intellectual efforts to understand and improve nature, the advent of new breeding techniques and rapid technological advancement has made breeding steps obvious. Plant variety laws indeed do not refer to the intellectual ingenuity, but to the efforts of the breeder. Art. 1 (iv) of UPOV, for example, explains that “breeder means the person who bred, or discovered and developed, a variety.” It is apparent in the text that breeders cannot “invent” a variety, but they can “breed” or “discover and develop.” These terms seem to differ from the concept of creativity as the “ability to produce or use original and unusual ideas.”10 Breeding activities follow a ladder innovative process, where any kind of plant improvement builds upon prior breeding activities. The US Plant Patent Act (US PPA), however, considers plant breeding activities as “invention.” Patents are granted to “Whoever invents or discovers and asexually reproduces any distinct and new variety of plant, including cultivated sports, mutants, hybrids, and newly found seedlings, other than a tuber propagated plant or a plant found in an uncultivated state, may obtain a patent therefor, subject to the conditions and requirements of this title.”11 The use of the term “invent” implies the use of thought for creating something new, “something that has never been made before.”12 The US PPA provides no explanations on the meaning of the term. It may, thus, be correct to suggest that plants “can be protected by patents if the genetic change – or the combination of changes – does not pre-exist in nature.”13 This argument can nevertheless be rebutted if one thinks that scientists cannot create the genome from scratch. The provision also refers to “discovers and asexually reproduces.” These words aim to cover those situations when a breeder finds a new plant and further breeds from it. Here again one can think that the breeder remains a discoverer, not an inventor despite the breeding work, since plants are a product of nature. As mentioned above, UPOV does not consider the breeder to be an inventor. US law, however, adopts another approach. The United States Patent and Trademark Office (USPTO) clarifies that “if one person discovers a new and distinct plant and asexually reproduces the plant, such person would be a sole inventor.”14 It appears thus that the term 8

9

10

11 12 13

14

Note that this paragraph uses ideas from chapter 4, section 4.1.1.7.2 “Moral and Ethical Concerns” of V Prifti, The Breeder’s Exception to Patent Rights: Analysis of Compliance with Article 30 of the TRIPS Agreement (Springer 2015). This is not claimed only in spiritual and shamanic circles, but it was also a central idea in the theory of evolution coined by Jean Baptist Lamarck. See RW Burkhardt Jr, The Spirit of System: Lamarck and Evolutionary Biology (Harvard University Press 1997). See the online Cambridge dictionary: . Accessed 15 July 2021. (35 US.C 161). . Accessed 15 July 2021. This has been suggested by Kock (n. 4) with respect to new breeding techniques-derived traits, related sequences, and plants. See the section “Inventorship” of “General Information about 35 U.S.C 161 Plant Patents” available at . Accessed 15 July 2021.

https://doi.org/10.1017/9781108989527.008 Published online by Cambridge University Press

Plant Variety Protection and Investment

89

“invention” is used as a legal fiction to indicate the activity of discovery of a new plant. The discovery of a new plant, however, needs to be accompanied by the improvement of the plant. Although the US legislator has chosen to consider plant breeding activities as an “invention,” as an application of human ingenuity, the formulation of legal provisions undeniably emphasizes the efforts of the breeder in the breeding process. These efforts are an “investment” in terms of financial, time, and physical resources. In Europe, the legal certainty of the patentability of plants, their parts and their varieties is highly discussed. Plant varieties are clearly excluded from patentability,15 but a plant variety can be indirectly patented when a patented DNA sequence is used in the breeding process.16 Many patents involving genetic sequences that affect plant varieties have been filed with the European Patent Office (EPO) in Munich and several have been challenged.17 While biotech companies claim that patents are necessary to recoup investments, some breeders and civil society organizations believe that patents on plant-related material strengthen the position of large companies to the expense of conventional breeding. The issue appears to be ongoing and on 8 July 2021, the German Federal Ministry of Justice and Consumer Protection invited patent offices, companies, the German Small Farmers’ Association “No Patents on Seeds” as well as professors and experts in the field to confront their views.18 While there were different positions on the effects of patents on small farmers, the central point of the discussion was the distinction between a technical and an essentially biological process and most stakeholders agreed on the need to clarify the term “essentially biological process.”19 It is not easy indeed to draw a line between breeding steps deemed as “technical” or as “essentially biological.” Since laws do not offer clear indications, patent offices may have the cumbersome task of providing legal certainty. In terms of the issue at hand, it is important to notice that the continuing controversy on plant patents emphasizes the role of IP in investment. This issue goes beyond legal questions such as “what can be patented,” and “what to exclude from protection,” but it encompasses important financial interests that can shape the future of plant breeding. Even though new breeding techniques have blurred the limit between genetically modified and nonmodified organisms,20 patents cover modified organisms inserted into plants whereas conventional breeding is concerned with “essentially biological processes”21 and products of these processes. In simple words, techniques that genetically modify organisms accelerate natural processes to quickly deliver the output and are

15

16

17

18

19 20

21

See art 53b of the European Patent Convention of 5 October 1973 as revised by the Act revising Article 63 EPC of 17 December 1991 and the Act revising the EPC of 29 November 2000. See also Rule 28.2 of its Implementing Regulations. The texts are available at . Accessed 16 July 2021. See also article 8.1 of the Directive 98/44/EC of the European Parliament and of the Council on the legal protection of biotechnological inventions, which extends patent rights to all biological material derived from the patented biological material when the derived material retains the same characteristics. Its article 9 further explains that patent protection embraces all material where patented genetic information is contained and performs its function. For more information on this point see the section “Patent Cases” and “News” on the website of “No Patents on Seeds” available at . Accessed 16 July 2021. The recorded event “Patentability of Plants and Animals: Scope for Action and Need of Reform?” is available online at . Last accessed 9 Aug. 2021. For more clarifications on this point, see the last hour of the conference in the above link. To be noted that some countries recognize new breeding techniques as GMOs. See Kock (n. 4). Indeed, the definition of GMO in Europe appears to be outdated and doesn’t take new developments into account. For a better understanding of “essentially biological” see the EPO Guidelines for Examination, Part G, Chapter II, sec. 5.4.2., available at . Accessed 17 July 2021.

https://doi.org/10.1017/9781108989527.008 Published online by Cambridge University Press

90

Viola Prifti

conducted in the lab, while breeders assist nature do its work in the fields or controlled environments. Here it should be noted that there is no provision in Europe that explicitly allows genetically modified organisms22 to be patented and excludes conventional breeding techniques. An invention is patented only when it is a technical solution to a technical problem.23 For example, a genetic sequence cannot be patented if it does not solve a problem such as draught intolerance or insect resistance. The invention needs to bring a technical progress and advantageous effect24 and all plant breeding techniques do bring such benefits. The organization “No Patents on Seeds” claims in its reports that too many patents are granted on conventional breeding techniques, thus hampering conventional breeding. Obviously, the investment on such techniques is significant and patents can act as an important tool to recoup the costs. In this case, the legal question is whether conventional breeding techniques are patentable. But the interests involved in this question are not only those of patentees and breeders, but of the whole society. V.II.B Patents The US patent system offers a good example of patent protection on plant varieties, plants, and parts thereof. There are different IP laws based on the types of seed and their use. Plants used for food purposes and seed-propagated plant varieties, for example, are protected under the 1970 Plant Variety Protection Act (PVPA), which grants rights to breeders. This law may be deemed similar to the plant breeder’s rights in Europe. The 1930 Plant Patent Act offers protection for asexually produced plants, that is, plants propagated by cutting or grafting, such as fruit trees and ornamental flowers. Patent rights on plant and parts thereof can be obtained under section 101 of the Patent Act, the most comprehensive form of protection that can be granted to seeds, genetic sequences, methods to produce plants, and so on. V.II.C Plant Breeder’s Rights Plant breeder’s rights are granted by the UPOV Convention on newly created varieties, which are distinct from any other variety and could propagate unchanged.25 This is the predominant form of protection for plant varieties in Europe, Asia, and many African countries. For the matter at hand, it is relevant to point out that although breeder’s rights may protect new varieties, the aim of the legislator is to help breeders recoup their costs. This can be supported by the fact that the UPOV Convention was modified in 1991 to take account of new genomic advancement and further strengthen breeders’ rights. In comparison to its previous 1978 version, the 1991 UPOV Convention contains new provisions that extend breeder’s rights to harvested material and to essentially derived varieties (EDVs) in its art. 14. The extension of rights to EDVs become necessary to prevent third parties from making minor changes to the variety. This is often termed cosmetic breeding and it can happen when slight mutations, such as a variation in shape or color, is caused by using biotechnology techniques such as mutation breeding, genetic modification, and repeated

22

23

24

25

Very few GMO crops are allowed in Europe and the Member States have discretionary power. For the EU framework on this point see . Accessed 16 July 2021. See Part F, Chapter II, Section 4.5 of the 2021 Guidelines for Examination of the European Patent Office, available at . Accessed 16 July 2021. Part G, Chapter I, Section 3 of the 2021 EPO Guidelines: . See articles 6–9 of the 1991 UPOV Convention.

https://doi.org/10.1017/9781108989527.008 Published online by Cambridge University Press

Plant Variety Protection and Investment

91

backcrossing.26 With the help of these techniques, biotechnologists can carry out plant mutations only in a few months, while breeders require from five to ten years to develop a variety. Breeders would have undoubtedly been in a disadvantaged position. Thus, the introduction of the concept of EDV in 1991 was deemed necessary as a protection mechanism.27 This makes it clear that the function of plant breeder’s rights is not to protect the simple act of creating new varieties, but to cover breeding costs. V.III PLANT VARIETY PROTECTION AND INVESTMENT

Euroseeds, a nonprofit international association that represents the breeding sectors, claims that most companies in the sector invest 15 percent of their annual turnover in plant breeding activities.28 A 2012 study reports that the EU average for R&D expenditure in the seed industry is 12.5 percent of the annual turnover.29 Undoubtedly, private investments are important in plant breeding, especially since public spending has sunk over the years.30 But it is not possible to bring here specific figures on plant variety protection by breeding companies since SMEs have no legal obligation to disclose such data. However, one can observe the general trend in R&D in agriculture and focus on multinationals, for which there is an obligation to provide data. A report of the United States Department of Agriculture (USDA) found that public R&D investments in agriculture in high income countries have fallen since 2009.31 Another USDA study reports that in 2007 Bayer was leading the private sector with $978 million followed by Syngenta and Monsanto with $830 and $770 million respectively. In the same year, the public sector in the USA had spent $456 million in crop science, while CGIAR (Consultative Group on International Agricultural Research) only $178 million.32 Overall, the annual private-sector food and agricultural R&D grew from $11.3 billion to $19.7 billion throughout 1994–2007. The most rapid increase in R&D was in crop breeding/biotechnology and this seems to be accompanied by a high level of market concentration throughout 1994–2009 in the sector. Economic analysis has found a self-reinforcing relationship between consolidation in the seed industry and less sponsored research.33 The consolidation in the seed industry can be also observed in patent concentration. Most patents in the sector are held by a small number of companies, as shown in Figures 5.1 and 5.2. “No Patents on Seeds” observes that these patents extend to conventional breeding. To circumvent legal prohibitions, companies use smart drafting of patents by carefully choosing the 26

27

28 29

30

31

32

33

For more, see Niels Louwaars et al., “Breeding Business: The Future of Plant Breeding in the Light of Developments in Patent Rights and Plant Breeder’s Rights” (Centre for Genetic Resources, the Netherlands 2009 CGN Report 2009–14). E Krieger, “EDV: A Protection mechanism, Not Plagiarism Prevention” (2021) 8(1) European Seed Magazine, available at . Accessed 16 July 2021. https://euroseeds.eu/subjects/research/. Accessed 16 July 2021. LEI Wageningen University, “Plant Reproduction Materials: A Dutch Motor for Export and Innovation” (2012) LEI Wageningen University and Research Center. Available at . Accessed 16 July 2021. DJF Eaton, (2013). “Intellectual Property Rights, International Trade and Plant Breeding” (PhD thesis, University of Wageningen NL). Available at . Accessed 16 July 2021. Paul W Heisey and Keith O Fuglie. “Agricultural Research Investment and Policy Reform in High-Income Countries” ERR-249 US Department of Agriculture, Economic Research Service, May 2018. For more information, see Keith O Fuglie, Paul W Heisey and John L King et al., Research Investments and Market Structure in the Food Processing, Agricultural Input, and Biofuel Industries Worldwide. ERR-130. US Dept. of Agriculture, Econ. Res. Serv. December 2011, 19 J Fernandez-Cornejo and D Schimmelpfenning “Have Seed Industry Changes Affected Research Effort?” (2004) 2(1) Amber Waves, United States Department of Agriculture, Economic Research Service, 14–19.

https://doi.org/10.1017/9781108989527.008 Published online by Cambridge University Press

92

Viola Prifti

fi gu r e 5 .1. Plant patent applications filed worldwide by main companies in the seed sector. Source: Patentdatenbank – Anmeldungen und Erteilungen – Kein Patent auf Leben! (kein-patent-auf-leben.de)

appropriate terms. In Europe, around 200 patents have already been granted despite the explicit legal prohibition of the patentability of conventional breeding techniques. This may harm the interests of breeders as well as consumers because these patents do not cover only plants and seeds, but they also extend to the food that may be produced. An example is the patent on bred barley granted to the international companies Carlsberg and Heineken, which extends to the beer produced with the protected barley.34 Patent concentration can be more concerning if one notes that several patents cover the same variety, thus hampering the free flow of germplasm. Based on a recent analysis, this trend increased in 2021 with 108 varieties with stacked proprietary traits out of a total of 881 varieties.35 What is the relationship between patents and investments in this case?

34

35

For more see Ruth Tippe, Johanna Eckhardt, and Christoph Then “Stop Patents on Our Food Plants! Research into Patent Applications Conducted in 2020 Shows How the Industry Is Escaping Prohibitions in Patent Law” (2021) available at . Accessed 17 July 2021. Civil protests against patents on beer were covered by Deutsche Welle Press. See K Wecker, “Storm Brews over Carlsberg, Heineken Beer Patent” (June 2017) available at www.dw.com/en/storm-brews-over-carlsberg-heineken-beer-patent/a-39161501>. For more information, see Kock (n. 4) 2–10.

https://doi.org/10.1017/9781108989527.008 Published online by Cambridge University Press

Plant Variety Protection and Investment

93

fig ure 5 .2. Patent applications on plant-related patents. Source: Patentdatenbank – Anmeldungen und Erteilungen – Kein Patent auf Leben! (kein-patent-auf-leben.de)

Obviously, it is the investment in patented plant traits that encouraged patents in the sector. But it would be too simplistic to find a linear relationship between investment, concentration in plant breeding, and restriction of plant germplasm. A proper study would need to consider other factors such as the geographical market, competition law, patent law, subsidies, the sociopolitical and economic context, infrastructure, and so on. Nevertheless, one can perceive that the current situation is not optimal and suggest a few improvements. While investments in the seed sector remain much needed in view of climate change phenomena and food supply, the patent system calls for more legal certainty to avoid that patent rights on biological material extend to plant varieties. This recommendation is valid only for Europe and for those countries which have chosen a double protection of plant varieties through patents and plant breeder’s rights. But a common suggestion for plant variety legislation is to assess whether laws perform their function. As Michael Kock notes “the argument that more IPR encourages more investment and innovation is overly simplistic in general but especially in the context of food and agriculture.”36 Patents and plant variety rights were conceived as a legal tool to temporarily limit the benefits that can accrue to society from accessing innovation in favor of the interest of the inventor to recoup his investment costs.37 As with any other law, IP laws should contribute to the betterment of society as a whole, not just to the interests of a few stakeholders. 36 37

ibid 3. See Prifti (n. 8) 66–74.

https://doi.org/10.1017/9781108989527.008 Published online by Cambridge University Press

94

Viola Prifti

Whether plant breeder’s rights and patent rights go beyond recovering the investment is a matter to be explored. There are not many studies that can offer evidence on this point. Those few studies that have dealt with this question do not fully support the incentive argument. For example, it was found that the increase in R&D in plant breeding in Argentina was due to economic policies rather than the introduction of intellectual protection.38 In any case, it would be too complicated to examine this question because not everything can be measured with numbers. For example, time and opportunity costs are not easy to evaluate. Moreover, the breeding sector is characterized by high uncertainty and is often subject to weather conditions. But one can look at laws encouraging plant breeding activities and assess whether the legal provisions benefit all stakeholders involved. The following section will further elucidate this point. V.IV PLANT VARIETY PROTECTION IN TRADE AGREEMENTS

As mentioned in Section V.II, the TRIPS Agreement gives countries freedom to adopt any kind of plant variety protection system provided that the system is effective.39 This provision seems to be ignored by free trade agreements (FTAs) between developed and developed countries, which require the latter to ratify the 1991 version of the UPOV Convention.40 This is the case of FTAs stipulated by the European Union (EU) with Asian and African countries as well as by the USA with Latin American countries. Art.12.42 of the EU–Vietnam FTA, for example, entitled “Plant Variety Rights” explicitly asks Vietnam to comply with the 1991 UPOV Convention.41 It was already mentioned that the 1991 UPOV Convention grants stronger rights to breeders in comparison with its 1978 version. At the same time, it weakens the so-called farmers’ privilege, that is, the right of farmers to save, resow, and exchange seed. While under the 1978 UPOV Convention farmers are free to carry out their traditional breeding practices, under the 1991 Convention such freedom becomes an option at the discretion of single countries. National laws implementing FTAs are often particularly cumbersome on traditional farming practices. This situation has raised social discontent and several negative effects of the 1991 UPOV Convention have been criticized as undermining domestic interests.42 The cases of Chile, Colombia, and Guatemala were particularly

38

39

40

41

42

W Jaffe and J van Wijk, “Intellectual Property Rights and Agriculture in Developing Countries” (March 7–8, 1995) 8. Proceedings of a Seminar on the Impact of Plant Breeder’s Rights in Developing Countries, Santa Fe´ de Bogota, Colombia. For the meaning of “effective” see Savita Mullapudi Narasimhan, “Towards a Balanced ‘Sui Generis’ Plant Variety Regime: Guidelines to Establish a National PVP Law and an Understanding of TRIPS-plus Aspects of Plant Rights” (2008) United Nations Development Programme. See also William Lesser, “An Economic Approach to Identifying an “Effective Sui Generis System” for Plant Variety Protection under TRIPS” (2000) 16(1) Agribusiness. An International Journal 96. Derek Eaton, Niels Louwaars, and Rob Tripp, “Intellectual Property Rights for Agriculture in International Trade and Investment Agreements: A Plant Breeding Perspective” (2006) 11, World Bank, Agricultural and Rural Development Notes. See Council decision (EU) 2019/753 of 30 March 2020 on the conclusion of the Free Trade Agreement between the European Union and the Socialist Republic of Viet Nam and The Free Trade Agreement between The European Union and the Socialist Republic of Viet Nam. The texts are available at . Accessed 17 July 2021. See, for example, Nappanee Supasiripongchai, “The Legal Protection of Breeder’s Rights for New Plant Varieties in Thailand: The Need for Law Reform Considering the International Convention for the Protection of New Varieties of Plants 1991” (2020) 23 (3–4) Journal of World Intellectual Property; see also GRAIN, “Trade Agreements that Impact Seed Laws in Africa” (2017) available at . Accessed 17 July 2021; BIOTHAI, “Impacts of the Thai–EU Free Trade Agreement on Plant Varieties, Biodiversity and Food Security” (2015) available at . Accessed 17 July 2021.

https://doi.org/10.1017/9781108989527.008 Published online by Cambridge University Press

Plant Variety Protection and Investment

95

remarkable.43 Farmers organized several protests and named proposals transposing FTAs plant variety provisions into national laws “Ley Monsanto” (Monsanto law) in order to indicate the bending of domestic interests to the demands of the multinationals.44 The domestic laws implementing the 1991 UPOV Convention were eventually revoked and are therefore currently noncompliant with the US FTAs. This is a reason for including Chile, Colombia, and Guatemala in the USA priority watchlist.45 Countries listed in the watchlist are identified by the Office of the United States Trade Representative and face the threat of unilateral trade sanctions or lawsuits. The legislative proposals implementing US FTAs in Latin American countries offer a good example in favor of the argument that laws on plant variety protection were exclusively driven by financial interests. This is because these laws seem to go beyond any ratio iure when they lay down criminal, civil, and administrative sanctions for acts that infringe the rights of the breeder without any exception for farming activities. Criminal sanctions for farmers are disproportionate in the context of plant breeding. They can put an unjustifiable burden on those who have traditionally engaged in seed-saving practices and might be unaware of changing laws. In particular, it would be challenging to eradicate social mores from indigenous cultures. Above all, it does not appear reasonable to make farmers criminals when their only aim is to support their families. For the sake of clarity, it should be noted that criminal sanctions are not imposed by FTAs, but by national laws. This point can be illustrated by the example of Guatemala. Guatemala has no plant breeder’s rights, but it provides for patents on plant varieties and parts thereof.46 Art. 275 (k) of the Guatemalan penal code sanctions the manufacture, production, sale, offering for sale, circulating, storing, or the unlawful possession of products covered by a patent, whereas art. 275 (l) penalizes use of a protected process or any of the above acts on products directly obtained by that process. While this provision may be justified to stop harmful competition, it is not clear how farming practices can damage patentees’ economic interests, especially when their markets differ and the protected varieties may be very different from farmers’ traditional varieties.47 The obligation to adopt the 1991 UPOV Convention appears even more unreasonable if one considers the fact that some developed countries continue to apply the 1978 version.48 Laws drafted for the benefit of the few inevitably harm the interests of the society as a whole in the long run. V.V FINAL CONSIDERATIONS ON PLANT VARIETY AS AN INVESTMENT TOOL

Whereas neoclassical economic theory claims that IP rights are a must to encourage R&D, new institutional economics views these rights necessary to reduce negative externalities in absence

43

44

45

46

47

48

For a detailed analysis of these laws see V Prifti, “An Answer to the Plant Variety Controversy in Chile” (2016) 19 (5–6) Journal of World Intellectual Property. David J Jefferson, “Development, Farmers’ Rights, and the Ley Monsanto: The Struggle Over the Ratification of UPOV 91 in Chile” (2014) 55(1) IDEA 39. See respectively pp. 39, 67, and 71 of the 2021 Special 301 Report, available at . Accessed 9 August 2021. See articles 93, 97, and 98 of the Law of Industrial Property (decree no 57/2000), available at . Accessed 17 July 2021. Prifti (2016) has analyzed plant varieties protected in Chile and explains that the market for these varieties is different from those of traditional varieties. See the letter of G Parmellin, Head of the Federal Department of Economic Affairs, Education and Research, Switzerland directed to Public Eye, a nongovernmental organization, dated 10 December 2020.

https://doi.org/10.1017/9781108989527.008 Published online by Cambridge University Press

96

Viola Prifti

of intellectual protection.49 In the first case, we can question the role of IP in encouraging R&D, especially when there are not many studies that offer evidence of this argument. In the second case, we can question the scope of the granted rights. Patent or plant breeder’s rights have different scopes of protection; the broader their scope, the larger the market for rightholders and fewer opportunities for new entrants. It was argued above that too many patent rights affect the same variety. This situation may lead to a patent thicket where breeders would need to incur high transaction costs to identify and negotiate licenses to be able to carry out their normal plant breeding activities. The final costs would have to be borne by consumers who will pay higher prices and probably will have less choice if patent blockages hinder breeding programs. Both these stakeholders will advocate for stronger rights and the legislator would need to find the optimal scope, length, and breadth of protection. This is undoubtedly a very difficult task. How can or should the legislator decide when the effects of rights are uncertain? It was mentioned earlier that morals can justify laws. But morals are in continuous evolution. Thus, there is the need to find a parameter of evaluation that transcends what could currently be perceived as “good” or “bad.” Such a parameter can be found in the concept of “ethics” understood as encompassing universally accepted values on human life that should guide what “behavior helps or harms sentient creatures.”50 Universally accepted values are found in UN documents such as the Universal Declaration of Human Rights (UDHR) and the International Covenant on Economic, Social, and Cultural Rights (ICESCR). Both these documents put emphasis on respecting human dignity and all rights of the “human family.” This reasoning can lead us to think that laws should serve all stakeholders to achieve societal well-being. In terms of the issue at hand, this means that one cannot strengthen IP rights at the expense of competitors and vice versa. While competitors who share the same market may need to be hampered through patent and plant variety rights, it seems excessive to prevent farmers from engaging in their traditional practices, especially since these practices are very important for preserving and enriching biodiversity. A broad base of genetic material would ultimately allow us to produce more and better food. It is, indeed, the significance of plant variety protection for food crops that should lead the legislator to pay special attention to plant variety provisions. In this context, article 25 of the UDHR and article 11 of the ICESCR may guide us in deciding when strong plant variety rights conflict with other societal interests.51 Both these articles state that food is a component of an adequate standard of life. Article 11, in particular, recognizes the right to food and to be free from hunger. One can argue that plant variety rights help breeders produce more varieties, thus, enhancing the right to food. This argument is valid when consumers can afford to pay for high food prices or when farmers in Europe can buy protected seed. But this is not the case for developing and least developed countries, where most farmers depend on subsistence farming. If patents and plant breeders can recoup the cost of their investment through sales, it appears unreasonable to profit from farmers and small breeders who barely meet their ends. If we recall article 1 of the UDHR as a guiding principle, we need to observe the precept that all human beings “should act towards one another in a spirit of brotherhood.” It is the “spirit of brotherhood” that impels the rich to alleviate the suffering of the poor and put the basis for sustainable development in our “human family.” In the long run, 49

50

51

Law and economics scholars argue that absent or poorly defined property rights do not allow for efficient bargaining. For more see Cooter and Ulen, Law and Economics (2016) Berkley Law Books, available at . Accessed 17 July 2021. R Paul, and L Elder, “The Thinkers Guide to Ethical Reasoning” (The Foundation for Critical Thinking 2013) cited in Prifti (n. 8). This paragraph draws upon ideas in previous work of the author. See Prifti (n. 8) 55–56.

https://doi.org/10.1017/9781108989527.008 Published online by Cambridge University Press

Plant Variety Protection and Investment

97

the benefits of such a choice surpass the limits on profit that some might endure in the short term. In this view, the attempts to limits countries’ choices to grant plant variety laws in line with their domestic interests is not justifiable. The adoption of the 1991 UPOV Convention and further restrictions on the right to save seed may make countries dependent on protected foreign seed and restrict the free flow of germplasm. Smallholders should be allowed to save and exchange seed as a private and noncommercial act.52 Small farmers in Europe are indeed allowed to save seed on their own farm.53 In contrast with many patent laws, art. 12 of the EU directive on the protection of biotechnological material allows breeders to apply for a compulsory cross-licensing. The conditions for obtaining such license are particularly cumbersome,54 but the provision is certainly a step forward in recognizing the importance of accessing plant innovations to allow the free flow of germplasm and improve current varieties. These provisions can appear reasonable in view of the function of plant variety rights as an investment tool. Flexibilities need to be introduced in the system to balance the rights between patentees and breeders. They should be able to recoup their initial investments and at the same time allow farmers to access protected seed to preserve traditional breeding practices. One important legal instrument that furthers the free flow of germplasm is the breeder’s exception to patent rights. France, Germany, the Netherlands, and Switzerland, have modified their patent acts to provide for such an exception. The exception is also found in art. 27 of the Agreement on a Unitary Patent.55 V.VI CONCLUSIONS

Plant variety protection may appear as a complex system of protection. The ever-evolving nature of biological material as well as differences between crops and their propagation methods are undoubtedly a challenge for legal systems that strive for legal certainty. Even though legal fictions help us create concepts to accommodate different interests, plant variety laws do not protect a genuine intellectual act of creation. Plant variety laws are envisaged to protect the efforts, time, and financial investment of breeders. These laws, however, may go against the objective of advancing the seed sector when patent rights cover plant varieties and thus block the free flow of genetic germplasm. To avoid undesirable consequences, countries need to be free to introduce flexibilities into the system in line with art. 27.3(b) of the TRIPS Agreement.

52

53

54 55

B De Jonge and P Munyi, “A Differentiated Approach to Plant Variety Protection in Africa” (2015) Wageningen Working Papers in Law and Governance, Law and Governance Group 2015/4. See art 11.1 of the Directive 98/44/EC of the European Parliament and of the Council on the legal protection of biotechnological inventions, and art 14 of Regulation (EC) No 2100/94. Prifti (2016) 7–8. Agreement on a Unified Patent Court (2013/C 175/01) available at . Accessed 17 July 2021.

https://doi.org/10.1017/9781108989527.008 Published online by Cambridge University Press

VI Software Protection under Copyright Law Interoperability and Protection of Program Interfaces Noam Shemtov

VI.I INTRODUCTION

Software plays a crucial role in most parts of today’s society. It is now an indispensable feature of the world of commerce, finance, industry and manufacture, education and research, medicine, government, entertainment, law and generally daily life. Furthermore, software stands at the heart of all disruptive technologies, such as artificial intelligence, automation and robotics, Internet of Things, biometric and digital identification, and virtual reality, to name but a few. As a legally protectable subject matter, software always proved to be challenging. None of our intellectual property regimes provide a perfect solution and intellectual property (IP) is better perceived as a tapestry of rights, some of which may be suitable for protecting certain facts of one’s software-based product or service, under certain circumstances. Other circumstances, even where it may involve one’s investment and effort being appropriated, are not always actionable under IP law either by design or omission. This chapter focuses on one IP branch and its adequacy in providing suitable protection against copyists,1 where the alleged copying takes place in relation to interoperability-enabling facets of computer programs, which are key for a thriving software market that attracts investment.2 The first branch of IP law that comes to mind in the context of software protection is copyright law. After debating the matter during the 1980s, the international community concluded that computer programs are to be protected by copyright law as literary works under the TRIPS Agreement.3 Copyright law plays an important role in protecting both software elements that are to be found ‘under the hood’, such as code and architecture, as well as visible elements, such as graphical user interface (GUIs).

1

2

3

For a comprehensive discussion the level of protection available for software elements under the different branches of IP see Noam Shemtov, Beyond the Code: Protection of Non-Textual Features of Software (OUP 2017). On the crucial role that interoperability and compatibility play in facilitating a vibrant and competitive software industry, see, eg, Mark Lemley and Pamela Samuelson, ‘Interfaces and Interoperability after Google v Oracle’ (2021) 100(1) Texas Law Review 44 under ‘The Value of Interoperability for Consumers and Creators’, C Doctorow, Adversarial Interoperability, ELECTRONIC FRONTIER FOUNDATION (Oct 2, 2019) at www.eff.org/deeplinks/ 2019/10/adversarial-interoperability>. See WTO Agreement on Trade Related Aspects of Intellectual Property Rights (TRIPS) Art. 10, para 1.

98 https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

Software Protection under Copyright Law

99

VI.II PROTECTING THE CODE

In general, protectability of computer program code is probably at present one of the least contentious areas of IP protection for software. It was the equivalence drawn between traditional literal text and the computer code, namely the instructions created by human authors to direct computer systems, that was at the basis of the decision to recognise such code as a literary work and a copyrightable subject matter under TRIPS. Drawing parallels between code and literary text as the basis for protecting software did not, however, go undisputed. There were convincing views to the contrary, according to which none of the existing IP branches suited the task for providing adequate and proportional protection for these types of work. The most articulated views in this respect were made public in what became known as ‘The Manifesto’.4 In the Manifesto, a group of prominent US IP academics maintained that neither copyright nor patents were suitable for providing calibrated and proportional protection for computer programs. It was argued that to avoid under-protection, which may lead to disincentivising investment in the creation of new programs, and over-protection, which may have a chilling effect on the creation of new software, a sui generis anti-cloning type of right was needed.5 The recommendation made in the Manifesto were not heeded by policymakers and reliance on existing and well-tested branches of IP was opted for instead. While it is of little value to evaluate the points made in the Manifesto as regards patent law,6 it may be useful to consider the extent to which copyright law either over-protects or under-protects the value embodied in software and, in this context, the extent to which it spurs investment in the creation of new works. While it is widely accepted that copyright law may protect code and visual elements in programs’ GUI as expressive aspects of such programs, the questions concerning Application Program Interfaces (APIs) and interfaces have not yet been answered determinatively both in the USA and in Europe.7 The significance of keeping interfaces open for the benefit of furtherance of technological development cannot be overstated. The ability of copyright holders to monopolise software elements that facilitate compatibility has been called into question. It is submitted that the legal position on this question is of paramount importance to the well-being of the software sector. Limiting the ability of actors in the software scene to reproduce compatibility-enabling aspects of software is likely to have an adverse effect on investment and growth in this field. Conversely, allowing later comers to re-implement interfaces of earlier programs in their code 4

5 6

7

Pamela Samuelson, Randall Davis, Mitchell D Kapor, and JH Reichman, ‘A Manifesto Concerning the Legal Protection of Computer Programs’ (1994) 94(8) Colombia Law Review 2308 ibid. This is so since patent law’s relationship with software has changed significantly since the days the Manifesto was written. Notwithstanding, the authors of the Manifesto’s general prediction that protection would sway between overprotection to under-protection is probably not without merit considering the cyclical movement in the scope of protection that we have seen in the last three decades. For a detailed discussion on the changes in patent law scope of protection in US jurisprudence in relation to software see Noam Shemtov, Beyond the Code: Protection of Non-Textual Features of Software (OUP 2017). In the USA the recent decision of the Supreme Court in Google LLC v Oracle America, Inc [1.141 S Ct 1183 (2021)] did not resolve this question. Indeed, a pending appeal before the CAFC in SAS Institute, Inc v World Programming Limited [96 F Supp 3d 1019 (ED Tex 2020), appeal docketed, No 2021-1542 (Fed Cir Jan 13, 2021)] addressed this very question (i.e., to what extent certain interface features could be subject to copyright protection). The position in the EU is not much clearer following the CJEU decision in SAS Institute Inc v World Programming Ltd C-406/10, regarding the coping of literal elements of APIs.

https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

100

Noam Shemtov

to enhance interoperability is likely to further innovation and promote key public policy considerations that underpin our copyright systems. VI.III OBSERVATIONS ON COPYRIGHT PROTECTION FOR SOFTWARE ELEMENTS IN THE US 8

The main reason for which scholars such the authors of the Manifesto were of the view that copyright is not ideally suited to protecting software was that the real value of computer programs, its crown jewels, are not to be found in the code but rather in its design and behaviour.9 The court in Whelan was so convinced by this contention that it was prepared to offer the plaintiff scope of protection under copyright law that set the high watermark in such cases. It essentially held that while the computer program general purpose and objective constitute unprotectable idea, everything else amounted to potentially protectable expression under copyright’s idea–expression dichotomy principle. Recognising the clear risk in over-protecting copyright owners in computer programs if the Whelan court’s approach is to be followed, in the subsequent Altai decision the 2nd Circuit Court of Appeal drastically scaled back the scope of protection offered under copyright law.10 It did so by endorsing Professor Nimmer’s complex ‘successive filtration’ test.11 The Court acknowledged that computer programs are utilitarian works and, as such, are eligible for thinner scope of protection under the principles laid down by the Supreme Court in Baker v Selden, and the limitations on copyright protection under S.102(b). Under the said successive filtration test three categories of materials were excluded from protectability: (1) features dictated by efficiency considerations, (2) featured dictated by external factors and, (3) features that are customary in the industry or are in the public domain. It is notable that after ‘filtering out’ all materials falling within the aforementioned three categories, there may be very little left to be protected beside that actual code. In fact, the court itself appeared to acknowledge as much.12 Notwithstanding the scaling back from Whelan by the Altai court, even this approach appeared to be excessive to some courts, who decided to pare back even further on the scope of protection offered to software’s functional elements. The case before the Court of Appeal for the 1st Circuit in Lotus v Borland initially involved a series of claims and counterclaims in the District Court of Massachusetts.13 In the first instance, a judgment was given in favour of Lotus deciding, inter alia, that Borland had copied Lotus’s menu command hierarchy and that the menu constituted a copyrightable expression.14 On appeal, the court had to determine whether the Lotus menu command hierarchy was a copyrightable subject matter. Borland admitted that it had copied Lotus’s menu command hierarchy. However, Borland maintained that the menu did not constitute a copyrightable subject matter since it fell within the list of excluded subject matter of S.102(b), which includes, inter alia, a system, process, method of operation, or 8

9

10 11 12 13 14

The author refers to the ensuing discussion as ‘observations’ since he does not purport to provide a comprehensive examination in this respect. Such comprehensive discussion goes beyond the scope of the present chapter. For example, see the 3rd Circuit Court of Appeal in Whelan Associates v Jasluw Dental Laboratory 797 F2d 1222 (3d Cir 1986), cert denied, 479 US 1031 (1987), where the court appeared to accept that the program’s crown jewels were to be found in some of its design and architectural elements rather than the code and instructions. Computer Associates International, Inc v Altai, Inc (1992) 982 F2d 693 (2nd Circ). ibid at 704. Supra 7, at 712. Lotus Development Corporation v Borland (1995) 49 F3d 807 (1st Circ). Such menu command hierarchy could be defined as an interface that enables a programmer accustomed to the Lotus software to seamlessly interact with Borland’s.

https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

Software Protection under Copyright Law

101

procedure. The court dealt with the primary question in the following order: first, the court rejected the applicability of Nimmer’s successive filtration test to the Lotus scenario since the case involved literal copying. Then the court turned to determine whether the menu command hierarchy fell within the list of the excluded subject matter of S.102(b).15 The court found the menu to be ‘a method of operation’ and hence not copyrightable, that is, it was held that Lotus’s menu command hierarchy was an unprotectable method of instructing a computer to perform spreadsheet functions. Thus, the court essentially enquired whether the part copied is of functional nature, that is, in this instance being a ‘method of operation’. Once this question was answered affirmatively, there was no need to enquire further as the part copied was not eligible to copyright protection and could therefore be copied with impunity as far as copyright law was concerned. Unlike the Altai court, the court in Lotus did not examine whether the menu command hierarchy was to be filtered out due to it being dictated or constrained by certain factors. Thus, it was irrelevant under this line of examination whether the said menu was a result of free and creative choices made by Lotus, or whether it was a result of various types of contains. If it clearly performs a functional objective, such as constituting a method of operation, it was rendered ineligible for copyright protection under S.102(b). The Lotus decision manifests an extreme reading of US copyright law and application of the idea–expression dichotomy in that regard. Under such an approach protection granted under copyright is largely ineffective against anyone and anything other than a straightforward plagiarist.16 It is fair to assume that if this approach would have taken hold among the different US Federal circuits, it may have had an unfavourable impact on generating incentives to invest in software-based products or services as it may possibly enable competitors to ‘borrow’ functional aliments of successful computer programs without consequences under copyright law. Bearing in mind that almost without exceptions, all program elements that are to be found ‘under the hood’ may be characterised as ‘systems’, ‘methods of operation’, or parts thereof, it could have triggered an open season for appropriation of functional program elements in the USA. VI.IV PROTECTING PROGRAM INTERFACES

Of particular importance to the innovation landscape in the present context are APIs. A useful definition thereto has been provided in the Google v Oracle litigation: Conceptually, an API is what allows software programs to communicate with one another. It is a set of definitions governing how the services of a particular program can be called upon, including what types of input the program must be given and what kind of output will be returned. APIs make it possible for programs (and programmers) to use the services of a given program without knowing how the service is performed. APIs also insulate programs from one another, making it possible to change the way a given program performs a service without disrupting other programs that use the service.17

While it is widely accepted that developers should be able to develop their own implementation of a previously created interface, a key question is whether it is also permissible for 15 16

17

Which, as we have seen, stems out and originates from the Supreme Court’s decision in Baker v Selden. As a matter of fact, even actual lines of code could be described as part of a method of operation, as they may be of a purely functional nature. 810 F Supp 2d 1002, 1005 (ND Cal 2011).

https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

102

Noam Shemtov

developers under copyright law to re-implement such earlier interfaces or part thereof in their independently created code. The Lotus analysis was explicitly rejected in the Oracle v Google dispute.18 It is suggested that the Oracle litigation provides a clear indication as to where on the spectrum between Lotus and Whelan current US judicial view is located. It may also be useful to assess whether the predictions of the authors of the Manifesto were indeed realised in this decision, in that it either results in overprotection or under-protection of computer programs. The dispute in Oracle involved, inter alia, Java’s application programming interfaces (APIs), 37 of which Oracle accused Google of replicating, including the declaring code and structure, sequence, and organisation (SSO) of these 37 API packages. The District Court found in this context in favour of Google on copyrightability grounds; namely, the court concluded that the above elements were not eligible to copyright protection, endorsing the Lotus analysis. The API at issue and the elements reproduced thereof were significant for programmers that were familiar with Java as it allowed them to interact easily with the Android platform created by Google. On Appeal, the Court of Appeal for the Federal Circuit (CAFC) reversed and found in favour of Oracle. Rejecting the Lotus approach, the court found that both elements of the Oracle claim, the literal copying claim in relation to the 11,500 lines of declaring source code and the non-literal copying claim in relation to the SSO of the 37 API packages, were eligible to copyright protection. Regarding the declaring code, the appellant court was of the view that the District Court erred in focusing on the options that were available to Google when it decided to reproduce the declaring code. The appellant court acknowledged that once Sun/Oracle created a package and then named it with a particular header, programmers who wished to use that particular package had to call it by its header’s name; hence, using the declaring code was necessary for compatibility purposes. However, this was of little relevance when deciding on copyright eligibility. The appellant court clarified that under the successive filtration test, the relevant point of time for assessing eligibility was when Sun/Oracle created the API packages, rather than when Google chose to re-implement it. At that point in time, there were no factors that compelled or dictated to Sun/Oracle to name the declaring code in the way it was named. In other words, the declaring code was named in that way because of free choices made by Sun/Oracle. This being so, it was clear that it was eligible for copyright protection; the facts that it was of functional nature and may be characterised as a ‘method of operation’ or system appeared to be neither here nor there. It is important at this point to emphasise that the court did not suggest that compatibility arguments endorsed by the District Court when finding the declaring code ineligible for copyright protection were of no relevance for evaluating infringement. The appellant court reasoned that such compatibility considerations should be factored in as part of a fair use analysis but not for assessing copyrightability. On the issue of copyrightability of the SSO, the appellant court rejected the District Court’s conclusion that it was ineligible to copyright protection since it was a system or method of operation. Citing with approval from Mitel,19 the appellant court stressed: ‘Section 102(b) does not extinguish the protection accorded a particular expression of an idea merely because that expression is embodied in a method of operation.’20 In contrast to the Lotus decision, what the appellant court essentially did was to read S.102(b) in a specific manner that allows a presiding court to maintain maximum flexibility in reaching a 18 19 20

Oracle Am, Inc v Google Inc (2014) 750 F3d 1339, 1364 (Fed Cir). Mitel, Inc v Iqtel, Inc (1997) 124 F3d 1366, 1372 (10th Cir). Oracle Am, Inc v Google Inc (2014) 750 F3d 1339, 1364 (Fed Cir), at 22.

https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

Software Protection under Copyright Law

103

conclusion on a given factual matrix, at the expense of certainty and predictability. It essentially treated the ‘procedure, process, system or method of operation’ part of S.102(b) as a subset of ‘idea’. The court explained that if a part of the program, such as a command structure, is original and survives the successive filtration test, it should not be regarded as an ‘idea’. This suggests that the divide is between ideas and expressions, and that the reference to ‘procedure, process, system or method of operation’ is merely a list of sub-categories of ‘ideas’. It follows that where an element of a computer program is not characterised as an idea because it survives the ‘successive filtration’, a S.102(b) enquiry must stop at that point and find the element at issue eligible for copyright protection. At this point, prima facie infringement has been established and the burden of proof shifts to the defendant, to show that under the specific circumstances of the case at hand, the complained use amounts to fair use. As aforementioned, in this context the appellant court found that compatibility arguments are relevant to a fair use analysis rather than to copyrightability. The court based its conclusion on this point on the trilogy of decompilation cases from the 1990s.21 The lower instance court in Oracle also referred to these three decisions and concluded that Oracle’s APIs were analogous to the interface protocols and procedures that were at issue in Atari, Sega, and Sony. However, this was criticised by the appellant court, according to which these cases were primarily about fair use and not about copyrightability. Hence, if there was a compatibility justification to support the copying of APIs in Oracle, such consideration may be relevant for a fair use analysis but cannot render those parts per se non-protectable.22 The case was then remanded by the Court of Appeal for the Federal Circuit to the court of first instance, where the jury decided that Google’s use qualified as fair use. Oracle appealed against this decision to the Federal Circuit Court of Appeal, where the decision was overturned and it was found in favour of Oracle, reasoning that ‘[t]here is nothing fair about taking a copyrighted work verbatim and using it for the same purpose and function as the original in a competing platform’.23 The Court of Appeal was clearly concerned with protecting the interests of rightsholders in expressive elements of works that are eligible for copyright protection and applied the fair use test in a manner that, in its view, reflected these concerns. As mentioned, the fact that not insubstantial verbatim copying took place, and that this was done to render a competing platform more attractive, appears to be highly significant in the court’s reasoning. Google appealed on both key points, arguing that (a) the APIs at issue were eligible for copyright protections, and (b) that in any event its copying of the declaring code qualified as fair use. The Supreme Court decided to grant Google’s petition for a certiorari and hear the case.24 With a 6:2 majority, the Court found in favour of Google, while failing to answer the copyrightability question, leaving it open. Instead, the Court chose to assume for the purpose of the discussion that the declaring code at issue from Java’s API was copyrightable, opting to focus on 21

22

23 24

Atari Games Corp v Nintendo of America Inc 975 F2d 832; Sega Enterprises Ltd v Accolade, Inc (1992) 977 F2d 1510; Sony Computer Entertainment, Inc v Connectix Corp (2000) 203 F3d 596 (9th Cir) The court opined that Google replicated Oracle’s declaring code and SSO to capitalise on the pre-existing community of programmers who were accustomed to using the Java API packages. But is it not this observation that renders the appellant court’s findings questionable? As the Computer & Communication Industry Association explained in its Amicus Curiea, ‘What could be better proof that something is a procedure, system, or method of operation than if a person can become “trained,” “experienced,” or “accustomed” to using it in the course of developing new works?’ [Amicus Curiae of The Computer & Communications Industry Association, Oracle Am, Inc v Google Inc (No 14–410) (November 7, 2014)]. Oracle Am, Inc v Google LLC, 886 F3d 1179, at 1210 (Fed Cir 2018). Google LLC v Oracle America, Inc, 1.141 S Ct 1183 (2021).

https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

104

Noam Shemtov

the question of fair use. On the latter point the Supreme Court overturned the decision of the Federal Circuit Court of Appeal, holding that Google’s copyright of the declaring code amounted to fair use primarily due to the transformative nature of using the code on a mobile platform, as well as the code’s real value residing in the fact that code developers became accustomed to it rather than any expressive value in it, which copyright aims to protect. The decision makes it clear that fair use is to play a key role in similar contexts. On this point the Supreme Court opined that ‘fair use can play an important role in determining the lawful scope of a computer program copyright’, by ‘providing a context-based check that keeps the copyright monopoly afforded to computer programs within its lawful bounds’.25 By shifting the goalposts from the S.102(b) type of assessment, the Supreme Court indeed opted for a context-based check that depends on the factual matrix that underpins the use made of the code by the defendant, rather than on the nature of the part taken. The latter type of assessment in theory should be examined in isolation from the motivation of the defendant, the type of use made of it by the defendant, and the impact that it may have on the market for the plaintiff’s program. Rather, a S.102(b) assessment focuses on the classification of the part taken as protectable or unprotectable subject matter, a determination that is to be made independently of the circumstances in which the defendant operates. It is submitted that resolving such disputes on the basis for Section 102(b) has the benefit of tailoring the courts’ response to the specific circumstances of each contested use, but this comes at the expense of certainty. Indeed, the Supreme Court’s refusal to address the copyrightability question may be interpreted, possibly by the CAFC as well as other regional courts, as a tacit approval for the latter’s view that API are prima facie copyrightable. As highlighted by Lemley and Samuelson, such a view was not previously shared by any other circuit courts.26 This may also lead to forum shopping by plaintiffs who seek copyright protection for their APIs, by simply including a patent claim in their lawsuit, which may give the CAFC jurisdiction over the complaint. VI.V PROTECTING SOFTWARE’S VARIOUS LAYERS UNDER COPYRIGHT LAW AT THE EU AND IN THE UK

In the BSA decision the CJEU was required to address for the first time the protectability of certain elements of computer programs under the Software Directive.27 Although the BSA decision was mainly concerned with the question of whether or not a user interface screen-display was a form of expression of a computer program and should therefore be protected as a part of a computer program under the Software Directive, part of the court’s reasoning is of relevance to the more general question of protectable and non-protectable subject matter.28 In BSA the CJEU articulated for the first time a concept not unlike the American merger doctrine. Referring to the originality of components of the graphic user interface (GUI), the court stated: ‘where the expression of those components is dictated by their technical function, the criterion of originality is not met, since the different methods of implementing an idea are so limited that the idea and the expression become indissociable [emphasis added]’.29 Thus, the court appears to suggest that 25 26

27

28 29

ibid at 3. Mark Lemley and Pamela Samuelson ‘Interfaces and Interoperability after Google v Oracle’ (2021)100 Texas Law Review 1. Directive 2009/24/EC of the European Parliament and the Council of 23 April 2009 on the Legal Protection of Computer Programs. Bezpečnostní softwarová asociace – Svaz softwarové ochrany v Ministry of Culture of the Czech Republic, C-393/09. ibid para 49.

https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

Software Protection under Copyright Law

105

where functionality-related constraints may lead to a limited number of choices available to a software engineer, even elements that reside on the expression side of the idea–expression divide may be held to be unprotectable. The court’s rationale for this conclusion was that such expressive elements are not original as they do not manifest an expression of the author’s intellectual creation in the sense of resulting from free and creative choices made by the programmer,30 which is somewhat different from the American rationale to the merger doctrine.31 Yet, it nevertheless leaves room for reasoning that where an element of a software product is created or expressed with a functional objective in mind (as is often the case) and where an adoption of such expression is one of a limited number of ways for achieving such functionality with any reasonable degree of success, such expression does not constitute copyrightable subject matter. Consequently, the opposite must also hold. Thus, where an element of a program is not drastically limited in the way it may be expressed, it is eligible for copyright protection notwithstanding the functional objective that it may have.32 Although not explicitly stated by the court to be so, there is little doubt that the court was guided by pro-competition considerations in introducing the said limitation. Acknowledging copyright protection for expressive elements where there is a very limited number of ways to express certain functional concepts is likely to result in the eventual monopolisation of the concepts themselves under copyright law,33 contrary to the idea–expression dichotomy principle. It follows that excluding such forms of expression from copyright protection is likely to keep such spaces open for competitors to reproduce to utilise a functional concept that is unprotectable under copyright law. This is likely to encourage a thriving and competitive software sector, as opposed to enabling first entrants into the market to monopolise such concepts and either foreclose it for further competition by second comers or engage in rent seeking; both outcomes are likely to stifle investment and innovation. Unlike BSA, the decision in SAS v WPL concerned the application of the idea–expression dichotomy in relation to various elements of computer programs.34 In this case WPL sought, and succeeded in creating, rivalling software that emulated the functionality of the SAS program. To create the said rivalling program, WPL obtained and studied a version of the SAS software known as the ‘learning edition’ which was subject to a licence. It is also noteworthy that it was not in dispute that WPL did not have access to the SAS source code. SAS claimed copyright infringement on three levels: (1) in SAS software, by creating the WPL software (‘program to program’ claim); (2) in SAS user manuals, by creating WPL software (‘manual to program’ claim); (3) in SAS user manuals, by creating WPL user manuals (‘manual to manual’ claim). The UK High Court referred several questions to the CJEU. The CJEU acknowledged that functionality, programming language and data file formats35 were not protectable expressions of computer programs. However, somewhat confusingly, the court suggested that while programming 30

31

32

33

34 35

This is an awkward reasoning. Of course, it all hinges on what is meant by ‘dictated’; an element may be ‘dictated’ by functional constraints in the sense of it being the only effective way for overcoming a functional problem. Designing such an element may clearly involve creativity. According to which protecting an expression under such circumstances would result in de facto protecting an idea, which, in turn, is contrary to the idea–expression dichotomy. In this context one should distinguish between a ‘functional objective’ and ‘dictated by functionality’. While the latter may result in a non-original expression, it is not necessarily the case in relation to the former. This is so since once the limited number of options to express the said concept is being exhausted, that concept may longer be ‘expressed’ and thus used, which in turn leads to a de facto monopolisation of this concept by the parties that were first to ‘express’ it. SAS Institute Inc v World Programming Ltd [2010] EWHC 1829 (Ch). A data file format essentially refers to the organisation and layout of data within a file.

https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

106

Noam Shemtov

languages and data file formats were not protected under the Software Directive, they may nevertheless be eligible for copyright protection under the Information Society Directive if they constitute the author’s own intellectual creation. Thus, while it is clear that the functionality of a program per se is not eligible to copyright protection irrespective of the legislative framework at issue, the same may not be asserted in relation to programming language or data file formats. Thus, not unlike the CAFC decision in Google v Oracle according to which the key question was whether the author of the declaring code was compelled to choose its final form due to technical or functional considerations, the CJEU suggested that programming language and the format of data files used in a computer program to interpret and execute application programs written by users and to read and write data in a specific format of data files may be eligible to copyright protection if they are the expression of the author’s intellectual creation under the Information Society Directive,36 rather than under the Software Directive. It appears that for the purpose of determining copyright eligibility, the essential factor is whether the computer program element at issue is an expression of the author’s intellectual creation in the sense of resulting from the exercise of free and creative choices.37 It is noteworthy that according to the CJEU ‘if a third party were to procure the part of the source code or the object code relating to the programming language or to the format of data files used in a computer program, and if that party were to create, with the aid of that code, similar elements in its own computer program, that conduct would be liable to constitute partial reproduction within the meaning of Article 4(a) of Directive 91/250’.38 Interoperability aside, it appears that the reason for which a third party may engage in such partial reproduction is not likely to play a role in determining whether or not the reproduced part is eligible for copyright protection.39 Furthermore, as shall be discussed below, it is questionable whether it could play any part at all in the overall assessment of copyright infringement. The CJEU in SAS further maintained that ideas and principles, which underlie any element of a computer program, could not be protected through a licensing agreement.40 Regarding manuals, the CJEU commented that those could be protected provided that the part taken thereof contains elements that are the expression of the intellectual creation of the author. In the court’s view, the keywords, syntax, commands, and combinations of commands, options, defaults, and iterations consisting of words, figures, or mathematical concepts, when considered in isolation, are not, as such, an intellectual creation of the author. However, through the choice, sequence, and combination of those words, figures, or mathematical concepts, the author may express his creativity in an original manner and such creativity may be eligible for copyright protection under the Information Society Directive. It is for the national court to evaluate whether the reproduction of parts of user manuals constitutes the reproduction of the expression of the intellectual creation 36

37

38 39 40

Directive 2001/29/EC of the European Parliament and of the Council of 22 May 2001 on the harmonisation of certain aspects of copyright and related rights in the information society. It is noteworthy that in the UK High Court, Justice Arnold asserted that the CJEU decision provides that: ‘copyright in a computer program does not protect either the programming language in which it is written or its interfaces (specifically, its data file formats) or its functionality from being copied’. It is submitted that while it is indeed the case that such elements are not protected as ‘copyright in computer programmes’ under the Software Directive’, the CJEU appears to highlight that both language and data format may be protected under the Information Society Directive. Hence, a conclusion in the present context that interfaces are simply not protected under copyright law in the EU is unwarranted. Arnold J acknowledged this possibility but decided that SAS did not please such claim under the Information Society Directive and that it was too late to introduce such claim at that stage of the proceedings, SAS v WPL [2013] EWHC 69 (Ch) para 36. SAS Institute Inc v World Programming Ltd, Case C‑406/10, para 43. As discussed below, the position on interoperability is somewhat more complex. In this case the restrictive licensing provision limited the use of the learning edition copy to ‘nonproduction purposes’. The English High Court found the WPL used the SAS copy for purposes that fell outside the scope of the licence.

https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

Software Protection under Copyright Law

107

of the author of the user manual. It is notable that some of these elements, such as commands and combinations of commands, may be viewed as interface components. Thus, as was the case in Google v Oracle, it appears that literal elements comprising interfaces may in principle be eligible for copyright protection. After the case was remitted back to the UK courts, Arnold J found in favour of the defendant on all accounts except for some limited literal copying in the ‘manual to manual’ claim. SAS appealed on all points except the lower court’s finding in relation to the ‘program to program’ claim. On appeal, the Court of Appeal upheld the Arnold J judgement, during which it made several interesting observations.41 One of the more interesting points was made in relation to the CJEU’s conclusion that the manuals could be protected as literary works to the extent that they expressed the author’s own intellectual creation. In this context, the Court of Appeal concluded that the Information Society Directive and the Software Directive have the same policy goals. Therefore, the CJEU’s conclusions regarding the operation of the idea–expression dichotomy under the Software Directive holds equally in the context of the Information Society Directive, irrespective of the fact the latter does not refer explicitly to the dichotomy principle. Since WPL took and implemented elements from the manuals such as statistical operations and mathematical formulas, rather than the description of such elements, the parts taken did not represent the expression of the intellectual creation that went into the development of the said manuals. Thus, such taking did not constitute copyright infringement. Would the same rationale be applicable to elements such as programming languages, data file formats, and algorithms? Although SAS did not appeal against the High Court findings regarding the program-to-program claim, if they would have done so, would the Court of Appeal have found that data file formats, irrespective of the creative effort that went into developing and expressing them, constitute an unprotectable idea under the information Society Directive? Considering the CJEU decision, it is far from clear that such rationale may be applied to elaborate literal copying of elements of a computer program, whether such elements are interface components. It is noteworthy that although a claim based on the Information Society Directive in relation to copyright eligibility of data file formats was not initially made by SAS and was therefore not considered by the court, Arnold J nevertheless referred to it in passing and stated: This claim also raises the question of originality, and whether a data file format is an intellectual creation. For this purpose, elements ‘differentiated only by their technical function’ must be disregarded [. . .]. What is required is something on which the author has stamped his ‘personal touch’ through the creative choices he has made [. . .]. It is open to evidence and argument as to whether data file formats such as SAS7BDAT satisfy this requirement.42

The proposition that data file formats, which are essentially interface elements, may be eligible for copyright protection under EU copyright jurisprudence was further bolstered in the subsequent UK High Court case of Technomed.43 The dispute’s main focus was the claimant’s contention that the defendant infringed their database right and copyright in their analysis and reporting system in the healthcare sector. One of the claims made by the claimant concerned an alleged infringement of copyright in a standardised XML format.44 Thus, the claimant claimed copyright in data 41 42 43 44

Although the court agreed with the Arnold J conclusion, on occasions it departed from its reasoning. SAS v WPL [2013] RPC 17 [2013] EWHC 69 (Ch) para 41. Technomed Ltd & Anor v Bluecrest Health Screening Ltd & Anor [2017] EWHC 2142 (Ch). ‘XML is a standard computer language for defining/representing structured data in a way which is partly selfdescribing using natural language terminology. It is not a data format, but a standardised abstraction which allows

https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

108

Noam Shemtov

formats written in XML language. After considering the evidence, the UK High Court concluded on this matter: ‘The XML Format is the product of Mr Fuller’s intellectual creation. Copyright subsists in the XML Format.’ Does this position accurately reflect the EU’s approach to copyright protection of interfaces? If so, what would be its likely impact on the software sector? These questions are discussed below via a hypothetical exercise of evaluating the outcome of a Google v Oracle dispute under EU copyright jurisprudence. VI.VI GOOGLE V ORACLE: A HYPOTHETICAL DISPUTE ASSESSED UNDER EU COPYRIGHT JURISPRUDENCE

This section examines a hypothetical scenario in which the Google v Oracle dispute was heard before EU courts, concerning its literal element. The purpose is twofold: First, it aims to evaluate the outcome of the Oracle complaints against Google regarding copyright infringement, should such complaints be filed before courts in the EU and adjudicated based on the EU jurisprudence. This should enable the reader to draw a direct comparison between the legal position in the EU with that in the USA. The discussion would then turn to evaluate the repercussions resulting from any differences between the two legal positions in relation to the well-being of the software sector. Before embarking on the exercise of hypothetically resolving the Google v Oracle dispute before EU courts, it may be useful to recall some key features of the relevant legislative framework in the EU. The Software Directive contains several provisions that are of relevance to the dispute at hand.45 Recital 10 reads as follows: The function of a computer program is to communicate and work together with other components of a computer system and with users and, for this purpose, a logical and, where appropriate, physical interconnection and interaction is required to permit all elements of software and hardware to work with other software and hardware and with users in all the ways in which they are intended to function. The parts of the program which provide for such interconnection and interaction between elements of software and hardware are generally known as ‘interfaces’. This functional interconnection and interaction is generally known as ‘interoperability’; such interoperability can be defined as the ability to exchange information and mutually to use the information which has been exchanged.

While the Recital above defines ‘interfaces’ and acknowledges their function and, implicitly, their importance, they do not comment on their eligibility for copyright protection. Recital 11 then states the idea–expression dichotomy principle and expands on which elements of a program may be classified as ideas; it does not directly reference interfaces. Finally, while Recital 15 appear to excuse reproduction of code done with a view to achieve interoperability, it is clear that it does so in relation to the practice of decompilation.46 Thus, it is submitted that the Directive may envisage that reproduction of code may be tolerated as part of a procedure that

45

46

flexibility in the kinds of data structure which can be represented, and in the choice of terminology and layout. Because of its flexibility, it is likely that independently designed XML schemata will differ markedly, even when describing essentially the same data.’ ibid at para 103 (emphasis added). Directive 2009/24/EC of the European Parliament and the Council of 23 April 2009 (emphasis added), On the Legal Protection of Computer Programs. Thus, not only does it explicitly refer to decompilation, but the reference to reproduction that may be excused – I followed by a reference to ‘translation’; hence, reproduction and translation as two subsequent phases in the process of decompliation.

https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

Software Protection under Copyright Law

109

aims to achieve interoperability, but only as an intermediary stage in a process intended to obtain the necessary information to achieve the said interoperability of an independently created program with other programs. As discussed below, it is arguable whether Recital 15 is intended to excuse the actual reproduction of code, not as an intermediary phase, but with a view to implement such code in another program to enhance interoperability. The actual provisions of the Software Directive would also be of little use to Google in our hypothetical examination. According to Article 1(2) ‘Protection in accordance with this Directive shall apply to the expression in any form of a computer program. Ideas and principles which underlie any element of a computer program, including those which underlie its interfaces, are not protected by copyright under this Directive.’ Reiterating the idea–expression dichotomy, this provision states, inter alia, that it is the ideas and principles that underlie interfaces which may be excluded from copyright protection, rather than the interfaces themselves – as expressed by code. This distinction is important. The provision further states that protection is available to expression in ‘any’ form. Thus, if the element at stake can be classified as ‘expression’, it may be eligible for copyright protection. It is also noteworthy that, unlike Section 102(b) of the US Copyright law, there is no explicit reference to ‘procedure, process, system of method of operation’; hence, it appears that there is less scope to argue, similarly to the court in Lotus,47 that while code that comprises an interface may be an expression, it is nevertheless excluded from copyright protection due to it being a ‘method of operation’. Article 5 concerns the legitimacy of the practice of reverse engineering, in the process of which a user may be required to perform the acts of loading, displaying, running, transmitting, or storing the program. Article 6 concerns a subset of reverse engineering, namely decompilation, where reproduction of the code and its translation are indispensable to obtain the information necessary to achieve the interoperability of an independently created computer program with other programs. As discussed above in relation to Recital 10, such reproduction of code carried out with a view of achieving interoperability appears to be permissible as an intermediary step, as part of a process of obtaining ‘information’ that is indispensable for achieving interoperability with an independently created program. This may be a further indication that in the end of the process what is expected to be created is an independent program that contains freshly created code, rather include thousands of lines of source code copied from the decompiled program. At last, Article 6 provides that it does not allow for the information obtained through decompilation to ‘to be used for the development, production or marketing of a computer program substantially similar in its expression, or for any other act which infringes copyright’.48 It is submitted that while the non-literal element of Oracle’s complaint concerning the structure, sequence, and organisation (SSO) of the 37 API packages may have failed to convince the court due to the operation of the idea–expression dichotomy principle whether under the Software Directive or the Information Society Directive, it is possible that under EU jurisprudence the part of Oracle’s complaint relating to literal copying, which concerned the declaring code (around 11,500 lines of source code) from Java’s API, may have been successful. First, the EU Software Directive, whether the recitals or the relevant provisions, do not appear to exclude from copyright protection interfaces themselves in relation to their expressive components such as source code. In this context, 11,500 lines of source code may be viewed as an expression of the programmer’s intellectual creation, the taking of which is thus likely to result in copyright infringement. 47 48

Lotus Dev Corp v Borland Int’l, Inc, 49 F3d 807 (1st Cir 1995). Directive 2009/24/EC, Art 6(2)(c).

https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

110

Noam Shemtov

As we have seen, unlike the legal position in the USA where it was fair use that came to Google’s rescue, the options open to a defendant under EU copyright law are much more limited and are reduced in the present context to the decompilation provision. However, as discussed above, neither the aim nor the language of this provision is likely to be of much assistance to a defendant in the present context. The reproduction of the declaring code was not an intermediate phase of decompilation, while the code itself was a product of the programmer’s free and creative choices. As highlighted by Lemley and Samuelson, the Supreme Court may have treated the Google v Oracle dispute differently from earlier interoperability cases49 since unlike those cases, most programs developed for the Android platform were not fully interoperable with other Java platforms.50 Lemley and Samuelson are of the view that this may explain why courts were prepared to find in earlier interoperability cases that interface elements that facilitate (complete) interoperability were not eligible for copyright protection, while the Supreme Court did not do so in Google v Oracle. Instead, the Supreme Court opted for the fair use route which, although perhaps not optimal, nevertheless provides a workable vehicle for courts aiming to promote choice and innovation in the software sector by allowing second comers to reproduce literal elements of interfaces with a view to enable a degree of interoperability. In contrast, we have seen that in the EU the literal reproduction of thousands of API source code lines is likely to lead to copyright liability. This is perhaps even more so if such reproduction is carried out in the EU for a purpose other than achieving complete interoperability. To conclude, the fact that elements of interfaces may be eligible for copyright protection, combined with the lack of an elastic judicial tool such as the fair use defence in the EU, is likely to result in a finding of copyright infringement should a scenario similar to Google v Oracle be heard before an EU court. While interfaces may be considered as eligible for copyright protection in both jurisdictions, in the USA courts have a greater degree for flexibility in determining whether literal reproduction of interfaces or part thereof ultimately lead to copyright infringement. This in turn allows USA courts to better safeguard the interest of second comers and carve out innovation-friendly spaces for such developers. VI.VII A BRIEF NOTE ON RESTRICTIVE LICENCE TERMS

We have seen that US copyright law may offer courts a more diverse toolbox under copyright law for adjudicating disputes regarding both literal and non-literal copying, which may enable courts to excuse the copying of literal interface elements, while EU courts’ capacity to do so is more limited. However, when it comes to restrictions imposed by contractual terms, it is defendants in the EU that may fare better.51 This is well illustrated by the SAS v WPL dispute in both jurisdictions. After failing to successfully sue WPL in the UK, SAS initiated proceedings in the USA. While it has so far failed in its copyright claim,52 it has been successful on its contract claims both at the 49

50

51

52

Eg, Sega Enterprises v Accolade 977 F2d 1510 (9th Cir 1992); Nintendo v Atari 975 F2d 832 (Fed Cir 1992); Sony Computer Entertainment, Inc v Connectix Corp 203 F3d 596 (9th Cir 2000) cert denied. Mark Lemley and Pamela Samuelson ‘Interfaces and Interoperability after Google v Oracle’ (2021)100 Texas Law Review 1. For a detailed discussion on the extent to which contractual mechanisms may be used in both jurisdictions to restrict acts that are permissible under copyright law, see Noam Shemtov, Beyond the Code: Protection of Non-Textual Features of Software (OUP 2017) ch 2. The District Court rejected SAS copyright claims in SAS Institute Inc v World Programming Limited et al., No 2:2018cv00295 - Document 465 (ED Tex 2020). At present, the case is being heard on appeal by the CAFC. It is

https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

Software Protection under Copyright Law

111

District Court and on appeal before the Court of Appeal for the Fourth Circuit.53 Breach of the ‘click through’ licensing terms, which sought to prohibit the licensee from developing a competing product (limiting the licensee to ‘non-production’ purposes), resulted in an award of total damages of $79.1 million in favour of SAS; this was confirmed by the Fourth Circuit. In conclusion, while US copyright law may prove to be more permissive that the EU one in the present context, such permissive scope may be trimmed down drastically via easily implemented contractual mechanisms. This may be contrasted with the outcome of the contractual claims in the EU, which were of a similar nature. The CJEU had little doubt in concluding: Consequently, the owner of the copyright in a computer program may not prevent, by relying on the licensing agreement, the person who has obtained that licence from determining the ideas and principles which underlie all the elements of that program in the case where that person carries out acts which that licence permits him to perform and the acts of loading and running necessary for the use of the computer program, and on condition that that person does not infringe the exclusive rights of the owner in that program.

Hence, The CJEU is clear that a lawful user of a program, who seeks to carry out in noncopyright infringing acts such as studying the program for ideas and concepts that underpin it, may do so as long it is irrespective of prohibitive licensing terms. In other words, this which is permitted under copyright law, for example black-box reverse engineering, may not be prohibited under contract. This position is quite sensible. Once the scope of permitted acts is determined under copyright law based on public policy considerations, it may not be derogated from by copyright holders via easily implemented contractual mechanisms such as ‘click through’ licences. VI.VIII CONCLUDING THOUGHTS

The outcome of the of the examination above, at least in relation to the scope of protection of program interfaces under copyright law, is somewhat surprising. Based on economic and political priorities, one would have expected the US position to favour copyright protection of interfaces and APIs, while the EU position to be somewhat more ambivalent in this regard.54 The analysis above demonstrates that this is not necessarily the case. The US Supreme Court decision in Google v Oracle suggests that the Court acknowledges that the question of copyright eligibility for API is complex and involves a matrix of nuances, hence an overriding rule of thumb according to which such elements are excluded from copyright protection may not be desirable. Instead, a more calibrated approach that enables a court to decide based on the facts of the case and the way public policy considerations play out

53 54

noteworthy that the case is being heard by the CAFC because SAS originally included patents claims in its lawsuit, only for those claims to be dropped early in the proceedings before the District Court. This tactic may indeed indicate the SAS view that it may fare better on the copyrightability issue before CAFC, perhaps considering the approach taken by the latter in Google v Oracle. SAS Institute, Inc v World Programming Ltd, No 19-1290 (4th Cir 2020). Dominant actors in the software industry predominantly originate from the USA, where the economy traditionally benefitted from the emergence and popularity of such actors. Such a situation may favour a more pro-copyright stance in the present context. In contrast, jurisdictions that may wish to challenge the dominance of the US companies and encourage the emergence of local champions may be expected to opt for an approach that seeks to limit the scope for protection available to such dominant actors and further adversarial interoperability, when it comes to interfaces and the significant role that these play in a competitive software sector.

https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

112

Noam Shemtov

in each instance, was considered more appropriate. As we have seen, such an approach may be utilised via the application of the fair use defence. While this approach is not without its critics,55 it may nevertheless provide courts with a tool for excusing literal re-implementation of preexisting APIs in one’s independently created source code. Regrettably, a tool having similar capabilities is not available under EU jurisprudence. Notwithstanding much of the fanfare that followed the CJEU’s decision in SAS v WPL regarding the non-copyrightability of APIs under EU law, we have seen that in fact the CJEU judgment may be read more narrowly, namely, that APIs (specifically data file formats) are not protected by copyright under the Software Directive. However, the court left open the possibility of protecting such elements under the Information Society Directive. Where that is the case, no fair use type of mechanism is available to EU courts. In fact, we have witnessed the consequences of the lack of fair use type of judicial tool in Technomed, where the court concluded that the defendant copied at XML formats at issue, infringing the copyright therein.56 It is suggested that in a Google v Oracle type of dispute, where the copying was literal, not quantitatively insubstantial, and did not lead to complete interoperability, may lead to a finding of copyright infringement under EU jurisprudence. While the space for engaging in adversarial competitive development in relation to APIs may be somewhat narrower in the EU, such space may not be derogated from via use of restrictive licensing terms. In contrast, the position is somewhat different in the USA as may be illustrated by the SAS v WPL breach of contract claims.57 The significance to APIs and the ability to re-implement them cannot be overstated in relation to interoperability. It is particularly so in the case of developers that are keen to keep their APIs under a proprietary cloak. This has been defined as adversarial interoperability, which one may choose to discourage or promote via its copyright rules on APIs copyrightability. As explained by Cory Doctorow: ‘For a really competitive, innovative, dynamic marketplace, you need adversarial interoperability: that’s when you create a new product or service that plugs into the existing ones without the permission of the companies that make them.’58 Digitisation and online platforms have revolutionised our public and private spaces. They affect the manner in which we interact socially with each other, the way we carry out business, and the speed of innovation. However, in some sectors it appears that the benefits reaped from such processes have been concentrated in a limited number of firms, which is likely to have an associated cost in terms of consumer welfare.59 The risk of such companies entrenching their dominants position in the marketplace, thus stifling competition and hindering market entry for potential competitors is a real one. While competition law is rightly seen as the primary vehicle for combating such undesirable developments, its inherent complexities and rather lengthy associated procedures, as well as their ex post nature, suggest that additional supplementary means for addressing such concentration of power may be justified. Indeed, in the recently approved text of the Digital Market Act, EU lawmakers agreed on the introduction of ex ante rules, that the largest messaging services, known as gatekeeper 55

56 57 58

59

Eg, Mark Lemley and Pamela Samuelson ‘Interfaces and Interoperability after Google v Oracle (2021)100 (1) Texas Law Review 43. Technomed Ltd & Anor v Bluecrest Health Screening Ltd & Anor [2017] EWHC 2142 (Ch), para 110. SAS Institute, Inc v World Programming Ltd No 19-1290 (4th Cir 2020). Cory Doctorow, Adversarial Interoperability, ELECTRONIC FRONTIER FOUNDATION (Oct 2, 2019) at . This is so due to the particular properties of digital industries, network effects, switching costs, data significance, and more; see, eg, J Crémer, Y-A de Montjoye and H Schweitzer, ‘Competition Policy for the Digital Era’ (2019) Luxembourg: Publications Office of the European Union.

https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

Software Protection under Copyright Law

113

platforms (such as WhatsApp, Facebook Messenger, or iMessage) will have to open up and interoperate with smaller messaging platforms, if they so request.60 However, it is clear that also outside the realms of gatekeeper platforms and messaging services, obstacles to adversarial interoperability take a toll on digital industries and soften competition. Admittedly, while such obstacles are not only facilitated by IP laws,61 IP barriers do play an important role in disrupting adversarial interoperability. Thus, it is suggested that in addition to a flexible and ever-evolving competition law regime, as well as the introduction of ex ante rules under certain circumstances as in the case of the DMA, EU policymakers and its judiciary would be well advised to simplify and clarify EU copyright rules regarding interfaces, specifically APIs.62 Hence, serving as an extra policy lever, an clear rule may be stipulated that due to their crucial role in facilitating connectedness between the different limbs of our digital world, APIs different layers, whether SSO or actual code, may not be considered as a protectable subject matter under copyright law irrespective of the amount of creativity that went into their production.63 Since such a measure will not necessarily aim to address actors that are at a gatekeeper level of dominance,64 it appears to be proportional and well-calibrated as it will merely seek to facilitate interoperability rather than to order it.

60

61

62

63

64

See . Mechanisms such as trade secrecy and repeated and frequent changes to technology may also assist in erecting barriers against adversarial interoperability. It is suggested that both the EU and the UK should clarify their copyright laws in a manner that favours adversarial interoperability, so that literal re-implementation of elements of APIs in one’s freshly created code resides clearly outside the scope of the copyright proprietor’s restricted acts. This should be the case whether ‘copyrightability’ is assessed under the Software Directive or the information Society Directive, as the case may be. We have seen that the DMA seeks to provide ex ante rules in relation to such actors in certain sectors.

https://doi.org/10.1017/9781108989527.009 Published online by Cambridge University Press

VII Bilski and the Information Age a Decade Later Michael J. Meurer

VII.I INTRODUCTION

About two decades ago the Federal Circuit threw open the doors of the US Patent Office to business method patent applicants. State Street announced that methods that yield a useful, concrete, and tangible result would be eligible for patent protection.1 This decision roughly coincided with the birth of e-commerce and an explosion of business method patents in the USA. About a decade ago the Supreme Court stepped back from State Street by installing screens that blocked applicants from patenting business methods claimed as abstract ideas. Bilski2 characterized claims to a method of hedging against energy price fluctuation risk as abstract, and therefore not eligible for patent protection.3 All nine justices supported this result, but their opinions revealed a significant split on the question of whether any patents on business methods should be permitted. Three justices joined Justice Stevens who called for categorical exclusion of business methods from the patent system.4 Three other justices joined Justice Kennedy who praised business inventions from this new “Information Age” and fretted that overly strong screens to eligibility established during the “Industrial Age” were no longer appropriate.5 While recognizing the method at hand was claimed too abstractly to be patent eligible, these justices seemed confident that the future would bring forth many business method inventions deserving of patents.6 The ninth justice, Justice Scalia, found the middle Abraham and Lillian Benton Scholar and Professor of Law at Boston University School of Law, [email protected]. I thank Michael Burstein and participants at the Suffolk IP Scholarship Workshop for their helpful comments. I also thank Austin Church and Dylan Welch for able research assistance. 1 State St Bank & Tr Co v Signature Fin Grp Inc, 149 F3d 1368, 1373 (Fed Cir 1999). 2 Bilski v Kappos, 561 US 593, 612 (2010). 3 Id. (“The patent application here can be rejected under our precedents on the unpatentability of abstract ideas.”) 4 Justice Stevens built a historical case that: A business method is not a “process.” Id. at 644. He also reinforced his case by reviewing patent scholarship and concluding: “I find it hard to believe that many of our entrepreneurs forwent business innovation because they could not claim a patent on their new methods.” Id. at 651. In a cautionary note, Justice Kennedy cited his concurrence in eBay which lamented that opportunistic patent litigation can be facilitated because “some business method patents raise special problems in terms of vagueness and suspect validity.” Id. at 608 (citing eBay Inc v MercExchange, LLC 547 US 388, 397 (2006) (Kennedy, J., concurring)). 5 Bilski, 561 US at 605. 6 Justice Kennedy wrote: “The machine-or-transformation test may well provide a sufficient basis for evaluating processes similar to those in the Industrial Age – for example, inventions grounded in a physical or other tangible form. But there are reasons to doubt whether the test should be the sole criterion for determining the patentability of inventions in the Information Age. As numerous amicus briefs argue, the machine-or-transformation test would create uncertainty as to

114 https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

Bilski and the Information Age a Decade Later

115

ground; he did not join the portion of Kennedy’s opinion discussing the Information Age.7 Nor did he join Stevens by embracing a categorical exclusion of business methods.8 Likewise, IP scholars have split on the question of whether business method patents are socially desirable and whether they should be permitted.9 Many scholars raised their voices soon after State Street, arguing that patents were not necessary as an incentive to induce investment in business method invention and that the patents would spawn opportunistic and anticompetitive patent litigation.10 Other voices responded to Bilski and Alice11 (reaffirming and extending Bilski three years later), expressing fears that development of business-related information technologies would be delayed by diminished patent incentives, and that start-ups in fields like fin-tech would particularly suffer.12 Given a decade of additional experience with business method patents in the USA, it’s a good time to revisit this debate. IP scholars remain divided on the question of whether business methods should be eligible for patent protection, but we have learned that there is a great appetite for patent protection of business methods in the Information Age, and neither Bilski nor Alice did much to slow the growth in patenting of business methods. We have also learned that business method patents are favorites of patent assertion entities. Despite solid evidence of social harm from opportunistic assertion of these patents, many scholars remain unconvinced that these practices justify eliminating patent protection of business methods. The cost–benefit analysis of business method patenting may have changed because the technological landscape for business method inventions is now quite different from a decade ago. Previous analyses had little to say about two new technologies now widely used to implement business methods – cloud computing and artificial intelligence. These new technologies are opaque to would-be imitators and business method innovators can be amply rewarded by trade secrecy, copyright, and other nonpatent means for appropriating innovation value.13 Contrary to Justice Kennedy’s assumption, the advent of an Information Age does not necessarily increase the social value of business method patents. This chapter comprises four parts. Section VII.I traces the path of the law of subject matter eligibility for business method inventions during this century. Section VII.II describes an explosion of business method patenting that has not faltered despite limitations imposed by Bilski and Alice. Section VII.III explains how business method innovators capture value from their the patentability of software, advanced diagnostic medicine techniques, and inventions based on linear programming, data compression, and the manipulation of digital signals.” 7 Id. (Scalia, J., concurring). 8 Id. (Scalia, J., concurring). 9 Compare John F Duffy, “Why Business Method Patents?” [2011] Stanford Law Review 1247, 1279–80 (“To the extent that a patent claim seems to fit within the rigors of this newly emerging field [of financial engineering], it will be more likely to be held patentable”) to Peter S Menell, “Forty Years of Wondering in the Wilderness and No Closer to the Promised Land: Bilski’s Superficial Textualism and the Missed Opportunity to Return Patent Law to Its Technology Mooring” (2011) 63 Stanford Law Review 1289, 1312 (“There is no reason to believe that ‘business methods’ have become a science or technology fitting the functional patent mold during the course of the past two centuries.”) 10 Rochelle Dreyfuss, “Are Business Method Patents Bad for Business?” (2000) 16 Santa Clara High Technology Law Journal 263, 275; Michael J Meurer, “Business Method Patents and Patent Floods” (2002) 8 Washington University Journal of Law & Policy 309; Bronwyn H Hall, “Business Method Patents, Innovation, and Policy” (UC Berkeley Econ. Dept. Working Paper No E03–331). 11 Alice Corp Pty Ltd v CLS Bank Intern, 573 US 208 (2014). 12 See Duffy, supra note 9, at 1263–69 (2011) (contending that the growth of operations research and financial engineering has made many business methods appropriate inventions for patent protection); Daniel F Spulber, “Should Business Method Inventions be Patentable?” (2011) 3 Journal of Legal Analysis 265, 272 (contending that patents on business method inventions support entry and growth of high-tech entrepreneurs). 13 See infra Section VII.III.B.

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

116

Michael J. Meurer

innovations using patents and other forms of intellectual property, and by using strategies that do not depend on intellectual property. Finally, Section VII.IV enumerates the social costs from business method patents and compares them to the incentive benefit from these patents. VII.II PATENT ELIGIBILITY OF BUSINESS METHODS

The US Patent Act offers limited guidance regarding coverage of business methods. Section 101 offers patent protection to new and useful processes. Section 100(b) unhelpfully defines process as “process, art or method, and includes a new use of a known process, machine, manufacture, composition of matter, or material.” The plain meaning of the statute is remarkably broad. It is easy to imagine that American courts and the Patent Office would recognize every new and useful business method as eligible for patent protection. At the turn of the century, after State Street v Signature Financial, that indeed seemed to be the case. For most of the aughts I did not cover the eligibility of business methods or software in my patent class on the belief that any such invention was eligible. American patent law changed course a decade later.14 In 2010 Bilski v Kappos rejected as ineligible a claim to a method of hedging against price fluctuations. During oral arguments in Bilski Justice Breyer jokingly puzzled over the boundary between eligible and ineligible business method inventions by asking whether he could get a patent if he invented “a great, wonderful, really original method to teach antitrust law that kept 80 percent of the students awake.”15 The unanimous decision made it clear that Breyer’s method falls on the wrong side of the line. The Supreme Court reinvigorated a judicially created exception to the statutory language that on its face apparently allows all method inventions to be patented.16 The exception bars patent claims directed to abstract ideas like the contested claims in Bilski. After Bilski, courts and patent prosecutors struggled to identify the boundary between claims directed to abstract ideas and claims that would pass muster as eligible applications of abstract ideas. Many thought that computer implementation of a business method was sufficient to make the method patent eligible.17 In 2014 the Court rejected that approach in Alice v CLS Bank.18 Alice created a two-part test of subject matter eligibility building on Mayo v Prometheus,19 a case decided after Bilski that addressed the eligibility of a method for optimizing a certain drug therapy. Step one asks whether the claim is directed to an abstract idea. If yes, then step two asks whether there are additional elements that impart an inventive concept and transform the claim 14

15 16

17

18

19

John Duffy observed that “in State Street, the Federal Circuit welcomed business method patents.” Duffy, supra note 9, at 1277. Duffy continues: “[But] in Bilski, the Supreme Court’s tone was utterly different. The Court accepted the patentability of business methods but it did so grudgingly, with the majority opinion even emphasizing that the law might not allow ‘broad patentability’ of such inventions. And the difference was more than just tone. In State Street, the Federal Circuit held unequivocally that the invention at issue there did fall within patentable subject matter. Bilski unequivocally held the opposite.” Id. at 1277–78. Daniel Crane, Antitrust (2014) The Supreme Court has repeatedly identified three judicially created exceptions to the statutory language: “Laws of nature, natural phenomena, and abstract ideas are not patentable.” Association for Molecular Pathology v Myriad Genetics, Inc 569 US 576, 589 (2013). Peter Menell resisted this view. “Merely implementing a process – such as running a business or entertaining an audience – on a machine should not thereby make the process or machine eligible for patent protection. The process must make a technological advance.” Menell, supra note 9, at 1312–13. Alice Corp Pty Ltd v CLS Bank Intern, 573 US 208, 212 (2014). For a recent iteration of this mode of analysis see Universal Secure Registry LLC v Apple Inc (Fed Cir 2021) (claims to methods of securing electronic payments ruled ineligible subject matter directed to abstract ideas; conventional computer implementation does not disclose an inventive concept). 566 US 66 (2012).

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

Bilski and the Information Age a Decade Later

117

into a patent eligible application of the abstract idea.20 The contested claims in Alice were directed to the abstract idea of using an intermediary to mitigate settlement risk in a financial transaction. Elements in the claims that added a generic computer implementation did not transform that idea into a patent eligible invention. The courts and the Patent Office have had ample opportunity to flesh out this test from Alice (often called the Mayo test); Alice has served as “the basis of nearly 1,000 court decisions.”21 Despite such intensive use, the test remains controversial, and outcomes are hard to predict. Talha Syed recently concluded what many others have said: “Everyone now knows there is an Alice two-step test, but no one knows quite what it means.”22 It is hard to draw a line between business method patents and other software implemented processes, but it seems clear that many business method claims have been assessed for eligibility post-Alice, and the test has been difficult to apply to this subset of process claims challenged as ineligible subject matter. Some judges are inclined to rely on their understanding that many business methods are not “technological” to exclude them from patent eligibility.23 In 2021 the Federal Circuit rejected as ineligible subject matter claims “directed to data privacy, customer loyalty systems, credit card fraud, transmitting and storing data, and retailer finder fees.”24 But other Federal Circuit decisions have been more permissive.25 Daryl Lim contends that: “though ‘labour’ or ‘investment’ in developing technology is generally insufficient to qualify, the Federal Circuit has used economic investment to justify its conclusion that the claimed invention was not ‘conventional, routine, and well-understood’ under Alice.”26 Patent prosecutors have adjusted patent disclosures and narrowed claim scope in business method applications to include technological implementations that go beyond the merely generic computer-related limitations appearing in claims in Alice. But prosecutors are unsure how far to narrow their claims. Federal Circuit Judge Newman worries: “inconsistency and unpredictability of adjudication have destabilized technologic development in important fields of commerce.”27 Like most commentators, I agree the law has been inconsistent and hard to predict, but in Sections VII.II through VII.IV, I will explain why I doubt that Alice has “destabilized technologic development.” Outside the USA the question of whether business methods should be protected by patents is also controversial, and application of eligibility rules is fraught.28 The Japanese 20 21

22

23

24

25

26

27 28

Alice 573 US at 217–18. Mark A Lemley and Samantha Zyontz, “Does Alice Target Patent Trolls?” (2021) 18 Journal of Empirical Legal Studies 47, 48. Talha Syed, “Owning Knowledge: A Unified Theory of Patent Eligibility” (2020) SSRN Working Paper . Ultramercial, Inc v Hulu, LLC 772 F3d 709, 721 (Fed Cir 2014) (Mayer, J, concurring) (“A rule holding that claims are impermissibly abstract if they are directed to an entrepreneurial objective, such as methods for increasing revenue, minimizing economic risk, or structuring commercial transactions, rather than a technological one, would comport with the guidance provided in both Alice and Bilski.”). Anthony J Fuga, “Top Section 101 Patent Eligibility Stories of 2021” (Dec 22, 2021) AIPLA Newstand . DDR Holdings, LLC v Hotels.com, LP, 773 F3d 1245, 1257 (Fed Cir 2014) (finding eligibility of a claim to an online advertising method). Daryl Lim, “Response: The Influence of Alice” (2021) 105 Minnesota Law Review Headnotes 345, 349 (discussing Exergen Corp v Kaz USA, Inc 725 F App’x 959, 966 (Fed Cir 2018)). Yu v Apple 1 F4th 1040, 1046–49 (Fed Cir 2021) (Newman, J, dissenting). Susan J Marsnik and Robert E Thomas, “Drawing a Line in the Patent Subject-Matter Sands: Does Europe Provide a Solution to the Software and Business Method Patent Problem” (2011) 34 Boston College International and Comparative Law Review 227 (describing conflicts across European courts and with the EPO regarding patent eligibility of business methods); Eugene F Derényi et al., “Protection of Business Method Patents Outside the

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

118

Michael J. Meurer

Patent Office was skeptical at first, but recently has been more receptive to business method patents.29 The European Patent Convention (EPO) explicitly excludes business methods “as such” from patentability under Article 52(2)(c).30 But business methods can be protected if they contain novel features that are “technical and solve a technical problem in a nonobvious manner.”31 The EPO found the technical effect requirement was satisfied in a case featuring a computer-implemented auction method.32 In contrast, the UK refused to grant a patent to a hedge fund “on a computer system that enables it to synchronize trades across multiple exchanges at the same time.”33 The method was not eligible subject matter because it avoided a technical problem rather than solved a technical problem.34 Similar to the USA, “considerable consensus exists that the [technical effect] rule in Europe is nebulous and that clarification is needed.”35 VII.III DID BILSKI AND ALICE DISCOURAGE BUSINESS METHOD PATENTING?

Bilski and Alice constrained business method patent prosecutors by reducing the potential scope of business method patent claims. On average, this sort of constraint increases the expected cost of prosecution and reduces the value of business method patents. If so, the result should be fewer business method patent applications because patents would no longer get filed on marginal business method inventions. Judge Moore assumed that Alice would have a significant effect on business method patents when she lamented “the death of hundreds of thousands of patents including all business method, financial system, and software patents.”36 Although she was looking at the fate of patents already granted, I expect she would have predicted the death of future business method patenting as well. It turns out that has not happened.

29

30

31

32

33

34 35 36

United States” (2009) 1(5) Landslide 18, 22 (reporting that in Japan business method patents are rejected on inventive step grounds more often than other types of patents). “The ambiguities associated with finance patents in the USA have also manifested elsewhere. European patent law explicitly excludes methods of doing business and finance from patent protection. But given the complexity of the definitions, some finance patents appear to have made it past these categorical exclusions. Meanwhile, Japan has shifted from one of the most skeptical patent offices regarding business methods to a much more permissive one: its rejection rate for these patents, of which finance constitutes a considerable number, fell from 92% in 2000 to 34% in 2012 through 2014 (Japanese Patent Office, 2019). Josh Lerner, Amit Seru, Nicholas Short, and Yuan Sun, “Financial Innovation in the 21st Century: Evidence from US Patents” (June 22, 2021) 63 SSRN Working Paper . See Robert E Thomas and Larry A DiMatteo, “Harmonizing the International Law of Business Method and Software Patents: Following Europe’s Lead” (2007) 16 Texas Intellectual Property Law Journal 1, 17. Susan J Marsnik and Robert E Thomas, “Drawing a Line in the Patent Subject-Matter Sands: Does Europe Provide a Solution to the Software and Business Method Patent Problem” (2011) 34 Boston College International and Comparative Law Review 227, 231–32; ReedSmith, Business Method Patents in Europe www.kazpatent.kz/sites/ default/files//business_method_patents_in_europe_en.pdf. Case T-258/03, Auction Method/HITACHI [2004] Official Journal of the European Patent Office (Apr. 21, 2004) 575, 587 (Technical Bd Appeal 3.5.01. Jonathan Browning, “Hedge Fund Renaissance Loses Bid to Patent Speedy Algorithm” (Apr. 21, 2004) Bloomberg Law. Id. Marsnick and Thomas, supra note 28, at 297. CLS Bank Int’l v Alice Corp Pty, 717 F3d 1269, 1313 (Fed Cir 2013) (Moore, J, dissenting in part).

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

Bilski and the Information Age a Decade Later

119

VII.III.A What Kinds of Business Methods Are Patented? Measuring the numbers, trends, and characteristics of business method patents is tricky because there is no consensus on how to define business methods and given a workable definition it is hard to identify patents containing claims that match the definition. Some commentary fails to distinguish business method patents from software patents. Such a distinction is easy to motivate by comparing the claims in Bilski directed to a method of hedging against price fluctuations that the applicant claimed without a software limitation to the claims in Alice directed to a method for mitigating settlement risk that the applicant claimed with generic software limitations. The Bilski patent is a business method patent but not a software patent. Even before Bilski, patented business methods typically featured software used to implement administrative tasks within a firm or to offer new services to customers (often in markets for financial products).37 Of course, most software patents do not cover business methods, and thus the analyst must be careful to exclude patents directed to nonbusiness methods implemented using software. Analysts use patent classifications, the text of patents, the identity of assignees, and other data to identify business methods.38 Most of the older studies identified US patents in Class 705: Data Processing: Financial, Business Practice, Management, or Cost/Price Determination as business method patents.39 This reasonable approach is underinclusive because of classification errors but especially because applicants sometimes disguise their application to avoid Class 705 and the increased scrutiny that sometimes has been given to business method patent applications.40 In addition to relying on USPTO (United States Patent & Trademark Office) classification, Lerner et al. took note of whether a patent was assigned to a financial institution, and examined the patent text with machine learning techniques to identify business method patents.41 Business methods can be organized into functional categories or by industry. Functions have been defined as: “new products or services (e.g., structured investments), new processes/procedures (e.g., risk management systems), and new organizations (e.g., internet banking).”42 This 37

38

39

40

41

42

“The majority of [financial sector] R&D is spent on software development and the majority of its R&D workers are programmers and software engineers. Using the definition of Bessen and Hunt (2007), four out of five business method patents are also software patents.” Robert M Hunt, “Business Method Patents and US Financial Services” (2011) 1 (Fed Rsrv Bank of Phila, Working Paper No 08-10/R) [hereinafter Hunt, Business Method Patents] (citing James Bessen and Robert M Hunt, “An Empirical Look at Software Patents” 16 Journal of Economic Management and Strategy 157). The US Patent Classification (USPC) was replaced by the Combined Patent Classification scheme in January 2013, and class G06Q is the new counterpart to class 705. Lerner et al., supra note 29, at 11. See, eg, Stefan Wagner, “Business Method Patents in Europe and their Strategic Use: Evidence from Franking Device Manufacturers” (2008) 17(3) Economic Innovation & New Technology 173 (using US patents in Class 705 that were also filed in Europe to study European business method patents); Megan M La Belle and Heidi Mandanis Schooner, “Big Banks and Business Method Patents” (2014) 16 University of Pennsylvania Journal of Business Law 431 (studying finance industry patenting in the USA). John R Allison and Emerson H Tiller, “The Business Method Patent Myth” (2003) 18 Berkeley Technology Law Journal 987, 1082. See, eg, Lerner et al., supra note 29 (using machine learning analysis of patent text to identify finance-related business patents). “[M]ost finance patents were classified under the current system within G06Q 40 (Finance; Insurance; Tax strategies; Processing of corporate or income taxes), a substantial number of blockchain and cryptocurrency patents were classified within H04L 09 (Cryptographic mechanisms or cryptographic arrangements for secret or secure communications).” Id. at 11. La Belle and Schooner, supra note 39, at 437 (deriving categories from Frame and White. W Scott Frame and Lawrence J White, “Empirical Studies of Financial Innovation: Lots of Talk, Little Action?” (2002) 3, available at . It is not commonly studied but a recent paper identified a significant number of organizational business method

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

120

Michael J. Meurer

type of categorization may be helpful when thinking about trade secrecy as a substitute for patent protection (more likely for back-office procedures that can be hidden like risk management). Most business method patents are assigned to corporations; thus, analysts have studied industry patenting patterns. “Business method patents are prevalent in the finance and information technology industries, but about thirty percent of the patents have been acquired by firms in manufacturing and trade.”43 Scholars have devoted particular attention to finance related patents. Lee and Soh identified certain terms as especially common in recent finance related patents: “auction marketplace, consumer authentication, asset allocation system, advisory service, and trading system.”44 Lerner et al. find that over 24,000 financial patents were granted in the USA before February 2019 that had application dates from 2000 to 2018.45 Immediately after State Street most of these patents were owned by computer makers and other large technology companies, but more recently there has been a growth in patenting by financial services companies and fin-tech start-ups.46 Before 2006 only Citigroup did much patenting in Class 705, but during 2007–2012, seven of the eight largest financial institutions did substantial patenting in Class 705.47 The portfolios of business method patents owned by these financial institutions are similar in size to the portfolios of large firms in other patent-intensive industries.48 Application of artificial intelligence to business methods is likely to significantly reshape the business method patent landscape.49 Artificial intelligence is widely used in the financial services, consulting, and advertising industries.50 Machine learning has been applied to business methods such as: automated customer service, customer recommendation engines, chatbots, marketing, pricing and price discrimination, accounting, procurement, investment choices, recruiting new employees, farm management, fraud detection, processing of loan applications, and equipment maintenance schedules.51 Lin and Rai report that the USPTO granted 6,583 artificial intelligence-related US patents since 2011, but they do not break out patents that are related to business methods.52

43

44

45 46

47 48 49

50

51

52

patents. Valery Yakubovich and Shuping Wu, “Is Organizational Innovation a Technology? Evidence from Patent Data” (February 2021) 1 (significant number of US patents granted that cover organizational innovation). Tian Heong Chan, Anandhi Bharadwaj and Deepa Varadarajan, “Business Method Innovation in US Manufacturing and Trade” (2021) SSRN Working Paper . Won Sang Lee and SoYoung Soh, “Identifying Emerging Trends of Financial Business Method Patents” (2017) 9 Sustainability 1. Lerner et al., supra note 29, at 2. La Belle and Schooner, supra note 39, at 474–76; Lerner et al., supra note 29, at 3 (“[T]he surge in financial patenting was driven by US information technology firms and those in other industries outside of finance”); Lee and Soh, supra note 44. La Belle and Schooner, supra note 39, at 472. Id at 471–72. “More than half of all AI-related patent applications have been published since 2013” citing WIPO Technology Trends 2019, Artificial Intelligence, at 13. Christian Rammer, Gastón P Fernández and Dirk Czarnitzki, “Artificial Intelligence and Industrial Innovation: Evidence from Firm-Level Data” (2021) 4, 9 SSRN Working Paper (finding German firms in financial services that used AI methods in 2018 accounted for one-half of sales in that industry, while for consulting and advertising services AI adopters accounted for one-quarter of sales). Forbes Technology Council, 15 Business Applications for Artificial Intelligence and Machine Learning (Sep 27, 2018) ; Monideepa Tarafdar, Cynthia M Beath and Jeanne W Ross, “Using AI to Enhance Business Operations” (Summer 2019) 37, 38 MIT Sloan Management Review. Yu-Kai Lin andArun Rai, “Patent Protection and Software Innovation: Evidence from Alice” (2021) SSRN Working Paper . The Quinn Emmanuel blog warns that: “[g] iven the limitations articulated in Alice and its progeny, it is unclear how many of the AI-related patents that have made their way through the US Patent Office would survive in eventual litigation.” Jordan R Jaffee et al., “The Rising

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

Bilski and the Information Age a Decade Later

121

VII.III.B Characteristics of Business Method Patents Bilski and especially Alice had significant effects on business method patents, but the headliner in this domain is not the Supreme Court, instead it is the Federal Circuit with its State Street decision.53 There were hardly any business method patents before State Street, and despite frequent eligibility invalidations after Alice,54 the number of business method patent applications and grants in the USA is still large.55 Looking specifically at finance related patents, Lerner et al. find that the share of granted finance patents in comparison to all granted patents in 2018 is only half of that share in 2013 (before Alice), but the share is roughly equal to the share in 2008 (before Bilski) when the boom was underway.56 Firms apparently believe business method patents are still valuable enough to justify incurring prosecution costs. This belief is justified if one understands Bilski, Alice, and their progeny as cases that constrained the freedom of prosecutors when they write patent claims.57 Certain claim language will be rejected on subject matter grounds if it is too abstract, but that does not mean that business method inventions are unpatentable. Recent empirical studies support the view that effective patent prosecutors responded to Alice by adjusting claim language in ways that possibly reduce patent value, but avoid subject matter rejection at the USPTO.58 Kesan and Wang split their data on business method patent applications into one set that was filed before Alice but examined after and another set that was filed after Alice.59 Business method applications filed before but examined after were four times more likely to be rejected.60 They conclude that patent applicants were successful at overcoming Alice for applications filed after Alice.61 VII.IV HOW DOES BUSINESS PROFIT FROM BUSINESS METHOD INNOVATION?

Business method innovation has flourished in the USA and elsewhere for decades before patents became a significant source of reward for these inventions. Evidently, inventors found other ways to capture enough profit from new business methods to cover their cost and motivate their

53

54

55

56 57 58

59 60

61

Importance of Trade Secret Protection for AI-Related Intellectual Property” QuinnEmmanuel (last visited Dec 30, 2021). New business method patent applications grew sharply after State Street, with about 11,000 new applications a year. Robert M Hunt, “Business Method Patents and US Financial Services” (2010) 8 Contemporary Economic Policy 322, 327. After Alice, the number of patent applications fell in bio-informatics, business methods, and software. Jay P Kesan and Runhua Wang, “Eligible Subject Matter at the Patent Office: An Empirical Study of the Influence of Alice on Patent Examiners and Patent Applicants” (2020) 105 Minnesota Law Review 527, 563. The finance and e-commerce subcategories of business methods garnered the most Alice and Sec. 101 rejections by examiners. The USPTO calculated filing trends for US business method patents over 1997–2017. Their data show 11,667 serialized filings in 2017 compared to 16,124 in 2014, and 9,122 in 2010. Measured instead in terms of RCE filings the levels are: 9,810 in 2017, 9,381 in 2014, and 8,739 in 2010. www.uspto.gov/sites/default/files/documents/FilingTrendsInBisinessMethods1997to2017.pdf. The data from Europe and Japan indicate business method patent grants are still common around the world. See supra note 29. Lerner et al., supra note 29, at 34. There is wide consensus that Alice narrowed the scope of patent protections for software. Lin and Rai, supra note 52. Lin and Rai used Alice as a natural experiment and found evidence that it caused a reduction in the scope of software patents. Lin and Rai, supra note 52, at 3. Kesan and Wang, supra note 54, at 38. Id at 47. Kesan and Wang used difference-in-difference analysis to show that the UPSTO implementation of Alice caused more rejections of the earlier applications under Sec. 101. Id at 43–44. Id at 57–58.

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

122

Michael J. Meurer

creation. That said, it is possible that some types of business method inventions would be neglected if not for the opportunity to patent. Perhaps this problem of under-reward will grow more serious in the Information Age because business methods inventions are growing more risky or more costly. A closer look at theory and evidence from the past two decades suggests that patents are usually not an important tool used by innovators to profit from new business methods. VII.IV.A How Important Are Patents? Researchers have not been able to demonstrate that patent availability causes or is even correlated with an increase in business method innovation.62 Although there is evidence that certain business method patents deliver value to publicly traded American firms that obtain them,63 they may not matter much for funding of high-tech start-ups. Taylor conducted a survey of “475 venture capital and private equity investors” to study the impact of patent eligibility law on investment decisions.64 He concluded that: “[i]nvestors overwhelmingly indicated . . . that the elimination of patents would either not impact their firms’ decisions whether to invest in companies or only slightly decrease investments in companies developing technology in the construction, software and Internet, transportation, energy, and computer and electronic hardware industries.”65 By way of contrast, patent eligibility rules mattered considerably to inventors in life science industries.66 Relatedly, research by Wagner and Cockburn suggests patents do not improve the survival prospects of start-ups:67 Interestingly for the debate about business method patents, we find that they have very little impact on [start-up] survival compared to patents classified in other classes. Based on this finding it can be argued that business method patents – on average – convey little economic value to the patentee. From a managerial perspective, it seems questionable whether benefits from patenting methods to conduct business outweigh the cost of patenting (cost of drafting the application, filing and examination fees, renewal fees and cost of enforcement).68

Instead of patents, business method innovators have used trade secrecy, contracts, and employment law to discourage suppliers, customers, and departing employees from making unauthorized use of business methods. Secrecy may effectively limit imitation of back-office administrative methods, for example, secret algorithms that are used for human resources management and marketing tasks. Trade secrecy has little or no role to play when business methods provide services that are revealed to customers, and therefore are not secret. For example, the creators of new 62

63

64 65 66

67 68

See Hunt, supra note 53, at 349. (“There is at present very little evidence to argue that business method patents have had a significant effect on the RandD investments of financial institutions”); Stefania Fusco, “The Patentability of Financial Methods: The Market Participants’ Perspectives” (2011) 45 Loyola of Los Angeles Law Review 1, 17 (surveying members of the financial industry and concluding “that patent protection has not been responsible for the innovation that occurred in the financial industry in the time between State Street and In re Bilski.”). Chan, Bharadwaj, and Varadarajan, supra note 43 (noting publicly traded American manufacturers gained 7 percent in market value after State Street if they owned patents in Class 705, while firms in retail, wholesale, warehousing, and transportation gained 25 percent in market value); Sarah Hinchliffe, “Class 705 Business Method Patents in the United States: A Study from 1998 to 2010” (2021) 69 Drake Law Review 73, 105–08 (stock market event study showing share value increased in the twenty day window centered on the grant of a Class 705 business method patent). David O Taylor, “Patent Eligibility and Investment” (2020) 41 Cardozo Law Review 2019, 2027. Id at 2066–67. Id at 2069 (finding “the most negatively impacted would be the pharmaceutical, biotechnology, and medical device industries”). S Wagner and I Cockburn, “Patents and the Survival of Internet-Related IPOs” (2010) 39 Research Policy 214, 226. Id at 217 (analyzing the effects of patents on firm survivability during dot.com boom of the late 1990s using data from collection of 356 firms that made IPO on NASDAQ between February 1998 and August 2001).

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

Bilski and the Information Age a Decade Later

123

financial products normally must comply with disclosure regulations that are incompatible with trade secret protection.69 The absence of effective patent or trade secret protection may be especially challenging for innovators in the insurance industry or other financial products that require expensive regulatory approval.70 The problem is that imitators can copy the innovation and avoid most of the expense of regulatory approval and thus free-ride on both the research and regulatory expenditure of the innovator.71 Nevertheless, extensive commentary suggests that firms have adequate incentive to create finance-related business methods absent intellectual property protection.72 Financial innovators rely on “reputational gains, tacit knowledge, and first mover advantages” to derive rewards from their innovations.73 Network effects are often present in financial markets and can arise via an interoperability requirement in many such services.74 Hunt observes: Firms in the financial sector “protect their innovations in ways like those observed among manufacturing firms. Historically, patents have not been a significant part of the story for these firms, and yet their absence has not prevented them from investing in new products (financial instruments) or the processes (e.g., trading platforms, pricing algorithms) required to offer them.75

The same range of nonIP incentives are likely responsible for motivating foundational business innovations that apply across all industries. Business scholars have compiled lists of the most significant management innovations of the twentieth century. Examples include: “the industrial research laboratory at GE, the use of capital budgeting and general metrics for evaluating division performance at DuPont, brand management at Procter & Gamble, organizational innovation at GM and Visa, and open source software development by Linux and others.” 76 These methods were invented without thought of patent protection and they 69

70

71

72

73

74 75 76

Christopher Petruzzi, Margueritte Del Valle and Stephen Judlowe, “Patent and Copyright Protection for Innovations in Finance” (1988) 17 Financial Management 66, 67. Robert M Hunt, “Business Method Patents and US Financial Services” (2011) 10 Federal Reserve Bank of Philadelphia Working Paper No 08-10/R . Petruzzi, et al., supra note 69, at 67 (creators of new financial products bear significant risk and expense and imitation is often cheap and easy); Gabriel Rauterberg, “Innovation in the Stock Market and Alternative Trading Systems” (Dec 2020) 13–14 (imitators can free ride on costly regulatory approval with respect to innovations in financial markets). See Hunt, supra note 53. (“Studies by Silber (1981) and Caskey (2003) present evidence that an established contract on one exchange enjoys an advantage in terms of liquidity that is often difficult to overcome when a similar contract is introduced on another exchange. Anderson and Harris (1986) argue that regulations that delay imitation by rival firms reinforce first mover advantages, increasing the rents associated with financial innovations. And among investment banks, there is evidence that first mover advantages play an important role in generating sustained profits from the introduction of new financial instruments (Tufano 1989)”). La Belle and Schooner, supra note 39, at 442 citing Tamar Frankel, “Cross-Border Securitizations: Without Law, But Not Lawless” (1998) 8 Duke Journal of Comparative & International Law 255; Petruzzi et al., supra note 70 (identifying lead time advantages that motivate financial innovation); Gary B Gorton and Ping He, “Economic Growth and Bank Innovation” (2021) 10 National Bureau of Economic Research Working Paper No 29326 (“In the period 1929–1941, banks innovated by developing methods of credit risk analysis and covenant design. During 1987– 2016, as loan maturity continued to increase, banks innovated to shift the risk to nonbank, institutional, investors. [. . .] Bank innovation that has resulted in these reallocations of risk are a very significant contributor to economic growth.”) Id at 27. Hunt, supra note 53 at 329–30. Hunt, supra note 37 at 9. Julian Birkinshaw, Gary Hamel, and Michael Mol, “Management Innovation” (2008) 33 Academy of Management Review 825, 829.

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

124

Michael J. Meurer

would not get effective patent protection under current law. Claims broad enough to block imitation would likely be characterized as abstract ideas – ineligible for patent protection.77 Her review of similar considerations led Dreyfuss to conclude patents are not needed to incentivize creation of business methods: But neither the free-rider nor the disclosure rationale justifies business method patents. Businesses are largely practiced in public. Accordingly, there is little need to especially encourage disclosure. Business methods are also hard to free ride on. They depend in strong ways on the social structure within the firms utilizing them – on compensation schemes, lines of reporting, supervising policies, and other business factors. Moreover, as we saw, sticky business methods are their own reward. With lock in, network effects, and even good old fashioned loyalty, lead time (the first mover advantage) goes a long way to assuring returns adequate to recoup costs and earn substantial profit. In sum, while business innovations are certainly desirable, it is not clear that business method patents are needed to spur people to create them.78

VII.IV.B Profiting from Business Method Innovation in the Information Age A skeptic may not be convinced that innovation incentives are still adequate without patent protection in the Information Age. Despite Alice, patent applications covering artificial intelligence are trending up. “More than half of all AI-related patent applications have been published since 2013,”79 and many of these applications relate to business methods.80 Perhaps aggressive patenting signals that business method innovation is becoming either easier to imitate or more costly and risky. There are two reasons these fears may be unfounded. First, many advances in artificial intelligence take place in an environment of collective invention in which tools and techniques have been widely shared. Second, other more specialized advances take place in an environment in which imitation is difficult even without patent protection. The term collective invention refers to historical episodes in which competing firms or inventors share research and development results.81 During the Industrial Revolution, profound advances from collective invention occurred in England in blast furnace and steam engine technologies.82 Open source development of software like the Linux operating system and the Apache web server arguably fits the collective invention model.83 Much of the research activity in machine learning also seems to fit the collective invention model.84 OpenAI, a nonprofit organization, makes AI tools and research widely available. There are open-source tools for 77 78

79 80

81 82

83

84

Most of these methods could not be protected by secrecy either. Rochelle Dreyfuss, “Are Business Method Patents Bad for Business?” (2000) 16 Santa Clara High Technology Law Journal 263, 275. WIPO, supra note 49, at 13. Iain M Cockburn, Rebecca Henderson, and Scott Stern, “The Impact of Artificial Intelligence on Innovation: An Exploratory Analysis” in Ajay Agrawal, Joshua Gans, and Avi Goldfarb (eds) The Economics of Artificial Intelligence: An Agenda (2019) 115, 132 ((finding 95 AI-related business software patents granted in the US from 1990 to 2014 – author’s calculation based on Table 4.5). Robert C Allen, “Collective Invention” (1983) 4 Journal of Economic Behavior & Organization 1. Alesandro Nuvolari, “Collective Invention during the British Industrial Revolution: The Case of the Cornish Pumping Engine” (2004) 28 Cambridge Journal of Economics 347. Alessandro Nuvolari, “Open Source Software Development: Some Historical Perspectives” (2004) Dept of Tech Mgmt, Technische Universteit Eindhoven, The Netherlands, Working Paper No 03.01. Many prominent AI researchers have insisted on retaining the right to publish their results when joining companies such as Baidu, Facebook, and Google. See Jack Clark, “Apple’s Deep Learning Curve” Bloomberg Business Week (Oct 29, 2015).

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

Bilski and the Information Age a Decade Later

125

curating data and training algorithms.85 The impact of open source is likely to be reduced cost86 for some kinds of business method innovation and less need for patent incentives. In the domain of propriety development of AI-based business methods, imitation is often difficult because the technology is opaque, and many of the inputs that would be needed to imitate are in short supply. Consider opaqueness first. Machine learning algorithms and other software implementations of innovative business methods often reside in the cloud, and they are protected as trade secrets.87 Patenting is not a profitable strategy “if the invention is for an AI algorithm that runs on a server that cannot be observed by the public, it may be impossible to tell which, if any, competitors are infringing on the technology.”88 Furthermore, “[d]ue to the prohibition on patenting abstract ideas, acquiring meaningful patents on artificial intelligence systems is not straightforward. Thus, companies are increasingly turning to trade secret protection to protect their AI-related intellectual property.”89 If there is no patent disclosure and steps to protect the secrecy of AI-based business methods make it hard to learn and copy, then imitators will need access to the inputs used by the innovator to develop the new business method. Attempts to imitate will fail if competing firms cannot access engineers with the right skills and the data used to train machine learning algorithms. There are reports of a “critical talent shortage”90 slowing the diffusion of AI, but it is hard to know how extensive and long-lasting that might be.91 More significantly, limits on data access may create durable barriers to imitation in many settings.92 Dominant firms in an industry naturally have bigger and more diverse customer databases. Smaller imitators who want to develop machine learning-based algorithms to support innovative marketing, advertising, and pricing practices may end up with inferior processes because of their smaller databases.93 85

86

87

88 89 90

91

92

93

Open source program Hadoop is widely used by business for distributed database management. Apache Hadoop TensorFlow is an open source platform for building machine learning models. TensorFlow www.tensorflow.org. Cockburn et al., supra note 79, at 140 (explaining how deep learning may reduce the marginal cost of research dependent on prediction). “[I]t is likely that most intellectual property generated in the United States today related to AI is being protected through the use of trade secrets.” Jordan R Jaffee et al., supra note 52. Id. Id. Holger Hürtgen, Sebastian Kerkhoff, Jan Lubatschowski, and Manuel Möller, “Rethinking AI Talent Strategy as Automated Machine Learning Comes of Age” (Aug 14, 2020) McKinsey Analytics. Christian Rammer, Gastón P Fernández, and Dirk Czarnitzki, “Artificial Intelligence and Industrial Innovation: Evidence from Firm-Level Data” (2021) 4 Leibniz Center for Economic Research Discussion Paper No 21-036. A durable advantage arises in part because algorithms get updated as new data becomes available. “[A]lgorithms are retrained as more data accumulates. Roughly a quarter of firms report refreshing their models daily, weekly, or monthly each. 13% of firms report having models that are not refreshed with new data.” James Bessen, Stephen Michael Impink, Lydia Reichensperger and Robert Seamans, “The Business of AI Startups” 2018) 18 Boston University School of Law & Economics Series, Working Paper No 18–28. It is not clear how much of an advantage flows from access to a larger set of data used to train an algorithm. Some reports describe significant quality increases derived from increasing the size of databases over a broad range, other reports suggest, in some settings, there is a quality plateau once a certain database size is reached. Also, researchers are discovering techniques that promise to reduce data needs required for effective training. See, eg, Martin J Willemink et al., “Preparing Medical Imaging Data for Machine Learning” (2020) 295 Radiology 4; Theophano Mitsa, “How Do You Know You Have Enough Training Data?” (Apr. 22, 2019) . Firms can protect both their data and algorithms using contracts and trade secret law. Bessen et al., supra note 92 at 19 (“To protect their access to data, startup firms who use customer data retain secondary reuse rights 52% of the time. To control the use of proprietary data between the firm and its customers, 83% of the firms use legal contracts that specify data uses. Additionally, firms use a variety of technical means to protect and control data access, including deidentification, encryption, passwords, access logs, and application program interfaces . . .”.)

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

126

Michael J. Meurer

In other settings, it is not size but expertise or exclusive access to specialized data that will give a firm an advantage in training an algorithm.94 AI start-ups and their funders are mindful of the possibility of building a durable business advantage by having special access to data.95 “Proprietary data – data that a firm can exclude others from using – is the most important type of data for AI start-up growth. [Bessen et al. use] a recent survey to show that AI start-up firms that use proprietary data receive more venture capital (VC) funding.”96 Finally, it is important to acknowledge that access to data does not create a barrier to imitation when the necessary data is provided by the government, available from an open source repository, or available at competitive prices in the market from data brokers.97 There are hundreds data brokers in the USA.98 They may be useful sources of data that encourages imitation for certain algorithms related to targeted advertising, background checks, credit, and risk mitigation.99 *** Thus far I have argued that business method innovation is incentivized largely by nonpatent factors like network effects, reputational and lead time advantages, and trade secrecy. Further, I have argued that in the Information Age it is increasingly difficult to get broad patent claims or detect infringement in the case of business method inventions. Why then are business method patent applications and grants still common? Often there is private value in business method patents that can be asserted in an opportunistic, anticompetitive, or other strategic way. These patents may not offer protection over a technology that the inventor intends to commercialize, but they may be used valuably to harass another firm when it introduces a new technology. Section VII.V describes the social costs of strategic patenting of business method patents. VII.V THE SOCIAL COSTS OF BUSINESS METHOD PATENTS

Strategic prosecution and assertion of business method patents creates multiple social costs. First, a patent “arms race” arose in the semiconductor industry in the 1990s, when semiconductor firms amassed defensive patent portfolios to deter competitor suits.100 La Belle and Schooner forecast that a similar arms race is emerging in the financial industry.101 Given the disruption to the industry caused by fin-tech, there is a risk that a patent détente will not last and 94

95

96 97

98

99 100

101

Outside of the business method context, IBM teamed with Sloan Kettering to gain access to 12 million pages of medical literature and patient case histories to train health care AI. Memorial Sloan-Kettering Cancer Center, IBM to Collaborate in Applying Watson Technology to Help Oncologists (Mar. 22, 2012) . “As such, using proprietary training data leads to less imitable products, positively impacting a startup’s ability to collect additional rents from the market and develop an initial competitive advantage in this nascent industry.” James Bessen, Stephen Michael Impink, Lydia Reichensperger and Robert Seamans, “The Role of Data for AI Startup Growth” (2021)13 Boston University School of Law Research Paper Series No 21–23. Id. Besssen et al., supra note 92, at 25 (“80% of startups use customer data and 63% use data available from third parties, including publicly available data. While data might pose a barrier to entry in some markets, like search, where large amounts of diverse data are needed, there are clearly many markets where it does not”). See Paul Boutin, “The Secretive World of Selling Data about You” NEWSWEEK (May 30, 2016) . See Federal Trade Commission, “Data Brokers: A Call for Transparency and Accountability” (May 2014). Bronwyn Hall and Rosemarie Ziedonis, “The Patent Paradox Revisited: An Empirical Study of Patenting in the Semiconductor Industry, 1979–1995” (2001) 32 Rand Journal of Economics 1. La Belle and Schooner, supra note 39, at 434; Megan M La Belle and Heidi Mandanis Schooner, “Fintech: New Battle Lines in the Patent Wars?” (2020) 42 Cardozo Law Review 277, 339–46.

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

Bilski and the Information Age a Decade Later

127

the industry could move to litigation battles.102 Second, low quality patents that are narrow in scope or possibly invalid can be used by established firms to slow or block new competitors.103 Amazon may have been practicing this strategy when it sought a preliminary injunction against Barnes & Noble to stop their online book sales at the start of the holiday shopping season.104 Third, strategic patenting creates a simple numbers problem – a deluge of applications slows examination and creates a thicket of patents that innovators may need to evaluate when they conduct patent clearance reviews.105 Fourth, substantial costs arise from weak or frivolous patent assertions intended to extract settlement payments from targeted firms.106 Many commentators apply the label “patent troll” to these asserters, which of course builds in a normative judgment. In what follows I will use the less freighted term patent assertions entities (PAEs) and provide a variety of evidence about the social costs from this practice. In theory, social gains rather than social costs could arise from PAE activity. A favorable narrative identifies PAEs as specialists who identify and purchase valuable patents and monetize them for the (direct or indirect) benefit of the original owners of those patents. The monetization process requires detecting users of the patented technology and negotiating a license payment backed up by the threat of litigation. These settlement payments are not social costs, but instead socially valuable transfer payments that support the inventive effort of the original patent owner. In practice, the favorable narrative breaks down because PAEs typically monetize low quality patents that are not associated with significant technological advances, and the targets of the assertions are socially valuable innovators vulnerable to a patent assertion because they introduced new technology.107 The threat of these assertions increases the cost of introducing a new technology and imposes a business method patent tax on innovators.108 102

103

104

105

106

107

108

Eugene Mar, and Ashleigh Nickerson, “Tips for Banks as USAA Check Deposit Patent Dispute Grows” (Jan 28, 2021) Law 360 (USAA owns patents covering remote check deposit. “Wells Fargo initially tried to challenge [three of] these patents under the covered business method review, arguing that the patents were invalid for claiming the abstract concept of taking a photograph. The PTAB dismissed the CBM petitions on the basis that USAA’s patents provided a technical solution for capturing images of a check remote deposit and thus fell into the ‘technological invention’ exclusion for CBM review.”) Michael J Meurer, “Controlling Opportunistic and Anticompetitive Intellectual Property Litigation” (2003) 44 Boston College Law Review 509; Ted Sichelman, “The Vonage Trilogy: A Case Study in ‘Patent Bullies’” (2014) 90 Notre Dame Law Review 543. Wagner, supra note 39, at 17-20, describes the anti-compeitive use of business method patents by Pitney-Bowes, the dominant firm in the franking machine market. Leslie Kaufman, “Amazon Sues Big Bookseller over System for Shopping” New York Times (Oct 23, 1999) . Gaétan de Rassenfosse and Alexandra Karin Zaby, “The Economics of Patent Backlog” (July 10, 2016) ; La Belle and Schooner, supra note 101 at 347. Daniel Harris Brean, “Business Methods, Technology, and Discrimination” (2018) Michigan State Law Review 307, 313–14. “Computer-implemented business practices are the clear favorite type of patent asserted by PAEs. Those kinds of methods-involving, eg online shopping, digital marketing, and payment processing-tend to be widely used by many successful businesses, allowing a single patent to be enforced against many such businesses to collect license or settlement fees from each. Making such methods largely ineligible for patent protection greatly diminishes the ‘in terrorem power’ of PAEs, albeit indirectly.” Id. NPEs acquire patents with vague claims and greater obviousness problems. See Josh Feng and Xavier Jaravel, Who Feeds the Trolls? Patent Trolls and the Patent Examination Process (Harvard University Working Paper 2016) (“NPE patent portfolios are disproportionately made up of patents that were granted by ‘lenient’ patent examiners, that is, examiners who spend relatively little time reviewing and narrowing patent claims.”) See James Bessen and Michael J. Meurer, Patent Failure: How Judges, Bureaucrats, and Lawyers Put Innovation at Risk (Princeton University Press 2008)144; La Belle and Schooner, supra note 39 at 450 (the increase in patent litigation has outpaced the increase in patent grants); Hunt, supra note 53, at 339 (business method patents are litigated at a higher rate compared to patents at a whole); Josh Lerner, “The Litigation of Financial Innovations”

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

128

Michael J. Meurer

Many commentators have noted the problem of low quality business method patents that may be asserted in socially harmful ways.109 Critics contend that the PTO grants many patents, including business method patents that lack novelty or are obvious.110 Other critics emphasize that lack of clarity in patent claims degrades the notice that patents should provide about scope of rights.111 The combined effect of these quality problems is that a firm may inadvertently stray within the bounds of a patent claim thinking the claim was invalid or that the claim would be read more narrowly. Increasingly, firms are taking defensive measures to mitigate potential harm from PAE assertion of business method patents.112 Economists have identified certain characteristics associated with low quality patents, and business method patents do not fare well in comparison to other patents. First, they are more often opposed at the European Patent Office,113 and they are more frequently litigated.114 Frequent challenges suggest these patents contain invalid claims or have uncertain scope.115 Second, they cite less nonpatent prior art.116 Patents on inventions that make strong technological advances tend (2008) National Bureau of Economics Research Working Paper No 729 (finding business method patents are litigated at a rate 27 times higher than for patents as a whole). 109 See, eg, eBay Inc v MercExchange, LLC 547 US 388, 397 (2006) (Kennedy, J, concurring)); John R Allison, Mark A Lemley, and Joshua H Walker, “Extreme Value or Trolls on Top? The Characteristics of the Most Litigated Patents” (2009) 158 University of Pennsylvania Law Review 1, 18 (finding software business method patents are overrepresented in their group of “most-litigated” patents). 110 Dan L Burk, “The Role of Patent Law in Knowledge Codification” (2008) 23 Berkeley. Technology Law Journal 1009, 1027 (tacit knowledge of management innovations cannot be used as prior art which makes it easier for business method inventions to jump the nonobviousness hurdle). 111 Bessen and Meurer, supra note 108, at 153 (finding that claim construction of business method patents was appealed to the Federal Circuit 6.67 times more often than the typical patent). But see Spulber, supra note 12 at 310–13 (addressing arguments that business method patents are intrinsically vague, and concluding that generally applicable standards are sufficient to weed out vague patents). 112 Tim Anderson, “Wells Fargo Patent Troll Case has Finance World all Aquiver so Barclays, TD Bank Sign up to Open Invention Network” The Register (Feb 15, 2021) . (Barclays Bank and Toronto-Dominion Bank Group are joining the Open Innovation Network to mitigate harm caused by patent troll lawsuits in the financial sector.) Eugene Mar and Ashleigh Nickerson, “Tips For Banks as USAA Check Deposit Patent Dispute Grows” (Jan 28, 2021) Law 360 (“Lastly, there will undoubtedly be renewed focus by banks on their vendor agreements, especially at times of renewal, to ensure that the vendor provides a robust indemnity provision along with ironclad warranties of no intellectual property liability. Conversely, such warranties and robust indemnity protection will likely come at a more expensive price, but that may still pale in comparison to expensive litigation that results in nine-digit damage awards and attorney fees in the millions.”) . Susanne M Hopkins, Patent Trolls Continue to Target Financial Institutions, But Change May Be Near, The Bankers Statement (Spring 2014) (advising financial institutions to seek indemnification from technology vendors to gain protection against patent trolls). 113 Wagner supra note 39, at 22. Business method patents are more likely to be opposed than other patents, even after controlling for the identity of the patent holder. 114 Bessen and Meurer, supra note 108, at 153 (finding that the rate of lawsuits filed per patent is nearly ten times higher for business method patents than the typical patent); Josh Lerner, Mark Baker, Andrew Speen and Ann Leamon, Financial Patent Quality: Finance Patents After State Street (Harvard Business School Working Paper No 16-068, 2015) (“finance patents are more likely to be litigated than non-finance patents, but increased academic citations appear to reduce that possibility relative to others”); Business method patents relating to financial innovations are especially likely to be the subject of litigation La Belle and Schooner, supra note 39 at 454. 115 More valuable patents also tend to get challenged more often, but business method patents and software patents generally tend to have low average value. Bessen and Meurer, supra note 108, at 153. 116 Lerner et al., supra note 114. (“We show that relative to two sets of comparison groups, finance patents in aggregate cite fewer nonpatent publications and especially fewer academic publications.”) (“In addition, it appears that patents assigned to individuals and associated with non-practicing entities (NPEs) cite less academic work than those assigned to non-NPE corporations. While not statistically significant due to the small number of academic citations in finance patents, we observe qualitatively similar patterns of under-citation when we restrict our analysis to finance patents held

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

Bilski and the Information Age a Decade Later

129

to cite more nonpatent prior art, while patents that are accrued for strictly strategic reasons may be prosecuted less carefully and thus cite less prior art. Third, they take longer to prosecute and they are older when asserted.117 These characteristics are associated with low quality if they reflect skepticism by examiners or the strategic choice to move the patent slowly through the system to surprise rivals when it is eventually granted. PAEs are willing to acquire and assert low quality patents because litigation costs tend to fall more heavily on alleged infringers during the early stages of litigation – this makes many targeted firms willing to make early settlement payments to PAEs.118 Though sometimes the target resists and a PAE that wants to maintain a reputation for being tough must litigate. A recent dispute between Innovation Science and Amazon is a good illustration.119 Innovation Science bundled eleven patents in a confusing mélange of assertions that a range of Amazon products used patented methods of making secure credit card payments over the Internet, displaying video transmitted over the Internet on televisions, and alerting Alexa users when a diaper needs to be changed.120 Most of the asserted claims were invalidated as ineligible subject matter,121 Amazon prevailed completely and was awarded fees from Innovation Science because of the frivolous nature of the assertions.122 A growing body of empirical research measures costs arising from opportunistic patent assertions and provides evidence that PAE activity depresses research and the performance of innovative firms. Some of the research I cite is specific to business method patents but most of the research addresses PAE activity. Since PAE activity overwhelming involves software patents, and since the effect of assertion of business method patents is likely not different from the effect of assertion of other sorts of software patents, the general evidence is quite relevant.123 One strand of research evaluates the impact of PAE activity on alleged infringers who are publicly traded firms. A prominent study found that public firms decrease their R&D on average by about 20 percent in response to PAE assertions.124 A second study finds that the

117

118

119

120 121 122

123

124

by individuals and NPEs, as opposed to non-NPE corporations. These findings raise questions about the quality of finance patents.”) However, Wagner finds that European business method patents tend to cite more prior art than other patents and examination tends to take almost a year longer. Wagner, supra note 39, at 13. Business method patents receive an average of two times as many citations as other patents, but it is unclear if patent holders are making more money from those patents. Id at 14. See Brian J Love, “An Empirical Study of Patent Litigation Timing: Could a Patent Term Reduction Decimate Trolls without Harming Innovators?” (2013) 161 University of Pennsylvania Law Review 1309, 1312; Michael Risch, “Patent Troll Myths” (2012) 42 Seton Hall Law Review 457, 490; Lauren Cohen, Umit G Gurun and Scott Duke Kominers, “Patent Trolls: Evidence from Targeted Firms” (2019) 65 Management Science 5461, 5470. There are other notable differences between patents asserted by practicing entities and nonpracticing entities. NPEs assert the same patent more often, their patents have more independent claims, and their patents are issued at times when the USPTO is especially busy. Cohen, Gurun and Kominers, “Patent Trolls” 5470–71. See La Belle and Schooner, supra note 39 at 459. (One common patent troll tactic is acquiring vague patents and broadly claiming infringement to extract licensing fees while avoiding litigation. Because patent trolls often do not manufacture products, they can litigate more aggressively because of the low countersuit risk and because their discovery costs are relatively low.) Innovation Science sued Amazon in both the Eastern District of Virginia, and the Eastern District of Texas. Jack Queen, “Amazon Alexa Devices Didn’t Infringe Patents, Jury Finds” (Sept 3, 2020) IP Law 360. Innovation Sciences, LLC v Amazon.com, Inc, 778 F App’x 859 (Fed Cir 2019). Va Innovation Scis, Inc v Amazon.com, Inc, 227 F Supp 3d 582 (ED Va 2017). Innovation Scis, LLC v Amazon.com, Inc, No 1:16-cv-00861, 2020 WL 4934272 (ED Va Feb 18, 2020); Queen, supra note 119. See James Bessen and Michael J. Meurer, “The Direct Costs from NPE Disputes” (2014) 99 Cornell Law Review 387, 413, 418. See Cohen, Gurun and Kominers, supra note 117 at 5477.

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

130

Michael J. Meurer

constraints imposed on patent prosecutors by Alice resulted in increased R&D by firms that faced a heightened risk of PAE suits.125 Another strand of research undercuts the view that business method patents promote hightech start-ups and thereby offer social benefits. Spulber conjectures that stringent patent protection of business methods would encourage entrepreneurs to create new business methods, thus decreasing firms’ reliance on corporate R&D and increasing the likelihood of innovative entrepreneurship.126 While some start-ups may benefit from PAE activity, many others are targeted as alleged infringers.127 These assertions disrupt start-up business plans and divert key personnel from essential research and management activity.128 Furthermore, PAE activity interferes with the funding of start-ups. Assertions are often timed to disrupt initial public offerings.129 A pair of econometric studies suggests that: PAE activity caused a 14 percent drop in venture capital funding over a five-year period;130 and the adoption of state antitroll laws “lead to a 4.4% increase in employment at high-tech startups.”131 The laudable effect of the antitroll law was attributed to “[i]ncreased access to financing, both venture capital and patent-backed lending. . .”132 In addition to state antitroll laws which punish bad faith patent assertions,133 there is limited evidence on four other reforms that mitigate harm from PAE activity. Econometric analysis finds that the eBay134 decision, which reduced the availability of injunctive relief and weakened the bargaining power of PAEs, reduced the magnitude of the patent tax on innovators.135 An econometric study of Alice found that affected patent claim scope shrunk, software firms did not lose share value, their sales increased, and they substantially increased their participation in open source projects.136 The authors conclude that narrowing patent

125

126

127

128

129

130

131

132 133

134 135

136

See Sridhar Srinivasan, “Do Weaker Patents Induce Greater Research Investments?” (2018) SSRN Working Paper . See Spulber, supra note 12 at 293–96. See also Hunt, “Business Methods Patents” supra note 37 at 10 (“[I]n certain areas of insurance . . . an innovating firm incurs the expense required to develop a new product and to obtain the necessary regulatory approvals. If the new product is successful, it is quickly imitated by competitors . . . In such an environment, the availability of patents may enable entry by new firms that do not own the complementary assets enjoyed by established firms.”) See Colleen Chien, “Startups and Patent Trolls” (2014) 17 Stanford Technology Law Review 461; La Belle and Schooner, supra note 101 at 348. See Colleen Chien, “Reforming Software Patents” (2012) 50 Houston Law Review 2 (software startups found that 41% reported “significant operational impacts” from patent troll lawsuits, causing them to exit business lines or change strategy); Robin Feldman, “Patent Demands and Startup Companies: The View from the Venture Capital Community” (2014) 16 Yale Journal of Law and Technology 236. See Robin Feldman and Evan Frondorf, “Patent Demands and Initial Public Offerings” (2015) 19 Stanford Technology Law Review 52 (surveying in-house legal staff at companies that have recently gone public and finding almost half of all responding companies received patent demands either shortly before their IPO or within a year following its completion). See Stephen Kiebzak, Greg Rafert and Catherine E Tucker, “The Effect of Patent Litigation and Patent Assertion Entities on Entrepreneurial Activity” (2016) 45 Research Policy 218. See Ian Appel, Joan Farre-Mensa, and Elena Simintzi, “Patent Trolls and Startup Employment” (2018) 133 Journal of Financial Economics 708, 708. Id. See Qian Huang, Grace King, and Tim Rawson, “Navigating the Landscape of Anti-Trolling Legislation” (June 2016) Pillsbury Law . eBay Inc v MercExchange, LLC, 547 US 388 (2006). See Filippo Mezzanotti, “Roadblock to Innovation: The Role of Patent Litigation in Corporate R&D” (2021) 67 Management Science 7362, s; Filippo Mezzanotti and Timothy Simcoe, “Patent Policy and American Innovation after eBay: An Empirical Examination” (2018) SSRN Working Paper . See Lin and Rai, supra note 52 at 3.

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

Bilski and the Information Age a Decade Later

131

protection for software could have both private and social benefits.137 An econometric study of the Second Pair of Eyes program in which the USPTO examined business method patents more carefully than other types of patents suggests the program succeeded in increasing the length of approved claims which is a proxy for reduced claim scope.138 Finally, mixed anecdotal evidence suggests either that the covered business method review (CBM) mitigated harm from PAE activity, or that it had little effect.139 The CBM was created by the America Invents Act for patents relating to financial products as a cheaper alternative to federal district court for invalidating business method patents.140 Supporters of the CBM believed it would weed out low quality business method patents and reduce the leverage of PAEs in patent litigation.141 A sunset provision terminated the program in September of 2020.142 An indication that the program was effective is that many PAE lawsuits involving financial patents appear to have been delayed until after the program was terminated.143 Some of the benefits of the program are also provided by other review mechanisms created by the America Invents Act, and it will take some time before enough data is available to rigorously evaluate the impact of the CBM on PAE assertion of financial patents. VII.VI CONCLUSION

American patent law missed an opportunity to carve out a categorical exclusion of business method patents in Bilski. But Bilski and Alice moved away from the laissez-faire approach of State Street. Did those cases arrive at an optimal eligibility rule? Did they change behavior much compared to State Street? Few commentators think the fuzzy doctrine embedded in the Mayo test is an optimal rule. And given the continued popularity of business method patenting, and the continued harm arising from PAE business method patent activity, the state of affairs has changed less from the dot.com era than many commentators claim.

137

138

139

140

141

142 143

Id at 14–15. This result complements the survey by Taylor who found that Alice did not have a significant negative effect on financing firms in the information and communication technologies: “most investors (62%) were not familiar with any of the Supreme Court’s eligibility cases, and even among investors with familiarity most (61%) had not changed their investment decision making after these decisions.” Taylor, supra note 66 at 2089. See Teruki Amano, “The Effect of USPTO’s Quality-improving Initiatives in 2000 on the Claim Scope of Business Method Patents” (2020) SSRN Working Paper . See La Belle and Schooner, supra note 39 at 459; “CBM Review: A Postmortem” National Review (Sept 3, 2020) . See Jarrad Wood and Jonathan R Stroud, “Three Hundred Nos: An Empirical Analysis of the First 300+ Denials of Institution for Inter Partes and Covered Business Method Patent Reviews Prior to in Re Cuozzo Speed Technologies” (2015) 14 John Marshall Review of Intellectual Property Law 112, 131; Matthew Bultman, “Banks Face Lawsuit ‘Frenzy’ After Business Patent Reviews End” (Apr. 13, 2021) Bloomberg Law . See Daniel Harris Brean, supra note 106 at 313. (“The creation of the CBM program was ostensibly motivated by two related factors: (1) skepticism concerning the quality and strength of many business method patents; and (2) the observation that patent assertion entities (PAEs), also known as ‘patent trolls,’ have wielded weak business method patents to obtain many settlement payments in mass litigation campaigns.”) See Bultman, supra note 139. Id (“Lawsuits against banks and e-commerce companies over financial services patents are piling up, following the expiration of a patent office challenge process that many saw as a potent defense mechanism against some litigation. Nearly three times as many patent suits have been filed against financial institutions such as JPMorgan Chase Bank NA and Bank of America Corp since August 2020 as in the previous eight months, Bloomberg Law data show. Ecommerce companies have also faced new patent suits.”)

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

132

Michael J. Meurer

Software as a service and machine learning implementation of business methods are making trade secrecy and other nonpatent sources of return on investment in business method R&D increasingly important, while the social costs of business method patents do not seem to be declining. The case for categorical exclusion of business method patents is stronger today than it was in 2010.

https://doi.org/10.1017/9781108989527.010 Published online by Cambridge University Press

VIII Pharmaceutical Patents and Evergreening Hazel V. J. Moir Centre for European Studies, Research School of Social Sciences, The Australian National University

VIII.I INTRODUCTION

The grant of a patent for a genuinely new chemical entity (NCE) provides for 20 to 25 years protection from generic competition, allowing the patent owner to set prices to recoup the cost of the invention. Evergreening is a strategy by which patent owners extend the life of a patent monopoly, surrounding an original inventive patent with numerous additional patents for modifications or variations to the original invention (secondary patents). The sub-set of secondary patents owned by the originator company are known as ‘evergreening patents’, that is, patents designed to further delay generic entry to the market. Only a small number of evergreening patents achieve this effect. Numerous articles on pharmaceutical marketing consider such patenting an important part of ‘lifecycle management’, ensuring continuing high profits from the original invention are kept ‘evergreen’. The brief for this chapter is to investigate the motivations which trigger evergreening and to explore the characteristics of evergreening patents, particularly their low degree of inventiveness. The motivation for patent evergreening for pharmaceuticals is simple – it is profitable. Very profitable. What is more complex is understanding why governments continue to allow behaviours which both undermine the purpose of the patent system and substantially add to the cost of healthcare. Pharmaceutical companies contribute substantially to politicians and political parties. They are also major lobbyists. Given that health trumps industry support, clearly additional factors must be at play. An important part of pharmaceutical lobbying is telling stories to support their interests. These stories are designed to prevent action to correct doubledipping and other problems in intellectual property policy. The central story, with respect to discouraging IP policy reform, concerns the alleged cost of developing new medicines, with the direct message that interfering with current arrangements will prevent the development of future new medicines. Earlier chapters have discussed the patent bargain – a bargain where society agrees to high prices during a patent monopoly in exchange to free access to the invention after 20–25 years. The subsequent spread of new knowledge through society can lead to a range of benefits – many competing products at lower prices and productivity increases from widespread adoption of the new knowledge. However, if patents are granted for ‘inventions’ that do not contain new knowledge, or inventions that would occur without the patent incentive, there is no benefit to society from the patent grant. As this chapter will illustrate, evergreening patents contain little

133 https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

134

Hazel V. J. Moir

or no inventiveness, lacking new knowledge. Some evergreening patents can considerably delay generic entry, at a very substantial cost to society. The chapter starts by considering the economics of the pharmaceutical industry – the cost of developing new medicines; who pays for this research and development (R&D); what pharmaceutical revenue is spent on and how profits are used. Attention is then turned to the patent system and the leading role the pharmaceutical industry has played in designing the critical features that allow secondary patents (Section VIII.II). The focus in Section VIII.III is on evergreening patents. Evergreening is closely intertwined with managing periods of data exclusivity, a subject addressed in Chapter III. Substantial research has already been done documenting patent evergreening, but there remain many unanswered questions. The chapter concludes by looking at the impact that the industry’s focus on evergreening patents has on the incentive to develop new medicines that address unmet health needs. Given the incentive structure in patent policy the industry has managed to achieve, it is far more profitable to extend market monopolies for existing medicines, and develop variants of these, than it is to undertake riskier research to develop totally new medicines. Further, medicalising conditions that are neither life threatening nor debilitating can be far more profitable than developing urgently needed new medicines. Numerous voices have called for substantial reforms to re-set the incentive structure for the pharmaceutical industry, but to date governments throughout the democratic world have been impervious to these. VIII.II PROFITS, SPENDING AND THE PHARMACEUTICAL INDUSTRY

The major lobbying story told by the pharmaceutical industry is the cost of developing new drugs. Their estimates are highly controversial. Useful perspectives on these cost estimates are the share of public financing in new drug research and the cost structure and profitability of the pharmaceutical industry. VIII.II.A The Cost of Developing a New Medicine In a systematic review of estimates of drug development costs, Morgan and colleagues found 13 studies.1 The cost estimates have two key elements – actual expenditure on R&D and the opportunity cost of the capital tied up during this lengthy process. However, as the pharmaceutical industry is one of the most profitable (see Section VIII.I.C), inclusion of the opportunity cost of capital is highly questionable.2 Their most important conclusion was that the estimates are unreliable, as for ten of the 13 studies the data were provided by unnamed companies and were about unspecified products. It is difficult, therefore, to accept these estimates as reliable averages. Despite this, they are regularly cited not just by the pharmaceutical industry but also in official reports. Originator pharmaceutical companies use estimates from Di Masi and colleagues to support their story of extremely high R&D costs for new medicines. The work of these authors has received substantial industry funding. The latest estimates from Di Masi et al.,3 like their 1 2

3

Steve Morgan and others, ‘The Cost of Drug Development: A Systematic Review’ (2011) 100 Health Policy 4. For a discussion of other aspects of the controversial inclusion of opportunity costs see Ayman Chit and others, ‘The Opportunity Cost of Capital: Development of New Pharmaceuticals’ (2015) 52 Journal of Health Care Organization, Provision, and Financing 1. Joseph A DiMasi, Henry G Grabowski and Ronald W Hansen, ‘Innovation in the Pharmaceutical Industry: New estimates of R&D Costs’ (2016) 47 Journal of Health Economics 20.

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

Pharmaceutical Patents and Evergreening

135

2003 estimates, are based on unnamed companies and unnamed compounds. Their estimates indicate a substantial increase in the cash cost of new drug development (from 2013 US$403 m to $1,395 m), a fall in clinical trial success rates and a relative increase in the cost of Phase III clinical trials compared to the earlier stages of drug development. If post-marketing R&D is included, cash costs increase to US$1,861 m. If the opportunity cost of capital tied up over the long period of drug development is included, the R&D cost is US$802 m (2003 estimate), US$2,558 m (2016 estimate) or US$2,870 m if post-market entry costs are included.4 Prasad and Mailankody find the cost of developing new cancer drugs to be only a quarter of the mean cost estimates advised by the pharmaceutical industry.5 The pharmaceutical industry’s cost estimates have been substantially criticised. Light and Warburton, for example, take the estimates apart to point out six fundamental flaws, including important biases in the data.6 Further there is substantial variation in cost, depending critically on the therapeutic area.7 Despite these criticisms and alternative estimates of costs, the industry estimates are repeatedly presented, including in official reports. VIII.II.B Public Contributions to the Costs of Developing New Medicine Most basic medical and biomedical research is publicly funded. A recent study on the role of US National Institutes of Health (NIH) in biomedical research funding found that the NIH contributed to research associated with each of the 356 new drugs approved by the US Food and Drug Administration (FDA) from 2010 to 2019, totalling US$230 billion. The scale of this contribution is not fully recognised – only 9 per cent of these drugs had patents that were explicitly based on the NIH research funding.8 The NIH contribution to research underlying the 136 first-in-class drugs and their biological targets was more than $190 billion dollars, an average of $1.4 billion dollars per first-in-class drug.9 Because of the long lead times from basic research to a useable medicine, if the cost of capital for this NIH funded basic research were added, the estimates would multiply substantially. There is general agreement that the largest R&D cost outlays for originator companies is for Phase III clinical trials, yet failure rates here, especially for genuinely new drugs, are high. Grainger suggests that were more appropriate statistical metrics used for choosing which potential drugs should proceed to Phase III trials, failure rates could be halved, saving billions in drug development costs.10 There are criticisms of the role of originator companies in clinical trials for new medicines, with evidence that authorship is not correctly reported, and that data can be manipulated to hide negative impacts.11

4 5

6 7

8 9

10

11

All these figures are in 2013 US dollars. Vinay Prasad and Sham Mailankody, ‘Research and Development Spending to Bring a Single Cancer Drug to Market and Revenues after Approval’ (2017) 177 JAMA Internal Medicine 1569. DW Light and R Warburton, ‘Demythologizing the High Costs of Pharmaceutical Research’ (2011) 6 BioSocieties 34. Morgan et al. (n 1); Aylin Sertkaya and others, ‘Key Cost Drivers of Pharmaceutical Clinical Trials in the United States’ (2016) 13 Regulation and Policy 117. Based on entries in the Orange Book (the FDA’s register of patents relevant to approved medicines). Ekaterina Galkina Cleary and others, ‘Contribution of NIH Funding to New Drug Approvals 2010–2016’ (2018) 115 Proceedings of the National Academy of Sciences USA 2329 at 46. David Grainger, ‘Why Too Many Clinical Trials Fail: And a Simple Solution That Could Increase Returns on Pharma R&D’ (2015) Forbes, . See, eg, Marcia Angell, ‘Industry-Sponsored Clinical Research: A Broken System’ (2008) 300 JAMA 1069.

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

136

Hazel V. J. Moir

VIII.II.C Pharmaceutical Industry Profitability and the Cost of Developing New Medicines Prasad and Mailankody usefully place their estimates of the cost of developing cancer drugs in the context of earnings from the new drugs.12 Using US Securities and Exchange Commission filings, they find that R&D costs are only about a quarter of industry estimates of mean R&D costs and that revenue from these new drugs far exceeds their R&D costs. Indeed, despite the alleged risks of pharmaceutical R&D, the industry is one of the most profitable industries.13 Spitz and Wickham provide an in-depth assessment of comparative pharmaceutical industry profits using data for 1988 to 2003 with an update to 2009. In particular, they investigated whether, compared to other industries with high R&D outlays, higher pharmaceutical profits were justified by higher R&D spending. They find that such R&D investment does not account for large ongoing profits in the pharmaceutical industry – profits which range from 2.5 to 37 times those of non-pharmaceutical industries.14 They also find that the higher profits are not spent on R&D. Others have also investigated how originator companies spend revenue and distribute net profits. In respect of expenditure, most major global originator companies continue to spend substantially on drug promotion (marketing), with some evidence that marketing expenditure is close to or exceeds outlays on R&D.15 A House of Commons Health Committee report in the UK noted that major pharmaceutical companies spend about twice as much on marketing as on R&D.16 A recent US National Academies of Sciences, Engineering and Medicine (National Academies) report on the cost of medicines in the USA commented that, while promotion costs are undisclosed, data for 12 large pharmaceutical companies for 2003 to 2015 found ‘expenditures on marketing and administration (a figure that includes executive pay) increased noticeably and exceeded research and development investments by up to 80 per cent’.17 While the originator company story on the costs of R&D suggests they need to heavily re-invest in developing new drugs, in fact US drug companies distribute, rather than re-invest, most of their profits.18 Since the introduction of the first blockbuster drug (cimetidine (Tagamet)) in the late 1970s, there has been substantial publicity for new medicines which achieve very large global revenues – so-called blockbuster drugs. Originally defined as those with global revenues in excess of US 1 billion, the current threshold would be US$ 2.5 billion. Naturally when such extremely profitable medicines come off patent, there is a very substantial incentive for originator 12 13

14

15

16

17

18

Prasad and Mailankody (n. 5). For data on profitability in the period 2006–2015 see US Government Accountability Office, ‘Drug Industry: Profits, Research and Development Spending, and Merger and Acquisition Deals’ 2017 at 16–19. Available at . Janet Spitz and Mark Wickham, ‘Pharmaceutical High Profits: The Value of R&D, or Oligopolistic Rents?’ (2012) 71 American Journal of Economics and Sociology 1 at 27. See for example, Marcia Angell, ‘Excess in the Pharmaceutical Industry’ (2004) 171 Canadian Medical Association 1451; Marc-André Gagnon and Joel Lexchin, ‘The Cost of Pushing Pills: A New Estimate of Pharmaceutical Promotion Expenditures in the United States’ (2008) 5 PLoS Medicine; Richard Anderson, ‘Pharmaceutical Industry Gets High on Fat Profits’ and Ana Swanson, ‘Big Pharmaceutical Companies are Spending Far More on Marketing than Research’ The Washington Post . UK House of Commons Health Committee, ‘The Influence of the Pharmaceutical Industry: Fourth Report of Session 2004–05’ (2005) at 25. Available at . National Academies of Sciences Engineering and Medicine, Making Medicines Affordable: A National Imperative (The National Academies Press 2018) at 90. William Lazonick and others, US Pharma’s Financialized Business Model, Institute for New Economic Thinking Working Papers 2017.

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

Pharmaceutical Patents and Evergreening

137

companies to use all possible strategies to first delay and then limit the impact of generic versions of these medicines. The end of the patent period for blockbusters is referred to as the patent cliff and has attracted substantial interest. VIII.III THE PHARMACEUTICAL INDUSTRY’S ROLE IN DESIGNING THE CURRENT PATENT SYSTEM

In 2020 in the USA the pharmaceutical/healthcare industry spent US$306 million on lobbying. This included nearly US$26 million from Pharmaceutical Research & Manufacturers of America (PhRMA) and between US$6 million and US$11 million by each of Novartis, Bayer, Merck, Gilead, AbbVie, Roche, Pfizer and Amgen.19 Expenditure on US lobbying has increased from $US69 million in 1998 – a more than four-fold increase. In the context of discussing the pervasive influence of the pharmaceutical lobbying in the UK, the House of Commons Health Committee noted that there had been no Select Committee report on the industry since 1914.20 While the pharmaceutical industry has played a lead role in the design of the patent system over time, the most significant features that benefit the industry are: allowing patents for substances which exist in nature; the change in the inventiveness test in the US Patent Act 1952 from inventive to ‘non-obvious’; the global spread of this low standard including through the Trade Related Intellectual Property (TRIPS) Agreement; and the gradual whittling away of the norm that patents should not be granted for methods of medical treatment. Also important is the introduction of new quasi-IP rights for clinical trial data, a right that arose out of political bargaining following a controversial US court decision over a pharmaceutical patent (see Chapter III). VIII.III.A Patenting Discoveries: Chemical Compositions Developing naturally occurring compounds and substances into marketable products involves, inter alia, identifying processes to produce these substances to a specific and reliable standard. Chemical product patents were granted by the mid to late nineteenth century in the UK, France and the USA, but they were not universally seen as desirable. For example, when Germany introduced its first patent law, it deliberately made pharmaceutical products non-patentable. This forced German chemical companies to focus on improving production processes. This regulatory regime is widely considered to be a key factor behind the German chemical industry’s rise to global dominance in the late nineteenth and early twentieth centuries.21 Prior to the TRIPS Agreement most countries did not allow patents for pharmaceutical products.22 Indeed in the early 1970s most European countries excluded pharmaceutical products from patenting. The European Patent Convention (EPC) was an important early mechanism for spreading pharmaceutical product patenting between countries. One of its impacts was the demise of Italy’s global leadership in generic medicine production.23 Since

19

20 21 22

23

Center for Responsive Politics, at . These data are available for the period 1998–2020 for each major lobbyist. UK House of Commons (n. 16) at 97–98. Graham Dutfield, Intellectual Property Rights and the Life Science Industries: A 20th Century History (Ashgate 2003). Kenneth C Shadlen, Bhaven N Sampat and Amy Kapczynski, ‘Patents, Trade and Medicines: Past, Present and Future’ (2020) 27 Review of International Political Economy 75 at 78–79. FM Scherer and Sandy Weisburst, ‘Economic Effects of Strengthening Pharmaceutical Patent Protection in Italy’ (1995) 26 International Review of Industrial Property and Copyright Law 1009. Generally, there is a dearth of analysis

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

138

Hazel V. J. Moir

1995, a condition of World Trade Organisation (WTO) membership has been the grant of product patents for pharmaceuticals. There is, however, substantial empirical evidence that the benefit to originator companies from this TRIPS requirement is a small fraction of the cost to consumers.24 Pharmaceutical product patents, like tariffs, are a very expensive policy choice. VIII.III.B Changing the Inventiveness Test: The 1952 US Patent Act While the evidence that the grant of patents for pharmaceutical products is welfare-reducing is compelling, the negligible standard for inventiveness allows the damage from such patents to be multiplied. The very low quantum of ‘inventiveness’ required for a patent grant allows evergreening patents and these can greatly extended periods of monopoly. Where they relate to blockbuster drugs, the healthcare costs are very high. The challenge in determining where the boundary of patenting lies is that, while inventiveness is a continuum from marginally different to significant breakthroughs, the decision to grant a patent is simply yes or no. Over decades, courts have struggled with a precise definition of inventiveness, usually when considering inventions at the margin of the yes/no decision. The US Supreme Court decision in 194125 that the critical test for inventiveness (and hence a patent) was a ‘flash of genius’ created a substantial problem for the pharmaceutical industry. The development of new medicines was already based on systematic assessment of a large number of potential compounds, that is, on large-scale investment. By 1945 this test had led to 64 per cent of challenged patents being deemed insufficiently inventive.26 The resolution to that problem was the US Patent Act of 1952, written largely by patent lawyers,27 a group which benefits significantly from a higher volume of patenting. Prior to this legislation the key decision was whether an alleged invention was sufficiently inventive to merit a patent. But the new 1952 US law turned the decision around – an alleged

24

25 26

27

of the EPC and its impacts. A particularly interesting set of stories underlie amendments to the EPC – especially those concerning revisions to change the legal norms to fit European Patent Office (EPO) practices. Using data on quinolones in the Indian market Chaudhuri and colleagues estimate the annual consumer welfare loss as US$144–450 million compared to a producer gain of just US$20–53 million (Shubham Chaudhuri, Pinelopi Goldberg and Panle Jia, ‘Estimating the Effects of Global Patent Protection in Pharmaceuticals: A Case Study of Quinolones in India’ (2006) 96 American Economic Review 1477). Using a dataset on 43 medicines in the Indian market Dutta estimates an average consumer welfare loss of US$9 million compared to a gain to patent owners of just US$1.4 million per medicine (Antara Dutta, ‘From Free Entry to Patent Protection: Welfare Implications for the Indian Pharmaceutical Industry’ (2011) 93 Review of Economics and Statistics 160). These two studies suggest that the benefit to producers from these monopolies is only some 11 to 15 per cent of the cost to consumers/health systems. There are, of course, welfare gains from the introduction of genuine and effective new medicines – but this depends on the enhanced therapeutic effect (Farasat AS Bokhari and Gary M Fournier, ‘Entry in the ADHD Drugs Market: Welfare Impact of Generics and Me-Too’s’ (2013) 61 Journal of Industrial Economics 339) and the price (Chirantan Chatterjee, Kensuke Kubo and Viswanath Pingali, ‘The Consumer Welfare Implications of Governmental Policies and Firm Strategy in Markets for Medicines (2015) 44 Journal of Health Economics 255). Using US data on generic entries during the period 1987 to 2008 for hypertension drugs, Branstetter and colleagues estimate consumer gains from earlier generic entry at US$92 billion compared to producer losses of some US$114 million – a massive difference (Lee G Branstetter, Chirantan Chatterjee and Matthew Higgins, ‘Regulation and Welfare: Evidence from Paragraph IV Generic Entry in the Pharmaceutical Industry’ (National Bureau of Economic Research Working Paper no 17188 (2011)). US Supreme Court, Cuno Engineering Corp v Automatic Devices Corp 1941. Herbert H Mintz and C Larry O’Rourke, ‘The Patentability Standard in Historical Perspective: “Invention” to Section 103 Nonobviousness’ in John F Witherspoon (ed) Nonobviousness: The Ultimate Condition of Patentability (The Bureau of National Affairs 1978) at 218, quoting 1955 evidence by the head of the US Patent and Trademarks Office. Giles S Rich, ‘Congressional Intent: Or, Who Wrote the Patent Act of 1952?’ in John F Witherspoon (ed) Nonobviousness: The Ultimate Condition of Patentability (The Bureau of National Affairs 1978).

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

Pharmaceutical Patents and Evergreening

139

invention was inventive unless it was obvious. This is quite a different matter; there is a vast gulf between non-obviousness and inventive. Imagine if deciding for finalists in a beauty contest was based not on beauty but on non-ugliness. Most entrants would be finalists. This is exactly what has happened with patent applications. The now very low ‘inventiveness’ standard for grant of a patent is detailed in Section VIII.III in relation to several types of evergreening pharmaceutical patents. The new law made it easy to obtain patents for very incremental inventions as examiners now had to ‘find and combine earlier published suggestions of the claimed invention to show that it is “obvious”’.28 As the US patent for swinging a swing sideways shows, this is not an easy task – there is not a market for writing down the obvious. Even further reductions in the US standard for inventiveness followed the establishment of the Court of Appeals for the Federal Circuit (CAFC) in 1982. There is a large literature documenting this.29 Other countries simply copied the new US non-obvious definition of inventiveness, starting with Japan in 1959. After several European countries had followed suit, the 1973 EPC built the US non-obviousness definition into European patent law. Kingston documents the role of the World Intellectual Property Organization (WIPO, whose largest funder was the USA) in spreading the US patent system globally.30 So, the test for patentability is now not whether the alleged invention is sufficiently inventive to merit a 20-year monopoly. It is not even whether there is any new knowledge disclosed in the application. In the most recent pluri-lateral treaty negotiations in which the USA was involved, the test is mandated as ‘whether the claimed invention would have been obvious to a person skilled, or having ordinary skill in the art, having regard to the prior art’.31 There is a wealth of limitations built into this legalese, resulting in an inventiveness standard so low that in many countries the bulk of granted patents do not meet the normal definition of inventive.32 Dutfield suggests there is a ‘pharmaceutical privilege’ with respect to how patent

28 29

30 31 32

William Kingston, ‘Why Harmonization Is a Trojan Horse’ (2004) 26 European Intellectual Property Review 447, 451. See, for example, James E Bessen and Michael J Meurer, Patent Failure: How Judges, Bureaucrats, and Lawyers Put Innovators at Risk (Princeton University Press 2008); Robert W Harris, ‘Prospects for Supreme Court Review of the Federal Circuit Standards for Obviousness of Inventions Combining Old Elements’ (1986) 68 Journal of the Patent and Trademark Office Society 66; Adam B Jaffe and Josh Lerner, Innovation and Its Discontents: How Our Broken Patents System Is Endangering Innovation and Progress, and What to Do about It (Princeton University Press 2004); Cecil D Quillen Jr, ‘Innovation and the US Patent System’ (2006) 1 Virginia Law and Business Review 210; Cecil D Quillen Jr, ‘Commentary on Bessen and Meurer’s Patent Failure: An Industry Perspective’ (2008) 16 Journal of Intellectual Property Law 57. Kingston (n. 28) at 453–54. Comprehensive and Progressive Agreement for Trans-Pacific Partnership (CPTPP) Article 18.37, footnote 30. The first qualification is that ‘prior art’ is only a sub-set of existing knowledge. For example, the Australian High Court deemed, in 2007, that knowledge of mortice locks was not relevant to determining inventiveness for rim-mounted locks (Lockwood Security Products Pty Ltd v Doric Products Pty Ltd (No 2) [2007] HCA 21). ‘Art’ means technology field, and if this is very narrowly determined, it reduces the inventiveness standard, as Bagley has demonstrated with respect to the USA (Margo Bagley, ‘Internet Business Model Patents: Obvious by Analogy’ (2001) 7 Michigan Telecommunications and Technology Law Review 253). In addition, the ‘skilled person’ (in the USA the person having ordinary skills in the art (PHOSITA)) is limited by further doctrines as to how they would or would not find something obvious. Until 2007, the PHOSITA was not allowed even ordinary creativity in assessing existing knowledge and the subject matter of the ‘invention’. This CAFC doctrine was eventually overturned by the Supreme Court in KSR v Teleflex, 127 SCt 1727 (2007). In Australia, the courts still use a test of whether the skilled person ‘would be led directly as a matter of course to try a particular approach with a reasonable expectation of success’. This test is extremely low (IP Australia, ‘Getting the Balance Right: Toward a Stronger and More Efficient IP Rights System’ 2009) at 12) and has entirely eliminated the need for any inventiveness (Charles Lawson, ‘Quantum of Obviousness in Australian Patent Laws’ (2008) 19 Australian Intellectual Property Journal 43).

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

140

Hazel V. J. Moir

offices treat pharmaceutical applications, regularly granting patents for applications with a low inventive step and questionable novelty.33 VIII.III.C Global Spread of US Patent Law: TRIPS and Bi-lateral Trade Treaties The role of a small number of large multi-national companies in the creation of the TRIPS Agreement is well known and is documented in detail both by Drahos and Sell.34 Kingston focuses on the role of the Intellectual Property Committee formed by the heads of Pfizer, IBM and Du Pont.35 The key companies used employer groups and direct lobbying to gain government support for their agenda in major trading countries/blocs – principally the USA, Europe and Japan. These are all major global pharmaceutical exporters. What was surprising was the role played by countries which clearly would not benefit from the proposed policies – countries such as Australia, which was a net importer of pharmaceuticals, yet was a ‘friend of intellectual property’ during the Uruguay Round negotiations. Brazil and India valiantly tried to save the world from TRIPS but were unsuccessful. TRIPS came into effect on 1 January 1995. Since then, there has been a proliferation of bilateral treaties pushing for yet further changes to patent and data protection policies, all to the benefit of originator pharmaceutical companies. Sell documents some key actors in these TRIPS-Plus negotiations.36 Generally these treaties limit the use of TRIPS flexibilities as well as extending originator privileges.37 The specific role of pharmaceutical companies in developing the text of the Trans Pacific Partnership Agreement has been documented.38 There was, however, disagreement within the industry body which advises the US government on trade negotiations on intellectual property rights – some members disagreed with incorporating TRIPS objectives and principles requiring balance.39 Overall, because of the time required for these new mandated patent policies to impact on drug prices and access to drugs, there is still much to be done in assessing the actual empirical impact of these changes to patent policy.40 VIII.III.D Use Patents: Having Your Cake and Eating It Too It is useful to remind ourselves of the strength of the privilege granted by a patent. A patent allows the owner to prevent others from ‘making, using, offering for sale, selling, or importing’ the product

33

34

35 36 37

38

39

40

Graham Dutfield, ‘Healthcare Innovation and Patent Law’s “Pharmaceutical Privilege”: Is There a Pharmaceutical Privilege? And If So, Should We Remove It?’ (2017) 12 Health Economics, Policy and Law 453. Peter Drahos (with John Braithwaite,), Information Feudalism: Who Owns the Knowledge Economy (Earthscan 2002); Susan K Sell, Private Power, Public Law: The Globalization of Intellectual Property Rights (CUP 2003). Kingston (n. 28). Susan K Sell, ‘TRIPS-plus Free Trade Agreements and Access to Medicines’ (2007) 28 Liverpool Law Review 41. Jayashree Watal, ‘Is TRIPS a Balanced Agreement from the Perspective of Recent Free Trade Agreements’ in Josef Drexl, Henning Grosse Ruse-Khan and Souheir Nadde-Phlix (eds) EU Bilateral Trade Agreements and Intellectual Property: For Better or Worse? (Springer 2014); Shadlen et al. (n. 22). Deborah H Gleeson and others, ‘How the Transnational Pharmaceutical Industry Pursues Its Interests through International Trade and Investment Agreements: A Case Study of the Trans Pacific Partnership’ in Alice De Jonge and Roman Tomasic (eds) Research Handbook on Transnational Corporations (Edward Elgar 2017). ITAC-15, The Trans-Pacific Partnership Agreement (Report of the Industry Trade Advisory Committee on Intellectual Property Rights (ITAC-15), 2015): 5. Available at . Shadlen et al. (n. 22).

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

Pharmaceutical Patents and Evergreening

141

(TRIPS Article 28).41 The breadth of this privilege prevents others from being ready to enter the market on day 1 after a patent expires, as tooling up (‘making’) is not allowed. The ban on making (as opposed to selling) also allows a patent holder to prevent a competitor from making a drug to export to a market where the patent is not in force.42 A major exception to this strong patent privilege is the so-called Bolar provision, which allows acts designed to gain marketing approval for a generic drug to take place before patent expiry.43 As marketing approval processes can be lengthy this ensures that additional months or years are not added to the 20–25 year patent term.44 Introduction of this policy in the USA reduced the delay in generic market entry from over three years to less than three months.45 However, a WTO dispute settlement panel has ruled that making supplies of the product so that they can be delivered to the market on day 1 after patent expiry is contrary to TRIPS as it is not a ‘limited’ exception.46 The Panel did not refer to the TRIPS objectives of balance (Article 7) in reaching this conclusion, and effectively granted all patent holders an additional monopoly of at least several months beyond the agreed 20-year patent term. The patent on the original NCE provides a monopoly on all uses of the new substance. Nonetheless there is a disturbing trend for patent offices to now grant additional patents for specific uses of previously patented substances. These are ‘method of treatment’ patents. One of the few globally agreed exclusions from patentability is ‘diagnostic, therapeutic and surgical methods for the treatment of humans or animals’ (TRIPS Article 27). Nonetheless such patents are now regularly granted for pharmaceutical substances, and in large numbers. Patent offices in major pharmaceutical exporting nations/blocs have had no difficulty in interpreting the law on this in such a way that when pharmaceuticals are used as a method of medical treatment they are excluded from the exclusion. For example, the original EPC clearly made methods of medical treatment non-patentable, though it allowed the patenting of substances or compounds for use in medical treatment (Article 52(4)).47 Article 54(4) allowed that such substances or compounds met the novelty requirement, provided that their use in a method of medical treatment had not previously been disclosed. But this only allowed the grant of patents for the compound itself. ‘Swiss type use claims’ were developed, by the Swiss Patent Office, as a legal form of expression that circumvented the prohibition on the first, second or subsequent medical use of a known substance or compound. The legal exclusion of methods of using pharmaceutical substances and compounds was eliminated in the 2000 version of the EPC through the addition of Article 54(5). The EPO was thus able to continue its practice of granting patents for specific uses of known compounds without any longer requiring convoluted legal language. 41

42

43

44

45

46

47

While TRIPS eliminated consumer and societal protections (such as the domestic working requirement to ensure the benefit of spillovers) from the patent regime, the breadth of the patent privilege was not simultaneously scaled back to ensure balance between innovator and consumer interests. T Harris, D Nicol and N Gruen, ‘Pharmaceutical Patents Review Report’ (2013) at 49–59. Available at . Jayashree Watal, ‘“Bolar” Exception to Patent Rights: Some Economic omplications’ (SCP Seminar on Exceptions and Limitations to Patent Rights). Available at . While the standard term for a patent is 20 years, in many countries governments allow up to five additional years of protection from competition for new chemical entities. In the European Union the so-called 8+2+1 policy allows up to 11 additional years of ‘protection’ for a range of pharmaceuticals. US Congressional Budget Office, ‘How Increased Competition from Generic Drugs Has Affected Prices and Returns in the Pharmaceutical Industry’ 1998) at chapter 4. WT/DS114/R at . The complainant was the European Community (EC), the defendant Canada. For all versions of the EPC see the webpage .

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

142

Hazel V. J. Moir

VIII.IV LIFECYCLE MANAGEMENT: EVERGREENING PATENTS

Clearly if a company has a profitable product and those profits will be reduced when a monopoly ends, there is a substantial incentive to seek to defer the end of the monopoly. Where the profits are substantial, as for small molecule blockbuster drugs,48 the incentives are very large. Originator companies use many different strategies to manage the lifecycle of their products: perhaps the most cost-effective is low quality evergreening patents.49 Rathod provides an excellent introduction to evergreening, covering not only its definition and the forms it takes, but assessing the extent to which statutory and judicial law has contributed to, or limited, the practice in a number of countries.50 Kesselheim and colleagues provide a review of the literature on market exclusivity periods provided by both patents and data exclusivity provisions and recent proposals to reduce or extend such periods.51 As noted above, the pharmaceutical industry has been instrumental in ensuring that the global patent system has a very low test for inventiveness. For example there is a US patent for crushing a drug and spreading it on applesauce52 and an Australian patent for making a fizzy version of a known compound.53 While not all evergreening patents are quite so absurd, the grant of patents for variants on the compound, production processes, formulations, dosages, new uses, packaging and segmented patient populations can make the process of generic entry at the end of the original patent term difficult, if not impossible. There is a large literature on evergreening patents providing insight into their characteristics, but a much smaller literature based on empirical evidence. The early empirical literature focuses on generic entry and the effective market life (monopoly period) of originator drugs. Much of this analysis is from the USA, which is the largest single pharmaceutical market in the world. Further, because of the Hatch–Waxman Act,54 FDA data allow such analyses.55 The average number of 48

49

50 51

52

53

54

55

There are substantial differences between older generation chemical-based drugs and newer generation protein-based biological drugs (biologics). In particular chemical-based (small molecule) drugs are easier to copy and production costs are relatively cheap. In contrast biologic drugs are harder to imitate and more expensive to produce. While the discussion in the text is largely based on small molecule drugs, the principles with respect to biologic drugs are the same. The greatest difference is that if, as biologic originators claim, it is not possible for biologics to be safely copied, then how is the patenting requirement for disclosure of the best method of production met? And if this requirement is not met, then how is the patent valid? For broader reviews of the range of lifecycle management strategies see Ove Granstrand and Frank Tietze, ‘IP Strategies and Policies For and Against Evergreening’ (2015) Centre for Technology Management working paper. Available at ; and Chie Hoon Song and Jeung-Whan Han, ‘Patent Cliff and Strategic Switch: Exploring Strategic Design Possibilities in the Pharmaceutical Industry’ (2016) 5 SpringerPlus 692. Sandeep Kanak Rathod, ‘Ever-greening: A Status Check in Selected Countries’ (2010) 7 Journal of Generic Medicines 227. Aaron S Kesselheim, Michael S Sinha and Jerry Avorn, ‘Determinants of Market Exclusivity for Prescription Drugs in the United States’ (2017) 177 JAMA Internal Medicine 1658. AS Kesselheim and MM Mello, ‘Medical Process Patents: Monopolizing the Delivery of Health Care’ (2006) 355 New England Journal of Medicine 2036 at 2039. Australian patent 712325 for the effervescent form of omeprazole or S-omeprazole. This was one of 53 low quality patents surrounding the original (529654) compound patent (Harris et al., n. 42 at 226). The 1984 US Drug Price Competition and Patent Term Restoration Act (Public Law 98-417) is informally known as the Hatch–Waxman Act. It authorises generic companies to use originator company clinical trial data to speed marketing approval through a process called an Abbreviated New Drug Application (ANDA). At the same time, it provides a four-year period during which clinical trial data for NCEs cannot be used for marketing approval. Companies are required to list all patents relevant to an approved drug on the FDA’s Orange Book. See the webpage . Though, as Feldman points out, they are very user-unfriendly and require very substantial resources to convert to a form that is useable for policy analysis (Robin Feldman, ‘May Your Drug Price Be Evergreen’ (2018) 5 Journal of Law and the Biosciences 590). The Purple Book, for biologic drugs, does not require relevant patents or exclusivities to be listed, so is even more unfriendly for policy analysis purposes.

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

Pharmaceutical Patents and Evergreening

143

patents per drug has increased over time, with more patents per drug where there was priority FDA approval or a patent term extension. These two variables are associated with blockbuster medicines – the 15 best-selling US drugs in 2002–04 have an average of five patents per drug, compared to an overall average of 2.5 in the late 1980s and 3.5 by 2005.56 Hemphill and Sampat also use FDA Orange Book data for 1985 to 2002 to show how the volume of evergreening patents has grown since the introduction of the Hatch–Waxman Act in 1984.57 In addition they show that generic challenges focus more on larger selling drugs, and on those evergreening patents which attempt the greatest extension in patent duration. The overall effect is that the effective market life – the period between FDA approval of an original new medicine and the first generic entry – has remained at about 12 years – in fact a longer duration than before the Hatch–Waxman Act.58 Kapczynski and colleagues, using Orange Book data for 1985 to 2005, confirm the substantial increase over time in evergreening patents in the USA.59 Their analysis focuses on the timing and type of evergreening patents. They find that evergreening patents are more common for drugs with large sales and tend to occur after FDA marketing approval. They also find that chemical variation patents add an average of 6.3 years to the nominal patent term,60 formulation patents (e.g., changes in tablet forms, dosage amount, and sustained release forms) add 6.5 years and method of use patents add 7.4 years. At a more granular level, Amin and Kesselheim investigate US evergreening patents for two HIV drugs, both owned by Abbott.61 They identified 82 granted patents and 26 applications covering compositions/formulations, processes, treatment methods and/or general patents. Overall, these patents added 12 years to the nominal patent term for ritonavir and lopinavir. Empirical data on evergreening in other jurisdictions are scarcer. In Canada mean patents per medicine are 40 – far higher than in the USA.62 In the period 1990–2000 in Australia there are approximately 12 evergreening patents among the 49 secondary patents attached to 15 of the costliest medicines.63 In its 2009 enquiry into the pharmaceutical industry the European Commission found substantial evidence of evergreening in Europe, also finding a larger number of patents for best-selling drugs.64 Abud et al. provide evidence of similar patent evergreening strategies in Chile.65 The USA is the only jurisdiction where there is an incentive for generic companies to actively challenge weak patents. The reward is 180 days before other generic companies can enter the 56

57

58

59

60

61

62

63 64

65

Lisa Larrimore Ouellette, ‘How Many Patents Does It Take to Make a Drug? Follow-On Pharmaceutical Patents and University Licensing’ (2010) 17 Michigan Telecommunications and Technology Law Review 299 at 314–17. C Scott Hemphill and Bhaven N Sampat, ‘When Do Generics Challenge Drug Patents?’ (2011) 8 Journal of Empirical Legal Studies 613. C Scott Hemphill and Bhaven N Sampat, ‘Evergreening, Patent Challenges, and Effective Market Life in Pharmaceuticals’ (2012) 31 Journal of Health Economics 327 at 336. Amy Kapczynski, Chan Park and Bhaven Sampat, ‘Polymorphs and Prodrugs and Salts (Oh My!): An Empirical Analysis of “Secondary” Pharmaceutical Patents’ (2012) 7 PLoS ONE e49470. The nominal patent term is the time between marketing approval for the originator brand and expiration of their last expiring patent. Tahir Amin and Aaron S Kesselheim, ‘Secondary Patenting of Branded Pharmaceuticals: A Case Study of How Patents on Two HIV Drugs Could Be Extended for Decades’ (2012) 31 Health Affairs 2286. Ron A Bouchard and others, ‘Empirical Analysis of Drug Approval–Drug Patenting Linkage for High Value Pharmaceuticals’ (2010) 8 Northwestern Journal of Technology and Intellectual Property 174. Andrew F Christie and others, ‘Patents Associated with High-Cost Drugs in Australia’ (2013) 8 PLoS ONE 1. ‘Finding More Patents per Medicine for Higher Volume Medicines (European Commission, Final Report: Pharmaceutical Sector Inquiry 2009) at 10–11). Available at . María José Abud, Bronwyn Hall and Christian Helmers, ‘An Empirical Analysis of Primary and Secondary Pharmaceutical Patents in Chile’ (2015) 10 PLoS ONE e0124257.

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

144

Hazel V. J. Moir

market. Without this reward there would be little incentive for a generic company to mount a patent challenge – the challenging company would bear the substantial legal costs while all generic companies would be able to enter the market once the patent was revoked. For originator companies, even very weak patents are profitable as market profits ‘easily outweigh the costs of patent litigation’.66 As Granstrand and Tietze point out with respect to omeprazole in the USA, delays in litigation can be very expensive for patients and health systems.67 A lack of proper incentives to challenge weak pharmaceutical patents creates substantial unnecessary costs to patients and national health systems. More recent empirical work on evergreening patents is either about particular medicines,68 extends analysis to other countries69 or addresses very specific issues, such as the role of Medicare Part D.70 The exception is Feldman’s investigation of whether evergreening practices are limited to a few firms or are a more general feature of the US pharmaceutical market. Her research, using Orange Book data for the period 2005–2015, indicates clearly that evergreening strategies – with respect to both patents and data exclusivity privileges – are the norm.71 Successful evergreening patents are those which delay generic entry. A key mechanism is shifting the market norm from the original patented product to a more recently patented variant. If the variant becomes the market norm, there is little point in a generic company producing the original product as prescribers and patients have moved on. Two of the most successful normshifting types of evergreening patent are those which cover improved methods of delivering the drug and those which provide an alternative compound, with the same effect, but marketed under a new name and covered by a new patent. The remainder of this Section looks at specific types of evergreening patents, commenting on the doctrines (patent policies) which give rise to such patents. VIII.IV.A Formulations: New Drug Delivery Methods With a substantial invention, the time necessary to develop a marketable product can be long. This applies in most technology fields, but the pharmaceutical industry has used the argument of lengthy commercialisation processes to first achieve a global duration of 20 years for all patents (TRIPS) and then to gain patent term extensions for pharmaceuticals in many countries. An essential part of the commercialisation process is ensuring the medicine is delivered in a manner that achieves high patient compliance. Formulating the drug to minimise negative side-effects and maximise continuation rates is thus integral to its commercial success. This is a normal part of the development of the invention into a commercial product. Yet patent offices around the world regularly provide additional patents for combining known delivery methods with known and patented substances. In Australia the patent for extended release version of the anti-depression drug venlafaxine kept generic drugs off the market for eight years before it was found to be invalid.72 The patent for an 66

67 68 69 70

71 72

M Burdon and K Sloper, ‘The Art of Using Secondary Patents to Improve Protection’ (2003) 3 Journal of Medical Marketing 226 at 238. Granstrand and Tietze (n. 49). Hazel V J Moir, ‘Exploring Evergreening: Insights from Two Medicines’ (2016) 49 Australian Economic Review 413. Abud et al. (n. 65). Geoffrey Sanzenbacher and Gal Wettstein, The Effect of Medicare Part D on Evergreening, Generic Entry, and Drug Prices (Boston College Center for Retirement Research WP 2019–18, 2019). Feldman (n. 55). The original version had very poor compliance because of side-effects and the larger market was only achieved with development of the extended-release version, so it was an essential part of converting the original invention into a commercial product.

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

Pharmaceutical Patents and Evergreening

145

enteric coating for the blockbuster drug omeprazole (Losec/Prilosec) was upheld in Australia, where it was effective in adding eight years to patent duration at substantial cost to taxpayers.73 The patent was granted in Europe, but a legal challenge led to its revocation in the UK for want of an inventive step.74 Lloyd investigates ownership of secondary patents on omeprazole in the USA and the EU and finds 485 secondary patents, with 29 per cent owned by the originator company (which was the leading applicant for such patents).75 Daidoji and colleagues specifically investigate new formulation patents, again using US data, and find that early versions of orally administered NCEs tend to take tablet or capsule forms.76 Later versions, particularly extended-release forms, create higher hurdles to generic entry and are a high proportion of additional formulation patents listed in the Orange Book. It is judicial interpretation of the non-obviousness requirement that allows apparently obvious combinations of known elements to be patented. Prior to the 1980s the ‘synergy’ doctrine ensured that combinations of known things did not merit a patent unless the combination gave an unexpected result or a result that was greater than the sum of the parts. This important doctrine was overturned by Australia’s High Court in 198077 and by the US Court of Appeals for the Federal Circuit (CAFC) in 1984.78 In the USA the CAFC approach was finally overturned by the Supreme Court in 2007,79 but despite an effort at reform by IP Australia,80 the synergy doctrine remains absent in Australia. The synergy doctrine has always been part of the EPO’s procedures but did not prevent grant of the omeprazole enteric coated patent that was revoked in the UK. Allowing patents for obvious and essential combinations that should be part of normal development of the initial patented substance has been extremely expensive in many countries. A clear fix to this problem would be to re-instate the synergy doctrine. VIII.IV.B Compound Variations: Enantiomers, Metabolites and More Louis Pasteur discovered molecular chirality in 1948 – that molecules could be separated into distinct right-handed and left-handed stereoisomers.81 Where these are mirror images they are referred to as enantiomers and the combination of the two is called a racemic mixture (racemate). In the pharmaceutical world many important and useful drugs are chiral compounds. Enantiomers can have quite different biological properties, and where a racemate is pharmacologically useful, often the utility is entirely ascribable to only one of the enantiomers. Some blockbuster medicines are racemates and have been patented initially as the racemate and subsequently as the pharmacologically active enantiomer. Perhaps the best known is omeprazole (Losec) and its enantiomer esomeprazole (Nexium). Another is citalopram (Cipramil) and

73 74 75

76

77 78 79 80 81

Moir (n. 68). The omeprazole case is discussed in detail in Granstrand and Tietze (n. 49). Mike Lloyd, ‘Evergreening by Whom? A Review of Secondary Patents for Omeprazole’ (2013) 2 Pharmaceutical Patent Analyst. Kengo Daidoji, Satoshi Yasukawa and Shingo Kano, ‘Effects of New Formulation Strategy on Life Cycle Management in the US Pharmaceutical Industry’ (2013) 10 Journal of Generic Medicines 172. Minnesota Mining & Manufacturing Co v Beiersdorf (Australia) Ltd [1980] HCA 9. ACS Hosp Sys Inc v Montefiore Hosp 732 F2d 1572, 1577 (Fed Cir 1984). KSR Int’l Co v Teleflex Inc 550 US 398, 407 (2007). See IP Australia (n. 32) at 12–13. Stereoisomers are ‘molecules with the same molecular formula and the same connectivity to other groups, but a different arrangement of atoms in three-dimensional space. . . . Enantiomers are pairs of stereoisomers which are mirror images of each other . . .’ (Carol Higginbotham, Introductory Organic Chemistry (Open Oregon 2021). Available at .

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

146

Hazel V. J. Moir

its enantiomer escitalopram (Lexapro). Effectively allowing a second patent for the enantiomer provides for two consecutive monopoly periods for the same medicine. Over the past 10–15 years there have been some judicial determinations that if the racemate has been disclosed, the enantiomer is obvious and a patent for it is therefore invalid. But such decisions have not always led to clear directions in patent examiner manuals that enantiomers are obvious. In reporting to the WIPO Standing Committee on the Law of Patents (SCLP) on enantiomers, Australia, for example, cites later judicial decisions that enantiomers lack novelty then goes on to refer to earlier decisions to posit that, in some circumstances, they may not lack novelty.82 Although the US Supreme Court increased the hurdle for inventiveness in 2007 in the KSR decision, it is clear from the USA report to the SCLP that US patent examiners have as yet no clear direction to find enantiomers obvious, given knowledge of the racemate. Case law from the EPO’s Boards of Appeal clearly declares that knowledge of the racemic mixture does not mean the enantiomer lacks novelty.83 However, the UK’s report to the SCLP states that guidelines for UK patent examiners say that ‘in most cases a single enantiomer is rendered obvious by prior disclosure of the racemate’.84 In some countries these and other chemical variants are specifically made non-patentable by statute.85 A review of a range of types of pharmaceutical products and processes for which patentability might be claimed was initiated by the World Health Organisation (WHO). In the resulting Working Paper, Correa recommends that single enantiomers not be patentable if the racemic mixture is known.86 Another important chemical variant is metabolites. Drugs are metabolised when ingested, so allowing a second patent for the metabolite is tantamount to providing two consecutive patent terms for the substance. An example is venlafaxine and its metabolite desvenlafaxine. Desvenlafaxine was never marketed in Europe – Wyeth (now Pfizer) withdrew its marketing application when the European Medicines Agency (EMA) asked for additional information. But the EMA report makes it clear that there was no evidence that desvenlafaxine had any greater efficacy than venlafaxine.87 In order to persuade clinicians to prescribe the replacement medicine, Pfizer used language such as ‘an evolutionary advance that allows some advantages in individual patients’.88 Nonetheless, Pfizer has had remarkable success in persuading prescribers to move to the new, more expensive, product.89 These and other variant chemical structures are the basis for originator companies developing replacement medicines for when a high-selling patented drug comes off patent. While originator companies argue improved therapeutic outcomes, there is rarely any clinical evidence of this. In a meta-analysis of studies on (es)citalopram Alkhafaji et al. find that the clinical benefits of escitalopram compared to citalopram remain uncertain.90 Despite this lack of evidence, in some countries, 82

83 84 85 86

87

88

89 90

WIPO, Further Study on Inventive Step (Part III) (Standing Committee on the Law of Patents, Thirtieth Session (SPC/30/ 4), 2019) at 22. Available at . See the webpage . WIPO (n. 82) at 23. Rathod (n. 50). Carlos M Correa, ‘Guidelines for the Examination of Pharmaceutical Patents: Developing a Public Health Perspective (2007) at 17. Indeed the European Medicines Agency’s Withdrawal Assessment Report states that ‘[r]elative to the mother compound venlafaxine, desvenlafaxine appears less effective with a similar safety and tolerability profile’ (Report EMEA/45054/2009, p. 3, available at . See the webpage . This very subtle language seems designed to imply that desvenlafaxine is an improvement, without actually lying. Moir (n. 68). Ali A Alkhafaji and others, ‘Impact of Evergreening on Patients and Health Insurance: A Meta Analysis and Reimbursement Cost Analysis of Citalopram/Escitalopram Antidepressants’ (2012) 10 BMC Medicine 142.

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

Pharmaceutical Patents and Evergreening

147

originator companies are very successful in shifting prescribing behaviour from the original medicine to the replacement medicine, entirely removing the market for cheaper generic products. This outcome is entirely legal, but such product switching adds considerably to healthcare costs and diverts resources from developing new treatments. As noted above, the WHO has attempted to develop guidance to ensure such variants are not patented. However, the clear lack of political willingness to attend to this problem remains and such evergreening patents continue to be granted. VIII.IV.C Me-Too Drugs: Statins Quite a different type of secondary patent is for ‘me-too’ drugs. These are close alternative medicines, with the patents usually owned by different companies, so technically they are secondary rather than evergreening patents. Although such secondary patents fall outside the scope of this article, a brief discussion is warranted as their story is important in understanding the overall landscape of evergreening and of competition in pharmaceutical markets. Me-too drugs can derive from simultaneous research by different companies. They can also be the result of a deliberate, but later, effort to develop a competing product where a blockbuster drug is showing large returns. The argument that these me-too variants have improved therapeutic value is stronger than for the evergreening variants discussed above. Li gives the example of ranitidine (Zantac), the first me-too drug following cimetidine (Tagamet), indicating advantages such as fewer negative side-effects and an improved half-life, which increases patient compliance.91 An economic study of attention deficit hyperactivity disorder (ADHD) drugs in the USA shows that some me-too drugs create new niche markets by introducing new, and preferred, delivery mechanisms, leading to increases in consumer welfare.92 This study also shows that other me-too ADHD drugs added no discernible value. More broadly, assessment of the value of me-too drugs would benefit from well-designed comparative clinical trials. These are scarce, and the current model for funding clinical trials does not encourage them. At some point, however, the question must be asked as to quite how many different statins are needed – or quite how many low-dose combination oral contraceptive pills.93 This raises issues of managing epidemiological research using Phase IV medical data – post-launch data on the use of drugs. Goldacre argues persuasively that in a country like the UK, with its National Health Service (NHS), anonymised data on patient health outcomes and prescription use could be made available for broader studies of relative efficacy in different circumstances.94 Such data would be invaluable to health authorities but are of little interest to originator companies. VIII.IV.D New Indications A popular current evergreening strategy focuses on new uses for old medicines, or uses of medicines for specific (preferably small) populations, to gain orphan drug benefits. Once a

91 92

93

94

Jie Jack Li, Blockbuster Drugs: The Rise and Fall of the Pharmaceutical Industry (OUP 2010) at 28–34. Bokhari and Fournier, ‘Entry in the ADHD Drugs Market: Welfare Impact of Generics and Me-Too’s’ (2013) 61 The Journal of Industrial Economics 339. A brief web search (9 March 2021) identified seven statins (Atorvastatin, Fluvastatin, Iovastatin, Pitavastatin, Pravastatin, Rosuvastatin, Simvastatin). A 2013 websearch of the US National Library of Medicine found 83 different brand names for low-dose combination oral contraceptive pills. Ben Goldacre, Bad Pharma (Fourth Estate 2012).

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

148

Hazel V. J. Moir

drug has been launched onto a market, substantially better data (referred to as Phase IV data) are available and effective analysis of these data can indicate additional conditions for which the medicine is effective, as well as providing better information on when the drug should not be used. Kapczynski et al. found that, in the USA from 1985 to 2005, new method of use patents added 7.4 years to the nominal patent term.95 There was an upward trend (2005– 2015) in the addition of new use codes to the Orange Book and among listed drugs with any Orange Book additions, the proportion with new use codes increased from 27 per cent in 2005 to 49 per cent in 2015.96 While understanding the full range of positive (and negative) effects of a medicine is extremely valuable to society, the current approach of allowing originator companies to take out additional 20-year patents for particular uses of known substances is not well balanced. To begin with, the original patent covers all uses of the new substance. So, a later additional 20year patent for a particular use effectively grants the patent twice, albeit for a narrower scope the second time. To the extent that clinical trials may be needed to confirm the new use, there may be a case for alternative funding or incentives. But to award a full 20-year patent is poor value. Public interest in what is now known as ‘drug re-purposing’ has increased substantially since the Covid-19 pandemic. As noted above Phase IV medical use data are critical to establishing when a medicine is efficacious and when it should not be used.97 So re-purposing again raises the question of how a society organises the use of the data that are (or potentially could be) available post-launch. In early 2021 the Therapeutic Goods Administration – the authority for approving drugs for use in Australia – undertook a consultation on the re-purposing of medicines.98 It is clear from the discussion paper that a critical issue is monitoring of post-launch data on the outcomes from using the medicine.99 Current regulatory arrangements mean that each approved use requires a sponsor, though in a smaller market the incentives for such sponsorship might be weak. Broader thinking about how best to manage post-launch drug monitoring might identify a range of improved options for ensuring a wide range of uses of medicines where substantial resources have already been spent on their initial development. A related form of evergreening is the approval of medicines not for new indications, but for new populations. While data exclusivity is the subject of Chapter III, it is worth noting here that, in the USA, seven-year data exclusivity provisions for orphan drugs extended the effective market life term for a third of small molecule drugs during the period 1985–2014.100 Pharmaceutical companies actively game the orphan drug exclusivity system, by ‘salami slicing’ the definition of orphan drug populations.101 It appears that little has been done to stop this practice.

95 96 97

98

99

100

101

Kapczynski et al. (n. 59). Feldman (n. 55). OECD, ‘Using Routinely Collected Data to Inform Pharmaceutical Policies; Analytical Report for OECD and EU Countries, 2019’). Available at . And in December 2019 the US FDA held a workshop on re-purposing off-patent drugs (). Therapeutic Goods Administration (Australia), ‘Repurposing of Prescription Medicines’ (2021). Paper now removed from website, but available from author. Ameet Sarpatwari and others, ‘Evaluating the Impact of the Orphan Drug Act’s Seven-Year Market Exclusivity Period’ (2018) 37 Health Affairs. Kapczynski et al. (n. 59).

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

Pharmaceutical Patents and Evergreening

149

VIII.IV.E Other Types of Evergreening Patents Many evergreening patents are simply for different dosages of a known medicine, though such ‘inventions’ can also be combined with other changes. Changing the presentation, for example from a tablet to a capsule, can also generate additional patents. Indeed, there are even patents for packaging the medicine. Trivial as these sound, they can be used to delay generic entry as they allow either shifts in the market norm and hence prescribing behaviour (dosages and forms) or consumer recognition (packaging). Combining two existing, known, medicines into a single medicine is also patentable. Examples are combining amlodipine with atorvastatin (Caduet), or clopidogrel with aspirin. Eli Lilly’s patent on the blockbuster drug fluoxetine (Prozac) expired in 2001, but it was able to obtain a patent for the combination of fluoxetine and olanzapine (Symbyax) which did not expire until 2017.102 To the extent that the company was able to shift the market to the ‘productline extension’ drug Symbyax, this effectively gave it an extra 16 years of patent life. Some of these combinations provide, in a single tablet or capsule, medicines which were previously often prescribed simultaneously for the same patient. A few anti-retroviral medicines are available in such fixed dose combinations (FDCs). Clearly this makes life easier for the patient, but does it merit a 20-year monopoly? Some such combinations simply add specific vitamins to the medicine. A US study measuring the time lag between approvals for single active ingredient and fixed dose combination (FDC) drugs found that FDCs for already registered drugs added a median of 9.7 years to the nominal patent life of the single active ingredient medicines in the combination.103 Another concerning form of evergreening is patents for devices for delivering medicines. Beall and colleagues find that, for four common conditions often treated using a device to deliver the medicine, there are unexpired patents for the drug/device combination in the USA.104 VIII.IV.F Taking Action to Reduce Evergreening Patents Overall, patent evergreening takes many forms. It is a pervasive practice and can substantially delay generic entry at considerable cost to health systems (taxpayers) and patients. Despite reports and inquiries into the availability and cost of medicines,105 governments appear reluctant to take effective action. In May 2019 the World Health Assembly adopted a resolution on access to medicines urging member states to enhance ‘transparency on pharmaceutical patents, clinical trial results and other determinants of pricing along the value chain from laboratory to patient’.106 If implemented, this could begin a much-needed process of rebalancing and reform. Doubtless, however, the well-resourced pharmaceutical industry lobby is ensuring that member states do as little as possible to give effect to this resolution. 102 103

104

105

106

Song and Han (n. 49). Jing Hao, Rosa Rodriguez-Monguio and Enrique Seoane-Vazquez, ‘Fixed-Dose Combination Drug Approvals, Patents and Market Exclusivities Compared to Single Active Ingredient Pharmaceuticals’ (2015) 10 PLoS ONE e0140708. Reed F Beall and others, ‘Is Patent “Evergreening” Restricting Access to Medicine/Device Combination Products?’ (2016) 11 PLOS One 1. For example, by the National Academies of Sciences Engineering and Medicine (n. 17); OECD, ‘Pharmaceutical Innovation and Access to Medicines’ (OECD Health Policy Studies, 2018); US Department of Health and Human Services, ‘Prescription Drugs: Innovation, Spending, and Patient Access’ (Report to Congress, 2016); and EU initiatives to improve access to medicines (at ). See the webpage .

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

150

Hazel V. J. Moir

A critical element of balance missing in most countries is an effective incentive to encourage generic companies to challenge weak evergreening patents. In countries where medicine prices are moderated by governments – for example through the UK’s NHS – the government could share a part of the savings with a company successfully challenging a weak patent and thus bringing forward in time the date of generic entry. Another concern here is, however, the tendency for large originator companies to acquire generic companies. This is a matter that should be referred to competition authorities for investigation. The discussion above has pointed to the range of patent doctrines which operate to ensure the grant of very weak patents. Another relevant factor is patent office practices, and ‘technical assistance’ to new patent offices has ensured that norms suitable for pharmaceutical-exporting regions are transferred to importing countries.107 A study comparing grant rates for primary and secondary patents in India, Brazil and Argentina compared to the USA, the EU and Japan finds that measures to restrict secondary patents in developing countries have very limited impact.108 Argentina’s measures to restrict secondary patents appear to be the most successful.109 Action to ensure that patent examiners properly understand issues of balance in patent policy could pay dividends in ensuring that evergreening patents are not issued. Understanding what policies and procedures have the greatest effect in achieving this is clearly fundamental. VIII.V CONCLUDING REMARKS: INCENTIVE STRUCTURES AND THE DEVELOPMENT OF NEW MEDICINES

The OECD (Organisation for Economic Co-operation and Development) reports increasing concerns among policy makers about the pharmaceutical innovation system. These concerns include the very high prices charged for some new medicines and for off-patent medicines, the degree of innovation of certain new treatments and the lack of innovation in areas of high unmet need such as anti-microbials, non-vascular dementia, and some rare diseases. Underlying these issues the OECD identifies a lack of trust between payers, civil society and the pharmaceutical industry. A particular issue with respect to evergreening is the impact this may have on investment to identify and develop genuinely new medicines to treat important unmet health needs. The OECD’s analysis of US data for 1980–2014 shows a sharp decline in the number of new drug approvals compared to pharmaceutical industry R&D expenditure. The decline was particularly marked in the period to the late 1980s and has declined more slowly since then.110 During the period 1988–2005 the FDA approved 938 non-biologic drugs.111 Of these, 380 were for NCEs, and 228 received priority FDA review. That is, just under a quarter of the ‘new’ drugs are for drugs which treat conditions where there is no satisfactory alternative therapy or where there is a significant improvement compared to other available drugs.112 Over the Peter Drahos, ‘“Trust Me”: Patent Offices in Developing Countries’ (2008) 34 American Journal of Law & Medicine 151. Bhaven N Sampat and Kenneth C Shadlen, ‘Secondary Pharmaceutical patenting: A Global Perspective’ (2017) 46 Research Policy 693. 109 This is confirmed by a number of presentations to the South Centre – Max Planck Global Forum on IP, Innovation and Access to Medicines, Munich, December 2019 (Vitor Henrique Pinto Ido and Viviana Muñoz-Tellez, ‘Report on the South Centre: Max Planck Global Forum on IP, Innovation and Access to Medicines’ (2020) 69 GRUR International 1130). 110 OECD, 2018 (n. 105) at 93. 111 Listed in the Orange Book, ie, excluding biologics, which are listed in the Purple Book. 112 Ouellette (n. 56) at 312. 107 108

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

Pharmaceutical Patents and Evergreening

151

following decade, from 2005 to 2015, novel drugs – drugs serving previously unmet medical needs – averaged just 13 per cent of FDA drugs approvals.113 For the same period, Feldman estimates that the majority of patents added to the Orange Book in any one year are for previously listed medicines, not for new medicines.114 Among the 2,534 drugs with patent additions only 23 per cent were newly approved medicines. This likely understates the proportion of new medicines as the level of analysis was by ‘new drug application’ not by specific molecule. Both these sources indicate that the bulk of FDA non-biologic drug approvals are now for previously approved chemical substances or modifications to existing drugs. As this chapter has demonstrated, most evergreening pharmaceutical patents have little, if any, inventiveness, if one uses an ordinary meaning of the word inventive. The low inventiveness standard derives from the 1952 US Patent Act and was designed to ensure that patents would be granted even for routine work. At that time pharmaceutical R&D was already based on largescale investment often for repetitive routine screening tasks. From this beginning the standard that ensures monopolies are granted simply for investment rather than for genuine additions to knowledge has spread worldwide. Pharmaceutical companies have been instrumental in this transition. Despite the substantial evidence to the contrary, the pharmaceutical industry continues to tell the story that high prices and ‘strong’ patent policy (low inventiveness standards) are essential to the development of new health treatments. But the evidence clearly shows that health benefits would be increased by eliminating, as far as possible, evergreening strategies by tightening the requirement for genuine inventiveness in the patent system. The critical issue for researchers is to understand why this available evidence has no traction with policy makers and to come up with new ways of telling the story to achieve evidence-based policy outcomes delivering better health outcomes.

113 114

US Government Accountability Office (n. 13) at 41. Feldman (n. 55) at 636–37.

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

https://doi.org/10.1017/9781108989527.011 Published online by Cambridge University Press

part ii

Culture and Entertainment

https://doi.org/10.1017/9781108989527.012 Published online by Cambridge University Press

https://doi.org/10.1017/9781108989527.012 Published online by Cambridge University Press

IX The Press Publishers’ Right under EU Law Rewarding Investment through Intellectual Property Stavroula Karapapa

IX.I INTRODUCTION

The press publishers’ right available under Article 15 of Directive (EU) 2019/790 on Copyright in the Digital Single Market (the ‘DSM Directive’) is one of the most recent additions to investment-driven intellectual property rights under EU copyright law. The right was introduced with a view to address the dramatic changes in the creation, distribution and consumption of news online. With the advancement of digital technologies and the Internet, news is no longer solely available directly from press publishers through their print editions and news websites but is increasingly accessed via other sources, such as news aggregators and social media platforms. The changes in news consumption trends have resulted in a drop of the revenues of press publishers since 2000, a progressive decline in the circulation of printed newspapers,1 and a dramatic increase of consumption of online news content. Because of these changes, the revenues and advertising sales of press publishers have substantially dropped despite their efforts and investment in making news accessible. It is to reward these efforts and this investment that a sui generis press publishers’ right was introduced initially at the level of EU Member States, such as Germany2 and Spain,3 and was later included in Article 15 of Directive (EU) 2019/790 on Copyright in the Digital Single Market. The right was envisaged as one that would improve the bargaining power of press publishers when negotiating licensing deals with online services and web platforms that re-use their content. But is the press publishers’ right necessary and, if so, appropriate towards stimulating and protecting the investments made by press publishers? This is a question that was heavily debated 1

2

3

Indicatively, between 2010 and 2014 it had declined by 17 per cent in eight EU Member States. See Commission Staff Working Document: Impact Assessment on the Modernisation of EU Copyright Rules, accompanying the document Proposal for a Directive of the European Parliament and of the Council on Copyright in the Digital Single Market and Proposal for a Regulation of the European Parliament and of the Council laying down Rules on the Exercise of Copyright and Related Rights applicable to Certain Online Transmissions of Broadcasting Organisations and Retransmissions of Television and Radio Programmes, SWD (2016) 301 final, Vol 1, 155 (14 Sept. 2016). Also see Gareth Price, ‘Opportunities and Challenges for Journalism in the Digital Age: Asian and European Perspectives’, Chatham House, August 2015. Sections 87f, 87g and 87h of the Urheberrechtsgesetz (German Copyright Act). For a commentary see Igor Barabash, ‘Ancillary Copyright for Publishers: The End of Search Engines and News Aggregators in Germany?’ (2013) 35(5) European Intellectual Property Review 243. Article 32 of the Ley de Propiedad Intelectual (Intellectual Property Law). For a commentary see Raquel Xalabarder, ‘Press Publisher Rights in the Proposed Directive on Copyright in the Digital Single Market’ Create Working Paper 2015/16, Dec. 2016, 17 et seq, .

155 https://doi.org/10.1017/9781108989527.013 Published online by Cambridge University Press

156

Stavroula Karapapa

at the EU level, and lobbying on the scope and effectiveness of Article 15 continues at national level. Even before the draft Directive was published,4 the proposed press publishers’ right received intense criticism from scholars across Europe,5 in that the right is unnecessary, undesirable, fundamentally misconceived and unlikely to achieve anything apart from adding to the complexity and cost of operating in the copyright environment. Indeed, as I will explain below, there is no hard evidence that the right can achieve its stated objectives, notably to facilitate rights clearance and enforcement in the press publishing industry. Quite on the contrary, the national implementations of press publishers’ rights before the launch of Article 15 were not as successful as expected. Article 15 was introduced despite these precedents. In this chapter, I will initially introduce the right offered under Article 15, outline its underpinning rationale, and explain how it introduces a layer of protection additional to copyright as a form of investment-driven intellectual property right. I will then elaborate on why the way in which the so-called newspaper crisis was addressed was not an appropriate response to issues of licensing and enforcement, which are the key sources of concern on the future of press publishing, and explore alternatives that could have offered a more suitable solution. IX.II THE PRESS PUBLISHERS RIGHT UNDER ARTICLE 15 OF THE DIGITAL SINGLE MARKET DIRECTIVE

The introduction of Article 15 was the outcome of a 2016 Commission public consultation on the role of publishers aiming to collect views on the desirability of a right for press publishers. The Impact Assessment on the Modernisation of EU Copyright Rules stated that: ‘[t]he shift from print to digital has enlarged the audience of press publications but made the exploitation 4

5

See, eg, Martin Kretschmer, Severine Dusollier, Christophe Geiger and P Bernt Hugenholtz, ‘The European Commission’s Public Consultation on the Role of Publishers in the Copyright Value Chain: A Response by the European Copyright Society’ (2016) 38(10) European Intellectual Property Review 591; Martin Senftleben, ‘Copyright Reform, GS Media and Innovation Climate in the EU: Euphonious Chord or Dissonant Cacophony?’ (2016) 5 Tijdschrift Voor Auteurs: Media-En Informatierecht 130–33; Ana Ramalho, ‘Beyond the Cover Story: An Enquiry into the EU Competence to Introduce a Right for Publishers’ (2017) 48(1) International Review of Intellectual Property and Competition Law 71; Reto M Hilty, Kaya Köklü and Valentina Moscon, ‘Position Statement of the Max Planck Institute for Innovation and Competition on the “Public Consultation on the Role of Publishers in the Copyright Value Chain”’ (15 June 2016) Max Planck Institute for Innovation and Competition, available at: See indicatively Lionel Bently et al., ‘Response to Article 11 of the Proposal for a Directive on Copyright in the Digital Single Market, entitled “Protection of Press Publications concerning Digital Uses” on Behalf of 37 Professors and Leading Scholars of Intellectual Property, Information Law and Digital Economy, 1 (5 Dec. 2016), (last access 8 Nov. 2017); also see P Bernt Hugenholtz, ‘Say Nay to the Neighbouring Right’ Kluwer Copyright Blog (14 Apr. 2016), (last accessed 8 Nov. 2017); Copyright Reform: Open Letter from European Research Centres (Feb. 24, 2017), ; Raquel Xalabarder, ‘Press Publisher Rights in the Proposed Directive on Copyright in the Digital Single Market’ Create Working Paper 2015/16, Dec. 2016, 17 et seq ; Matthew Karnitschnig and Chris Spillane, ‘Plan to make Google Pay for News Hits Rocks’ id; Christophe Geiger, Oleksandr Bulayenko and Giancarlo F Frosio, ‘The Introduction of a Neighbouring Right for Press Publisher at EU Level: The Unneeded (and Unwanted) Reform’ (2017) 39(4) European Intellectual Property Review 202; Martin Senftleben, Maximilian Kerk, Miriam Buiten and Klaus Heine, ‘New Rights or New Business Models? An Inquiry into the Future of Publishing in the Digital Era’ (2017) 48(5) International Review of Intellectual Property and Competition Law 538; Stavroula Karapapa, ‘The Press Publishers’ Right in the European Union: An Overreaching Proposal and the Future of News Online’ in E Bonadio and N Lucchi (eds) Non-Conventional Copyright: Do New and Atypical Works Deserve Protection? (Edward Elgar 2018) 316–39.

https://doi.org/10.1017/9781108989527.013 Published online by Cambridge University Press

The Press Publishers’ Right under EU Law

157

and enforcement of the rights in publications increasingly difficult. In addition, publishers face difficulties as regards compensation for uses under exceptions.’6 To address this issue, the proposal for a Directive on Copyright in the Digital Single Market included provisions for a related right for press publishers (Article 11) and a share in the authors’ compensation from remunerated copyright exceptions and limitations attributable to publishers in general (Article 12). Other options were also subject to consideration at the time, such as an amendment of Article 5 of the Enforcement Directive,7 with the effect of enabling press publishers to bring proceedings to enforce copyright in materials of which they are the identified publishers.8 These did not reach consensus and the result of these preliminary discussions was a single right for press publishers available under Article 15 of the Digital Single Market Directive. Article 15 gives press publishers the exclusive right to authorise the reproduction and making available to the public – as described in the Information Society Directive 2001/29/EC (the ‘Information Society Directive’) – of their press publications for online uses carried out by information society service providers, such as search engines, news aggregators and media monitoring services. It is an ancillary or neighbouring right to copyright enabling press publishers to negotiate new or improved licensing terms with relevant information society service providers. Exclusions to the scope of the right apply. In particular, the right does not cover private or non-commercial uses of press publications by individual users. It also does not cover acts of hyperlinking and the use of individual words or very short extracts of a press publication. The exclusion of private or non-commercial uses of press publications by individual users aligns with the copyright exception of Article 5(2)(b) of the Information Society Directive, which applies to the press publishers’ right according to Recital 57 of the Digital Single Market Directive. Questions could emerge as to the private and non-commercial nature of relevant uses, especially uses that are carried out in social media, as is often the case with news items, for example re-posting a news story on a social medium with ads appearing next to the post.

6

7

8

Commission Staff Working Document: Impact Assessment on the Modernisation of EU Copyright Rules, accompanying the document Proposal for a Directive of the European Parliament and of the Council on Copyright in the Digital Single Market and Proposal for a Regulation of the European Parliament and of the Council laying down Rules on the Exercise of Copyright and Related Rights applicable to Certain Online Transmissions of Broadcasting Organisations and Retransmissions of Television and Radio Programmes, SWD(2016) 301 final, Vol 1, 155 (14 Sept. 2016) at 5.3.1. Directive 2004/48/EC of the European Parliament and of the Council of 29 Apr. 2004 on the Enforcement of Intellectual Property Rights, OJ L 157 (30 Apr. 2004). See indicatively Lionel Bently et al., ‘Response to Article 11 of the Proposal for a Directive on Copyright in the Digital Single Market, entitled “Protection of Press Publications concerning Digital Uses” on behalf of 37 Professors and Leading Scholars of Intellectual Property, Information Law and Digital Economy’ 1 (5 Dec. 2016),