Challenges to Assumptions in Competition Law 2021932424, 9781839109072, 9781839109065

Citation preview

Challenges to Assumptions in Competition Law

Challenges to Assumptions in Competition Law Edited by

David Bosco Professor, Aix-Marseille Université, France

Michal S. Gal Professor and Director, Center for Law and Technology, University of Haifa, Israel

ASCOLA COMPETITION LAW

Cheltenham, UK • Northampton, MA, USA

© Editors and Contributors Severally 2021

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or otherwise without the prior permission of the publisher. Published by Edward Elgar Publishing Limited The Lypiatts 15 Lansdown Road Cheltenham Glos GL50 2JA UK Edward Elgar Publishing, Inc. William Pratt House 9 Dewey Court Northampton Massachusetts 01060 USA A catalogue record for this book is available from the British Library Library of Congress Control Number: 2021932424 This book is available electronically in the Law subject collection http://dx.doi.org/10.4337/9781839109072

06

ISBN 978 1 83910 906 5 (cased) ISBN 978 1 83910 907 2 (eBook)

Contents List of contributorsvii Introductionviii PART I 1

NEW CHALLENGES IN THE DIGITAL SPHERE

Cooperation, dependence, and eviction: how platform-to-business coopetition relationships should be addressed in mobile telephony ecosystems Frédéric Marty and Julien Pillot

2

Technological tying: unbundling the assumptions Stephen Dnes

3

Understanding the role of agricultural data on market power in the emerging Digital Agriculture sector: a critical analysis of the Bayer/Monsanto decision Can Atik

PART II

2 22

41

NEW CHALLENGES TO EMERGING COUNTRIES

4

Merger control in China’s digital economy: challenges and prospects80 Wei Han and Yajie Gao

5

The adoption of specialised competition tribunals in Latin American countries: transplants and commonalities among them Claudia O’Kane

6

Where influence lies in the International Competition Network Christopher Townley

7

Cartels’ little helpers: a comparative study of the case law regarding the facilitators of collusion in Europe, United States and South America Andrés Calderón

v

106 124

161

vi

Challenges to assumptions in competition law

PART III NEW INSTITUTIONAL CHALLENGES 8

Public interest considerations in European merger control regimes 184 Oliver Budzinski and Annika Stöhr

9

Formalism, fairness and freedom of contract: abuse of dominance in the UK courts and the business ‘Achilles heel’ of objective justification Barry J. Rodger

10

206

Using trade tools to counteract anticompetitive conduct within global value chains: competition chapters in trade agreements238 Galyna Kostiukevych

Index257

Contributors Can Atik, PhD Candidate at Tilburg Law School with affiliation to the Tilburg Law and Economics Center (TILEC) David Bosco, Professor, Aix-Marseille Université, France Oliver Budzinski, Professor of Economic Theory, Institute of Economics, Institute of Media and Mobile Communication, Ilmenau University of Technology, Germany Andrés Calderón, Professor of Law at Universidad del Pacífico, Director of the Legal Clinic for Freedom of Information and Transparency, Peru Stephen Dnes, Lecturer in Law, New College of the Humanities at Northeastern University, London, UK Michal S. Gal, Professor and Director, Center for Law and Technology, University of Haifa, Israel Yajie Gao, PhD candidate at the Centre for Commercial Law Studies, Queen Mary, University of London, UK Wei Han, Associate Professor, Executive Director, Competition Research Center of the University of Chinese Academy of Social Sciences, China Galyna Kostiukevych, European University Institute, Firenze, Italy Frédéric Marty, CNRS – GREDEG – University of Côte d’Azur, France Claudia O’Kane, Centre for Commercial Law Studies, Queen Mary University London, UK Julien Pillot, INSEEC Business School, France Barry J. Rodger, The Law School, University of Strathclyde Glasgow, UK Annika Stöhr, M.Sc., Institute of Economics, Ilmenau University of Technology, Germany Christopher Townley, King’s College London, UK

vii

Introduction Challenges to assumptions at the basis of competition law: there are at least two reasons to justify inviting ASCOLA members to direct their academic research towards this theme. First, whilst the assumptions at the basis of competition law have been challenged throughout the years, globalization and digitalization have significantly increased the need to ensure that the assumptions we rely upon – whether with regard to substantive, procedural or institutional features – are fit for purpose. There is certainly no need to elaborate upon the profound upheavals in the world economy in recent years, particularly due to the globalization of trade and the considerable importance of digital companies in world trade. The Coronavirus pandemic has since exacerbated the dominance of some digital companies in world trade. Throughout the world, competition law has been implemented to regulate such markets and to organize the globalization of transactions. Yet competition law was designed and built for a society that is very different from the globalized and digitalized society we know today. The very foundations of competition law are therefore subject to an unprecedented test. Accordingly, competition authorities around the world are exploring the limits of antitrust and the concepts that underlie its application. Everywhere, similar questions arise: should the usual methods of delimiting the relevant market be reviewed? Can the price criterion still be relied upon or given much weight to identify an infringement of competition? Must dominant firms share the assets they have built up in order to allow effective competition? Almost everywhere around the globe, the same concepts are used as a basis for reasoning. This made ASCOLA an excellent framework for fomenting a global reflection on the important challenges competition law faces: its members come from Europe, the Middle East, the United States, Russia, China, India, Africa, Australia; they come from wherever there is competition regulation, that is to say, from almost everywhere in the world. The second set of reasons that made the theme relevant focuses on the fact that ASCOLA is an academic society composed exclusively of researchers. The upheavals in competition law are so significant that an in-depth analysis has become necessary. The sharing and exploration of ideas enabled by ASCOLA creates an in-depth and necessary dialogue on the law’s underpinnings. The book is divided into three parts, which set out three sets of challenges for competition law. First, the digital economy and innovative sectors are viii

Introduction

ix

discussed. Frédéric Marty and Julien Pillot explore the two-sided eco-system of mobile telephony to show the limits of the current tools of competition law, in particular the qualifications for abuse of dominant position. Stephen Dnes studies the practice of technological tying and its competitive implications. He questions the rules resulting from the Microsoft litigation by placing them in a modernized framework of reflection (in particular the case of Google’s integration of maps). Can Atik offers a critical analysis of the Bayer/Monsanto decision by questioning the role of data in the agricultural sector. In the second part, the book includes studies on the challenges faced by emerging countries. Wei Han and Yajie Gao present the application of merger control in China in the digital sector. Claudia O’Kane proposes an institutional reflection on the adoption of specialized tribunals in Latin American countries, based on the Mexican experience. Christopher Townley discusses the influence that Western large companies and competition authorities can exert within the framework of the International Competition Network. Andrés Calderón takes the reader to South America and explores how competition authorities treat cartel facilitators. In the third part, the book focuses on institutional challenges. Oliver Budzinski and Annika Stöhr show how public interest considerations are taken into account in European merger control regimes. Barry J. Rodger returns to the foundations of abuse of dominance by exploring the jurisprudence of the British courts. Galyna Kostiukevych discusses the use of international trade instruments to combat anti-competitive behaviour suffered by participants in the global supply chain. David Bosco and Michal Gal Aix-en-Provence and Haifa, August 2020

PART I

New challenges in the digital sphere

1. Cooperation, dependence, and eviction: how platform-to-business coopetition relationships should be addressed in mobile telephony ecosystems Frédéric Marty and Julien Pillot INTRODUCTION The current debate on the power of dominant firms in the digital industry is mainly based on the concentration of economic power they achieve and its possible irreversibility. Two issues are of particular concern: (1) How did these firms get such a market power and the ability to upkeep and expand their dominant positions? (2) Are the markets in which these firms compete still challengeable, or would the control over users’ data constitute impenetrable barriers to entry? The competition agencies’ capacity to prevent these competitive risks through their M&A control policies, or to effectively sanction strategies aimed at foreclosing existing or even potential competitors from the market, is frequently challenged. In a way, dominant firms in the digital markets are suspected to own financial and technical resources, as well as nowcasting capabilities, likely to prevent them from any potential competitive threat. Then, any company whose products or services would be likely to contest, even in the long run, the dominant firm market shares would enter a kill zone. A counter-intuitive result should then be highlighted: ecosystem insiders would be particularly exposed to this risk. Such a situation may seem paradoxical. Indeed, whether the platform is open or closed, it is basically a two-sided market. On the one side, it needs a critical mass of users and, on the other side, it seeks for major industrial partners and developers. Both are required in order to fully exploit cross-network externalities. The literature shows how system integrators stand to gain from creating 2

Cooperation, dependence, and eviction

3

a cooperative ecosystem.1 Such cooperation involves contractual mechanisms to distribute financial incentives to ecosystem members, the pooling of (technical) resources and implicit non-competition commitments. For instance, Gawer and Henderson2 show that Intel avoided developing products that were potentially competitive to those offered by its complementors,3 unless the performance of the latter was unsatisfactory or if it appeared necessary to strengthen incentives for innovation through a competitive threat. As Zhu and Liu4 note: ‘By committing to not competing with them Intel could encourage these complementors to make [substantial sector specific] investments.’ In this light, how can we appreciate entry decisions on a market segment previously occupied by a complementor? Is it for substituting its own services to replace elder ones seen as not efficient enough, or does it pertain to a two-stage non-cooperative strategy? After an initial cooperative phase aiming at detecting promising (or potentially disruptive) third-parties’ developments, the keystone player may replace them with its own products and services (or take control of the complementor) in a second phase. Our purpose is to analyze these coopetitive strategies.5 How to assess the competitive effects generated by such relationships? The case of a mobile operating system (MOS) illustrates how agreements that financially and technically organize such complementarities may have pro- and anticompetitive effects. Financial incentives or free provision of resources to complementors, two practices at the heart of the European Commission’s Android decision,6 can be explained by a search for efficiency. As such, these practices could be defended on the basis of two-sided platform’s constraints. Competition law 1 Annabelle Gawer and Michael A. Cusumano, Platform Leadership: How Intel, Microsoft, and Cisco Drive Industry Innovation (Harvard Business School Press, 2002). 2 Annabelle Gawer and Rebecca Henderson, ‘Platform Owner Entry and Innovation in Complementary Markets: Evidence from Intel’, (2007) Journal of Economics and Management Strategy, 16(1), pp. 1–34. 3 Complementors are independent companies developing complementary products that are interoperable with the technical system developed by the keystone company. Marco Iansiti and Roy Lieven, The Keystone Advantage. What the New Dynamics of Business Ecosystems Means for Strategy, Innovation, and Sustainability (Harvard Business School Press, 2004), p. 272. 4 Feng Zhu and Qihong Liu, ‘Competing with Complementors: An Empirical Look at Amazon.com’, (2018) Strategic Management Journal, 39(10), pp. 2618–2642. 5 Frédéric Marty and Thierry Warin, ‘Innovation in Digital Ecosystems: Challenges and Questions for Competition Policy’, (2020) Cahier de Recherche du CIRANO, 2020s-10, February. 6 Google Android (case AT.40099) European Commission Decision, C(2018) 4761 final, 18 July 2018.

4

Challenges to assumptions in competition law

does not sanction business-to-business (B2B) cooperation when it results in promoting economic efficiency or innovation. Hence, in the Android decision, the pre-installation of applications and the remuneration paid to complementors could have been found to be procompetitive rather than considered as part of a foreclosure strategy. The same goes for anti-fragmentation agreements. These could have been justified by safety objectives (both for the app developer and the OS owner) as well as interoperability objectives. To that extent, contractual terms made to third-parties (app developers, terminal manufacturers, and mobile network operators) could be analyzed as part of a logic of complementary developments within an ecosystem. The latter is both a two-sided structure in the sense of its business model – having to link several sides to unlock value (thus possibly justifying pre-installations and financial incentives) – and in the sense of its technical architecture (requiring rules to guarantee compatibility of developments, such as anti-fragmentation agreements). In a nutshell, each of the three practices that led the European Commission to impose a fine of more than 4 billion euros on Google (profit-sharing, pre-installation of applications and anti-fragmentation agreements) could have been defended on the basis of efficiency, with regards to two-sided platforms’ economic and technical specificities. At the opposite end, these ecosystems may give rise to exclusionary or exploitative abuses. Most third-party developers are yet in a situation of economic dependence vis-à-vis the platform’s owner. Thus, considering these (inter)dependence relationships may be essential to assess the likelihood of a foreclosure attempt (under Article 102 TFEU), as well as the competitive damages that may result from notified concentrations. Finally, attention should be paid to the effects of remedies imposed to bring a potential infringement to an end, both in the context of merger control and abuse of dominance cases. Remedies could accidentally harm consumers and competition if they result in the loss of some of the gains flowing from firms’ coordination. Similarly, remedies may have insufficient effects, especially if they are pronounced too late, when the integrated firm has fewer incentives to share investments (and value) with its complementors. Our chapter is structured as follows. First, we introduce how cooperative ecosystems are set up in two-sided platforms, and how complementors benefit from the sharing of technical resources and from financial incentives. Second, we show that this technological and competitive symbiosis can result in economic dependence. Third, we focus on the risks complementors have to assume. We illustrate these risks both by the distortion of competition flowing from asymmetries in data access, and by buy-out strategies aiming at neutralizing potential competitors. At last, in the light of the EU Commission decision in Android, we discuss these coopetitive business models under the prism of

Cooperation, dependence, and eviction

5

the damage theory established by the Commission and question the effects of competitive remedies.

1.1

COOPERATIVE ECOSYSTEMS AND TWO-SIDED MODELS: HOW TO ANALYZE P2B RELATIONSHIPS IN THE CASE OF MOBILE OPERATING SYSTEMS?

Whether a MOS is open like Google Android or closed like Apple iOS, it implies various stakeholders, including app developers, and relies on cooperative strategies, including from the integrating company itself. Such an organization allows actors to share investments and competences to co-construct a value network. The keystone firm pools boundary resources for the benefit of all, including its complementors. The MOS is designed as a two-sided structure both in terms of business model and technical architecture. Such platforms rely on externalities between developers and users. It is then rational for the platform owner to subsidize developers to reduce their technical and financial barriers to entry, and to encourage them to adopt the MOS. Each new developer contributes to the enhancement of the ecosystem as a whole, and benefits from it in return. 1.1.1

The Provision of Boundary Resources

According to Vezzoso,7 digital platforms are much more than mere demand aggregators, but have to be seen as catalysts for a system of complementary innovations. An Android-like platform is based on technical and institutional mechanisms that aim to promote and guarantee indirect network externalities between the different users of the platform. In a way, a third-party developer is also a customer of the platform: both access to the market and effective development of the product depend on the services provided by the platform owner. In other words, a platform is not reducible to a matching algorithm but is also a modular production network relying on complementary resources and investments as Moore (2006) has shown.8 As such, a platform’s ability to attract key complementors is a key success factor. The MOS owner has to generate enough trust in the capacity to create collective value to encourage third-party investments in specific assets. In the 7 Simonetta Vezzoso, ‘Open Digital Platforms and Antitrust: A More Technological Approach’, in Oliver Budzinski and Julius Haucap (eds), Recht und Ökonomie (Nomos, 2020). 8 James F. Moore, ‘Business ecosystems and the view of the firm’, (2006) The Antitrust Bulletin, 51(1), pp. 31–75.

6

Challenges to assumptions in competition law

case of Android, which came on the market later than Apple’s iOS, it was positively a sine qua non condition for catch-up. Such modularity obviously generates efficiency gains. It reduces barriers to entry and ensures devices and services interoperability. Meanwhile, it limits costs and interface problems. As noted by Wen and Zhu:9 ‘By providing efficient matching or development kits, such platforms have also significantly lowered the barriers for many small firms or individuals to innovate and to market their products and services.’ The openness of the platform has to be analyzed both in terms of resources availability (IP, data, open source licenses, etc.) and of access policy.10 Start-ups entering the ecosystem obtain resources to interact with all other actors (Application Programming Interfaces – APIs and software development kits – SDKs). This allows the platform owner to attract innovative companies and ensure interoperability, while preserving its control over the ecosystem. The platform therefore provides its ecosystem members with resources to ensure interoperability of services, compatibility of developments and economies of scale and scope.11 This provision is essential to make possible the construction of a complex ecosystem involving heterogeneous and uncoordinated actors. According to von Hippel and Katz,12 these boundary resources transfer design capabilities to users. These resources include every contractual or technical asset likely to interface third-parties’ products and services with those of the platform.13 When available free of charge to new entrants, such assets drastically reduce barriers to entry, coordination and transaction costs: ‘[Their] use ... is aimed at lowering the often-large development and commercialization costs to new operations, therefore helping to create wider network effects ...’14 (in other words, they secure and facilitate third-party investments, which generate

9 Wen Wen and Zhu Feng, ‘Threat of Platform-Owner Entry and Complementor Responses: Evidence from the Mobile App Mark’, (2019) Strategic Management Journal, 40(9), pp. 1336–1367. 10 Simonetta Vezzoso, n. 7. 11 Ben Eaton, Silvia Elauf-Calderwood, Carsten Sørensen and Youngjin Yoo, ‘Distributed Tuning of Boundary Resources: The Case of Apple’s iOS Service System’, (2015) MIS Quarterly, 39(1), pp. 217–243. 12 Eric von Hippel and Ralph Katz, ‘Shifting Innovation to Users via Toolkits’, (2002) Management Science, 48, pp. 821–834, 824. 13 Ahmad Ghazawneh and Ola Henfridsson, ‘Balancing Platform Control and External Contribution in Third-Party Development: The Boundary Resources Model’, (2013) Information System Journal, 23, pp. 173–192. 14 Kristian Lauslahti, Juri Mattila and Timo Seppälä, ‘Smart Contracts – How will Blockchain Technology Affect Contractual Practices?’, (2017) ETLA Reports, 68, January, p. 7.

Cooperation, dependence, and eviction

7

complementary assets in return).15 It allows each app to adjust automatically to changes resulting from decentralized and uncoordinated decisions by all ecosystem members.16 However, this model of complementary investments involves heterogeneous firms with distinct bargaining power. The integrating firm, which acts as a gatekeeper and coordinator for third-party investments, is in a pivotal position.17 It is both a business provider and an orchestrator of technical developments.18 1.1.2

An Attractive Business Model for All Its Stakeholders

Beyond the abovementioned technical dimensions, integration into an ecosystem can also be financially attractive for complementors. The keystone player also has incentives to subsidize its complementors. For instance, in Android, the Commission fined Google for having ‘made payments to certain large manufacturers and mobile network operators on condition that they exclusively pre-installed the Google Search app on their devices’.19 In this logic, the MOS owner redistributes to third-parties some of the network externalities generated jointly. As mentioned above, incentives can also be provided by making boundary resources available free of charge. This profit-sharing is mutually beneficial since it strengthens network effects. According to Lauslahti et al.:20 ‘Digital platform owners mostly benefit from sharing boundary resources with third-parties by capitalising on split revenue business models.’ Surplus redistribution among the ecosystem members can also give rise to price protection mechanisms. The US Supreme Court ruling in Apple v Pepper

15 David J. Teece, ‘Profiting from Technological Innovation: Implications for Integration, Collaboration, Licensing and Public Policy’, (1986) Research Policy, 15, pp. 285–305. 16 Ahmad Ghazawneh and Ola Henfridsson, ‘Balancing Platform Control and External Contribution in Third-Party Development: The Boundary Resources Model’, (2013) Information System Journal, 23, pp. 173–192. 17 Robin Mansell, ‘Platforms of Power’, (2015) Intermedia, 43(1), pp. 20–24. 18 Orla Lynskey, ‘Regulating “Platform Power”’, (2017) LSE Law, Society and Economy Working Papers, 1/2017. 19 Google Android (case AT.40099) European Commission Decision, C(2018) 4761 final, 18 July 2018. 20 Kristian Lauslahti, Juri Mattila and Timo Seppälä, ‘Smart Contracts – How will Blockchain Technology Affect Contractual Practices?’, (2017) ETLA Reports No. 68, January, p. 7.

8

Challenges to assumptions in competition law

illustrates this point.21 The question was whether transactions between app developers and end users are carried out under Apple’s effective control.22 This case sheds light on the nature of the links between developers and ecosystem managers.23 An amicus curiae submitted to the Supreme Court shows the ambiguity of the relationship between the different stakeholders, which covers both an economic dependency relationship and a mechanism for redistributing the rent within the ecosystem. If an app developer is likely to accept unbalanced competition practices, is it because it fears some retaliation measures from the integrated firm, or because it also benefits from redistribution mechanisms? In this case, app developers are not likely to seek damages for potential antitrust injuries from the exercise of Apple’s monopsony power. Developers who bring suit against Apple risk jeopardizing their access to the App Store, as Apple can seek retribution against them in any of a myriad of different ways. In addition, many app developers may be unwilling to sue Apple for its monopolization of the app distribution market due to a perception that Apple may manipulate the pricing and sale of the apps in ways that ultimately benefit the developers.24

Indeed, as Vaheesan25 highlights, the rule imposed by Apple on the App Store, to impose prices ending in 99 cents has the effect of increasing developers’ revenues by limiting price competition. For the monopsonic firm (Apple’s App Store de facto is), it is also a way to redistribute part of its income to developers. The same intuition is highlighted by Posner26 who draws a parallel with the Microsoft case: ‘in exchange for agreeing to exclusionary provisions in their contracts with Microsoft, the manufacturers received various benefits, including discounts, cooperation in development, and greatly enhanced computer sales via the continuous upgrading’.

21 US Supreme Court, Apple Inc v Pepper et al., Certiorari to the US Court of Appeal for the Ninth Circuit, 17-204, 13 May 2019. 22 Walid Chaiehloudj, ‘Abuse of Dominant Position and Big-Tech: An Apple Store Case After the Android Affair?’, (2018) Concurrences, 4-2018. 23 The Supreme Court held that app users (e.g. indirect purchasers) could directly sue Apple for an allegedly anticompetitive commission overturning its Illinois Brick case law (Illinois Brick Co. v Illinois, 431 U.S. 720 (1977)). 24 Sandeep Vaheesan, Brief of Amicus Curiae Open Markets Institute in Support of Respondents, (2018) US Supreme Court, case 17-2004, Apple Inc v Robert Pepper et al., October, p. 3. 25 Sandeep Vaheesan, n. 24. 26 Eric Posner, Brief of Amicus Curiae Open Markets Institute in Support of Respondents, (2018) US Supreme Court, case 17-2004, Apple Inc v Robert Pepper et al., October, p. 18.

Cooperation, dependence, and eviction

1.2

9

A SYMBIOSIS COVERING A SITUATION OF ECONOMIC DEPENDENCE?

However, to access the ecosystem, complementors must accept the risk of swimming between sharks.27 Complementors’ dependence comes from both the lack of interoperability between competing ecosystems, and the strong market concentration since platforms are naturally winner takes all markets.28 Briefly, platforms are coopetition markets in which firms share their resources to create value but compete to maximize their profits.29 With respect to the incomplete contract theory, complementors – who make specific investments – are particularly exposed to a risk of contractual hold-up.30 The MOS is then a contracting nexus involving heterogeneous stakeholders in a value creation process. However, even if the coopetition model prevails, the bargain power is asymmetrically distributed among actors since only the integrator benefits from the regulatory power.31 Complementors rely on the pivotal firm both to develop their services and to access the market. The risks complementors have to assume are all the more important as the pivotal firm can collect data on their activity to compete with them ultimately. For instance, in September 2018, the EU Commission has announced the opening of a proceeding to investigate the way Amazon uses data related to the activity of independent merchants on its marketplace to adjust its own offers.32 Indeed, the pivotal firm owns a gatekeeper position it could take advantage of to reduce the visibility of the complementary offers through, for example, the manipula-

27 Luis Diestre and Nandini Rajagopalan, ‘Are All Sharks Dangerous? New Biotechnology Venture and Partner Selection in R&D Alliances’, (2012) Strategic Management Journal, 33(10), pp. 1115–1134. 28 Feng Zhu and Marco Iansiti, ‘Entry into Platform-based Markets’, (2012) Strategic Management Journal, 33(1), pp. 88–106. 29 Adam Brandenburger and Barry J. Nalebuff, Co-opetition (Currency Doubleday, 1997). 30 Oliver E. Williamson, Market and Hierarchies: Analysis and Antitrust Implications (Free Press, 1975). 31 Kevin J. Boudreau and Andrei Hagiu, ‘Platform Rules: Multi-Sided Platforms as Regulators’, in Annabel Gawer (ed.), Platforms, Markets, and Innovation (Edward Elgar, 2009), pp. 163–191. 32 Dirck Auer, ‘The Amazon Investigation and Europe’s “Big Tech” Crusade’, Truth on the Market, October, available at https://​truthonthemarket​.com/​2018/​10/​21/​ the​-amazon​-investigation​-and​-europes​-big​-tech​-crusade/​, accessed 16 December 2020.

10

Challenges to assumptions in competition law

tion of its ranking.33 The pivotal firm can also buy promising start-ups before it becomes a potential threat, like Facebook’s takeover of Instagram.34 Then, such platforms place complementors in a double situation of technical and economic dependence vis-à-vis the integrator.35 Technically, the platform owner can exclude third-party offers from the application store. Economically, exclusion strategies can be achieved through less favorable access conditions or, even more sharply, by bundling a substitute offer at no additional cost as a standard feature of the platform. In platform-to-business cases, the Commission has concerns about unfair contractual clauses or practices, the opacity of trading conditions and the lack of clear criteria for resolving disputes involving third-parties and the platform owner. The Commission therefore aims to define ‘principles of preventing abuse of market power and ensuring that platforms that serve as a gateway to downstream market do not become gatekeepers’.36 Platform owners have the unilateral power to deny (or to significantly impede) access to the market to third-party developers. This can be achieved, for instance, through dereferencing, modification of contractual conditions, discriminating between offers, etc. Such unilateral power results in the ability to impose unbalanced contractual conditions.37 In this regard, Vaheesan38 shows that Apple benefits from regulatory capacities over its complementors. It ‘includes being made subject to the particular, and arbitrary, political and moral rules that Apple uses to regulate the nature of the apps that are available for sale to iPhone customers. And it includes being made subject to arbitrary, discriminatory manipulation in the act of shopping for and buying apps.’ If Apple imposes unbalanced contractual terms to 33 Such as Zynga’s position after the reduction in the number of games appearing on Facebook’s newsfeed. 34 Li Zhuoxin and Agarwal Ashish, ‘Platform Integration and Demand Spillovers in Complementary Markets: Evidences from Facebook’s Integration of Instagram’, (2017) Management Science, 63(10), pp. 3438–3458. 35 The Commission notes that economic dependence mainly occurs when the ‘asymmetry between the relative market strength of a small number of leading platforms – not necessarily dominant in the sense of competition law – is combined with a highly fragmented supply-side of many small business users’. European Commission, Final Report on the E-commerce Sector Inquiry, 2017, COM(2017) 229 final, 10 May, p. 25. 36 European Commission, Impact Assessment Accompanying the Proposal for a Regulation on Promoting Fairness and Transparency for Business Users of Online Intermediation Services, 2018, SWD(2018) 138 final, 26 April. 37 Patrice Bougette, Oliver Budzinski and Frédéric Marty, ‘Exploitative Abuse and Abuse of Economic Dependence: What Can We Learn from an Industrial Organization Approach?’, (2019) Revue d’Economie Politique, 129(2), pp. 261–286. 38 Sandeep Vaheesan, n. 24, p. 6.

Cooperation, dependence, and eviction

11

third-parties, its retaliation capacities could also prevent it from potential lawsuits: ‘Developers who sue Apple for antitrust violations run the risk of being removed from the App Store and losing their access to end users – a threat that is not entirely theoretical.’39 The same argument can be found in the amicus curiae produced in the same case by Eric Posner,40 which takes up on this point a reasoning already held by Hovenkamp:41 ‘Many direct purchasers are “highly unlikely to sue” because they think they would be better off permitting an antitrust violation to continue rather than risking their relationship with the alleged violator.’ In most traditional coopetition models, companies cooperate upstream to create value and then compete downstream. In platforms models, downstream competition is distorted as long as the platform owner also controls the market access conditions. The platform is by itself an access lock and almost the only link between the ecosystem members and end users.

1.3

FROM SYMBIOSIS TO KILL ZONE: WHY PIVOTAL FIRMS CAN BE LED TO EXCLUDE THEIR COMPLEMENTORS?

Pivotal firm and complementors are then engaged in a complex process of value co-creation, but the former benefits from a strategic position it may take advantage of. In this third section, we explore market conditions in which platform owners have some strong incentives to exclude their complementors. 1.3.1

Dissymmetry in Data Control that Can Lead to Exclusionary Strategies

The ‘cooperative’ resource pooling mechanisms described above enable pivotal firms to detect, at a very early stage, promising new products or services, as well as potential disruptive innovation that could ultimately undermine their dominant positions. In this context, takeover bids are the first means to prevent competitive threats in the long term.42 Exclusionary strategies are the second. Microsoft, for example, is alleged to have excluded Netscape and Real Network respectively from the internet browser and music player markets through those Sandeep Vaheesan, n. 24, p. 12. Eric Posner, n. 26, p. 16. 41 Herbert Hovenkamp, ‘The Rationalization of Antitrust’, (2003) Harvard Law Review, 116(3), pp. 917–944. 42 Thibault Schrepel, ‘Predatory Innovation: The Definite Need for Legal Recognition’, (2018) SMU Science and Technology Law Review, 21, pp. 19–73, p. 49. 39 40

12

Challenges to assumptions in competition law

kind of practices.43 Such exclusionary practices may be analyzed as leverage strategies, or platform envelopment strategies, as referred to in the management science literature. Shared resources and users’ relationships allow the platform owner entry into the complementor’s market.44 Most of the time, the complementor has no chance in remaining competitive when a substitute offer is directly provided by the platform.45 Control over data – but also broader client and technology knowledge – gives the platform owner the ability to implement nowcasting strategies:46 ‘Nowcasting also represents a potent data-based weapon, not previously available for monopolies, to monitor new business models in real time. The nowcasting radar can help some dominant firms identify nascent competitive threats.’ The gradual replacement of offers developed by third-parties by homemade services is an attempt by the platform owner to capture the existing or future value.47 As described above, Microsoft’s bundling strategy gave it the opportunity to extend its dominant position from the tying market (Windows) to two tied markets (the ones of Netscape and Real Player). Predatory innovations pursue the same goal.48 Such strategies lie in ‘the modification of a technological platform and the technical design of a product – which are aimed at removing the compatibility of third-party technologies with those of a dominant firm, or at impairing competing technologies operation’.49 In one case, exclusion is achieved through the voluntary degradation of interoperability. In the other, it is through the launch of a homemade product. Such a move would impair the competitiveness of existing complementary goods, and deter potential rivals to enter the market. The platform owner can also exclude a complementor by degrading the performance of its offer, for instance in delaying the transmission of information needed for updates,50 or via demoting strategy (as reproached to Google in the Google Shopping case). The EU Commission suspects Amazon to engage in such practices towards the independent resellers of its marketplace. Amazon could exploit aggregated commercial data, as well as unique sourcing capabilities, to detect and then self-distribute the bestsellers. Such an information Simonetta Vezzoso, n. 7. Thomas R. Eisenmann, Geoffrey Parker and Marshall W. Van Alstyne, ‘Platforms Envelopment’, (2011) Strategic Management Journal, 32(12), pp. 1270–1285. 45 Feng Zhu and Qihong Liu, n. 4. 46 Maurice E. Stucke and Allen P. Grunes, ‘Data-Opolies: The US Point of View’, (2017) Concurrences, 2-2017. 47 Joseph Farrel and Michael L. Katz, ‘Innovation, Rent Extraction, and Integration in System Markets’, (2000) Journal of Industrial Economics, 48(4), pp. 413–432. 48 Thibault Schrepel, n. 42, p. 45. 49 Thibault Schrepel, n. 42, p. 22. 50 Thibault Schrepel, n. 42, p. 28. 43 44

Cooperation, dependence, and eviction

13

asymmetry is incompatible with the principle of fair competition.51 As stated by the Commissioner: The question here is about the data, because if you as Amazon get the data from the smaller merchants that you host … do you then also use this data to do your own calculations? What is the new big thing, what is it that people want, what kind of offers do they like to receive, what makes them buy products? These concerns relate to the fact that Amazon acts as both a retailer in its own right and a platform for other retailers, which allegedly constitutes a ‘conflict of interest’.52

This risk is also highlighted by Zhu and Liu.53 The latter, based on an empirical study on Amazon’s marketplace, shows that ‘Amazon’s entry strategy is likely premised on acquiring new information after forming partnerships with third-party sellers.’ 1.3.2

Exclusion through Anti-fragmentation Agreements

Exclusionary practices can also rely on anti-fragmentation agreements, as described above.54 Admittedly, such agreements can be defended on the basis of interoperability and security. But, as counterparts for the access to boundaries resources shared by the platform owner with no additional costs, these agreements55 are also a powerful tool to prevent any disruption risk. For example, within the Android ecosystem, terminal manufacturers had to accept the Mobile Applications Distribution Agreement (MADA) to get an Android license. Basically, MADA does not allow manufacturers to select apps on a discretionary basis, and even forces the pre-installation of several Google apps such as Search or Maps.56 From an economic perspective, such a requirement makes sense as the platform is a two-sided one and the end users operate Google’s services for free. Then, Google needs a critical mass of users to get a return on investment. With MADA comes AFA, for anti-fragmentation agreements, which aim at impeding development of alternative versions of the MOS, called Android

Dirck Auer, n. 32. M. Verstager, press conference of 18 September 2018. For a perspective and discussion of this case see Auer n. 32. 53 Feng Zhu and Qihong Liu, n. 4, p. 2620. 54 Frédéric Marty and Julien Pillot, ‘With Uncertain Damage Theory Come Unpredictable Effects of Remedies: Libres Propos on the Android Case’, (2018) Competition Policy International Antitrust Chronicle, December. 55 Such as CDD (Compatibility Definition Document) and CTS (Compliance Test Suite). 56 Frédéric Marty and Julien Pillot, n. 54. 51 52

14

Challenges to assumptions in competition law

forks. While third-parties in principle allowed modifying Android’s source code (openly accessible) to create forks, in practice devices manufacturers had to commit not to sell even a single smartphone running on an Android fork if they want to pre-install Google’s proprietary apps on their devices. When MADA can be justified by the economics of two-sided models, the rationale of AFA lies in the prevention of incompatibility issues. However, such a practice also reduces incentives to develop potential competitive offers to Android. Thus, the control over boundary resources gives the platform owner a strong (and defendable) competitive advantage.57 1.3.3

Buyouts Strategies to Exclude Potential Competitors

Not only are the major players in the digital economy continually fueling their growth through the acquisition of promising start-ups at a frenetic pace,58 but also these buyouts often involve companies that are part of their ecosystem. This is a usual practice in the biotechnology industry where innovation is mainly pushed by small and medium-sized firms. Indeed, in that industry, major companies – that are reluctant to assume the R&D costs and uncertainties – prefer to buy companies (and patents) once these risks have been removed. However, things are different in the digital world where such acquisitions seem much more motivated by the need to prevent a potential competitive threat, rather than the necessity to support and fund the growth of promising start-ups.59 Indeed, in digital markets, even the GAFA’s (Google, Amazon, Facebook and Apple) dominance can be disputed by a challenger offering a disruptive product or service,60 such as Nokia in the telephony market.61 Some buyouts can then be seen as a weapon to prevent any disruption risks. Facebook’s takeover of Instagram and the acquisition of Picasa by Google, as an attempt to respectively remain dominant in the social network market and to promote Google Photos, are two perfect examples. In a way, once a promising start-up has reached a certain stage of development, it would enter a kill zone. The

Simonetta Vezzoso, n. 7. Kevin A. Bryan and Erik Hovenkamp, ‘Start-up Acquisitions, Error Costs, and Antitrust Policy’, (2020) University of Chicago Law Review, 87, pp. 331–356. 59 Eric Posner and Glen Weyl, Radical Markets – Uprooting Capitalism and Democracy for a Just Society (Princeton University Press, 2018). 60 Clayton M. Christensen, Michael E. Raynor and Rory McDonald, ‘What Is Disruptive Innovation?’, (2015) Harvard Business Review, December, p. 5. 61 Timo O. Vuori and Quy N. Huy, ‘Distributed Attention and Shared Emotions in the Innovation Process: How Nokia Lost the Smartphone Battle’, (2016) Administrative Science Quarterly, 61, pp. 9–51. 57 58

Cooperation, dependence, and eviction

15

concept comes from Microsoft’s practices towards its complementors in the 1990s, which, to some extent, could lead to an ‘embrace, extend and extinguish’ strategy.62 Every promising complementor is particularly exposed to a buy-out risk: ‘they will eat their own children to live another day’.63 Third-parties can barely deal with such buy-out attempts, particularly because traditional response strategies, such as price competition, low-cost approaches or better user experience, are almost inefficient in the present case,64 if not impossible.65 Indeed, two-sided platforms have a zero price business model and competition on the tied markets is based on quality. As the platform owner is vertically integrated and benefits from the gatekeeper advantage, it represents a very credible threat of market exclusion for complementors.66 With that in mind, the latter are all the more likely to accept a friendly takeover approach. Besides, several structural factors increase the likelihood of kill zones as compared with the 1990s. First, numerous start-ups have the overall objective of being taken over by a leading company. Second, Tech Giants benefit from huge capacities in data capture and exploitation that give them the ability to detect potential competitive threats much earlier than in the past. Third, dominant firms own an unprecedented stock of scarce resources, including engineers. As most of the start-ups just cannot offer comparable salary conditions, they often appear less attractive than Tech Giants. Finally, takeovers of complementors are also the result of competition among platforms. In the Apple-Shazam M&A case,67 the Commission questioned how access to Shazam’s data could distort competition both on the music streaming market and the one for devices. As a vertically integrated group, Apple could be encouraged to downgrade Shazam’s music recognition algorithm for competitors’ customers.68 The Apple-Shazam case shows how

62 Noah Smith, ‘Big Tech Sets up a “Kill Zone” for Industry Upstarts’ (Bloomberg Opinion, 7 November 2018). 63 ‘American Tech Giants Are Making Life Tough for Start-Ups’, The Economist, 2 June 2018. 64 Benjamin Edelman and Damien Geradin, ‘An Introduction to the Competition Law and Economics of Free’, (2018) Competition Policy International Antitrust Chronicle, September. 65 Damien Geradin, ‘What Should EU competition Policy Do to Address the Concerns Raised by Digital Platforms’ Market Power’, (2018) working paper, available at https://​ssrn​.com/​abstract​=​3257967, accessed 16 December 2020. 66 Friso Bostoen, ‘Online Platforms and Vertical Integration: The Return of Margin Squeeze?’, (2018) Journal of Antitrust Enforcement, 6(3), pp. 355–381. 67 Case M.8788 Apple-Shazam, European Commission, 6 September 2018. 68 Nicolo Zingales, ‘Apple/ Shazam/ Data is Power, But Not a Problem Here’, (2018) Competition Policy International, December.

16

Challenges to assumptions in competition law

one takeover can hinder competition when it results in granting the dominant firm an essential asset or decisive informational advantage. One could wonder if requirements usually held in merger control procedures are well designed to assess how the concentration of massive (and initially fragmented) users’ data is likely to distort competition in the long term.

1.4

WHAT ROLE FOR COMPETITION RULES ENFORCEMENT?

Strategic relationships between platform owners and complementors are complex. As coopetition can meanwhile generate procompetitive and anticompetitive effects, such cases have to be undertaken with the greatest care by competition agencies, whether it be for assessing the damage to innovation and competition or for designing efficient remedies. 1.4.1

How to Assess the Potential Damage to Innovation and Competition?

1.4.1.1 Fair expansion strategies or exclusionary abuses? Most complementors are fully aware about the risks associated with two-sided models when the platform owner is (or could be) integrated both vertically and horizontally. They could then adjust their ex ante behavior and strategy in order to avoid the kill zone, e.g. deter the platform owner from extending its market.69 If the threat of integration is high, R&D investments shift to other markets and functionalities as the risk of contractual hold-up increases. Moreover, the effective entry of the platform owner is often beneficial to consumers in the short term as it reduces transaction costs and underinvestment bias.70 However, uncertainties about the future conduct of the platform owner are likely to dissuade third-parties from entering into long-term relationships and making specific investments. Such a situation may reduce further incentives to innovate.71 As two-sided platforms are mostly operated at no cost on the end user side, exclusion of complementors through vertical integration may not reduce the consumer welfare if the platform owner provides more efficient or less expensive substitutes. However, such a practice harms innovation as it 69 Such a result has been established for independent vendors on the Amazon platform. Feng Zhu and Qihong Liu, n. 4. 70 Wen Wen and Zhu Feng, n. 9. 71 Such as in Microsoft (2004), in which the Commission stated that ‘restrictions should not create disincentives to compete with Microsoft, or unnecessarily restrain the ability of the beneficiaries to innovate’.

Cooperation, dependence, and eviction

17

restricts freedom of choice as well as preventing the emergence of alternative technological trajectories. With such unbalanced relationships, what are the options open to complementors? Niche strategies are a first option. Non-cooperative strategies that consist in restricting data sharing with the platform owner are a second one.72 However, the sustainability of the latter option is somewhat hard to believe since the platform owner often has control over end users data. In any case, the platform owner is able to exclude from the market a complementor that could eventually prove to be a potential competitor. Such a practice could challenge the contestability of the market and give rise to irreversible dominant positions. Then, despite obvious information asymmetries, competition agencies have to balance pro and anticompetitive effects and determine how exclusionary practices could harm competition, innovation and consumers in the long term. 1.4.1.2 In M&A cases In M&A cases, several questions may be asked. First, what are the revenue and the market share of the target? Most of the targeted start-ups are not desirable because of their performance indicators, but for their technologies, their patents and sometimes their brands and services by themselves. In that respect, usual structural indicators (such as HHI) say little about potential competitive risks.73 Those are all the higher as platform owners are structurally focused on the future markets and compete in the search for new and low end market footholds. They engage corporate ventures in frontier technological areas, and attack peripheral markets.74 As platform owners follow mimetic conglomerate

Feng Zhu and Qihong Liu, n. 4, p. 2637. Traditional indicators should be completed by, for example, the stock market capitalization of the target (as proxy of the future consolidated cash flow) future or the mass and quality of data the consolidated group will control. The German reform on merger control that came into effect on 31 March 2017 was grounded on these principles. In addition to the traditional market share criteria, the value of the consideration paid in return for the transaction is now taken into account. If the amount exceeds €400 million, the deal is subject to preliminary checks, with no regards to the parties’ revenue or market shares. Heike Schweitzer, Julius Haucap, Wolgang Kerber and Robert Welker, ‘Modernization of Abuse Control for Companies with Market Power’, (2018) DICE Consult, Düsseldorf. More generally, should we in the case of innovation markets consider that patents or R&D expenditures can be adopted as indicators of the likelihood of the consolidated group to distort competition in the future? Richard J. Gilbert and Steven C. Sunshine, ‘Incorporating Dynamic Efficiency Concerns in Merger Analysis: The Use of Innovation Markets’, (1994) Antitrust Law Journal, 63, p. 569 et seq. 74 Nicolas Petit, Big Tech and the Digital Economy: The Moligopoly Scenario (Oxford University Press 2020). 72 73

18

Challenges to assumptions in competition law

expansion schemes, customer lock-in into silos and market foreclosure are two serious risks in the long term that antitrust agencies have to take into account. At this point, one may argue that a specific regime should be implemented for every M&A project that involves a dominant firm and one complementor. But, vertical integration is not problematic in itself. As mentioned above, in the biotech industry, it is usual to let small and medium-sized firms endorse most of the R&D expenditures, and to integrate the successful innovators afterwards. Thus, adopting a specific regime could distort established investment schemes in some industries. Similarly, in IT sectors where there are strong needs for service convergence and integration, reinforcing merger control proceedings may finally prove harmful to end users. However, as we have seen above, excessive market concentration can result in technological monopoly and huge switching costs for end users likely to foreclose markets in the short term, and restrict the development of alternative technological trajectories in the long term. How to balance pro and con effects in a context of still incomplete and asymmetric information? To that extent, the French competition agency proposes to implement an ex post control procedure for mergers.75 Such control already exists in some EU Member States, including the UK and Sweden, ‘where particular grounds exist’, e.g. when the competition agency considers that the nature of merger calls for an observation period to scrutinize its impact on the economy. Two questions can then be asked. First, how long should this observation period be? Second, how to deal with the legal uncertainty that such observation period introduces? On that particular point, the French competition agency proposes to publish clear guidelines and to limit the period in which such an ex post control could be implemented from six months to two years. Even a shortened period may seem very long in digital industries. 1.4.2

What are the Effects of Competitive Remedies?

In short, ecosystems are mainly coopetitive, and involve heterogeneous actors in a process of collective (and shared) value creation. Such a context makes it very difficult to assess the economic pros and cons of M&A or unilateral expansion strategies when a platform owner is involved. The same goes for designing efficient remedies, e.g. remedies likely to preserve the contestability of the markets without harming end users.

75 On 20 October 2017, the French Competition Authority launched a public consultation to modernize and simplify French merger control regime. See the public document at: https://​www​.aut​oritedelac​oncurrence​.fr/​sites/​default/​files/​consultation​_c​ oncentrati​ons20oct17​.pdf, accessed 13 January 2021.

Cooperation, dependence, and eviction

19

For example, in Android, the EU Commission required Google to bring some anticompetitive practices to an end (pre-installation of Google’s home services, including Search, Chrome and Play Store; financial incentives conditioned to exclusive installation of Google Search, and anti-fragmentation agreements). However, the EU Commission did not ignore the two-sided nature of the market, as it states: ‘the decision does not prevent Google from putting in place a reasonable, fair and objective system to ensure the correct functioning of Android devices using Google proprietary apps and services, without however affecting device manufacturers’ freedom to produce devices based on Android forks’. To comply with the Commission’s injunctions, Google proposed a set of remedies, including: • Disjoint distribution of Google Play’s Search and Chrome applications; • Disjoint and non-exclusive distribution of free licenses for Search and Chrome; • Full licensing for Google Play and eight other services (including Gmail, Google Maps ...), with price discrimination regarding the range of the device. The end of free access is intended to compensate the loss of the two-sided effect described above; • Removal of anti-fragmentation clauses. Then, manufacturers will be able to develop devices running on Android fork, as long as they inform consumers about the potential lack of compatibility with apps developed for the official MOS. Although these remedies do not challenge the MOS model per se, they still partially undermine it. For example, financial incentives towards manufacturers and phone operators could be seen as a way to redistribute part of the co-created value. Forcing Google to put an end to that practice may then have undesirable effects, such as increase in price in the device market, especially for entry-level products. Similarly, putting an end to anti-fragmentation clauses may incite Google to replace Android by a new MOS. Such a move would impose huge switching costs both for undertakings and end users. It is no accident if Google is already working on a next generation MOS, expressly confirmed under the codename ‘Fushia’. But with no need to catch-up a leading edge technology this time, as it was in 2007 when Apple disrupted the smartphone market, would Google maintain an open architecture for Fushia? Consequently, not only may such remedies not prove efficient, but they also may arrive too late while several Asian original equipment managers (OEMs) (such as ZTE and Huawei) presently work on their proprietary MOS. Interestingly, it is not antitrust proceedings but international trade policy considerations (especially tensions between the United States and China) that

20

Challenges to assumptions in competition law

encourage OEMs to enter the MOS market. Nevertheless, from the economic perspective, the more OEMs, the more fragmented are investments, the more important are development costs to avoid interoperability issues. Moreover, regardless of the EU Commission decision in Android and how US jurisdictions will settle in Apple v Pepper, it is almost fuzzy to forecast how relationships between MOS owners and complementors – especially native app developers76 – will evolve in the near future. With the progressive emergence of Progressive Web Apps (PWAs), interdependence between platform owners and complementors could no longer be taken for granted. PWAs are less costly, as these are not specific to one given ecosystem, and entail less update costs than native apps. PWAs are also somewhat user friendly since they save storage space and offer offline functionalities. The more PWAs will be widespread, the less the symbiosis model will be relevant. For all these reasons, the efficiency of remedies is nearly impossible to attest from an ex ante basis. The nature of the tradeoff will almost exclusively depend on how firms will adjust their business and technology strategies following the proclamation of remedies, and how they would have behaved in their absence in the long run. As competition agencies have few – if any – control on these issues, the risk of moral hazard looks extremely – and perhaps unwisely – important.

CONCLUSION MOS are the keystones of very complex B2B relationships between the platform owner and complementors. Such coopetition relationships in two-sided markets have both technical and business justifications. P2B cooperation has the advantage to share innovation risks and investments among several firms, and to create shared value in return. However, such a market organization confers a strong structural advantage on the platform owner as it benefits from a position of market gatekeeper, and controls most of the users’ data. The platform owner could take advantage of its specific market position to engage in some exclusionary practices, such as unfair envelopment strategies, likely to hurt innovation.77 While such practices do not harm consumer welfare in the short term (as services are most likely to be proposed at no additional cost once integrated by the platform owner), they 76 At an early stage, the development of native apps required numerous third parties and entailed significant costs (particularly with regard to update constraints). Nevertheless, these native apps have not always met expectations, offering sometimes low user experience and value added compared to a website with responsive design capabilities. 77 Damien Geradin, n. 65.

Cooperation, dependence, and eviction

21

could restrict freedom of choice in the long term as they prevent the emergence of alternative technological trajectories. There are ongoing discussions on the standard that competition agencies should apply to cases that involve two-sided markets. For example, the effective competition standard developed by Marshall Steinbaum and Maurice Stucke encompasses not only the consumer welfare dimension, but also seeks to protect/preserve competitive marketplaces, undertakings with no market power and more generally atomistic competition, free access to the market, and future opportunities for innovative firms.78 However, it is not clear whether such a standard could be efficient in two-sided digital markets. Indeed, two questions remain. First, how can we characterize any market power if the relevant market itself cannot be unambiguously defined? As Oliver Budzinski and Annika Stöhr note, market power is a distributed and relative phenomenon whose definition requires a case-by-case basis. Perhaps competition laws are less appropriate than contract laws to deal with that kind of case.79 Second, is it worth extending the special responsibilities incumbent on dominant firms to every complementor and, more largely, every member of its proprietary ecosystem? We have shown that a platform owner exercises quasi-regulatory power over its ecosystem as it can unilaterally determine and change the prices and market access conditions. It can also influence complementors’ investment decisions. In a way, as long as rival ecosystems exist, most of the situation of economic dependence results from (more or less) efficient and objective strategic choices. However, when a single and natural standard emerges, complementors have no more exit options. At this point, competitive risks have unquestionably to be taken into consideration regardless of the complex nature of the market organization or the potential risks flowing from imperfect remedies.

78 Marshall Steinbaum and Maurice Stucke, ‘The Effective Competition Standard – A New Standard for Antitrust’, (2020) University of Chicago Law Review, 86, pp. 595–623. 79 Oliver Budzinski and Annika Stöhr, ‘Competition Policy Reform in Europe and Germany – Institutional Change in the Light of Digitalization’, (2019) European Competition Journal, 15(1), pp. 15–54.

2. Technological tying: unbundling the assumptions Stephen Dnes 2.1 INTRODUCTION This chapter investigates the values underlying technological tying law. As a specific example, it considers the interaction of bundling law and Google’s integration of search and other functionality. For example, in relation to search and maps, or search and product comparison, one could consider whether: 1. As a dominant company in a related market, Google was free to combine maps and search into a new integrated product, i.e. the law on tying and new product designs; 2. The shift in the industry towards an integrated model of maps and search (the ‘competitive fringe’ of similar products) demonstrates the superior efficiency of an integrated model, and if so what the correct specification of the evidence required to meet competition law duties is; 3. Whether, factually, the decision of Google to integrate search and other functionality met this standard: in turn determining whether anti-competitive effects exist within the meaning of the law. At each stage of this analysis, questions of values and assumptions arise, not only as to the trade-offs between rivalry and efficiency, but at a more fundamental level as to the specification and identification of relevant competition. As developed below, there is a contrast between (1) deferring to the market to drive innovation through successive monopoly – in broad strokes, the position adopted by federal US antitrust law – and (2) maintaining a competitive fringe using competition law to manage adverse impacts on rivals. In both cases, the values underlying the choice are expressly efficiency-based, but differences exist in application: for the US system, the Schumpeterian belief that successive monopoly will maximise efficiency, and for the EU, that lowering barriers to entry will maximise productive efficiency, at least in the short term. Interestingly, this position may have changed slightly in 2018: with reference to the Google investigation, EU Commissioner Vestager has suggested 22

Technological tying: unbundling the assumptions

23

that the values underlying the investigation may be driven as much by fairness as by efficiency.1 Looking to the US, the latest merger enforcement may not be driven by express references to fairness, but a more subtle move towards greater attempts at vertical merger enforcement may herald similar developments by which vertical unbundling of the large platforms occurs.2 In summary, contrasts in underlying assumptions are in tension, and appear not lightly to be resolved. This chapter explores these differences in underlying values through the following structure. First, it assesses the argument that EU law is relatively strict in distinguishing in context whether integration reflects innovation, or problematic monopolistic behaviour. EU law is seen thereby to suggest only limited and context-sensitive deference. An important element of this assessment is that the balance between harm to competition and innovation must be proportionate under EU law: the balance struck must give effect to Article 102’s aims. This relationship directly engages the underlying values of Article 102. The position under US law is contrasted, and this is seen to be more permissive, in particular following the judgment of the Court of Appeals for the D.C. Circuit in Microsoft II, and related circuit court litigation showing significant deference to integration where there is an arguable case for efficiency benefits. The chapter then applies the contrasting legal positions to a contemporary example: Google’s integration of product functionality, which sits under the rules developed in relation to Microsoft. Particular attention is paid to the integration of maps, which was the focus of the English High Court’s analysis in Streetmap.EU v Google.3 Finally, the chapter assesses economic literature on product integration, but finds that the more general literature on vertical integration may depend on assumptions that are not as clearly applicable to online markets: Technology cases might be plausible exceptions to empirical results suggesting welfare enhancements from a number of vertical restraints. The chapter therefore concludes that, even on an effects-based approach, there appear to be differences in underlying assumptions about market dynamics that may require attention as case-law develops, not least given the global scope of many technology markets.

1 See e.g. ‘Fairness and Competition’, Speech delivered by Commissioner Vestager, 25 January 2018. 2 See e.g. U.S. v AT&T et al. Case 1:17-cv-02511 (D.D.C.) (challenge to vertical merger between AT&T and Time Warner; merger upheld, 12 June 2018; decision affirmed, 26 Feb 2019; case file available at https://​www​.justice​.gov/​atr/​case/​us​-v​-att​ -inc​-directv​-group​-holdings​-llc​-and​-time​-warner​-inc, accessed 16 December 2020). 3 Streetmap.EU v Google [2016] EWHC 253 (Ch).

24

Challenges to assumptions in competition law

2.2

LEGAL BACKGROUND TO TYING SEPARATE PRODUCTS

2.2.1

General Position on Integrating Products

Tying and bundling occur where companies with a dominant market position for a particular product (A) require consumers to use another product (B) in order to be allowed to use product A. Put another way, access to the dominant market is conditioned on the use of a secondary product where the company has less market power. The competitive effects of tying and bundling have been debated considerably, a noted criticism being that in certain straightforward markets it would be as simple, and perhaps more successful, to increase prices in the dominant product market (the ‘one monopoly rents’ theory arguing that monopoly profits can be obtained once and only once from dominant market A). In many markets, however, the position is more complex and the effect of tying and bundling can have significant anti-competitive potential. This is especially true where tying can be used to distinguish between groups of consumers with different consumer preferences, to charge them different effective prices. This can allow a monopolist to charge certain groups higher prices and extract more monopoly profits than would occur absent the tie. Example: Heavy users of photocopiers derive more value from copiers than light users and would pay a higher price for the product than light users. However, the manufacturer might be constrained from charging a higher price for the copier because of the potential loss of light users. One way around this problem would be to require the purchase of other products from the company, such as toner and repairs, and to mark up the cost of these supplies above their competitive price. The increase in the price of supplies affects the heavy user more than the light user, and the heavy user effectively pays a higher price. Normally, the expensive supplies would encourage switching to other copier manufacturers. However, if competition in the copier market is not strong, the copier manufacturer might succeed in selling copiers despite the policy of tying consumables. In this way, an effectively higher total price is charged to the high volume user who uses a large quantity of the marked-up supplies.

In the case of Google, tying search and maps might result in the effective price charged for the product (e.g. data collection, adverts) being higher and quality being lower than it would otherwise be in the absence of market distortion. It can also allow different user groups to be segmented (e.g., those accessing via horizontal search, and those accessing mapping sites direct, who may have different demand profiles).

Technological tying: unbundling the assumptions

25

Reflecting these concerns, it is illegal under EU competition law for a dominant company without objective justification: 1. to condition the provision of one product on the provision of another where, 2. it gives no choice to obtain the tying product without the tied product, 3. where the result of the practice is to foreclose competition.4 Prominent examples of bundling occur where formerly separate technological products are integrated. Many examples have taken place over the years, including cars and radios, cameras and telephones, and formerly disparate software products. Where the result of the practice is to distort competition, the combination is illegal under the ban on dominant tying and bundling. 2.2.2

Deference to Technological Tying in the Case of New Products

In the case of new product design, sometimes called ‘technological tying’, many competition law systems show a degree of deference towards innovation and US competition law is no exception. For instance, early cases on computer product integration under US antitrust law were notably deferential, see e.g. California Computer Products, Inc v IBM Corp, 613 F2d 727,744 (9th Cir. 1979) (holding that even a monopolist has ‘the right to redesign its products to make them more attractive to buyers – whether by reasons of lower manufacturing cost or improved performance’). The underlying thought has been crisply summarised by Areeda and Hovenkamp, whose leading treatise states that ‘product superiority is one of the objects of competition and cannot be wrongful, even for a monopolist’.5 A well-known example of deference to product integration under US antitrust law exists in the D.C. Circuit’s decision in U.S. v Microsoft Corp (‘Microsoft II’) in which the court held that integrated products, i.e. those providing a combined offering the consumer could not create after separate purchase, would be immune to challenge for not being separate products.6 Significantly, the court chose to apply a standard under which ‘the question is 4 Case T-201/04 Microsoft para 842. See also Case T‑30/89 Hilti v Commission [1991] ECR II‑1439, upheld in Case C‑53/92 P Hilti v Commission [1994] ECR I‑667 (tying of nails and nail guns), and Case T-83/91Tetra Pak v Commission [1994] ECR II‑755, upheld on appeal in Case C‑333/94 P Tetra Pak  v  Commission [1996] ECR I‑5951 (‘Tetra Pak II’) (tying cartons and carton machinery). 5 Phillip E. Areeda and Herbert Hovenkamp, Antitrust Law: An Analysis of Antitrust Principles and Their Application (2nd edn., Little Brown, 1996), § 781a at 268. 6 United States v Microsoft Corp, 147 F3d 935 (D.C. Cir. 1998).

26

Challenges to assumptions in competition law

not whether the integration is a net plus, but merely whether there is a plausible claim that it brings some advantage’.7 2.2.3

Technological Tying under EU Law

By contrast, EU competition law looks much more to market context and especially market structure to decide whether technological tying is illegal. The early cases cited above, such as Tetra Pak (carton machines and supplies) and Hilti (nail guns and nails) placed little weight on the impact of product design decisions on innovation. Even in the computer technology field, where a degree more deference can be advocated, the EU courts have been very sensitive to the precise context of the practices in terms of their impact on competition for other, related, products. In Microsoft, the Court of First Instance (CFI) did note that there is scope for a single product to evolve from ‘what initially appear to be separate products’8 which may subsequently be regarded as forming a single product because of ‘constant and rapid evolution’9 of IT products. However, in contrast with the approach of the D.C. Circuit in Microsoft II, little deference was extended to technological tying. The following factors articulate a significantly stricter approach: 1. The General Court stressed that the separate products question is to be assessed with regard to the ‘factual and technical situation that existed at the time when … the impugned conduct became harmful’, that is, the point at which the decision to combine the product is made.10 Thus, subsequent changes to consumer preferences after the tie are unlikely to excuse technological tying in the absence of an objective basis to adopt the tie at the point when it is made. This significantly narrows the scope to argue that a tie is innovative in a dynamic sense. 2. The General Court expressly stated (citing Hilti) that separate products can exist, even where consumer demand for them is complementary. Thus, the fact that there was no demand for Windows without Media Player did not bar a finding of separate products, since purchasers might well wish to obtain a number of products together, but from different suppliers. The Court pithily noted that ‘the fact that most client PC users want their client PC operating system to come with word-processing software

9

Ibid. at 948. Emphasis added. Microsoft n. 4 above at para 913. Ibid. at para 913. 10 Ibid. at para 914. The point was repeated in the context of remedies at para 942. 7 8

Technological tying: unbundling the assumptions

27

does not transform those separate products into a single product for the purposes of Article 82 EC’.11 3. The existence of even a relatively small group of consumers seeking copies of Windows without Media Player for workplace purposes pointed strongly away from a finding of a single, integrated product.12 4. Significantly, indirect as well as direct evidence of consumer demand was considered. Indirect evidence included aspects of the nature and features of the software, their historical development, the state of the market, and commercial practice. Indirect evidence noted by the court included: (1) distinctions between operating systems and software, (2) the existence of independent suppliers of streaming software, (3) the integration of Microsoft media streaming products with other operating systems, such as Apple’s, (4) the existence of independently downloadable software, (5) separate marketing activities, (6) different licensing of Media Player and Windows, and (7) the decision of some consumers to use competing products despite the bundling.13 5. The Court expressly rejected the argument that technical integration formed part of the ‘normal and necessary’ ‘constant improvement’ of products, instead looking to the purpose and technical constraints in the market which suggested that two separate products had been combined.14 6. Bundling by other operating system providers did not defeat a finding of separate products because: a. Microsoft provided no evidence that the other operating system providers had bundled the products at the time it chose to do so, and b. Some chose to allow a selection of players or no player to be installed.15 7. Findings of ‘commercial usage’ were in any event held to be suspect where one company held a 95 per cent market share.16 8. Even where a ‘natural link’ exists between products, or where combining them is consistent with commercial usage, combining products can still be found abusive unless the combination is objectively justified.17 Ibid. at para 922. Ibid. at para 925. 13 Ibid. at paras 926 to 935. 14 Ibid. at paras 935 to 937. 15 Ibid. at para 941. 16 Ibid. at para 940. 17 Ibid. at para 942. (‘even when the tying of two products is consistent with commercial usage or when there is a natural link between the two products in question, it may none the less constitute abuse within the meaning of Article 82 EC, unless it is objectively justified (Case C-333/94 P Tetra Pak II, paragraph 293 above, paragraph 37)’). 11 12

28

Challenges to assumptions in competition law

Many factors therefore played a role in the finding that Media Player and the Windows operating system constituted separate products. Importantly, the analysis by the CFI was sensitive to a wide variety of contextual factors which were given significant weight wherever they suggested that some separate demand might exist.18 The factors almost work to seek out potential separate demand, showing quite a different focus from that of the D.C. Circuit’s approach to show deference where a plausible efficiency claim exists. One important result is that the consideration of efficiency claims occurs later on and as a (rarely if ever successful) defence once a prima facie case against the practice has arisen. This is far less deferential than the approach of the D.C. Circuit to credit plausible benefits to prevent a claim arising in the first place. Instead, a wide range of contextual factors are considered including the state of demand at the time of the decision to combine the products, and the potential scope for others to offer the product separately. One might summarise the approach under EU law as requiring a strong claim that combining the product is necessary to achieve a pro-competitive aim, i.e. objective justification of tying provided that there is some evidence that separate products existed.

2.3

OBJECTIVE JUSTIFICATIONS, PROPORTIONALITY AND EFFECTIVENESS

The essence of Google’s objective justification claim, as made in Streetmap, is that combining maps and search is efficient and therefore justified. Under EU law, there is a strong claim that efficiency justifications for tying must be proportionate, because proportionality is a general principle of EU law interpretation. The General Court has summarised the principle as requiring a measure to be (1) suitable; (2) necessary; and (3) not excessively burdensome in relation to the desired aim.19 There is relatively recent authority for the proposition that all EU competition law regulation must be proportionate.20

18 The use of indirect evidence to show potentially separate demand was also seen in recent cases on aftermarkets, see e.g. T-427/08 CEAHR v Commission (existence of independent repairers taken to demonstrate separate demand for spare parts). 19 Case T-70/99 Alpharma Inc. v Council of the European Union, para 324. 20 Case C-441/07 P Alrosa (as observed by the parties and by the Advocate General in point 42 of her Opinion, although Article 9, unlike Article 7 of Regulation No. 1/2003, does not expressly refer to proportionality, the principle of proportionality, as a general principle of EU law, is none the less a criterion for the lawfulness of any act of the institutions of the Union, including decisions taken by the Commission in its capacity of competition authority).

Technological tying: unbundling the assumptions

29

The proportionality principle can often have a limiting effect on EU regulation. For instance, it would require a product regulation to be the least onerous option to achieve a given aim. By the same token, it can be argued that substantive limitations on commercial freedom under competition law must be the least onerous restrictions that still achieve the same aim. Although rarely argued expressly in proportionality terms, areas such as mandatory licensing show a concern not to regulate competition more than is necessary to achieve a given aim.21 The decision of the CFI in Microsoft to treat claimed efficiency improvements as objective justifications fundamentally reverses this position. Consider a mandatory licensing obligation. A duty to deal is a departure from a starting point of freedom to choose trading partners, and would need to be the least onerous effective option from the perspective of the regulated company if it is to comply with the proportionality principle. By contrast, an objective justification is an exception to a general ban on abuse of a dominant position. The legal starting point is that dominant companies are banned from tying separate products. This creates an entitlement on the part of competitors and consumers to benefit from separate sales. However, a narrow exception exists where efficiency is proved. The proportionality position is thus exactly reversed: the entitlement to separate sales is to be broadly construed, and exceptions to this are only permitted to the extent necessary. The relevant analysis is no longer finding the least onerous option for the company, but to find the one that provides the narrowest exception to the rule against tying from the consumer/competitor perspective. This follows from the decision in Microsoft to treat efficiency as a question for objective justification, i.e. as a defence rather than an element of the prima facie case. Although not expressly couched in proportionality terms, this approach to analysis is clear from the treatment of objective justifications in Microsoft at paragraphs 1144 to 1167. In assessing Microsoft’s claimed efficiencies from combining products, the CFI noted: 1. As a threshold matter, it was the failure to offer a plain version of Windows that was at issue; it was entirely possible to offer an efficient combined product provided that a non-bundled version was also available (para 1149); 2. Claimed efficiencies from the uniform presence of the combined product could just as easily suggest concerns about widespread harm from bundling as much as they could suggest widespread benefit (para 1151);

21 See e.g. necessity analysis in Case C-7/97 Oscar Bronner (‘economic indispensability’ requirement).

30

Challenges to assumptions in competition law

3. Even if standardisation of the product in its bundled form were efficient, other means besides a problematic tie by a dominant firm and its attendant competitive risks existed to achieve these (para 1152); 4. The alleged benefits seemed to be available without this specific (and extensive) tie (para 1154). The tie did not seem ‘essential’ to realise asserted efficiencies in code development (para 1157); 5. Many claims surrounding faster running had not been proven as a matter of evidence (para 1160) and in many cases seemed ‘speculative’ (para 1166). The above reasoning has a great deal in common with the proportionality approach. The court notes in (1) that the step needed to avoid potential competitive harm was small, and in (3), (4) and (5) that the tie had not been shown to be necessary to achieve these. The approach differs from the approach of the D.C. Circuit to defer to even potential efficiency savings from separate product integration. In effect, potential competitive harm in adjacent markets is prioritised over the possibility that a new integrated product might be useful, but have an impact on adjacent markets: the dominant company is required to show that it managed this risk, a potentially high bar that puts the innovator under a duty to manage the impact of its innovations. It is not enough that they just are useful to some users. This approach can be seen elsewhere, including in the underlying legislation, namely Regulation 1/2003. The reversal of the proportionality burden also follows from Recital 5 of that Regulation (emphasis added): In order to ensure an effective enforcement of the Community competition rules and at the same time the respect of fundamental rights of defence, this Regulation should regulate the burden of proof under Articles 81 and 82 of the Treaty. It should be for the party or the authority alleging an infringement of Article 81(1) and Article 82 of the Treaty to prove the existence thereof to the required legal standard. It should be for the undertaking or association of undertakings invoking the benefit of a defence against a finding of an infringement to demonstrate to the required legal standard that the conditions for applying such defence are satisfied. This Regulation affects neither national rules on the standard of proof nor obligations of competition authorities and courts of the Member States to ascertain the relevant facts of a case, provided that such rules and obligations are compatible with general principles of Community law.

The approach articulated in Recital 5 strongly suggests that it is for dominant companies such as Google to show that their objective justifications are proportionate departures from the general ban.

Technological tying: unbundling the assumptions

31

Microsoft states the same point in slightly more equivocal terms: It must be borne in mind, as a preliminary point, that although the burden of proof of the existence of the circumstances that constitute an infringement of Article 82 EC is borne by the Commission, it is for the dominant undertaking concerned, and not for the Commission, before the end of the administrative procedure, to raise any plea of objective justification and to support it with arguments and evidence. It then falls to the Commission, where it proposes to make a finding of an abuse of a dominant position, to show that the arguments and evidence relied on by the undertaking cannot prevail and, accordingly, that the justification put forward cannot be accepted.22

Although the statement is not as strong as that in Recital 5, the fundamental position is the same, namely that the defendant has to prove the justification. 2.3.1

Effectiveness Principle

Article 102 TFEU must be interpreted consistently with the effectiveness principle. The effectiveness principle plays a similar role to the requirement under proportionality that a measure be ‘suitable’, which implies a threshold of suitability. Importantly, the effectiveness principle is somewhat broader in requiring attention to be paid to the wider context of law in its interpretation. Where interpretive ambiguity exists, preference is to be given to an interpretation that gives effect to the provision in question. In Toshiba, the court summarised the point as follows: In interpreting a provision of EU law, it is necessary to consider not only its wording, but also the context in which it occurs and the objectives pursued by the rules of which it is part (see, in particular, Case 292/82 Merck [1983] ECR 3781, paragraph 12; Case C‑162/09 Lassal [2010] ECR I-0000, paragraph 49).23

Lassal elaborates with specific reference to competing interpretations: In that sense, the enacting terms of an EU act are indissociably linked to the reasons given for it, so that, when it has to be interpreted, account must be taken of the reasons which led to its adoption (Case C‑298/00 P Italy v Commission [2004] ECR I‑4087, paragraph 97 and the case-law cited, and Sturgeon and Others, paragraph 42). Likewise, the Court held that, where a provision of EU law is open to several interpretations, preference must be given to the interpretation which ensures that

Microsoft n. 6 above at para 1144. Case C-17/10 Toshiba and others v Úřad pro ochranu hospodářské soutěže, para

22 23

73.

32

Challenges to assumptions in competition law

the provision retains its effectiveness (see Sturgeon and Others, paragraph 47 and the case-law cited).24

There are several possible interpretations of competition law duties. These general principles require an interpretation that gives effect to the underlying purpose of Article 102 TFEU in its legislative context.25 2.3.2

Balancing Risks to Innovation and Risks to Foreclosure in Light of Effectiveness and Proportionality Principles

Importantly, the effectiveness requirement means that whatever the weightings adopted to innovation and competition concerns, the context of Treaty and other legislative provisions safeguarding competition suggests that the balance struck must effectively safeguard competition. As a result, the balance therefore seems to be in favour of interpretations safeguarding competition when weighing innovation concerns. Thus, where a small step could be taken to prevent a substantial risk to competition, without seriously diminishing incentives to innovate, a proportionality approach would suggest that the balance should be struck so as to require this. The general principles above require the balance to be struck in a proportionate way that gives effect to the law in its context as in the following examples. Example I: Microsoft Media Player. Potential harm to competition is high if competing media players are excluded from the market. Claimed efficiencies seem small and can be obtained through less restrictive means. The balance suggests no justification. Example II: IBM disk drive cases. Potential harm to competition exists, but is perhaps more limited (e.g. others might supply disk drive components to IBM if they are more efficient). Claimed efficiencies are very large, restriction-specific and verifiable (increased storage capacity). There is more scope to argue justification.

It is exactly this type of weighing that the CFI undertook in Microsoft and is notably different from the deferential approach of the D.C. Circuit in Microsoft II. Additional arguments can be made in favour of a precautionary approach, in addition to those listed above from proportionality and effectiveness.

Case C‑162/09 Lassal [2010] ECR I-0000 paras 49–51. Emphasis added. This includes Regulation 1/2003, whose many recitals make frequent reference to effective application. 24 25

Technological tying: unbundling the assumptions

2.4

33

APPLICATION TO GOOGLE’S INTEGRATION OF MAPS

The legal context on separate products and proportionality can now be applied to Google’s decision to integrate maps and search: 2.4.1

Integrating Maps and Search: Separate Products Analysis

Under the rules above, Google could claim that maps and search are no longer separate products, and therefore that it was free to create an integrated search and mapping product. Drawing on the guidance from Microsoft, strong arguments can be made against the claim that search and maps have become a single product. EU law is relatively strict on the separate products question, and many factors from Microsoft suggest that a finding of separate products is likely: • At the time of integration many independent map providers existed; • Demand for maps and searching for locations may be complementary, but this does not bar a finding of separate products as the complements could be separately sourced; • Even a relatively small number of users of an independent product can amount to sufficient consumer demand for a finding of separate products; • Indirect evidence of consumer demand for separate mapping products, such as supply side practices including licensing, financing, and advertising, can be credited and point towards separate products; • There is no technological constraint requiring combination of the products; • There is no obviously proportionate objective justification for doing so (see below). Applying the Microsoft analysis therefore strongly suggests a finding of separate products under EU law, whereas an integrated product market might follow if laying emphasis on integration benefits as occurs under U.S. law. Even at the level of product market definition, therefore, underlying assumptions appear to affect analysis. 2.4.2

The Competitive Fringe and Comparisons with Other Providers

Many other providers have integrated maps and search products. This raises the prospect whether widespread use prevents a finding of separate products, and whether the conduct is objectively justified as discussed at 2.4.3 below. The argument that a finding of separate products is inappropriate where a practice is widespread is weaker than might first be presumed. This follows

34

Challenges to assumptions in competition law

from the fact that it is entirely possible that a widespread practice is inefficient where it is undertaken by a dominant company. The Microsoft court was very much alive to this possibility, and many aspects of its analysis suggest that even quite widespread use by others will not defeat a finding of separate products: • Microsoft bars assessment for separate products with hindsight. The paramount importance of assessing the state of the market at the time the bundling occurred, and not later (para 914). In the case of Google, the decision to integrate came early and in a market in which many providers had not done so. This alone is fatal to the claim, a fair reading of Microsoft barring judgment of product integration questions with the benefit of hindsight (especially at para 914). • Other bundles might differ, so the practice of non-dominant players can be misleading. Everyone might bundle a product, but the Microsoft court was careful to stress that other bundles by non-dominant players might contain a different mix of products and be much more responsive to consumer demand (para 922). In fact, this might be evidence that the products are separate. • Widespread bundling could still be anti-competitive. If a practice is widespread, it is perhaps less likely to be inefficient. It is nonetheless entirely possible that widespread bundling could be inefficient; in fact, it might be a greater problem than one firm undertaking the practice (para 1151). • Comparisons with competitors are misleading where market shares are very high. In the pithy words of the Court, ‘It is difficult to speak of commercial usage in an industry that is 95% controlled by Microsoft.’ (para 940): • The incentives faced by non-dominant companies may be different and thus misleading. • Even if they are similar, there might be incentives to replicate the activities of a dominant player in some circumstances, no competitive pressure to offer different map/search bundles because of consumer expectations based on the abusive conduct of the dominant player. • Even a small competitive fringe that behaves differently will defeat a more widespread industry practice for the tying analysis; again, as it might be evidence that the tie is simply successful in its exclusionary effects (para 941). • Evidence of the existence of other players goes to exclusion analysis, but exploitation analysis is still potentially problematic. Other players might exist, but Google might not be constrained by them and still use its tie to exploit consumers, e.g. through differentiating tied and non-tied

Technological tying: unbundling the assumptions

35

demand. Exclusionary and exploitative effects do not necessarily coalesce; evidence of exclusion speaks primarily to the exclusionary aspect. Once again, the relatively strict approach in Microsoft suggests that claims that innovation points towards a finding of combined products will receive relatively little weight under EU law. Although the suggestion that courts should defer to innovators may have a degree of intuitive appeal, the close analysis undertaken by the court in Microsoft suggests that, in the specific context of a dominant company, care is needed to consider the full context in which the integration took place. In other words, the risks of foreclosure from dominance are receiving more weight, and although risks of chilling innovation through regulation are weighted, they receive far less weight than under US law. Whether that is a difference of principle, or of interpretation, is a potent question: on one view, the difference is simply in interpretation of foreclosure, but on another the EU is seen as having a relatively precautionary approach that may reflect more protective or interventionist values. 2.4.3

Proportionality Arguments and Objective Justifications

Relatively widespread integration of maps and search could also be taken to be evidence of efficiencies giving rise to objective justification of the practice. Much of the same analysis directly above would suggest caution in finding objective justification, as the adoption of a practice by other companies following its adoption by a dominant company does not necessarily mean that it is efficient. Additional arguments can be advanced based on the specific context of objective justifications as a defence to be construed proportionately. As discussed above, the EU competition rules seek to prevent distortion of competition, and Article 102 TFEU specifically seeks to balance concerns about competition with concerns about innovation in a proportionate way. The following points can be made in this context if applying EU law, which might differ under the arguably more permissive US rules: • Google must show objective justification on a proportionate basis. The status of objective justifications as a matter for companies to prove has an important implication: once a prima facie case exists against Google, it is for Google to demonstrate the proportionality of its potentially distortive combination of products. The claimed efficiency gain must be shown to have been achieved using the least restrictive means. • What was the extent of the efficiency asserted? Microsoft requires a balance to be struck weighing innovation and risks to competition from

36

Challenges to assumptions in competition law

tying. There may be a degree of efficiency in combining search and map functions (although whether it is greater than lost efficiency from rivalry in mapping is an open question). At the same time, the innovation is relatively small, obvious and seems unlikely to be deterred by application of the tying rules. Unlike the integrated computer systems in the IBM cases, for instance, it does not seem necessary to pursue map and search integration on such an exclusionary basis to achieve the desired new product. • Less restrictive means to achieve that efficiency abound. Google could have used many less restrictive means to achieve the same efficiency than making itself the sole integrated provider. As noted in Microsoft, it was entirely open for Google to offer an integrated product; the objection lies in failing to offer a non-integrated option (para 1151). An option to show alternative providers, or a selection of providers, or even a changeable default, would all have less scope to distort competition than integrating only the Google product and would in no way bar the efficiencies Google might have achieved by integrating its product alongside a non-integrated option. • Incentive effects of such a duty are likely to be ambiguous. It might be argued that the incentive to innovate is diminished by the potential for other providers to benefit from Google’s integration. First, this does not necessarily follow: Google might in fact benefit from partnering with the most efficient map provider. The decision not to do so might suggest that monopolistic activity is occurring in the map market, e.g. lesser quality mapping geared to Google’s own purposes. • Microsoft requires dominant companies to strike a very specific balance between risks of harm to competition and risks of harm to innovation. A defence based on the argument that dominant players are entitled to integrate products without regard for competitive effects disregards the legal position articulated in Microsoft. The position is that a balance must be struck between returns to integration by dominant companies, and competitive risks posed by it. Google arguably faced significant potential harm to competition in the effect that leveraging its strong market position was likely to have on the separate mapping market when maps and search were integrated. A small and simple step in providing a choice of mapping providers would have preserved this innovation at only a very modest cost to the return to innovation (perhaps even a positive one if the alternative map provider was better or more popular). Whether this was disproportionate will require careful analysis. In addition to proportionality analysis, an argument can be made that an interpretation that would allow Google not to heed its special duty as a dominant company to pay attention to potentially distortive effects of its conduct would

Technological tying: unbundling the assumptions

37

frustrate the purpose of Article 102 TFEU. As Article 102 TFEU is primarily concerned with the regulation of dominance and its attendant risks, rather than the encouragement of innovation, interpretation must give effect to this choice of priorities. Thus, an interpretation that robs Article 102 TFEU of its effectiveness by placing a dominant company under no duty regarding product integration considerations should be rejected as a poor choice of interpretation when several more effective interpretations are available. A strong case therefore exists that on a close reading of the balance struck in Microsoft between potential competitive harm and innovation, Google was required to pay considerable attention to the competitive risks its dominant position entailed under Article 102 TFEU, in particular to consider small but effective steps to diminish those risks – in effect, reversing the burden of proof. In any event, it is for Google to show evidence that the harm to competition its practices were likely to cause was strictly necessary to achieve asserted efficiencies, in contrast with the US position deferring to such claims.

2.5

ECONOMIC ANALYSIS

To return to the initial question regarding values and aims, it is interesting to consider how the above judgments correspond to the question of economic efficiency. As noted at the outset, there is no necessary conflict between economic efficiency and any of the rules above: they simply have different emphasis on short-term harm to rivalry and longer term innovation; and even these may not always be in tension. In short, the rules appear to emphasise different points within a framework that is based primarily on a concept of efficiency. There is extensive economic literature on tying and bundling in the context of technological products. Microsoft itself generated a significant level of discussion; some, being part of lobbying efforts, perhaps needs to be read with a degree of care. Many respectable commentators have supported the position that the product integration in Microsoft was anti-competitive. Perhaps the most remarkable and prominent voice was that of Robert Bork in the United States, a voice very much associated with efforts to decrease antitrust scrutiny. Yet in Microsoft’s case, Bork argued in favour of intervention and suggested that banning tying practices was correct.26

26 See e.g. Harry First, ‘Bork and Microsoft: Why Bork was Right and What we Learn about Judging Exclusionary Behavior’ (paper presented at 2013 symposium at Yale Law School).

38

Challenges to assumptions in competition law

There is a range of analysis with debate over the correct outcomes even within a framework that considers economic efficiency alone. For example, an influential 1990 paper by Michael Whinston, ‘Tying, Foreclosure and Exclusion’,27 pushed back against influential arguments that suggested limits to the scope for anti-competitive tying practices. This more modern analysis suggests that there is much greater scope for anti-competitive effects from tying where products are differentiated or where goods are purchased in different quantities. Indeed, Whinston commented on Microsoft following the 1990s litigation in ‘Exclusivity and Tying in US v Microsoft: What We Know, and Don’t Know’.28 The article concluded that, at least in the early 2000s, there was limited consensus on the likely welfare effects of tying, albeit that significant changes had occurred since the time of the Chicago school dominance in the 1980s in the heyday of the one-rents theory. A more recent resource spanning both theoretical and empirical literature is Abbott and Wright, ‘Antitrust Analysis of Tying Arrangements and Exclusive Dealing’.29 After a detailed analysis of the legal and historical context of tying, the authors cite two influential works suggesting limited empirical evidence of harm from tying.30 There is also empirical work to complement the theoretical pieces. A conference organised by the Swedish competition authority advocated that more empirical work is needed in this area.31 Such work as there is shows mixed empirical results. An influential 2005 study notes that in many cases, vertical restraints appear to be at least neutral, but that there are some notable instances of vertical restraints harming consumer welfare as well.32 The results are also from more ‘traditional’ industries and do not, on the whole, come from technology cases. The authors are, however, careful to note that results are neither comprehensive nor robust enough to prompt a wholesale change of approach, that cases do exist in which negative effects are seen, and that much depends on what is classified as a negative effect for the purpose of the empirical studies. American Economic Review (1990), 80(4), pp. 837–859. The Journal of Economic Perspectives (Spring 2001), 15(2), pp. 63–80. 29 George Mason University Law and Economics Research Paper Series No. 08-37. 30 Michael A. Salinger, ‘Business Justifications in Tying Arrangements’ (2005), in Wayne D. Collins (ed.), Issues in Competition Law and Policy (ABA, 2006); James C. Cooper, Luke M. Froeb, Dan O’Brien, and Michael G. Vita, ‘Vertical Antitrust Policy as a Problem of Inference’, Int’l J. Indus. Org. (2005), 23, 639. 31 The Pros and Cons of Vertical Restraints, Swedish Competition Authority, 2008. 32 F. Lafontaine and M. E. Slade, ‘Exclusive Contracts and Vertical Restraints: Empirical Evidence and Public Policy’, Journal of Economic Literature, (2005), available at https://​economics​.ubc​.ca/​files/​2013/​05/​pdf​_paper​_margaret​-slade​-exclu sivecontracts-verticalrestraints.pdf, accessed 16 December 2020. 27 28

Technological tying: unbundling the assumptions

39

This leaves open the important question of high-tech industries, in which there appears to have been less empirical work. As is noted in a working paper by Kuhn, Stillman and Caffarra, the dangers of bundling are more pronounced where network effects are strong. With specific reference to the Microsoft case they note: However, we explain that intervention against bundling is most justified when the evidence suggests that the motivation for the bundling is to affect the intensity of competition in the future. As discussed below, industries in which network effects are important seem especially conducive to such bundling. This in turn suggests that competition authorities should pay special attention to the competitive effects of bundling for instance in industries such as software, where network effects are unquestionably important.33

In summary, the empirical literature does not so far appear to have arrived at firm conclusions on technological tying, meaning that a policy element to the discretions inevitably remains. Although there are some suggestions that this might be modelled in terms of empirical effects, there is limited experience with doing so, which in turn implies that significant, value-based judgment calls are involved in tying policy at least for the foreseeable future.

2.6 CONCLUSION The above discussion suggests that there are significant differences in underlying values and assumptions, in particular in the essential question of whether market dynamics should receive sufficient deference that short-term increases in barriers to entry are tolerable. This shows up most clearly in the burden of proof, which was seen to be on the government under US law, whereas the burden is on the dominant company to show justification under EU law. Insofar as this shows a difference in underlying assumptions, it may be possible to develop further empirical literature to examine which approach drives more innovation as a matter of fact. Whichever approach is favoured, it is striking that both levels of intervention can be supported with reference to efficiency-based concepts; differences may lie in application more than in any fundamental difference of approach. Indeed, as technology markets have moved towards greater consolidation, and, perhaps, greater public concern over consolidation, it seems that both the EU

33 Kai-Uwe Kuhn, Robert Stillman, and Cristina Caffarra, ‘Economic Theories of Bundling and Their Policy Implications in Abuse Cases: An Assessment in Light of the Microsoft Case’ (November 2004). Available at SSRN: https://​ssrn​.com/​abstract​=​ 663404.

40

Challenges to assumptions in competition law

and US systems have moved towards a greater focus on vertical unbundling, and on maintaining competition within, as well as for, the ‘stack’ of products. It will be interesting to contrast this position with that developing in other jurisdictions, which may provide further insight into preferences on error costs, as many jurisdictions struggle to balance risks and benefits of technological change, both within competition law, and more generally for society at large. In particular, it can be noted that the pending Google litigation at the EU courts therefore raises important points of principle and will make a lasting mark at the level of principle in the context of a rich existing debate.

3. Understanding the role of agricultural data on market power in the emerging Digital Agriculture sector: a critical analysis of the Bayer/Monsanto decision1 Can Atik 3.1 INTRODUCTION Although there might be a perception that agriculture is a rural activity that is far from technology,2 modern agricultural production has become densely data-driven with advanced Big Data implementations3 thanks to the proliferation of the Internet of Things (IoT) and cloud computing technologies.4 Agricultural data (ag-data) is collected via various methods and used in agricultural decision-making processes such as implementing sensors in fields to track and prevent crop diseases, using drones and satellite images to observe crop developments, or using IoT infrastructure in glasshouses to 1 This chapter has been presented at the 14th ASCOLA Conference on 27–29 June 2019 in Aix-en-Provence, France and subjected to double blind peer review. This research has been conducted with the support of the postgraduate scholarship program of the Ministry of National Education, Republic of Turkey. 2 See the ever-present perception that agriculture is even behind the industry, ‘industrialization’ or ‘modernization’ of agriculture theories after the 1980s, and the EU regulations’ implicit adoption of these developments at Francis Synder, New Directions in European Community Law (1st edn., Weidenfeld and Nicolson, 1990), pp. 109–119. 3 Michael E. Sykuta, ‘Big Data in Agriculture: Property Rights, Privacy and Competition in Ag Data Services’ (2016) 19 International Food and Agribusiness Management Review 57, p. 58. 4 Harald Sundmaeker et al., ‘Internet of Food and Farm 2020’, in Ovidiu Vermesan and Peter Friess (eds), Digitising the Industry – Internet of Things Connecting the Physical, Digital and Virtual Worlds (River Publishers, 2016), pp. 132–133; Sjaak Wolfert et al., ‘Big Data in Smart Farming – A Review’ (2017) 153 Agriculture Systems, pp. 69–75.

41

42

Challenges to assumptions in competition law

enhance coherence in management of the production.5 It is expected that Big Data will positively affect productivity, sustainability, food safety, efficiency of resources and waste management.6 Moreover, current developments are regarded as just the first step of a big revolution in the operation, management, and structure of the entire agricultural value chain.7 In line with these developments, the Digital Agriculture sector (the DAs) has emerged based on ‘the collection of data and information about farms with the aim of providing tailored advice or aggregated data to farmers’.8 Within this particular sector, Agricultural Technology Providers (ATPs)9 offer various data-driven agronomic services for farmers to help them become better decision-makers from production to marketing stages of their profession.10 While Digital Agriculture practices create significant benefits,11 this data-driven transformation has brought about some competition law concerns. For instance, the European Commission (the Commission) reacted to the Bayer/Monsanto merger12 and implemented remedies13 due to inter alia the Digital Agriculture aspects of the concentration.14 Although there have been some attempts to discuss the case and the possible effects of Digital Agriculture on the agricultural inputs sector from a vertical integration perspective,15 the role of ag-data on providing services and on 5 Adam Lesser, ‘Analyst Report: Big Data and Big Agriculture’ (GIGAOM, 2014) https://​gigaom​.com/​report/​big​-data​-and​-big​-agriculture/​ accessed 5 January 2021; Wolfert et al., n. 4, p. 73; Krijn J. Poppe et al., ‘Information and Communication Technology as a Driver for Change in Agri‐food Chains’ (2013) 12 EuroChoices 60, pp. 60–63; Krijn J. Poppe et al., ‘A European Perspective on the Economics of Big Data’ (2015) 12 Farm Policy Journal 11, pp. 11–12. 6 Poppe et al., n. 5, p. 18. 7 Ibid., p. 12. Additionally, this development is not limited to food production, i.e. this may also have significant effects on various industries that are dependent on the agricultural raw material. 8 See the definition at Case No. COMP/M.8084 – Bayer/Monsanto, Commission Decision (29 May 2018), para 2442. 9 See at Sykuta, n. 3, p. 58, footnote 1. 10 Wolfert et al., n. 4, p. 74. 11 Poppe et al., n. 5, p. 11. 12 Which is the first investigation on the emerging Digital Agriculture sector so far. See at Bayer/Monsanto, n. 8, para 2555. 13 Ibid., para 3030 and the subsequent paras. 14 Ibid., para 2555. 15 Which predominantly focus on seeds/pesticides markets and reinforcement of concentration risks with Digital Agriculture implementations. See, for instance, Ioannis Lianos and Dmitry Katalevsky, ‘Merger Activity in the Factors of Production Segments of the Food Value Chain – A Critical Assessment of the Bayer/Monsanto Merger’ (2017) UCL-CLES Policy Paper Series: 2017/1; Maurice E. Stucke and Allen P. Grunes, ‘An Updated Antitrust Review of the Bayer-Monsanto Merger’ (2018) The

The role of agricultural data in the emerging Digital Agriculture sector

43

market power of ATPs in the emerging DAs with a horizontal competition perspective has not been discussed enough so far. Therefore, this study aims to fill this gap in the literature by exploring data-driven market power in the DAs and also critically analyzing the Bayer/Monsanto decision in this regard. In particular, it will be argued that the Commission’s considerations regarding the role of ag-data on market power might not be considered as comprehensive enough16 since it did not provide an extensive analysis about distinguishing reasons for the emergence of data-driven market power in this new sector from the perspective of the economic characteristics of ag-data even though it touched upon various potential elements in the text such as network effects and first-mover advantage.17 As having a well-rounded understanding of data-driven dominance is of prime importance for possible future DAs cases, this chapter aims to identify (1) the sector dynamics and the ag-data’s distinctive role on market power in the DAs; (2) the indispensable elements of an optimal market power assessment based on economic characteristics of ag-data in the sector; and (3) to what extent the Commission’s considerations in the Bayer/Monsanto decision is compatible with them. In particular, this study puts forward that the most prominent element of data-driven entry barriers in the sector is switching costs stemming from legal and technical obstacles that prevent free-flow of data in the DAs. This is exacerbated by lack of data substitutability due to cross-dependency of farmers and ATPs in the data collection process. Even though these issues were mostly neglected in the decision, they should be taken into account during market power assessments in the possible future DAs cases. The remainder of the chapter is organized as follows. Section 3.2 will convey the essential information regarding the Bayer/Monsanto decision to provide a baseline for the rest of the analysis. Section 3.3 will attempt to explore the economic characteristics of ag-data in order to generate an in-depth insight into the reasons for the emergence of data-driven market power in the DAs by also discussing the Commission’s considerations in the Bayer/Monsanto decision.

Konkurrenz Group White Paper; Tom Verdonk, ‘Planting the Seeds of Market Power: Digital Agriculture, Farmers’ Autonomy, and the Role of Competition Policy’, in Leonie Reins (ed.), Regulating New Technologies in Uncertain Times (Springer, 2019), pp. 112–115. 16 Maybe because it was the first case that separately evaluated the emerging DAs so far. See at Bayer/Monsanto, n. 8, para 2555. Although the Commission touched upon precision farming in the Dow/DuPont decision (See at Case No. COMP/M.7932 – Dow/DuPont, Commission Decision (27 March 2017), para 246), the DAs was not under the investigation in this case. 17 Bayer/Monsanto, n. 8, paras 2830–2846.

44

Challenges to assumptions in competition law

Section 3.4 will provide an overall evaluation regarding the Commission’s data-driven market power assessment in the Bayer/Monsanto decision, and it will put forward some suggestions for future cases. In light of this, Section 3.5 will conclude by formulating an optimal market power test for the DAs.

3.2

BAYER/MONSANTO MERGER CASE

There has been a merger wave amongst the agricultural giants within the last decade. In 2013, Monsanto acquired The Climate Corporation mainly because of its data sets by paying nearly USD 1 billion.18 After this data-driven takeover in the US, Monsanto attempted to acquire Syngenta in July 2014, but this attempt was not successful due to the rejection from the target company.19 However, these activities started a broader merger trend, including mega-mergers such as ChemChina/Syngenta,20 Dow/DuPont,21 and finally Bayer/Monsanto.22 This merger wave mainly decreased the number of market players in the agricultural inputs sector, but the DAs was also affected because nearly all the merging conglomerates had Digital Agriculture operations as well.23 However, the Commission provided a standalone investigation about Digital Agriculture related concerns only in the last one, the Bayer/Monsanto merger.24 3.2.1

Bayer/Monsanto Case at a Glance

According to the Commission, the transaction would create the largest globally integrated seed and pesticide player, and this would have significant impacts on price and innovation in several markets, including seeds, pesticides, and also Digital Agriculture.25 In particular, the concern was that this merger

18 Tony Danova, ‘Big Data Is Worth $1 Billion To Agricultural Giant Monsanto’ (Business Insider, 2013) https://​www​.businessinsider​.com/​monsanto​-buys​-climate​ -corporation​-for​-1​-billion​-2013​-10​?international​=​true​&​r​=​US​&​IR​=​T accessed 5 January 2021. 19 Lianos and Katalevsky, n. 15, p. 2. 20 Case No. COMP/M.7962 – ChemChina/Syngenta, Commission Decision (5 April 2017). 21 See Dow/DuPont, n. 16 above. 22 See Bayer/Monsanto, n. 8 above. 23 Lianos and Katalevsky, n. 15, p. 3, footnote 7. 24 See at Bayer/Monsanto, n. 8, para 2555. 25 ‘Mergers: Commission clears Bayer’s acquisition of Monsanto, subject to conditions’ (2018) https://​ec​.europa​.eu/​commission/​presscorner/​detail/​en/​IP​_18​_2282 accessed 5 January 2021 [‘Press release’ henceforth].

The role of agricultural data in the emerging Digital Agriculture sector

45

‘would have strengthened Monsanto’s dominant position on certain markets, where Bayer is an important challenger of Monsanto’.26 The Commission’s investigation identified particular competition concerns about Digital Agriculture.27 It focused on the loss of potential competition in Europe between Bayer’s recently launched Xarvio and Monsanto’s FieldView platform, the leading platform worldwide, which was about to be launched in Europe.28 As a response to this, Bayer offered a set of commitments29 to license a copy of its worldwide current offering and pipeline on Digital Agriculture to BASF (The BASF Divestment Package)30 in order to maintain competition by allowing BASF to replicate Bayer’s position in the European Economic Area (EEA).31 However, after this initial commitment, Bayer requested to replace its original commitment with complete divestment of its Digital Agriculture assets and products to BASF with a condition of a temporary license back to these assets and products on 11 April 2018, and the Commission approved this by considering it was sufficient to maintain the race to become a leading supplier in Europe in this emerging field.32 3.2.2

The Commission’s Relevant Market Definition and the Broader Digital Agriculture Sector

Services that ATPs provide for farmers are divergent, and there are various markets in the DAs. Some companies offer services in pest and disease modelling, some focus on satellite data, others have weed monitoring services with cameras, or there are crop modelling and nitrogen optimization services, and some ATPs have image recognition services for disease identification.33 Markets may also vary according to input types such as herbicides, insecticides, fungicides, fertilizers, plant growth regulators or seeds.34 Additionally, markets may be further distinguished according to crop groupings such as broadacre crops, e.g. corn, wheat, or barley.35 While being aware of this market Ibid. Inter alia the Commission’s particular concerns in (a) seeds and traits, (b) pesticides, and (c) other initial concerns that were not confirmed during the investigation such as innovation in biological pesticides and bee health. See more at ibid. 28 Ibid. 29 Other commitments were about (a) vegetable seeds, (b) broadacre seeds and traits, and (c) pesticides. See more at ibid. 30 See more detail at Bayer/Monsanto, n. 8, para 3046 and subsequent paras. 31 Press release, n. 25 above. 32 Ibid. 33 See more at Bayer/Monsanto, n. 8, Annex – 3, p. 44. 34 Ibid., para 2576. 35 Ibid., paras 2577–2578. 26 27

46

Challenges to assumptions in competition law

variety, the Commission’s concerns focused on the market ‘for the provision of digitally-enabled prescriptions of fungicides for broad acre crops in the EEA’36 in the Bayer/Monsanto merger. The crucial point of discussion in the market definition part is related to whether the provision of digitally-enabled prescriptions can be considered under the same market with traditional agronomic advisory services, as the Notifying Party (Bayer) argues that they are in the same broader market.37 However, the Commission rejected this argument by considering demand and supply sides of the related services that differentiate these two markets significantly because the demand requires much more detailed and customized services whereas supply is much more complicated, which requires significantly different capabilities such as accessing huge data sets and having advanced algorithms, compared to traditional agronomic advisory services.38 Regarding the relevant geographic market, the Commission stated that it was determined as ‘national’39 by considering several elements such as language differences, different farming practices, government regulations, types of crops, local crop diseases and farmers’ ATP preferences that have regional focuses.40 As this study mainly focuses on market power, there will not be a substantial discussion about market definition, but it is necessary to state that the approach of the Commission on market definition appears reasonable.41 The Commission is aware of the divergent services for farmers in the DAs. It identified the sector in general as well as the particular relevant and geographic markets in detail. However, it is difficult to argue that the data-driven market power assessment in the Bayer/Monsanto decision is as comprehensive as in the relevant market considerations to identify the most delicate parts of the matter.

Ibid., para 2612. Ibid., paras 2557–2561. 38 Ibid., paras 2562–2578. 39 Ibid., paras 2583 and 2593. 40 Ibid., paras 2583–2592. 41 Indeed, there is a trend that attaches less importance to market definition, and more significance to theories of harm and identification of anticompetitive strategies in data-driven markets investigations. See Jacques Crémer, Yves-Alexandre de Montjoye and Heike Schweitzer, ‘Competition Policy for the digital era – Final Report’ (Publications Office of the European Union, 2020), pp. 3–4 [‘EU Report’, henceforth]. On the other hand, it is important to note that some competitors had concerns regarding other Digital Agriculture markets as well. See at Bayer/Monsanto, n. 8, para 2598. 36 37

The role of agricultural data in the emerging Digital Agriculture sector

3.2.3

47

Competitive Assessment: Digital Agriculture Aspects of the Concentration

Within the competitive assessment part, the Commission starts with the concerns raised during the market investigation. According to the questionnaires conducted by the Commission with stakeholders, it was highlighted that the merger would result in higher prices, narrower choices and less innovation in the sector.42 In particular, competitors were worried about the acceleration of network effects and possible predatory pricing strategies in the post-merger period by bundling agricultural inputs and Digital Agriculture services.43 The risk of increased entry barriers for small-scale start-ups was also underlined by arguing that the new company would control too many IP rights and data.44 These risks drove competitors to a general concern about the dominance of the merged entity in Europe that would influence all the rival services.45 The Notifying Party argued that the merging parties were not leaders in the sector and their existing Digital Agriculture activities were related to different stages of the agricultural process, and thus, they did not overlap each other as Monsanto did not provide services in the EEA, and no evidence showed it would enter in a timely manner.46 Additionally, the Notifying Party asserted that merging parties’ access to data was not unique and the combination of data sets via the merger would not give them a competitive advantage,47 and there was no basis for first-mover advantage or network effects concerns.48 Also, it put forward that there were enough competitive constraints by rivals in the sector, there was no plan to cease innovation activities in the post-merger period, market investigation of the Commission was misleading, and the market was not defined based on widely accepted market studies.49 However, the Commission rejected these arguments by listing and comparing the existing and upcoming products of the merging entities, and discussing Monsanto’s capabilities, strategic position, plans and possible entry to the EEA.50 Based on this comparison, it concentrated on the elimination of

Bayer/Monsanto, n. 8, para 2594. Ibid., para 2599. 44 Ibid., para 2600. 45 Ibid., para 2601. 46 Ibid., paras 2602–2606. 47 Ibid., para 2603. This study strongly opposes this allegation as the substitutability of ag-data can be very limited or not possible in some situations. See more discussion in Section 3.3.2.6. 48 Bayer/Monsanto, n. 8, para 2609. 49 Ibid., paras 2607–2611. 50 Ibid., paras 2614–2699. 42 43

48

Challenges to assumptions in competition law

potential competition51 by considering that the merging parties had significant strengths and capabilities that constituted significant competitive constraints to each other.52 Eventually, the Commission concluded that the transaction was likely to lead a significant impediment to effective competition due to the elimination of an important competitive constraint in the post-transaction period.53 The Commission could not obtain reliable data about market shares of the merged entities as the sector is new and still emerging.54 Therefore, the competitive assessment was conducted based on the following main layers of substantial analysis. In the first layer, in addition to the existing and upcoming products comparison as the indication of potential competition,55 the Commission considered the rivals’ internal reports and their views along with the internal documents of merging parties to demonstrate that they might have imposed important competitive constraints to each other in the absence of the transaction.56 For instance, the view of BASF was stressed in the decision in a way that Monsanto was the leading player in Digital Agriculture globally, and Bayer was also the leading player in the EEA.57 Accordingly, the Commission also concluded that Monsanto was the leader in the sector globally and likely to become the leader in the EEA, and Bayer was ‘ahead of the other competitors with regard to digitally-enabled prescriptions of crop protection products in the EEA’.58 In order to support the conclusion declared in the first layer, the Commission provided additional reasons. The first reason was that the merging parties’ superior knowledge about their conventional seeds and pesticides businesses could generate competitive advantage in the post-merger period.59 Second, as access to raw data is one of the fundamental pillars in digitally-enabled prescription services, merging parties’ data collection networks would help them to generate better prescriptions thanks to their superior access to the public, third-party, customer and proprietary data sets.60 The text provided the merging parties’ data access capabilities such as their data collection partner Ibid., para 2612, and subsequent paras. In particular, para 2631. Ibid., para 2700. 53 Ibid., para 2872. As will be discussed below, although there are accurate considerations in this analysis, there are also significant missing points or potentially misleading evaluations for future DAs cases, if not this one. See more discussion in Section 3.3.2. 54 Bayer/Monsanto, n. 8, para 2613. 55 Ibid., paras 2614–2699. 56 Ibid., paras 2700–2708. 57 Ibid., paras 2705–2706. 58 Ibid., para 2710; see also para 2735. 59 Ibid., paras 2712–2714. 60 Ibid., para 2715. 51 52

The role of agricultural data in the emerging Digital Agriculture sector

49

ships with third parties including John Deere, AGCO and others in addition to their farm data sets61 from their customers, and also in-house research and proprietary data about their input products.62 After listing the related capabilities, the Commission again referred to rivals’ opinions regarding the combination of these data sets. For instance, it was indicated that data aggregation in the hands of the merged entity might generate clear advantages compared to rivals63 with better algorithms trained through these superior data sets, or a wide variety of data could enable Bayer/Monsanto to expand their services to other crop types and services.64 This was summarized as the ability to offer services and improve them as well as to develop new services compared to rivals.65 The third factor was the merging players’ powerful agronomic engines and data processing capacities.66 Therefore, the Commission considered that the creation of prescriptions by combined capabilities and data sets of the merging giants would be superior compared to products of those who have relatively limited access to these required data layers.67 Thus, the decision refused the Notifying Party’s arguments about not having unique access on proprietary (about seeds and pesticides) and third-party data sets, and the absence of any competitive superiority over rivals.68 It concluded that the distinctive capabilities of the merging companies generate a competitive advantage, and the merger would eliminate close potential competition as well as significant competitive constraints in the market.69 This study agrees with the related considerations to a large extent. However, the Commission did not provide a standalone analysis regarding the replicability or substitutability of these data sets70 even though this is one of the most important parameters when assessing how concerning a data combination could be from the competition point of view.71 Additionally, the conclusions

61 However, the text did not discuss details, importance or the distinctive role of farm data in the sector. 62 See Bayer/Monsanto, n. 8, paras 2720–2723. 63 However, there is not any discussion about how serious this concentration would be, what the separate roles of these data sets on market power are, whether they are replicable or not. 64 Bayer/Monsanto, n. 8, para 2726. 65 Ibid., para 2727. See detailed discussion in Section 3.3.2 below. 66 See Bayer/Monsanto, n. 8, paras 2729–2734. However, data’s importance on providing prescription services was still stressed as the decisive one. See para 2733. 67 Ibid., para 2734. 68 Ibid., para 2728. 69 Ibid., paras 2734 and 2738. 70 Especially, it underestimated the farm data component throughout the analysis. 71 See the importance of determining replicability/substitutability of data when assessing data-driven market power at EU Report, n. 41, pp. 49 and 101–105; see also

50

Challenges to assumptions in competition law

were mostly supported by rivals’ opinions, but depending too much on competitors’ concerns might not be considered as the optimal analytical criteria to understand the role of ag-data on market power in the DAs. Instead, the Commission could have provided more detailed economic characteristics analysis based on the sector realities and the facts of the particular case.72 The third assessment layer in the decision focused on the existing competitors of the merging parties. The Commission concluded that very few players (such as DowDupont, ChemChina/Syngenta or BASF)73 might have comparable capabilities, but even they could not exercise a sufficient degree of competitive pressure on the merged entity in the post-merger period based on its investigation conducted with questionnaires, conference calls and also direct interactions in a trade fair.74 The reason behind this conclusion is strictly related to data-driven prescription services’ dependency on a wide variety of data.75 In particular, the Commission underlined the importance of proprietary data76 on agricultural inputs and capacity of data processing tools,77 and emphasized the difficulty to reach an equal capability with the merging entity for rivals78 by also considering that the non-integrated companies such as small start-ups or software developers did not have the required genetic and performance insight about inputs in terms of scale and/or scope to provide digitally enabled prescriptions equal with integrated giants.79 The Commission also Autorité de la concurrence and Bundeskartellamt, ‘Competition Law and Data’ (2016), pp. 44–47 [‘Franko-German Report’ henceforth]. 72 The fourth assessment layer touched upon some of them, but the critical ones were neglected. See Section 3.3.2 for a more detailed discussion in this regard. 73 Bayer/Monsanto, n. 8, para 2758. However, it concluded that they were unlikely to exercise competitive pressure in the post-merger period effectively. When evaluating the other integrated giants’ possible competitive constraints such as DowDupont, ChemChina-Syngenta or BASF, the Commission, concluded that they were either not active in the EEA or did not have the comparable capabilities for the particular market under investigation. See more at paras 2787–2829. Moreover, even though they had enjoyed the equal capabilities to compete, they would have faced the first-mover advantage of the merging entity and eventual network effects. See paras 2817 and 2826. 74 Ibid., paras 2739–2744; see also paras 2827–2829. 75 Ibid., para 2746. 76 The decision focused too much on proprietary data sets (data about seed or pesticide performance). See at ibid., paras 2470 and 2745–2829. The Commission’s assessment is correct but does not represent the whole picture in the sector as data-driven market power is not limited to proprietary data sets. See more discussion in Section 3.3.2.1 below. 77 Bayer/Monsanto, n. 8, para 2747. 78 Ibid., para 2748. 79 Ibid., para 2750 – see also para 2760; for software companies, there is an additional limitation regarding farmers’ data as they are further away from farmers compared to integrated inputs giants such as merging parties. See paras 2782–2786.

The role of agricultural data in the emerging Digital Agriculture sector

51

checked the large distributors and agricultural equipment companies, but concluded that the former ones had the same limitations with the small start-ups in terms of the required capabilities.80 Even though the latter ones have certain services in the broader DAs, machine producers are collaborators rather than competitors in the data-driven prescription services.81 Exclusive access to proprietary data sets, for sure, is a highly relevant parameter to assess the effect of data accumulation on the market power of merging parties, especially in the inputs (seed and pesticide) prescription markets as in the case of the merger. However, it can be argued that the Commission overfocused on the proprietary data advantage, and neglected the role of farmers’ data and the distinctive nature of the users (farmers) and ATPs’ relationships on market power.82 In the fourth layer of the assessment, the Commission mentioned the first-mover advantage and network effects as the main characteristics of the DAs (including the specific market under investigation) and argued that these were likely to decrease potential post-merger competitive constraints further.83 The Commission reached this conclusion by investigating merging parties’ internal documents,84 but did not provide a detailed explanation about the reasons for this conclusion. It particularly referred to an internal document containing Bayer’s positive feedback loop expectations, i.e. the more subscribers (farmers) in the Digital Agriculture services means more authentic products, which in turn attracts more farmers as additional users of the services.85 The Commission connected positive feedback loops with the first-mover advantage by stating that first-comers could create a quality advantage compared to late-comers, and this might constitute entry barriers as the users would not have any incentive to switch new services86 by again referring to internal documents of Bayer and Monsanto about the importance of quantity and quality of farmers’ data to develop better services.87

Ibid., paras 2763–2768. Ibid., paras 2769–2779. 82 The Commission mentioned the importance of accessing farmers’ data to compete in the market in general. Ibid., paras 2783 and 2786. However, it did not discuss farm data’s relevance to market power assessments in the sector in detail. See extensive discussion on the matter in Section 3.3.2. 83 Bayer/Monsanto, n. 8, para 2830. 84 Ibid., para 2831. 85 Ibid., para 2836. 86 Ibid., para 2837. 87 Ibid., paras 2838–2840. However, these natural user preferences might not be the correct base of the theory of harm for the decision. This paper argues that focusing on data related lock-ins might have been a more solid approach when considering the sector conditions. See below Section 3.3.2 in general, and especially Section 3.3.2.9. 80 81

52

Challenges to assumptions in competition law

Understanding the economic characteristics of ag-data is the most delicate part of the market power assessment, but the Commission frequently used internal documents or rivals’ opinions in its discussion instead of providing an in-depth analysis about the sector conditions from the perspective of the economic characteristics. Additionally, the Commission’s very brief assessment in this layer reflects only a part of the story.88 The next sections will provide a comprehensive analysis regarding the role of ag-data on market power in the DAs. In particular, the chapter will demonstrate that even though the first-mover advantage is the most prominent characteristic of the DAs, the reason behind this advantage is more related to switching costs, which mainly derive from farmers’ data related to lock-in situations due to legal and technical barriers that prevent free data flow in the sector. This already high switching cost is exacerbated by the lack of substitutability of farm data because of the distinctive cross-dependency of farmers and ATPs in the costly data collection process. The last layer of the substantive assessment in the decision is about whether Bayer will continue the same innovation efforts in Digital Agriculture in the post-merger period. Based on the market investigation and the internal documents of the merging parties, the Commission concluded that the pre-merger separate innovation efforts were unlikely to be continued as they would be merged in the post-merger period.89 As a general consideration, even though the concerns and large parts of the reached conclusions are not far from the sector realities, it can be argued that especially the data-driven market power considerations could have been much more analytical based on generating insights from a comprehensive economic characteristics discussion instead of overreliance on stakeholder ideas as discussed in detail below.

88 For instance, the Commission used quotes from individual calls with KWS and DowDupont about switching costs and network effects in order to complement its main position on the importance of the first-mover advantage and network effects, but the Commission did not provide a separate switching costs analysis. However, especially KWS’s statements on switching costs contain very crucial hints about the most delicate parts of the data-driven entry barriers in the DAs that were not covered deservedly by the Bayer/Monsanto decision; ‘It is difficult to switch from one platform to another, since the industry is not able to agree on one common data protocol (joint data format), therefore there is high incentive for the farmer to decide on only one platform. Even though farmers keep the ownership of provided data and they can contractually request that their data are returned to them, from the technical point of view, such data are not compatible with another platform and can therefore not be easily transferred to another platform from a practical point of view.’ See at ibid., paras 2742–2843. 89 Ibid., paras 2847–2862.

The role of agricultural data in the emerging Digital Agriculture sector

3.3

EXPLORING DATA-DRIVEN MARKET POWER IN THE DIGITAL AGRICULTURE SECTOR

3.3.1

The Concept of Market Power

53

The market power90 of a company is a highly significant parameter of competition law analysis for all kinds of cases because it may lead to detrimental consequences that contradict the main aim of the EU competition law, i.e. consumer welfare.91 Accordingly, the Commission’s guidelines on the assessment of horizontal mergers92 consider creation or strengthening of a dominant position via mergers as a primary form of such competitive harm.93 In the conventional market power assessment, the first determinant is the market shares of the merging undertakings.94 It is considered that very large market shares – 50 percent or more95 – might be a significant indicator of dominance although there might be exceptional circumstances, and the importance

90 The level of market power that leads to dominance is described in the United Brands v Commission case as ‘a position of economic strength enjoyed by an undertaking, which enables it to prevent effective competition being maintained on a relevant market, by affording it the power to behave to an appreciable extent independently of its competitors, its customers and ultimately of consumers’. Case T-27/76 United Brands Company and United Brands Continental v Commission ECLI:​EU:​C:​1978:​22, para 65. 91 Richard Whish and David Bailey, Competition Law (Oxford University Press, 2018), p. 25; see some other risks and potential sufferers from data-driven dominance at Wolf Sauter, ‘A Duty of Care to Prevent Online Exploitation of Consumers? Digital Dominance and Special Responsibility in EU Competition Law’ (2019) Journal of Antitrust Enforcement, pp. 11–13. 92 Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between undertakings, OJ C 31, 5.2.2004 [‘Horizontal Merger Guidelines’, henceforth]. 93 Ibid., para 2. Dominance ‘provides an important indication as to the standard of competitive harm that is applicable when determining whether a concentration is likely to impede effective competition to a significant degree’. See para 4. 94 Market shares provide a useful first indication for the Commission regarding the market structure and the relative importance of the various undertakings active on the relevant market. See at Case 85/76 Hoffmann-La Roche & Co. v Commission [1979] ECLI:​EU:​C:​1979:​36, paras 39–41; Case C-62/86 AKZO v Commission [1991] ECLI:​EU:​C:​1991:​286, para 60; Case T-30/89 Hilti v Commission [1991] ECLI:​EU:​T:​ 1991:​70, paras 90, 91 and 92; Case T-340/03 France Télécom v Commission [2007] ECLI:​EU:​T:​2007:​22, para 100. Basically, it is calculated based on the ratio of sales of a company compared to the total sales in the market. See more at the EU Report, n. 41, p. 48. 95 Horizontal Merger Guidelines, n. 92, para 17; AKZO v Commission, n. 94, para 60; Case T-221/95, Endemol v Commission, [1999] ECLI:​EU:​T:​1999:​85, para 134, and Case T-102/96, Gencor v Commission, [1999] ECLI:​EU:​T:​1999:​65, para 205.

54

Challenges to assumptions in competition law

of the market shares might vary from market to market.96 However, with the datafication of the economy, these exceptional situations have become more frequent because conventional assessment tools struggle to perform their function.97 Especially in dynamic and fast-growing markets with short innovation cycles, market shares might not represent the market power correctly.98 In such cases, entry barriers in the relevant markets are stated as helpful to assess the market power more accurately.99 Although one may consider that services are not provided free of charge in the DAs100 and, thus, the market shares might represent the market power to a certain extent,101 focusing solely on market shares might still be deceptive in such a highly dynamic and still emerging data-driven sector. Also, this dynamism might not be limited to the infancy period of the sector as there is no constant stability regarding the roles and powers of players due to the everlasting innovative developments, i.e. market players’ positions are changing in line with IoT and Big Data advancements.102 Ongoing adoption of Smart Farming practices by farmers might also influence this instability. Indeed, estimations for the EEA-wide revenues were not stable at the time of the transaction in the Bayer/Monsanto decision.103 Accordingly and accurately, the Commission did not focus on market shares in Bayer/Monsanto by saying that there was See Hoffman-La Roche, n. 94, para 41. Especially in data-driven online markets. See at Rupprecht Podszun and Stephan Kreifels, ‘Data and Competition Law’, in Vanessa Mak, Eric Tjong Tjin Tai, and Anna Berlee (eds), Research Handbook in Data Science and Law (Edward Elgar Publishing, 2018), p. 191. 98 Case T-79/12 Cisco Systems Inc. and Messagenet SpA v Commission, [2013] ECLI:​EU:​T:​2013:​635, para 69. Digital platforms are considered as one of the prominent examples of this, i.e. lack of market shares’ effectiveness in market power assessment. See EU Report, n. 41, pp. 48–50; See, for instance, the Commission’s consideration at Case No. COMP/M.7217 – Facebook/WhatsApp, Commission Decision (3 October 2014), para 99. See more at Inge Graef, ‘Market Definition and Market Power in Data: The Case of Online Platforms’ (2015) 38 World Competition: Law and Economics Review 473, p. 503. 99 Especially, in the substantial assessment of data-driven mergers. See EU Report, n. 41, p. 116. 100 See generally at Sundmaeker et al., n. 4; See Bayer’s business model, for instance, at Bayer/Monsanto, n. 8, para 2547; See also the FieldView as an example at ‘Pricing’ (FieldView, 2018) https://​climate​.com/​pricing accessed 10 February 2020. See more information about the business model of the sector in Section 3.3.2.5 below. 101 If the market shares can be calculated properly. 102 Wolfert et al., n. 4, p. 75; Sundmaeker et al., n. 4, p. 147. 103 Bayer/Monsanto, n. 8, para 2448; Even though the Commission did not investigate Digital Agriculture in the Dow/DuPont decision, it touched upon the issue when explaining industry trends and new technology developments as ‘it is still unclear which player will emerge as a leader in this area’. See at Dow/DuPont, n. 16, para 246. 96 97

The role of agricultural data in the emerging Digital Agriculture sector

55

no reliable data about market shares.104 Moreover, it can be argued that even though there had been reliable data about market shares, this might still have been misleading to solely depend on as they may represent only a part of the broader picture because of the determinative role of data on the market power of ATPs beyond the ratio of sales of a company compared to the total sales in a market, as explained in detail below. However, market shares can be useful when calculated with an innovative approach. For instance, the German competition authority, Bundeskartellamt, assessed the market shares based on the number of users in the platform compared to the total users in the relevant market in its Facebook decision.105 However, this method might not fit in well with the market shares assessment in the DAs because the number of users might not directly represent the data advantage or market shares of ATPs as one ATP might have more data-driven advantage/income with fewer users compared to their rivals if the subscriber farms conduct large-scale activities, and thus, upload more data and pay more subscription fees. Therefore, focusing on the number of acres for which an ATP is providing services compared to the total acres in the relevant market might be a more useful criterion to identify.106 On the other hand, as farmers have a profoundly weak position against ATPs,107 which have standard ‘take it or leave it’ terms and conditions that are closed to negotiations,108 farmers have very limited power to decide on already collected data sets, and they are highly dependent on ATPs and machine producers.109 Therefore, countervailing buyer power might have much less meaning when investigating the market power of a company in the DAs compared to market shares. For this reason, the main focus should be on entry barriers when exploring data-driven market power.

Bayer/Monsanto, n. 8, para 2613. See Case Summary of Facebook decision, B6-22/16, p. 6 https://​ www​ .bundeskartellamt​.de/​SharedDocs/​Entscheidung/​EN/​Fallberichte/​Missbrauchsaufsicht/​ 2019/​B6​-22​-16​.pdf​?​_​_blob​=​publicationFile​&​v​=​4 accessed 5 January 2021. 106 Of course, this has to be complemented with a broader data-driven market power assessment by focusing on entry barriers due to explained dynamism in the sector. 107 Sundmaeker et al., n. 4, p. 144; Verdonk, n. 15, pp. 118–119; Mihalis Kritikos, ‘Precision Agriculture in Europe – Legal, Social and Ethical Considerations’ (EPRS | European Parliamentary Research Service, 2017), p. 39. 108 See more in Section 3.3.2.1. 109 See more in Sections 3.3.2.1 to 3.3.2.6. 104 105

56

Challenges to assumptions in competition law

3.3.2

Assessing Data-driven Entry Barriers in the Digital Agriculture Sector

3.3.2.1 Switching costs There are significant reasons that make switching from one ATP to another highly difficult or costly for farmers. The question of ‘can farmers take the historical data with them if they move to another supplier?’110 is critical in this regard. As KWS (one of the stakeholders that the Commission had individual calls with) hinted,111 there are two main reasons for data-driven lock-in concerns in the DAs, i.e. legal ambiguity in data rights and technical barriers to transfer data. To understand the reasons for this, it is vital to know first, what the ag-data covers, and what the legal status of its main proportion is. Agricultural data The notion of agricultural data or agronomic data covers different types of data sets that are relevant for the Digital Agriculture services. It can be analyzed from two different perspectives: (1) its collection sources and function, as well as (2) its legal status.112 1.

Types of ag-data according to collection sources – There are three main data types in the DAs according to their sources and function:113 (1) farm data (collected from farms via sensors, machines or directly by farmers for tailor-made agronomic prescriptions); (2) complementary data (such as weather, satellite and other environmental data, including precipitation events, evapotranspiration, and heat unit accumulation);114 and (3) proprietary data (such as data of an agricultural inputs company about its agronomic products, the results of the field tests and other exclusive information).115 In order to provide farmers with Digital Agriculture services,

110 See EIP-AGRI ‘Data Revolution: Emerging New Data Driven Business Models in the Agri-Food Sector’ https://​ec​.europa​.eu/​eip/​agriculture/​sites/​agri​-eip/​files/​eip​ -agri​_seminar​_data​_revolution​_final​_report​_2016​_en​.pdf accessed 5 January 2021, p. 12 [‘EIP-AGRI report’, henceforth]. 111 Bayer/Monsanto, n. 8, para 2842, table 170. However, the Commission did not discuss this issue apart from citing the view of KWS. 112 Although the decision explored the importance of agricultural data sets, it neglected their legal status. 113 Bayer/Monsanto, n. 8, para 2453. 114 Keith Coble et al., Advancing U.S. Agricultural Competitiveness with Big Data and Agricultural Economic Market Information, Analysis, and Research (FARE, 2016), p. 3. See also another type of categorization at the same source. 115 See Bayer/Monsanto, n. 8, para 2453.

The role of agricultural data in the emerging Digital Agriculture sector

57

different combinations of these data sets might be needed according to the type of services.116 2. Applicable legal regime to ag-data – In its decision, the Commission did not discuss the legal status of ag-data or any possible legal regime applicable to it even though this may substantially affect the competitive assessment.117 In particular, it is critical to know whether the collected ag-data can be considered as personal or non-personal. This categorization is essential to determine data control and portability (or existence of any) regime in the DAs. As ‘farm data’ is more related to the mentioned lock-in problem of farmers compared to the other two components of the broader notion of ag-data, it deserves particular attention. The so-called ‘farm data’ is further categorized in the literature, namely (a) process-mediated data;118 (b) machine-generated data;119 as well as (c) human-sourced data,120 and the main proportion of data in the sector is considered as machine-generated non-personal data.121 Additionally, the complementary data and proprietary data might not represent personal characteristics as they are related to the environment and performance of agricultural inputs which might not easily be associated with a natural person. Indeed, paragraph 9 of the Regulation on a framework for the free-flow of non-personal data in the European Union gives ag-data as an example of non-personal data.122

Ibid., paras 2453 and 2724. Regardless of its reason, i.e. whether it was a conscious choice or negligence, it has to be noted that this is a problematic silence because, for instance, determining whether there is an explicit right to data portability for the related data sets might affect the data substitutability assessment, which is also absent in the Commission’s analysis to a large extent, as discussed in detail at below Section 3.3.2.6. 118 Which covers purchasing inputs, feeding, seeding, applying fertilizer, taking an order, etc. See at Wolfert et al., n. 4, p.74. 119 Which is generated by IoT systems such as sensors and smart machines. See ibid. 120 It is based on human experiences, which have been recorded in books and works of art, and later in photographs, audio and video. This data is generally not well structured and stored. Therefore, in Big Data related discussions, human-sourced data is rarely mentioned except marketing part of the agricultural value chain. See at ibid. 121 Ibid. See also Kritikos, n. 107, p. 15; See EU Report, n. 41 above, p. 87. This gives soil data as an example of ‘strictly non-personal’ data; See also Podszun and Kreifels, n. 97, p. 193. 122 ‘…Specific examples of non-personal data include … data on precision farming that can help to monitor and optimise the use of pesticides and water.’ See Regulation (EU) 2018/1807 of the European Parliament and of the Council of 14 November 2018 on a framework for the free flow of non-personal data in the European Union, OJ L 303, 28.11.2018, pp. 59–68 [‘Free flow of non-personal data regulation’, henceforth]. 116 117

58

Challenges to assumptions in competition law

On the other hand, even if these data sets are considered as personal, the applicability of the General Data Protection Regulation (GDPR)123 framework might not be effective enough in the sector. It must be recalled that the only beneficiaries of data protection rules under the GDPR are natural persons.124 Therefore, farms, which are run by legal persons (both small- and medium-sized enterprises (SMEs) and corporations), cannot institutionally benefit from the GDPR framework. Here, one may think that workers of these farms might enforce their rights as natural persons, but this can be highly burdensome and confusing as workers are not a part of the contractual relationship with ATPs, and it might be highly complicated to determine which worker collects which data. Also, enforcement would fail if the related workers leave the job. Moreover, the free-flow of non-personal data regulation does not provide any binding data portability right125 for non-personal data sets. It only suggests the creation of sector-specific voluntary codes of conduct.126 Indeed, the joint EU Code of Conduct on agricultural data sharing was released by a coalition of the EU agri-food associations on 23 April 2018 in Brussels.127 However, although this is a noteworthy endeavor to mitigate legal ambiguities in the sector, it is not a binding regulation.128 It is rather a guideline, which might be voluntarily followed by the stakeholders in the sector. Additionally, the superiority of contractual freedom prevails despite the proposed rules: ‘unless otherwise agreed in the contract, the data originator (farmer) has the right to transmit this data to another data user’.129 The same emphasis of ‘unless stated in the contract’ is repeated when discussing the portability of data.130 Considering farmers’ weaker position vis-à-vis ATPs while they are entering into agreements,131 it

123 Regulation (EU) No. 2016/679 of the European Parliament and of the Council 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 (‘GDPR’) [2016] OJ L 119/1. 124 See Article 1 of the GDPR. 125 Which lets the right holders transfer their data sets from one company to another. 126 Free flow of non-personal data regulation, n. 122, Article 6. 127 See more at ‘EU Code of conduct on agricultural data sharing by contractual agreement’ (Copa and Cogeca et al., 2018) https://​www​.cema​-agri​.org/​ images/​publications/​brochures/​EU​_Code​_of​_conduct​_on​_agricultural​_data​_sharing​ _by​_contractual​_agreement​_update​_2019​.pdf accessed 5 January 2021 [‘EU Code of Conduct’, henceforth]. 128 Ibid., p. 4. 129 Ibid., p. 9. 130 Ibid., pp. 9–10. 131 Sundmaeker et al., n. 4, p. 144; Verdonk, n. 15, pp. 118–119; Kritikos, n. 107, p. 39.

The role of agricultural data in the emerging Digital Agriculture sector

59

can be argued that the EU Code of Conduct does not present a helpful approach for the data lock-in problem, especially from the enforcement and functionality perspective.132 Data-driven lock-ins in the Digital Agriculture sector due to legal ambiguity Connected to the explained legal ambiguity, one of the most prominent discussions in the literature is whether data belongs to farmers, data collectors (if not farmers), ATPs, landowners or even financial lenders.133 This uncertainty makes farmers concerned as they might not have discretion about the data collected from their farms.134 This concern was also mentioned in the EIP-AGRI workshop in 2016 that ambiguity of data ownership and lock-in situations might detrimentally affect the development of the sector.135 Even though some ATPs try to relieve farmers’ concerns by stating ‘yes, you own the data’ in their discourse, their terms and conditions generally contain ultimate rights in favor of ATPs.136 For instance, some ATPs restrict the usage of the collected ‘farm data’ in another field of the same farmer if there is no contractual relationship for this second field.137 Similarly, usage of farm data (collected by, for instance, ATP 1) by another service provider (for instance, ATP 2) is sometimes prohibited by the terms and conditions after the contractual relationship ends with the ATP 1.138 As a more robust and relevant

132 See more detailed discussion at Can Atik and Bertin Martens, ‘Competition Problems and Governance of Non-personal Agricultural Machine Data: Comparing Voluntary Initiatives in the US and EU’ (Joint Research Centre of the European Commission | Digital Economy Working Paper 2020-07-JRC121337, Seville 2020) https://​ec​.europa​.eu/​jrc/​sites/​jrcsh/​files/​jrc121337​.pdf accessed 5 January 2021. 133 Coble et al., n. 114, p. 6; See also Joan K. Archer and Cordero A. Delgadillo, ‘Key Data Ownership, Privacy and Protection Issues and Strategies for the International Precision Agriculture Industry’ (2016) Proceedings of the 13th International Conference on Precision Agriculture, p. 3. 134 Sundmaeker et al., n. 4, p. 144; Leanne Wiseman, Jay Sanderson and Lachlan Robb, ‘Rethinking Ag Data Ownership’ (2018) 15 Farm Policy Journal 71, pp. 71–72. 135 See EIP-AGRI report, n. 110, p. 14. 136 Kritikos, n. 107, p. 17; See also Monty Guild and Tony Danaher, ‘Big Data Comes to the Farm’ (Financial Sense, 2014) https://​www​.financialsense​.com/​ contributors/​ guild/​ big​ -data​ -farm accessed 5 January 2021; Matt McIntosh, ‘Data Ownership Questions – and Why They’re Important’ (Future Farming, 2018) https://​ www​.futurefarming​.com/​Tools​-data/​Articles/​2018/​10/​Data​-ownership​-questions​-​ -and​-why​-theyre​-important​-340743E/​accessed 5 January 2021; Jop Esmeijer et al., ‘Data-driven Innovation in Agriculture: Case Study for the OECD KBC2-Programme’ (2015) TNO report R10154, p. 27. 137 Sykuta, n. 3, pp. 68–69. 138 Ibid.

60

Challenges to assumptions in competition law

example for this study, Monsanto’s FieldView platform139 makes a clear distinction between personal and non-personal data in its terms and conditions.140 For personal data, it provides a separate set of terms and conditions that contains a list of what is considered as personal data.141 Related provisions on non-personal farm data start by saying that farmers and farms own the data, but the rest of the text shows that the data can only be shared with other FieldView users or platform partners of FieldView, not with rivals.142 There is no portability provision outside of the platform. Moreover, the terms and conditions state that both hardware and software are licensed, not sold.143 The hardware includes the FieldView Drive, the hard disk that collects and stores all data.144 Since there is no undisputed and explicit data portability right for farmers, it is difficult to see how farmers could transfer their data to another platform when they desire to do so. Apart from ATPs’ ultimate power on the collected data, there are similar concerns in the agricultural machinery side because agricultural machinery and vehicles also have data collection tools.145 It is stated that multinational machinery giants apply end-user license agreements (EULA) regarding the usage of machinery and data collection procedure (in which farmers have minimal access after collection), and in these agreements, there are particular provisions to let machine producers block the machines if farmers do not abide by the data collection guidelines in the scope of EULA.146 For instance, John Deere goes beyond and claims that they do not sell the tractors. Instead, it claims that farmers only get ‘an implied license for the life of the vehicle to operate the vehicle’ when they pay for the tractor.147 By saying that, Deere can also claim that the data (generated by the tractor) is also owned by Deere based

139 The core concern in the Bayer/Monsanto merger was the concentration of this platform’s capabilities with Bayer’s Xarvio. See Section 3.2.1 above. 140 ‘Climate Fieldview™ Terms of Service’ (Climate.com, 2020) https://​climate​ .com/​fieldview​-terms​-of​-service accessed 5 January 2021. 141 Ibid., para 1.2 – Personal Data. 142 Ibid., para 4, in particular, para 4.3 – Sharing Data with other Climate FieldView Users and Platform Partners. 143 Ibid., para 1.5 – Limited License. 144 Ibid., para 3.1 – Lease of Equipment. 145 EIP-AGRI report, n. 110, p. 10. 146 See, for instance, John Deere’s practices at ‘Vendor Lock-in, DRM, and Crappy EULAs Are Turning America’s Independent Farmers into Tenant Farmers’ (boingboing.net, 2018) https://​boingboing​.net/​2018/​03/​08/​you​-are​-the​-product​-5​.html accessed 5 January 2021. 147 Kyle Wiens, ‘We Can’t Let John Deere Destroy the Very Idea of Ownership’ (Wired.com, 2015) https://​www​.wired​.com/​2015/​04/​dmca​-ownership​-john​-deere/​ accessed 5 January 2021.

The role of agricultural data in the emerging Digital Agriculture sector

61

on the law of property principles.148 Additionally, agricultural vehicles are highly expensive, and inherently constitute very high fixed costs for farmers. This alone has the potential to lock farmers in existing machine providers during the time for paying debts of the agricultural machinery149 and even maybe after the financial burden due to the need for technological support and maintenance. If the data dependency is added on top of that, it is highly difficult to speak about farmers’ freedom of choice. This might speculatively mean that farmers might become only tractor drivers for agricultural machinery producers to collect ag-data in the scope of the ‘Smart Farming’ practices. Technical challenges in interoperability and standardization In addition to the legal uncertainty, there are also technical barriers regarding the adaptability of one infrastructure to another system requirement.150 It is argued that hardware/software systems are sometimes intentionally designed to block moving data to rival systems in order to nudge farmers to buy the whole system from the same company.151 There are very limited data standards as a result of this so-called ‘brand protection’ strategy, and therefore, technical interoperability barriers prevent further innovation and technological advancement due to lack of interconnection of farming equipment and tools.152 Sundmaeker and others assert that this proprietary architecture in technical compatibility of systems and lack of data standards are the core reasons for the absence of widespread interoperability.153 Accordingly, Lianos and Katalevsky argue that lack of interoperability might bring about farming solution clusters where services cannot be changed with other companies’ services, and this can be detrimental for existing and potential competition as well as innovation in the sector.154 To put it another way, interoperability and standards are critical for farmers to ensure their autonomy to choose the best services from the wide range of alternatives for each particular need in the farm instead of depending on one integrated bundle of services from a single company.155 In the scope of the right to the fruits (fructus) element of ownership right. Kritikos, n. 107, p. 7. 150 Sundmaeker et al., n. 4, pp. 142–143; See also at Esmeijer et al., n. 136, pp. 24–25; Copa–Cogeca /European Farmers European Agri-Cooperatives, ‘Main Principles Underpinning the Collection, Use and Exchange of Agricultural Data’, QJ(16)2689:6 – DA/FG/mvs, p. 3 [‘Copa–Cogeca’, henceforth]. 151 Guild and Danaher, n. 136; Kritikos, n. 107, p. 19. 152 Kritikos, n. 107, p. 19. See also at EIP-AGRI report, n. 110, p. 14. 153 Sundmaeker et al., n. 4, p. 143. 154 Lianos and Katalevsky, n. 15, p. 1. 155 To give an example, farmers should be legally and technically able to use their soil data collected through a soil analysis company in order to get seeding prescription services from another company that also needs soil data sets to provide services. This 148 149

62

Challenges to assumptions in competition law

As a general consideration, switching costs in the sector have a distinctive and strong role in creating entry barriers when considering legal and technical reasons together. Furthermore, by taking into account farmers’ weaker position vis-à-vis ATPs in behaving autonomously (they cannot negotiate on or change standard terms and conditions,156 to be more precise) and highly imbalanced contractual provisions regarding the control of collected data,157 already explained high switching costs render the first-mover ATPs ultimate decision-makers on third parties’ access to data regardless of the wills of farmers. This, in turn, could make them the gatekeepers of the DAs beyond locked-in data sets’ significant competitive advantage for the first-movers over their existing rivals and new entrants, who suffer data shortage to compete with the established ATPs. Even though the Commission cited a stakeholder view that touches upon the interoperability challenge in the sector,158 it did not mention the legal ambiguity in applicable rules to non-personal ag-data. More importantly, it did not provide a standalone analysis regarding the switching costs in the Bayer/Monsanto decision. Instead, the Commission primarily focused on network effects and considered the first-mover advantage in the sector as a result of these effects.159 However, switching costs160 are the core reason for the first-mover advantage in the DAs compared to the relatively limited role of network effects161 that need to be considered together when assessing data-driven market power in the DAs. Beyond the explained legal and technical barriers that prevent free data flow, there are additional factors that potentially exacerbate the impacts of data-driven switching costs in the sector. The following sections will provide an in-depth discussion about them.

would also mitigate farmers’ switching costs beyond interoperability of various services. If there are data standards and interoperable infrastructures, then farmers can also change their seeding prescription services company or soil analysis company later on by keeping their previous data sets usable. 156 Kritikos, n. 107, pp. 1 and 39; See also, Ashley Ellixson and Terry Griffin, ‘Farm Data: Ownership and Protections’ (2017) AREC Fact Sheet | FS-1055, p. 7; See at Jeremy de Beer, ‘Ownership of Open Data: Governance Options for Agriculture and Nutrition’ (2016) Global Open Data – GODAN Papers, p. 14. 157 Sykuta, n. 3, pp. 68–69. 158 See Bayer/Monsanto, n. 8, para 2842, table 170. 159 See ibid., paras 2843–2844. 160 Deriving from legal and technical barriers that result in the data-driven lock-in problem in the sector, as discussed above. 161 That will be discussed in detail in Section 3.3.2.9.

The role of agricultural data in the emerging Digital Agriculture sector

63

3.3.2.2 Importance of historical farm data Historical farm data is essential for ATPs not only to develop their algorithms/ tools but also to provide tailor-made competitive services for the targeted farms.162 The Commission mentioned this briefly, but concentrated on the first aspect.163 However, the importance of historical data sets is not limited to training algorithms. Historical farm data sets are also needed to generate effective agronomic predictions based on retroactive patterns.164 Therefore, even the rivals, which have already advanced data processing tools or wide proprietary/ complementary data sets, might not be able to compete with first-mover ATPs, which exclusively control historical farm data, once farmers are locked-in due to these data sets. For this reason, accessing historical data is crucial for new entrants and existing market players, and this particularity of the sector should be taken into account specifically as a factor that further deteriorates the explained data related lock-in concerns in the DAs. In some cases, the so-called velocity of data might also be the critical element when analyzing data’s role on market power whereas in other instances, having historical data sets might be the core asset to give rise to market power. Sometimes both of them may have equal importance. It should be noted that the importance of historical data is valid predominantly for ‘farm data’ and possibly ‘proprietary data’ sets of ATPs while real-time access to accurate complementary data, such as weather data or satellite data, might be more relevant as this component of ag-data inherently requires velocity. In this regard, the data-driven market power analysis should be conducted by separating different data types required for the particular Digital Agriculture service and also identifying the importance of each especially by taking into account the relevance of the velocity of data or historical data sets to generate the competitive services in the relevant market. In the case of Bayer/Monsanto decision, the Commission successfully separated the related data sets in the descriptive section,165 but did not provide the suggested standalone discussion especially regarding the importance of historical farm data sets.

See at Sykuta, n. 3, p. 69; See also Sundmaeker et al., n. 4, p. 143. See Bayer/Monsanto, n. 8, para 2570. See also general statements regarding data accumulation and its importance for improving the tools, i.e. positive feedback loops at paras 2830–2846. 164 Keith H. Coble, et al., ‘Big Data in Agriculture: A Challenge for the Future’ (2018) 40 Applied Economic Perspectives and Policy, pp. 87 and 91. 165 See Bayer/Monsanto, n. 8, para 2453. 162 163

64

Challenges to assumptions in competition law

3.3.2.3 Importance of data combination Interrelated to the statements above, the combination of different types of data can be precious for providing accurate Digital Agriculture services,166 and the Commission successfully explored this sector characteristic and the special need for data combination in the data-driven prescription services.167 Particularly, it linked the possibility of creating a competitive advantage with unique access to proprietary and third-party data sets168 and considered the importance of field testing and proprietary data sets as the necessary element to provide services in the markets for agronomic prescription services.169 Also, locked-in farm data sets of the existing customers have an influential role in entry barriers in the sector, as explained above. It might even be argued that the farm data sets are the most critical part of the agricultural prescriptions because they make the services tailor-made that is the most distinctive characteristic of the DAs. Therefore, the importance of each data component (especially, the role of farm data on the required combination) should be determined and assessed in detail in an optimal market power test in the DAs. 3.3.2.4 Costs of ‘Smart Farming’ and scarcity of alternative data collection sources Adoption of the so-called ‘Smart Farming’ is strictly related to costly equipment and machinery investments in addition to the role of subsidies and tax benefits as financial incentives.170 For this reason, relatively developed farms might be more expected to switch to data-driven farming from traditional farming. For even these farms, there are still doubts whether these benefits can quickly compensate for the required investments171 despite the significant benefits of Digital Agriculture.172 Additionally, when farmers need to expand ‘Smart Farming’ practices towards their second and/or third lands or barns, they have to invest nearly the same amount once again for setting the data

Sundmaeker et al., n. 4, pp. 143–144; Copa–Cogeca, n. 150, pp. 2 and 5. Bayer/Monsanto, n. 8, para 2733. 168 Ibid., para 2728. 169 Ibid., para 2805. 170 Esmeijer et al., n. 136, p. 13; see also ibid., para 2838 for the importance of the high adoption rate in the sector. 171 See for example at Iria Soto et al., ‘The Contribution of Precision Agriculture Technologies to Farm Productivity and the Mitigation of Greenhouse Gas Emissions in the EU’ (Joint Research Center, 2019), pp. 21–22. 172 See the remarkable differences between conventional production and data-driven production in field trials, for instance, in terms of crop yield, soil fertility loss, quality of crops, input consumption percentages at ‘Internet of Food and Farm 2020’ (iof2020eu, 2020) https://​www​.iof2020​.eu/​communication​-materials/​iof2020​-booklet​ -2019​-highres​.pdf accessed 5 January 2021. 166 167

The role of agricultural data in the emerging Digital Agriculture sector

65

collection infrastructure as well as service fees of ATPs as much as the first investment. Therefore, it can be argued that ATPs’ data collection capacity is naturally interconnected with farmers’ intentions to switch to ‘Smart Farming’, which is strictly related to the mentioned fixed and marginal investment costs. In this environment, farmers are at the center of the data collection stage in the fields,173 and Digital Agriculture services are up to ‘farmers’ ability to automatically upload their field data’.174 This distance from fields decreases ATPs’ control on the data collection stage. The mentioned distance between ATPs and data collection in the fields is eliminated to a large extent by IoT devices, which collect data and automatically send it to databases of ATPs, but the help of IoT facilities to eliminate the physical distance sometimes has limitations due to a farm’s distance to required infrastructure. For instance, due to internet connection problems, data transfer from field to ATPs might be interrupted.175 This can be extended to electricity, defect of the sensors, and other essential infrastructural and technical problems. All these challenges have to be overcome by farmers. In this regard, it can be put forward that ATPs’ distance to data collection stage is also related to: (1) farms’ distance to essential infrastructures, and connected to this; (2) their heavy responsibility regarding (a) setting the IoT infrastructure in farms, (b) keeping the data collection tools in working condition, and (c) solving problems regarding the infrastructure. All these costs significantly influence data collection alternatives, and in turn, the value of data accumulation (especially, locked-in farm data sets) in the hands of first-movers in the sector. Therefore, understanding the value of already collected farm data is important for data-driven market power assessment beyond the direct role of costly infrastructures when changing ATPs due to the technical incompatibility of systems, as explained above. 3.3.2.5 Business model and distance to data effect The reason for farmers’ data collection costs is also related to the business model of the DAs. Although the Commission stated in the Bayer/Monsanto decision ‘business models in digital agriculture are not stable’,176 the critical 173 EU Code of Conduct, n. 127; as well as responsible for the quality and credibility of data. See at Copa–Cogeca, n. 150, p. 2. 174 Bayer/Monsanto, n. 8, para 2688. However, the Commission did not focus on this that much. 175 See more at Brian E. Whitacre, Tyler B. Mark, and Terry W. Griffin, ‘How Connected Are Our Farms?’ (2014) 29 (3) The Magazine of Food, Farm, and Resource Issues 1; and Zerina Kapetanovic, ‘Farmbeats: Empowering Farmers with Low-Cost Digital Agriculture Solutions’ https://​www​.microsoft​.com/​en​-us/​research/​video/​ farmbeats​-empowering​-farmers​-with​-low​-cost​-digital​-agriculture​-solutions/​?OCID​=​ msr​_video​_farmbeats​_facsum​_tw​#!related​_info accessed 5 January 2021. 176 Bayer/Monsanto, n. 8, para 2991.

66

Table 3.1

Challenges to assumptions in competition law

Business model of the Digital Agriculture sector

Services for Users

Data Collection

Main Income of the

Primary

Secondary

or Customers

Awareness and

Businesses

Purpose

Purpose of Data

Role of Data in

of Data

Collection

Services

Collection

Data-driven

Data collection

Via subscription

To provide

To train

digitally enabled

is up to farmers’

fees for data-driven

tailor-made

algorithms, fix

agronomic /

explicit intention,

agronomic services.

agronomic

errors, enhance

solutions /

investment and

Business income

services for

services

prescriptions/

practice

mainly depends on

customers

Network effects

observation

Data is the core

the size of acres

(farmers)

enabled by the

services

input for services

and number of

scale and scope of

customersa

datab

Notes: a See Section 3.3.2.8; b See Section 3.3.2.9.

information to identify the business model of the DAs might be that Digital Agriculture services (including agronomic prescription services) are not free of charge,177 and the service fees consist of the primary business income of ATPs.178 Additionally, it can be observed that data is not an externality of service usage. Instead, farm data is the center of the agronomic prescription services as an irreplaceable input in such a way that farmers intentionally set the data collection infrastructures to extract and send data to ATPs in order to get tailor-made data-driven solutions from ATPs by paying subscription fees.179 This chapter argues that there is a novel relationship between farmers and ATPs due to the special value of farm data. On the one hand, there is a distance between farmers and already collected farm data sets due to legal and technical barriers, as discussed above. Farmers do not have any explicit legal or practical position to control data after the data collection stage (farmers’ distance to data). On the other hand, ATPs have limited discretion on farmers’ costly data collection investments, and thus, their data collection capacity is interrelated with farmers’ technological transformation. This cross-relation cannot be easily overcome by ATPs alone. Thus, their distance from data collection stage (ATPs’ distance to data) increases the value of farm data sets for ATPs, which 177 Joddy L. Ferris, ‘Data Privacy and Protection in the Agriculture Industry: Is Federal Regulation Necessary?’ (2017) 18 Minn. J. L. Sci. and Tech. 1, p. 314. 178 See generally at Sundmaeker, n. 4; See Bayer’s business model at Bayer/ Monsanto, n. 8, para 2490 and 2547; See also the FieldView as an example at ‘Pricing’ (FieldView) https://​climate​.com/​pricing accessed 5 January 2021. 179 Notifying Party also stated the ATPs’ dependency on the farm data collection process operated by farmers. See at Bayer/Monsanto, n. 8, para 2688.

The role of agricultural data in the emerging Digital Agriculture sector

67

in turn, exacerbates the effects of the data related lock-in problem from the perspective of entry barriers for new entrants and expansion barriers for existing rivals as a result of the problem of switching costs for farmers. ATPs’ distance to data due to farmers’ centrality in data collection and its costs for farmers might bring about the question of whether this can constitute a significant countervailing bargaining power for farmers. However, this chapter is cautious about this possibility because farmers have equal or even (arguably) more incentive to collect data and receive Digital Agriculture services to reach more efficient and productive farming solutions. In this regard, farmers might not effectively use the data collection argument to bargain with ATPs because farmers demand technology services from ATPs by knowing that they will invest in data collection infrastructure and pay for Digital Agriculture services unless they have large operations with considerable data potential that can particularly draw ATPs’ attention to consider exceptional pre-contractual bargaining processes.180 Moreover, large operations of farmers might not directly affect ATPs’ claim on collected data sets since companies have other options to bargain with farmers such as offering discounts for the given services. Therefore, it can be stated that ATPs’ distance to data does not necessarily increase (especially relatively small) farmers’ bargaining power as they do not have any level playing field to change standard terms and conditions of the Digital Agriculture services. Instead, this makes farm data highly valuable for ATPs and increases the anticompetitive potential of already aggregated data sets in the hands of limited ATPs. Moreover, ATPs’ distance to data does not mean that farmers have control over already collected data sets as already explained above in the scope of farmers’ lack of control after data collection stage, i.e. farmers’ distance to data. To mitigate farmers’ distance to data problem, cooperatives can come to mind as an effective mechanism that may increase farmers’ bargaining power. As already discussed in the case law,181 farmers’ cooperatives can purchase agricultural inputs wholesale in order to benefit from discounts, and they can impose restrictive rules on their members ‘to forbid them to participate in other forms of organized cooperation which are in direct competition with it’.182 Therefore, one can consider whether a similar form of organization could be a solution to ease farmers’ distance to data after the data collection stage by entering into agreements with ATPs as a powerful entity. This could be a promising attempt to increase countervailing buyer power of farmers 180 However, in the scope of this research, no such situation has been encountered in the literature or sector investigation. 181 See at Gøttrup-Klim e.a. Grovvareforeninger v Dansk Landbrugs Grovvareselskab AmbA, ECLI:​EU:​C:​1994:​413. 182 Ibid., para 45.

68

Challenges to assumptions in competition law

vis-à-vis ATPs,183 but this chapter is also skeptical about the possible success of such an endeavor. As the nature of the DAs is based on tailor-made solutions/suggestions/services to individual farms, it might be challenging to organize a collective agreement. Additionally, as explained in the above paragraph, ATPs might offer other alternatives such as reduced prices instead of renouncing their exclusive de facto control on collected data sets. More importantly, this type of collective agreements might constitute another set of switching costs for individual farmers when they desire to change the service providers during the process due to potentially restrictive rules of these cooperations. However, there are initiatives to create intermediary data pooling cooperatives to store collected farm data and open-up these data sets to ATPs according to farmers’ requests later on.184 These initiatives might be useful to mitigate concerns on switching costs, but, still, established first-movers do not have any explicit incentive to let farmers change the existing practice in the sector. This makes the distance to data effect (especially from the farmers’ side) more concerning as the market forces might not be able to correct the failure easily without external intervention. Nonetheless, the potential of the explained pooling intermediary initiatives might still be valid for new ‘digital farmers’, as they can choose ATPs accordingly. 3.3.2.6 Lack of data substitutability The level of data substitutability has been emphasized by various reports as an indicator of data-driven market power, i.e. if competitors cannot easily access the same required data, then substitutability level is low, and there is a higher level of concern and vice versa.185 Therefore, it is stated that data sets of the undertaking(s) under investigation and available substitutable data for their rivals should be taken into account when assessing market power.186 This issue has long been discussed by the Commission in online platforms cases, but with a different perspective, as a relieving factor for data combinations via mergers such as in Google/DoubleClick,187 Telefónica UK/

183 Even though such an initiative has not been encountered in the scope of this research. 184 See, for instance, ‘JoinData – The Future of Smart Farming’ (Join-data.nl, 2019) https://​www​.join​-data​.nl/​en accessed 5 January 2021; see more about the platform at ‘Data Sharing in the Agricultural Sector | Support Centre for Data Sharing’ (Eudatasharing.eu, 2019) https://​eudatasharing​.eu/​examples/​data​-sharing​-agricultural​ -sector accessed 5 January 2021. 185 EU Report, n. 41, p. 49; Franko-German Report, n. 71, pp. 44–47. 186 Ibid. 187 Case No. COMP/M.4731 – Google/ DoubleClick, Commission Decision (11 March 2008), paras 269–272 and 365–366.

The role of agricultural data in the emerging Digital Agriculture sector

69

Vodafone UK/Everything Everywhere/JV,188 Facebook/WhatsApp189 because of the availability of substitutable data for competitors. For instance, it is stated in the Facebook/WhatsApp case that ‘there will continue to be a large amount of Internet user data that are valuable for advertising purposes and that are not within Facebook’s exclusive control’.190 The ‘exclusive control’ emphasis was also placed in the Microsoft/LinkedIn merger.191 It can be observed that the Commission is lenient about the data concentrations unless they are exclusively controlled by the merging entities. Due to the sector-specific conditions in the DAs, the substitutability of data argument, which is repeated in online platforms cases, should be reconsidered cautiously in possible DAs cases. It can be put forward that substitutability level of farm data sets can be very low in the DAs due to inter alia the data-driven lock-ins, high data collection costs and scarcity of data collection alternatives which are encapsulated by the notion of distance to data effect (from both ATPs’ and farmers sides) in this chapter as explained above. Apart from farm data, a substitutability assessment might also be meaningful for proprietary data sets (data about the performance of agricultural inputs) as they are naturally exclusively produced and controlled by input producers. Therefore, there could be a test to assess the substitutability level of each data set under investigation.192 A substitutability assessment is not only related to the same data sets, but also relevant for economies of scope enabled by data. The combination of different data sets (including farm data, proprietary data and complementary data) is one of the prominent characteristics of the DAs.193 In this regard, it can be argued that substitutability of required data variety, i.e. whether providing a competitive service with slightly different combinations might also be pos-

188 Case No. COMP/M.6314 – Telefónica UK/Vodafone UK/Everything Everywhere/ JV, Commission Decision (4 September 2012), para 543. 189 Facebook/WhatsApp, n. 98, paras 188–189. 190 Ibid., para 189. 191 Case No. COMP/M.8124– Microsoft/LinkedIn, Commission Decision (6 December 2016), para 180. 192 Beyond the market power assessment, the Commission could consider imposing data portability remedy for locked-in farm data sets in order to mitigate concerns regarding high entry barriers or data access obligations for proprietary data sets in favor of new entrants or existing rivals, especially when such data is the pre-requisite for doing business in certain markets. 193 Bayer/Monsanto, n. 8, para 2733. As this was determined by the Commission, it can be acknowledged that the Commission considered the issue (importance of data combination) in the decision properly even though it over-focused on proprietary data advantage of integrated giants.

70

Challenges to assumptions in competition law

sible or not, should also be one of the elements in data-driven market power test in the DAs. In the Bayer/Monsanto decision, the Notifying Party claimed that there was no unique access for proprietary or third-party data, but the Commission did not directly respond to this and provided a general statement that merging parties have special capabilities that put them in an advantageous position compared to rivals.194 Although the Commission mentioned the importance of access to ag-data for entering into a market in the DAs195 or discussed the advantage to offer, improve or develop new services with superior access to proprietary and complementary data sets,196 the substitutability levels of related data sets were not discussed at all in the decision. More importantly, the role of farm data on market power was underestimated to a large extent. In this respect, this silence in the Bayer/Monsanto decision regarding the substitutability of data sets is worth further investigation to identify whether it means a diversion from the older precedents based on online platforms cases by realizing the sector conditions, or whether it was entirely ignored in the case by considering the remedy, The BASF Divestment Package197 that would keep Digital Agriculture operations of the merging parties separate anyway. Regardless of the reason for this silence, it should be noted that market power assessments in the sector should not neglect the substitutability levels of related data sets (especially the farm data component) in light of the particular markets’ conditions. 3.3.2.7 Practical rivalrousness of farm data Just like other data types, farm data is non-rivalrous from a technical perspective because multiple players can use identical farm data at the same time. However, it can be argued that the profitability of data collected from farms might be rivalrous in practice due to the novel business model of the DAs. Farm data cannot be used profitably by more than one company at the same time in the same market for agronomic prescription services because income comes from farmers’ subscription fees for the tailor-made agronomic services,198 not data-driven targeted advertisement, for instance. Therefore, replicated data would not be directly profitable by any other company unless the farmer replaces the existing service provider, and starts paying the fees. The only benefit to access farm data from non-contractual farms might be training the algorithms or providing shadow services to compete with existing ATPs by convincing farmers with more accurate and better services than the 196 197 198 194 195

Ibid., para 2728. By referring Monsanto’s considerations on the matter. See at ibid., para 2452. Ibid., paras 2715–2728. Ibid., para 3046 and subsequent paras. See Section 3.3.2.5 above.

The role of agricultural data in the emerging Digital Agriculture sector

71

existing provider. In this regard, due to the practical rivalrousness, farm data might only be indirectly profitable for rivals in the same market, but technically non-rivalrous data can still be valuable to use in connected markets that require the same farm data. There might be an exception here for the non-rival use of ‘farm data’ by rivals at the same time. For instance, there are field information services like AcreValue that process various data including soil type and crop yields, and other relevant historical data to help farmers choose a correct field to buy or rent.199 In this market, farm data can be used by all the rival companies at the same time to provide the best service to customers because of the peculiarity of this particular service. Similarly, replicability and non-rivalrousness discussion might be more meaningful for complementary or proprietary data sets because more than one company can use them to make a profit at the same time in the same market as all the ATPs need to complement these data sets with the farm-specific data of their customers. In this regard, it is important to note that the replicability or non-rivalrousness of data should be discussed carefully in market power analysis of the DAs cases. The Commission should take into account the type of the aggregated data and the particular business model of the relevant market in the broader DAs and should be highly cautious when implementing its precedents from online platforms cases based on the argument of abundance, replicability or non-rivalrousness of data, especially when it comes to farm data. 3.3.2.8 Economies of scale in providing agronomic prescription services The Commission stated in the Bayer/Monsanto decision that the economies of scale is a ‘key element to ensure the competitiveness and viability. Therefore, the development of services for broad acre crops is more attractive’ compared to prescriptions for smaller-scale farming such as fruits and vegetables.200 In this regard, it can be inferred that the data-driven competitive advantages of ATPs might be different in different prescription markets for different crop groupings according to the scale of the farming activities. In the crop groups that are grown in smaller-scale fields, a data-unit might be more important (especially for training algorithms) because of the limited total acres as the potential data collection sources due to the nature of the particular agricultural production. If the related market is dominated by few players, this might result in higher entry barriers for the new entrants who suffer from data shortage to

199 ‘Acrevalue - Granular’ (Granular, 2020) https://​granular​.ag/​acrevalue/​ accessed 5 January 2021; Similar data is also used for scouting services. See Bayer/Monsanto, n. 8, para 2489 (3), and 2505. 200 Ibid., para 2577.

72

Challenges to assumptions in competition law

train algorithms compared to markets that provide Digital Agriculture services for large-scale production that naturally result in more total data to feed data processing tools of market players. Furthermore, this may make comparisons of data aggregation levels between different DAs markets difficult. For instance, it may be misleading to take data aggregation levels in a large-scale market as a benchmark for a small-scale market when evaluating how concerning a data concentration is. In this regard, the scale of the farming activities, its reflection on the data collection scale for agronomic prescription services, and the structure of the particular market under investigation should be explored in detail when assessing the role of data on market power in the DAs cases. For instance, data aggregation in the hands of limited players might be more concerning in the markets which provide services for small-scale agricultural production compared to markets that have more data potential due to the large-scale agricultural production. 3.3.2.9 Positive feedback loops instead of direct network effects As discussed in detail above when explaining the business model and distance to data effect, farmers’ primary objective to implement data collection infrastructure and to take paid tailor-made suggestions/prescriptions from ATPs is more related to making better farming decisions and increasing their productivity rather than reaching other farmers or their data directly. Therefore, an ATP having a vast network of farmers does not necessarily attract other farmers directly. In this sense, it is very difficult to argue that direct network effects are one of the main economic characteristics in the DAs. Consequently, focusing on direct network effects might be misleading in market power assessments in the DAs. In the Bayer/Monsanto merger decision, the Commission focused on data-driven positive feedback loops even though it used ‘network effects’ as a broader notion that typically covers both direct and indirect network effects. The Commission considered that first-movers might have a quality advantage compared to late-comers, and thus, the users would not have any incentive to switch new services.201 It is reasonable to argue that having a large number of farmers (as customers) means having a wide variety of farm data,202 which can be used to develop algorithms and services, and thus, to attract others with high-quality services. In this sense, indirect network effects deriving from the Ibid., para 2837. However, it is important to note that this kind of generalization might not always be valid because sometimes limited customer farms may mean more data and subscription income for ATPs if they conduct large scale farming activities. See the suggestion for an innovative market share calculation for the DAs cases in Section 3.3.1 above. 201 202

The role of agricultural data in the emerging Digital Agriculture sector

73

scale and scope of data advantage are one of the non-negligible factors when assessing entry barriers in the DAs, and the Commission did it in its decision properly. However, in the same paragraph, the Commission considered network effects as the main reason for the first-mover advantage and entry barriers.203 This chapter opposes this consideration by virtue of two main reasons, especially from the perspective of the broader DAs. First, there is a robust first-mover advantage in the sector, but the core reason for this is more related to switching costs, as explained above. Second, presenting consumer accumulation deriving from better services as a competition problem seems a bit awkward because it is a natural consequence rather than explicit and direct harm for consumers even though this may affect the entry barriers after all. Therefore, it can be argued that constructing the theory of harm based on switching costs might be a more sound base in the DAs cases as there is a definite problem if the users are locked in the low-quality services compared to innovative and better alternatives in the market.

3.4

CONSIDERATIONS AS TO THE BAYER/ MONSANTO DECISION AND SUGGESTIONS FOR FUTURE CASES

Building on the in-depth discussion above, the following issues can be stated as prominent regarding the Commission’s data-driven market power assessment in the Bayer/Monsanto decision and should be particularly taken into account in future market power inquiries in the sector. First of all, the Commission did not consider market shares in its assessment as it could not reach reliable data.204 Indeed, focusing solely on market shares might be misleading to estimate the market power in the dynamic and still emerging DAs. Additionally, when considering farmers’ weaker position vis-à-vis ATPs and ‘take it or leave it’ type of terms and conditions, countervailing buyer power might not be a meaningful competitive constraint as well.205 Therefore, the core element to investigate market power in the DAs should be based on entry barriers. However, still, innovative methods to estimate market shares might be helpful to complement entry barriers analysis, such as focusing on the number of acres for which an ATP provides services compared to the total available acres in the market.206 See Bayer/Monsanto, n. 8, para 2837. Ibid., para 2613. 205 See Sections 3.3.1 and 3.3.2.1. 206 See more in Sections 3.3.1 and 3.3.2.8. This can be adapted according to the conditions of the market under investigation. For instance, the Commission can check the 203 204

74

Challenges to assumptions in competition law

Regarding the assessment of the economic characteristics, the Commission over-relied on stakeholders’ ideas via questionnaires and internal documents throughout the text instead of providing in-depth analytical discussions to justify its position. Apart from the method, the Commission indicated network effects as the main reason for the first-mover advantage in the sector.207 However, the core reason for the first-mover advantage mainly derives from the switching costs compared to the relatively limited role of indirect network effects stemming from positive feedback loops.208 Therefore, the Commission’s overemphasis on network effects209 by neglecting to discuss switching costs210 might be problematic, and it should be revised in future investigations.211 As a more general consideration, instead of arbitrarily mentioning various economic characteristics in the analysis, a checklist, which covers all the potential characteristics, can be created to provide an overview to see which factors exist and which are absent, which reasons intensify existing factors, and to explain why the particular characteristics are chosen for market power analysis as the reasoning of the decision. There are definite factors that should be kept in mind as they have the potential to exacerbate the role of switching costs in the sector; • The concentration of farm data can be highly problematic due to the explained switching costs in the DAs, but the Commission underestimated this and focused on proprietary data sets in the Bayer/Monsanto decision.212 It also did not mention the cruciality of historical farm data sets for farmers when changing service providers.213 It can be argued as a general point

number of cows for which an ATP provides data-driven animal health solutions compared to the total cows being traced and served in the same market. 207 See Bayer/Monsanto, n. 8, para 2837. 208 See the overall discussion in Section 3.3.2. 209 See Bayer/Monsanto, n. 8, paras 2843–2844. 210 It did not provide a standalone discussion on the lock-in problem in the sector. 211 Especially, it may affect the remedies and their effectiveness regarding addressing competitive concerns. For instance, if it had identified the switching costs in detail, the Commission could have considered data access or data portability obligations in favor of rivals and farmers, respectively, in addition to the divestment package. 212 See Bayer/Monsanto, n. 8, paras 2715–2728. Maybe the reason for this is related to the fact that Monsanto had not yet started its operations in the EEA, and thus, did not have large farm data sets in Europe at the time of the merger. However, as the main focus of the Commission was on elimination of potential competition, Monsanto’s potential farm data coverage after initiating the EEA operations in full capacity might have been discussed. 213 It only mentioned historic data sets when explaining that data aggregation is key for training/improving tools. Ibid., para 2570. Again, switching costs discussion could have been made by considering the post-operational period of Monsanto in the EEA.

The role of agricultural data in the emerging Digital Agriculture sector

75

that the importance of data combination and historical farm data sets in the context of switching services should be paid attention when assessing entry barriers in the DAs while not neglecting the role of real-time access to complementary data sets (about weather or satellite images). Additionally, market power assessments in the DAs should first categorize the relevant data sets under investigation and determine their potential importance for given services. The Commission separated different data types,214 and mainly explained the importance of proprietary data sets,215 but failed to provide a detailed discussion about distinctive competitive effects of each data set and especially, neglected the farm data analysis in the relevant market. • Adopting Smart Farming is a costly transformation for farmers due to high fixed and incremental costs in addition to farmers’ maintenance responsibilities and subscription fees of digital services.216 Thus, farmers become at the center of the data collection stage. They need tailor-made prescriptions/ suggestions/solutions, and ATPs need farm data not only to make a profit but also to develop their algorithms and services. This cross-relationship requires a novel way of thinking when considering market power in the DAs cases, which was absent in the Bayer Monsanto decision. As this is another factor that can affect switching costs assessments in the DAs, entry barriers investigations should take into account the explained distance to data effect, i.e., whether the level of ATPs’ control on data collection stages is high or low, and to what extent farmers control the already collected data sets (the intensity of data-driven lock-in situations). • Data-driven lock-ins due to legal and technical problems, data collection costs and scarcity of data collection sources influence the substitutability of data in the sector. However, the Commission kept silent on the substitutability of ag-data, unlike its previous decisions in online platforms cases.217 This could mean a diversion from the older precedents by considering the distinctive conditions in the sector, or the Commission might have just neglected this discussion in this case. Regardless of the reason for this silence, it should be noted that the substitutability levels of related ag-data sets under investigation should be one of the essential elements of the market power assessments in the sector.

Ibid., para 2453. Ibid., para 2733; and somehow complementary data sets. See at para 2728. 216 See Sections 3.3.2.4 and 3.3.2.5. 217 The Commission has been lenient about data concentrations in online platforms cases by considering the abundance of available/substitutable data for rivals. See Section 3.3.2.6. 214 215

76

Challenges to assumptions in competition law

• Connected to the substitutability discussion, it is also essential to understand the practical rivalrousness of farm data from the perspective of directly making a profit out of it. Farm data from non-contractual farms might not be used to generate direct profit by ATPs unless they convince the farmers to switch their existing service providers, excluding some exceptional situations in the DAs.218 In this regard, farm data can be practically rivalrous, unlike proprietary and complementary data sets that can be used by various ATPs in the same market in a directly profitable manner. As the Commission did not put emphasis on the business model of the sector and the role of farm data on market power in the Bayer/Monsanto decision, it did not identify this issue as well. However, it can be argued that the suggested substitutability of data analysis should also cover the potential practical rivalrousness of farm data sets in each case in accordance with the particular business model and other features of the relevant market under investigation. • The economies of scale in providing agronomic prescription services are rightfully determined and considered in the Bayer/Monsanto decision, but the relationship between the importance of a data-unit and the risks of dominance in different DAs markets that provide services to different scales of agricultural production is absent in the Commission’s analysis. In this regard, it can be argued that data aggregation might be more concerning in Digital Agriculture markets for small-scale agricultural production where the total scale of data would also be limited, compared to the markets that provide services for large-scale agricultural production where the total data would naturally be more. Therefore, the scale of farming activities, which are served under a particular DAs market, should also be determined during the substitutability/replicability of data assessment as this is relevant for the market power of ATPs under investigation, especially from the perspective of training the algorithms.

3.5 CONCLUSION This study has aimed to investigate the role of agricultural data on market power in the DAs, and, in particular, to critically analyze the Bayer/Monsanto decision in light of the distinctive features of the sector in order to identify indispensable elements of an optimal market power test for the DAs. In the scope of this investigation, it can be concluded that the core focus when assessing market power in the sector should be on entry barriers in light

See more in Section 3.3.2.7.

218

The role of agricultural data in the emerging Digital Agriculture sector

77

of an in-depth economic characteristics analysis.219 The main novelties in this regard derive from very high legal and technical barriers preventing free data flow, the vitality of data combination, the special importance of historical farm data sets, significant data collection costs for farmers, cruciality of data collection for ATPs, and lack of data substitutability, especially for farm data sets.220 These characteristics have a collective potential to significantly give rise to the market power of those who hold data exclusively. Interrelated to this, there is a sector-specific cross-dependency between farmers and ATPs that has been explained as the distance to data effect both from ATPs’and farmers’ sides.221 Due to the distinctive nature of the sector, ATPs have limited control over data collection stages, which makes farm data highly valuable for them. At the same time, farmers lack control after the collection stage due to legal ambiguity about the applicable (or existence of any) legal framework on data rights and interoperability problems as technical barriers to transfer data when farmers decide to change existing service providers. In such an environment, switching from one ATP to another is very costly and sometimes nearly impossible for farmers due to very harsh lock-in situations, and, thus, there is a solid first-mover advantage for ATPs who have already reached a certain amount of users. Although switching costs are the main reason for entry barriers and there are no direct network effects in the sector,222 positive feedback loops as indirect network effects also contribute to these already high entry barriers in the emerging DAs. As the DAs covers a wide variety of data-driven services, each can be considered as different markets. Therefore, the roles of farm data, proprietary data and complementary data on the market power of service providers might vary according to the specific conditions of each particular market. For instance, field biomass measurement services might be predominantly dependent on satellite data223 while irrigation products or spraying prescription services might primarily need climate data and farm data beyond satellite data224 or some services such as farm management products might solely use data pro219 In addition to the helpful potential of the innovative methods such as focusing on the ratio of acres that an ATP provides services for compared to the total acres in a market. See Section 3.3.1. 220 They are mostly neglected in the Bayer/Monsanto decision except for the vitality of data combination. See more in Section 3.3.2 in general and Section 3.4 above. 221 See Section 3.3.2.5. 222 Because the main aim of farmers is receiving tailor-made agronomic services, not reaching other farmers or their data directly. See Sections 3.3.2.5 and 3.3.2.9. 223 See, for instance, Monsanto’s Field Health Imagery product at Bayer/Monsanto, n. 8, para 2505. 224 See, for example, AquaTEK at ibid., para 2508; and Field Manager at ibid., para 2538.

78

Challenges to assumptions in competition law

duced in the target farms.225 In this regard, understanding the particular role of each data set in the related market under investigation is of crucial importance. Therefore, all cases should be investigated specifically according to their facts, conditions, and particular business model of each market in addition to the sector-wide relevant factors discussed above when assessing market power in the DAs cases. Building on the substantial discussion above, it can be put forward that optimal market power test should include the following steps: (1) determination of required data types for the given service in the market under investigation, and their legal status; (2) assessment of each data set’s distinctive importance for generating the related agronomic service/solution; (3) analysis of the substitutability levels of each data set by considering the applicable (or existence of any) legal regime as well as the distance to data effect (from both farmers’ and ATPs’ sides),226 especially for farm data component;227 and (4) a comprehensive discussion about how concerning the data concentration as a result of the merger would be and the reasons for the related concerns in light of an in-depth economic characteristics and business model analysis.228 Thus, the formulation of remedies would become more effective and targeted to the specific reasons for the related concerns in the relevant market. With regards to the suggestions for future research, there is a need for particular attention to switching costs deriving from data lock-ins that create very high entry barriers in the sector. Therefore, future research can analyze the adequacy of existing legal frameworks to address the lock-in problem in the DAs. In such harsh sector conditions, it might even be discussed whether regulatory intervention is necessary beyond the competition law enforcement in order to eradicate at least the legal barriers (preferably mitigate technical barriers as well) to facilitate free data flow with a cost-efficient and fair regulatory framework without hindering data-driven innovation in the sector.

225 For instance, VitalFields provides record-keeping services to its customer farms. See at ibid., para 2509. 226 I.e., assessing the intensity of data-driven lock-in situations and the level of ATPs’ control on data-collection stages, and thus, value of farm data accumulation for ATPs. See Sections 3.3.2.5 and 3.3.2.6. 227 The Commission determined the data types, but neglected the legal status and substitutability levels of related data sets in its analysis. It identified the proprietary data sets’ importance in detail when providing services, but underestimated the distinctive role of the farm data component in the Bayer/Monsanto decision. See the details of these and other criticisms in Section 3.4 above. 228 A checklist can be created for a more systematic analysis.

PART II

New challenges to emerging countries

4. Merger control in China’s digital economy: challenges and prospects1 Wei Han and Yajie Gao 4.1 BACKGROUND By June 2020, the number of Chinese netizens had reached 940 million, with an internet penetration rate of 67.00 per cent, 36.25 million more than that of March 2020. Till the end of the first half of 2020, the turnover obtained in China’s online retailing market was as high as CNY 5.1501 trillion. China has been the biggest online retailing market for seven consecutive years since 2013. With the popularity of mobile internet, the internet economy, having data and algorithm as both the key elements and driving force, has been developing exponentially in China. Until June 2020, the mobile netizens in China had reached 932 million, while as many as 99.20 per cent of all Chinese netizens had obtained access to the internet via mobile phones.2 During the first half of 2020, China’s mobile payment achieved CNY 196.98 trillion, with a year-on-year increase of 18.61 per cent and having topped the global list for three consecutive years. During the fight against COVID-19, Chinese government services platform has introduced the mobile internet-based ‘Health QR Code’, having collected 623 million pieces of information and provided services to 600 million people. It has brought great help to the prevention and control of the pandemic. What is worth mentioning is the new fulcrum to revitalize China’s economy – e-commerce livestreaming. Acting as the most emblematic example of dynamic economies, e-commerce 1 This chapter draws heavily from a previous article: Wei Han and Yajie Gao, ‘Challenges and Prospects for Merger Control in China in the Digital Economy’ (2019) March Competition Policy International; and is updated on the basis of the latest development of China’s anti-monopoly legislation and enforcement since the first publication of the article. 2 Office of the Cyber Security and Information Committee of the CPC Central Committee and Cyber Space Administration of China, The 46th China Statistical Report on Internet Development (2020) 1–5, available at http://​www​.gov​.cn/​xinwen/​2020​-09/​ 29/​5548176/​files/​1c​6b4a2ae06c​4ffc8bccb4​9da353495e​.pdf accessed 3 January 2021.

80

Merger control in China’s digital economy

81

livestreaming has become one of the fastest growing internet applications, which has made great contribution to the alleviation of poverty. Statistics have it that by June 2020, China’s e-commerce livestreaming had attracted 309 million users.3 The official annual report published by the Anti-Monopoly Bureau (‘AMB’) of the State Administration for Market Regulation (‘SAMR’) has it that, in 2019, for all the 103 initiated anti-monopoly investigations, 46 cases have been closed, with a total penalty and confiscation of illegal gains of CNY 320 million. The AMB investigated 36 mergers which should have been notified and made 18 administrative punishment decisions accordingly. Five hundred and three mergers were notified to the AMB, 462 of which were registered and 465 of which were closed.4 Vibrant mergers and acquisitions (‘M&A’)5 have become one of the most salient features of China’s digital economy.6 M&As between leading undertakings7 from certain sectors have sprung up constantly, naming mergers between Ctrip and Qunar, two top online traveling companies in 2015, Didi Chuxing’s (‘Didi’) acquisition of Uber China (‘Uber’) in 2016, Alibaba’s acquisition of Damai in 2017, and Meituan’s acquisition of Mobike in 2018 as classic examples. What is worth mentioning is that only limited digital M&As have been under anti-monopoly review of the Chinese competition authority8 during more than the last ten years since the Anti-Monopoly Law of Ibid. Anti-Monopoly Bureau (‘AMB’) of the State Administration for Market Regulation (SAMR), The Annual Report of China’s Anti-Monopoly Enforcement (2019) 2–3, available at http://​www​.gov​.cn/​xinwen/​2020​-12/​25/​5573435/​files/​19​ 5171fdee02​4615933c10​d57f141171​.pdf accessed 4 January 2020. 5 Please note that ‘M&A’ and ‘concentration’ are used interchangeably all through this chapter. 6 In accordance with the 2019 M&A Report released by China Renaissance on 27 February 2020, available at https://​mp​.weixin​.qq​.com/​s/​OypVz5uj51​_zVnJ2yCU4Ig accessed 28 February 2020, big Chinese internet companies were relatively cautious when it came to M&As in 2019, while those in telecommunication, media and technology industries were still more active than others. Science and technologies, finance, culture, entertainment, content, e-commerce and retailing were among the focus of big Chinese internet companies. Taking Tencent and Alibaba as typical examples, besides science and technologies, both also put emphasis on the finance industry. Furthermore, Tencent put more sources in culture and entertainment, while Alibaba preferred logistics. 7 Please note that, in China, ‘undertaking’ is less commonly used than ‘business operator’ if translated literally and directly from Chinese to English. Broadly speaking, both of them refer to the subject of competition law or anti-trust law. 8 Please note that the competence to review concentrations which lead or may lead to elimination or restriction of competition of the Ministry of Commerce (‘MOFCOM’) has been transferred to the new agency – the State Administration for 3 4

82

Challenges to assumptions in competition law

the PRC (‘AML’) came into effect,9 because of the turnover-based notification threshold, structural arrangement based on the variable interest entity (‘VIE’)10 and for other reasons. Relevant cases include the conditional clearances of Walmart’s Acquisition of 33.60 per cent shares in Newheight Holding, which concerned online direct sales business in 2012,11 a merger between Dow Chemical Company (‘Dow’) and E.I. Du Pont De Nemours and Company (‘Du Pont’) in 2017,12 and Bayer Aktiengesellschaft, Kwa Investment Co.’s (‘Bayer’) acquisition of Monsanto Company (‘Monsanto’), which involved the digital agricultural market, in 2018.13 Only few concentrations have officially been reviewed by the Ministry of Commerce (‘MOFCOM’), the previous agency in charge of merger control in China, although much more influential M&As in the digital economy have triggered widespread doubt as regards their compliance with anti-monopoly law, such as Didi’s acquisition of Uber. Ever since the institutional reform of the Chinese government in 2018, merger control is no longer within the competence of MOFCOM, while the ‘AMB’ affiliated with the SAMR has become Market Regulation (‘SAMR’). For the following, no matter it is the main text or footnotes, MOFCOM and SAMR would be referred to when necessary. In other words, the explicit reference to MOFCOM or SAMR corresponds with the cases it has dealt with in practice. 9 In comparison, as for the latest development, the five cases conditionally cleared by the Chinese competition authority in 2019 cover industries of semiconductor, shipping, medicine, automobile and others. Zhenguo Wu, ‘Devote to Fair Competition and Serve for Reform and Development – General Description of the Anti-Monopoly Work in 2019’ [2020] available at http://​www​.cicn​.com​.cn/​zggsb/​2020​ -02/​20/​cms123961article​.shtml accessed 28 February 2020. 10 Due to lack of explicit legislation, administrative decision and judicial judgment with respect to the legality of variable interest entity (‘VIE’) in China, the writers of this chapter try to define it only for the understanding of what obstacles the Chinese competition authority has been confronted with in practice when trying to obtain competence to review concentrations. VIE is also called ‘protocol control’, through which foreign investors obtain controlling power of domestic undertakings via a series of agreements instead of through equity investment, for the reasons of preferable treatment to foreign investment, circumvention of domestic control over certain industries (such as internet and relevant services, as well as telecommunications) not totally open to foreign investors, and convenience to raise money overseas in the past. 11 MOFCOM, Decision of 14 August 2012, Public Announcement No. 49 [2012], available at http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201303/​20130300058730​.shtml accessed 28 February 2020. 12 MOFCOM, Decision of 2 May 2017, Public Announcement No. 25 [2017], available at http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201705/​20170502568075​.shtml accessed 28 February 2020. 13 MOFCOM, Decision of 13 March 2018, Public Announcement No. 31 [2018], available at http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201803/​20180302719123​.shtml accessed 28 February 2020.

Merger control in China’s digital economy

83

the only central department to enforce anti-monopoly laws. Considering merger control plays an essential role, what attitudes the SAMR poses towards digital concentrations is of great interest both at home and abroad. This chapter aims to present relevant challenges and to predict the future of merger control in China’s digital economy. Section 4.2 discusses notification thresholds. Special issues in the competition assessment of digital economy are analyzed in Section 4.3. Section 4.4 focuses on remedies, while prediction of the future development and main issues to be solved are presented in Section 4.5. Section 4.6 concludes.

4.2

NOTIFICATION OF CONCENTRATIONS

4.2.1

The VIE Structure

Though the advancement of digital economy has facilitated digital transformation of almost every industry, the most vibrant market participants are still internet companies who started their businesses on the basis of the internet from the very beginning. A large percentage of Chinese internet companies have a connection with VIE, which is one of the prominent reasons why barely any concentration in the internet industry has been reviewed by MOFCOM in more than the last decade. Since the Chinese authorities’ attitude towards VIE used to be unclear,14 MOFCOM was concerned that its anti-monopoly review of concentrations involving VIE structure per se might be understood by the

14 For example, see MOFCOM, Notes on the Foreign Investment Law of the PRC (Draft for Comments) (2015), 3(3) Handling of Protocol Control, ‘Since the Foreign Investment Law will have come into effect, it applies to (foreign) investment to be made in the form of protocol control. As for (foreign) investment already made in the form of protocol control before the Foreign Investment Law comes into effect, if it would still concern industries not totally open to foreign investment, three proposals have been put forward by both theoretical and practical circles as regards how to regulate them: For foreign-invested enterprise implementing protocol control, (a) if it declares to the competent foreign investment department of the State Council that it is de facto controlled by a Chinese investor, the said foreign-invested enterprise may continue to retain the structure of protocol control, and relevant parties may continue with their business activities; (b) it shall apply to the competent foreign investment department of the State Council for recognition of being de facto controlled by a Chinese investor; after the competent department has recognized that the foreign-invested enterprise is de facto controlled by a Chinese investor, it may continue to retain the structure of protocol control, and relevant parties may continue with their business activities; and (c) it shall apply to the competent foreign investment department of the State Council for market entry license, while the competent department shall, in conjunction with other departments concerned, make a decision by taking into comprehensive consideration of the de facto controller of the foreign-invested enterprise and other factors’. Available at http://​

84

Challenges to assumptions in competition law

public as recognizing the legality of VIE indirectly. As a result, MOFCOM declined to accept any notification of concentrations, especially those involving internet companies having a connection with the VIE structure. It partly explains why most of the internet concentrations have not been notified to MOFCOM, in accordance with official sources. The three punishment decisions released by the SAMR on 14 December 2020 have cleared the uncertainty. More specifically, Alibaba Investment, China Literature (whose de facto controller is Tencent Holding) and Hive Box (affiliated company of SF Holding) were all fined CNY 500,000, the maximum of administrative penalty prescribed by the AML,15 for not notifying their mergers, even if the SAMR held that each of the mergers would not impede effective competition.16 In the following press conference, the SAMR made it clear that the AML applies to every market player, including the internet industry, even if it has some new attributes. The SAMR introduced that it was the first time to impose administrative punishment on illegal implementation of mergers concerning the VIE structure. All the three cases are connected with VIE: either the undertaking under investigation was established on the basis of VIE or the target controls domestic operating entity through VIE. The SAMR made it crystal clear that the AML applies to mergers concerning VIE. In other words, VIE is no longer the excuse for internet companies to circumvent the notification obligation under Chinese anti-monopoly law.17 4.2.2

Notification Threshold

4.2.2.1 Limitation of the turnover-based notification threshold The notification threshold of merger control in China is based on turnover, which might not be a proper benchmark due to special attributes of business models in the digital economy. For example, as for the multi-sided market, the

tfs​.mofcom​.gov​.cn/​article/​as/​201501/​20150100871010​.shtml accessed 28 February 2020. 15 Standing Committee (‘SC’) of the National People’s Congress (‘NPC’), ‘Anti-Monopoly Law of the PRC’ [2008] Article 48, available at http://​www​.gov​.cn/​ flfg/​2007​-08/​30/​content​_732591​.htm accessed 4 January 2021. 16 SAMR, ‘Administrative Punishment Cases’, available at http://​www​.samr​.gov​ .cn/​fldj/​tzgg/​xzcf/​ accessed 3 January 2021. 17 SAMR, ‘Person in Charge from the AMB of the SAMR Answers Questions Raised by Journalists Concerning the Administrative Punishment of Three Mergers Which Should Have Been Notified: Alibaba Investment’s Acquisition of Yintai Business, China Literature’s (whose de facto controller is Tencent Holding) Acquisition of New Classics Media and Hive Box’s (affiliated company of SF Holding) Acquisition of Zhongyou Zhidi S&T’ (2020), available at http://​www​.samr​.gov​.cn/​xw/​zj/​202012/​ t20201214​_324336​.html accessed 3 January 2021.

Merger control in China’s digital economy

85

model widely used in the digital economy, usually services provided by one side of the platform are free of monetary payment, which not only makes it more complicated to define the relevant market, but also triggers challenges for application of the turnover-based notification threshold. In fact, it is quite normal for digital undertakings to keep in deficit at the starting stage. Nevertheless, they could still impose non-negligible effects on the market. Just as Didi’s acquisition of Uber in 2016 shows, and as Didi, in response to public doubt, stated: ‘Until now, neither Didi nor Uber has made any profits. Turnover obtained by Uber in China in the last financial year did not meet the notification threshold.’18 In general, the turnover-based notification threshold could not effectively screen out the potentially problematic concentrations in the digital economy. 4.2.2.2 Introduction of the transaction-value-based threshold The transaction-value-based notification threshold has been introduced in certain jurisdictions, such as Germany and Austria,19 as well as South Korea.20 As for China, in accordance with the currently valid laws and regulations, the SAMR could only obtain the competence to review concentrations, the aggregate turnover of whose undertakings concerned have not met the turnover-based threshold, through residual jurisdiction in accordance with Article 4 of the Provisions of the State Council on the Threshold for Notifying Concentration of Undertakings21 (‘Provisions on Threshold for Notification’) promulgated in 2008. What is worth mentioning, on 2 January 2020, the draft amendment to the AML was released by the SAMR (‘SAMR Draft

18 21Jingji, ‘Didi Responded to Questions Raised by the MOFCOM with Regard to the Notification Obligation: Notification Threshold Has Not Been Met’ (2016), available at http://​m​.21jingji​.com/​article/​20160802/​herald/​29​df68176ebf​184049383a​ 973596a939​_zaker​.html accessed 28 February 2020. 19 Bundeskartellamt and Bundeswettbewerbsbehörde, Guidance on Transaction Value Thresholds for Mandatory Pre-Merger Notification (Section 35 (1a) GWB and Section 9 (4) KartG) (2018), available at https://​ www​ .bundeskartellamt​ .de/​ SharedDocs/​Publikation/​EN/​Leitfaden/​Leitfaden​_​Transaktio​nsschwelle​.pdf​?​_​_blob​=​ publicationFile​&​v​=​2 accessed 28 February 2020. 20 Competition Policy Bureau of Korea Fair Trade Commission, The complete revision of the Monopoly Regulation and Fair Trade Act for the first time in 38 years (2018), available at http://​www​.ftc​.go​.kr/​solution/​skin/​doc​.html​?fn​=​d683​557915e470​ 4e9db250e2​a6ae49f66d​eea3658917​3a4539a300​a322d23592​&​rs​=/​fileupload/​data/​ result/​BBSMSTR​_000000002402/​accessed 28 February 2020. 21 State Council, ‘Provisions of the State Council on the Threshold for Notifying Concentration of Undertakings’ [2008], available at http://​fldj​.mofcom​.gov​.cn/​article/​ c/​200811/​20081105917434​.shtml accessed 28 February 2020.

86

Challenges to assumptions in competition law

Amendment to the AML’) to solicit public opinions.22 Article 24 of the SAMR Draft Amendment to the AML authorizes the SAMR to set and/or adapt the notification threshold according to the economic development level and industry scope. If finally approved by the Standing Committee (‘SC’) of the NPC, introduction by the SAMR of the transaction-value-based notification threshold would be expected. Section 4.3 below lists typical issues in the substantive competition analysis of concentrations in the digital economy, covering data, privacy, innovation and leveraging effect.

4.3

SUBSTANTIVE COMPETITION ASSESSMENT

4.3.1

Openness of Data

4.3.1.1 Data disputes With the prevalence of smart phone and the concept of ‘big data’, both hardware terminals and software companies have paid increasing attention to the value which might be brought by data. Uncertainty still exists as regards the ownership of user data, such as the dispute between Huawei Technologies Co. (‘Huawei’) and Tencent Holdings Ltd. (‘Tencent’). Honor Magic, the first concept mobile phone released by Huawei Honor, could provide its users with intelligent recommendations through making use of the WeChat chat content. Tencent held that Huawei had captured its own data and infringed WeChat user privacy. Huawei responded that it was the users who were entitled to the data, neither WeChat nor Honor Magic was. Before processing user data on the Honor Magic devices, Huawei had obtained authorization from users. The dispute ended with the settlement by the Ministry of Industry and Information Technology (‘MIIT’).23

22 SAMR, ‘The SAMR Solicits Public Opinions on the Draft Amendment to the Anti-Monopoly Law of the PRC’ (2020), available at http://​www​.samr​.gov​.cn/​hd/​ zjdc/​ 202001/​ t20200102​ _310120​ .html accessed 28 February 2020. Afterwards, the SAMR’s draft still has to be submitted to the Department of Justice and the Standing Commission of the NCP for further deliberation and final approval. 23 Xinjie Yang, ‘The Ministry of Industry and Information Technology (“MIIT”) Responds to Data Dispute between Huawei and Tencent: Investigation Undergoing While Collection of Data by Relevant Enterprises Urged to Abide by Laws and Regulations’ (thepaper.cn, 8 August 2017), ‘As regards the dispute between Tencent and Huawei with respect to the newly introduced function, in order to protect user personal information, the MIIT will abide by the “Provisions on Telecommunications and Protection of Internet User Personal Information” and other laws and regulations so as to urge enterprises to strengthen internal management, self-regulate collection and use of user personal information and protect legitimate rights and interests of users. As for disagreement and dispute between information and communications enterprises, the

Merger control in China’s digital economy

87

Fight for data is much more common between internet companies in China, and the People’s Courts have also resolved certain questions, even if none of the relevant disputes has been resolved in accordance with the AML. Sina Microblog alleged that Maimai, a Chinese real-name business social platform, had breached the Developer Agreement and its normal permission, under which Maimai could only capture name, gender, avatar and email address of the users. In practice, it was alleged that Maimai had also illegally obtained educational and occupational information, telephone numbers as well as other information, access to which could only be obtained under advanced permission. Besides, Maimai was also believed to have matched the telephone number with Sina Microblog avatar. Sina Microblog held that Maimai’s conduct constituted unfair competition. Both the Beijing Haidian District People’s Court and the Beijing Intellectual Property Court (‘BJ IP Court’) sustained Sina Microblog’s observations and determined that Maimai’s conduct constituted unfair competition. The BJ IP Court observed that an undertaking had property right over the data it collected and accumulated in the sense of anti-unfair competition law, if it had invested efforts and resources in this regard. In other words, user data could be made use of, licensed and made profits from, just like other assets. Nevertheless, exercise of property right of data shall be restricted by the personal right of the data subject. The court also put forward a three-step license principle, namely authorization by users + authorization by the platform + authorization by users. Furthermore, personal data could only be collected legitimately, reasonably and necessarily. Whether relevant undertakings’ conduct harms consumer rights and interests is an element which shall be taken into account when determining whether it constitutes infringement.24 Other similar data disputes include those between Dianping and Baidu for the illegal migration of the addresses of, and comments on, a certain restaurant from Dianping to Baidu Map. The Shanghai Intellectual Property Court (‘SH IP Court’) recognized the positive effects brought by the innovative business model introduced by Baidu, such as upgrading consumer experience and enriching consumer choices to some extent. Nevertheless, the SH IP Court also observed that it had broken the bottom line, which would not only harm the legitimate rights and interests of Dianping but might also destroy other undertakings’ incentive to invest in information collection, thus destroying MIIT would proactively coordinate and guide industrial self-regulation so as to create sound market order for mass entrepreneurship and innovation’, available at https://​ www​.thepaper​.cn/​newsDetail​_forward​_1756038 accessed 28 February 2020. 24 (2016) JING 73 MIN ZHONG No. 588, available at http://​ wenshu​ .court​ .gov​.cn/​website/​wenshu/​181107ANFZ0BXSK4/​index​.html​?docId​=​49854fde619a4 7d7b772a71d000fcf00 accessed 28 February 2020.

88

Challenges to assumptions in competition law

the normal industrial ecosystem, having side effects on competition order, and further damaging consumer rights and interests in the end. It constituted unfair competition.25 More recently, Tencent sued TikTok26/Duoshan27 in Tianjin Binhai New District People’s Court and applied for a behavioral injunction in February 2019. Tencent claimed that it held legitimate rights and interests over user avatars and nicknames, while Duoshan had illegally obtained WeChat/QQ avatars and nicknames of its users through TikTok, and requested the court to order Duoshan to ‘stop using WeChat/QQ users’ avatars and nicknames in its products and immediately delete all the WeChat/QQ data retained in its terminal server’. In March 2019, the court issued an injunction verdict, sustaining Tencent’s observation.28 What Tencent fights for may not necessarily be the rights and interests of users’ avatars and nicknames themselves but being concerned that the WeChat friend list would be copied and pasted to TikTok/ Duoshan at large scale.29, 30 4.3.1.2 Concentration for more data Operation of most online platforms in China are driven by data. Since data, especially personal information, play an increasingly essential role in business, a large percentage of digital concentrations involve or are even undertaken for the reason of data integration. As for the two concentrations which have gained widespread attention all over China, Alibaba’s acquisition of Eleme

25 (2016) HU 73 MIN ZHONG No. 242, available at http://​ wenshu​ .court​ .gov​ .cn/ ​ l ist/ ​ l ist/ ​ ? sorttype​ = ​ 1 ​ & ​ n umber​ = ​ L F8LZ7SM​ & ​ g uid ​ = ​ 4 e4aa9b1 ​ - bcae ​ - 1ee254ef​ -8de5b561752a​&​conditions​=​searchWord+​(2016)​%E6​%B2​%AA73​%E6​%B0​%91​ %E7​%BB​%88242​%E5​%8F​%B7+​AH+​+​%E6​%A1​%88​%E5​%8F​%B7:​(2016)​%E6​ %B2​%AA73​%E6​%B0​%91​%E7​%BB​%88242​%E5​%8F​%B7 accessed 28 February 2020. 26 TikTok, also known as Douyin (抖音) in Chinese, is a media application through which users could create and share short videos. 27 Duoshan (多闪) is a short video social application released by TikTok on 15 January 2019. 28 (2019) JIN 0116 MIN CHU NO. 2091, available at http://​wenshu​.court​.gov​ .cn/​ w ebsite/​ w enshu/​ 1 81107ANFZ0BXSK4/​ i ndex​ . html​ ? docId​ = ​ 1 bc6e2edab324 8b09030aa470163d9e7 accessed 28 February 2020. 29 Weiwei, ‘The Fight between Toutiao and Tencent Continues, Beijing Business Today’ (epaper, 21 March 2019), available at http://​epaper​.bjbusiness​.com​.cn/​site1/​ bjsb/​html/​2019​-03/​21/​content​_422998​.htm​?div​=​-1 accessed 28 February 2020. 30 Another influential dispute is that between SF Express and Cainiao. Both parties complained that the other party had tried to collect data concerning user privacy more than necessary. After two days since the incident broke out, both parties’ data channels were closed. Until late 2 June 2017, the State Post Bureau of the PRC settled the dispute.

Merger control in China’s digital economy

89

and Meituan’s acquisition of Mobike in 2018, data integration might probably be what the acquirers were looking for. Competition agencies from the main jurisdictions globally pay close attention to data, which we believe that the SAMR would also take into account in its future enforcement, including determining whether the data could trigger input foreclosure.31 In data-driven concentrations, huge controversies still exist as regards whether a certain data set held by one undertaking could be recognized as essential input, and whether competitors could obtain similar data from other sources. Besides, as for the request to open access to data, special attention shall be paid to the potential chilling effect data openness might have on innovation and the potential side effect on privacy. 4.3.2 Privacy Theoretically speaking, privacy could be recognized as a non-price parameter of competition. A firm with certain market power may not offer its customers a better deal in terms of privacy if it faces no pressure from competitors. Or the firms in a concentrated market might tacitly coordinate to avoid competition on privacy.32 For a concentration, if one of the undertakings concerned has strong marker power, data privacy might concern competition. If two horizontal market players compete in privacy as an aspect of product quality, their merger could be expected to reduce quality.33 4.3.2.1 Widespread privacy infringement calls for sound regulation In November 2016, The Internet Rule of Law Research Centre affiliated with the China Youth University of Political Studies, together with the Cover Institute released the Report on Chinese Individual Information Security and Privacy Protection (‘Report’) on the basis of millions of questionnaires in

31 MOFCOM, ‘Interim Provisions on Assessment of the Impact of Concentrations on Competition’ [2011], Article 7, ‘A concentration is likely to raise the entry barriers for the related market. An undertaking may, by making use of the market control gained or enhanced through concentration, make it more difficult for other undertakings to enter the related market by controlling production factors, sales channels, technological strength, and essential facilities, etc’, available at http://​fldj​.mofcom​.gov​.cn/​article/​ c/​201109/​20110907723357​.shtml accessed 28 February 2020. 32 Organization for Economic Co-operation and Development (OECD), Considering Non-price Effects in Merger Control – Background Note by the Secretariat (2018) 34–35, available at https://​one​.oecd​.org/​document/​DAF/​COMP(2018)2/​en/​pdf accessed 28 February 2020. 33 Autorité de la concurrence and Bundeskartellamt, Competition Law and Data (2016) 24, available at http://​www​.aut​oritedelac​oncurrence​.fr/​doc/​re​portcompet​ itionlawan​ddatafinal​.pdf accessed 28 February 2020.

90

Challenges to assumptions in competition law

China for the first time. The Report had it that more than 70 per cent of the participants acknowledged the seriousness of information leakage; 26 per cent had received at least two to three spam short messages every day; 20 per cent had received at least two to three harassing calls per day; as many as 81 per cent had the experience that the other side of the unknown call had knowledge of his/her own name, employer and other personal information; 53 per cent complained that they had been harassed continuously by certain advertisers only for the reason that they had visited web pages, or personal information left behind; the percentage for unwelcome advertisement or fraud due to leakage of information on house rent, house purchase, car insurance, entrance of higher level of education and others was as high as 36 per cent.34 With the rapid development of mobile internet, a large percentage of mobile applications (‘App’) have not only brought huge convenience but also triggered privacy concern. As regards financial App, for example, in accordance with the (Report on) 2019 Mobile Financial App Privacy Policy Transparency Test, transparency of almost 80 per cent of the financial Apps’ privacy policy failed the test, while 60 per cent of the Apps did not disclose rules on collection of sensitive personal information to consumers. Mandatory and/or excessive collection of personal information was really prevalent. What is worse was that some Apps even send provident fund account and passwords of the customers to their business servers without encryption, which had exposed sensitive personal information to potential cyberattack.35 4.3.2.2 Relevant laws and regulations have been consecutively promulgated to better protect personal information and privacy In China nowadays, effective provisions in relation to personal information can be found in various branches of law, such as Criminal Law of the PRC, Civil Code of the PRC (‘Civil Code’), Consumer Rights and Interests Protection Law of the PRC (‘Consumer Rights and Interests Protection Law’), Cyber Security Law of the PRC (‘Cyber Security Law’), and E-Commerce Law of the PRC (‘E-Commerce Law’). The draft of Personal Information Protection Law (‘PIPL’) has been submitted to the 22th Meeting of the 13th SC of the NPC in

34 Central School of Communist Youth League of China, ‘China Youth University of Political Studies Releases the First Report on Chinese Individual Information Security and Privacy Protection in China’ (23 November 2016), available at http://​news​ .cyu​.edu​.cn/​xyyw/​hzjl/​201611/​t20161123​_78640​.html accessed 28 February 2020. 35 Lin Jiang, ‘Test on Privacy Policy of One Hundred Financial Apps: 60 per cent Do Not Disclose Rules of Sensitive Personal Information Collection’ (m.mp.oeeee. com, 15 March 2019), available at https://​m​.mp​.oeeee​.com/​a/​BAAF​RD00002019​ 0315144629​.html​?layer​=​3​&​share​=​chat​&​isndappinstalled​=​0 accessed 28 February 2020.

Merger control in China’s digital economy

91

October 2020 for deliberation. If enacted by the SC of the NPC, PIPL would become the first systematic Chinese law to protect personal information.36 What is worth mentioning is the preliminary foundation for the establishment of personal information protection policy laid down by the Cyber Security Law, which came into effect as of 1 June 2017. Acting as the basic law for management of China’s cyber security, the Cyber Security Law has systematically established a series of rules and requirements, having protection of personal information as one of the aspects to which it attaches great importance. For the first time, definition of ‘personal information’ is provided at the level of law,37 while the basic rules to protect personal information are confirmed systematically by law.38 For the first time, exceptions for prohibition of providing personal information to third parties have been explicitly provided.39 For the first time, administrative obligations for breach of personal information protection rules have been established by law.40 In order to facilitate implementation of relevant policies prescribed by the Cyber Security Law, a series of supporting regulations have been enacted by the Cyberspace Administration of China (‘CAC’) to provide more detailed guidance, such as the Security Assessment Measures on Migration of Personal Information and Essential Data to Overseas41 and the Security Assessment Measures of Cyber Product and Service.42 Furthermore, the Information Security Technology – Personal Information Security Specification (‘Personal Information Security Specification’), the first recommended national standard in China to protect personal information, has also been formulated by 33 experts who are not only entitled to promulgate policies and technical standards, but also have rich practical experience.43 36 NPC, ‘Draft of Personal Information Protection Law First Comes to the Public – Concentration on Prominent Issues Concerning Protection of Personal Information and Implementation of the Obligation to Protect Personal Information’ (2020), available at http://​www​.npc​.gov​.cn/​npc/​c30834/​202010/​56​9490b5b76a​49c292e64c​416da8c994​ .shtml accessed 3 January 2021. 37 The law here refers to the legislation at the supreme level of the Chinese legislative system. For more details, please find ‘Chapter II Laws’ of the Legislation Law of the PRC. 38 Articles 76(5), 22, 37, 41, 42, 43, 44 and 45 of the Cyber Security Law of the PRC (‘Cyber Security Law’). 39 Article 42 of the Cyber Security Law. 40 Chapter 6 Legal Obligations of the Cyber Security Law. 41 Available at http://​www​.gov​.cn/​xinwen/​2019​-06/​13/​content​_5399812​.htm accessed 28 February 2020. 42 Available at http://​www​.cac​.gov​.cn/​2017​-02/​04/​c​_1120407082​.htm accessed 28 February 2020. 43 Available at https://​isecurity​.neusoft​.com/​data/​download/​GB​-T35273​-2017​.pdf accessed 28 February 2020.

92

Challenges to assumptions in competition law

The Personal Information Security Specification has a set rules for collecting, saving, applying and other links of personal information management, having filled the gap of practical standards in relation to domestic protection of personal information.44 Due to differences in background and standpoints of members of the drafting group, the Personal Information Security Specification is the result of controversies and compromises between academics and enterprise experts. 4.3.2.3 Administrative authorities, people’s courts and enterprises have taken actions In recent years, forced license, over-authorization and excessive collection of personal information by mobile Apps have triggered public complaint. In order to solve the serious problem, the CAC together with the MIIT, the Ministry of Public Security and the SAMR, initiated special rectification work in 2019, so as to maintain the legitimate rights and interests of Chinese netizens and respond to the requirements put forward by the Cyber Security Law and Consumer Rights and Interests Protection Law.45 In the end of 2019, Measures for Recognizing Illegal Collection and Utilization of Personal Information by Apps (‘Measures’) were released,46 to provide guidance for the authorities’ enforcement, the App operators’ self-correction and the netizens’ social supervision. The Measures provide detailed rules on how to recognize ‘not disclosing rules on collection and utilization’, ‘not clearly stating the purpose, method and scope of personal information collection’, ‘utilizing personal information without obtaining users’ consent’, and ‘collecting personal information irrelevant to the service provided, in breach of the principle of necessity’, etc. During the latest round of institutional reform at provincial level, as many as eight provincial authorities have established big data management bureaus, including Guizhou, Shandong, Chongqing, Fujian, Guangdong, Zhejiang, Jilin and Guangxi.47 With successful experience obtained since commencement of the first internet court in Hangzhou in August 2017, the third Meeting of 44 Meijin Yi, ‘Comments on the Information Security Technology – Personal Information Security Specification’ (Baidu Public Policy Research Institute, 26 January 2018), available at https://​mp​.weixin​.qq​.com/​s/​kw​ZgLUNc7D9N​PGuW2qIWYA accessed 28 February 2020. 45 CAC, Special Rectification Work on Potential Illegal Collection and Utilization of Personal Information by Apps Initiated (2019), available at http://​www​.cac​.gov​.cn/​ 2019​-05/​23/​c​_1124532020​.htm accessed 28 February 2020. 46 CAC, GUO XIN BAN MI ZI [2019] NO. 191, available at http://​www​.cac​.gov​ .cn/​2019​-12/​27/​c​_1578986455686625​.htm​?scene​=​2​&​clicktime​=​1577667572​&​enterid ​ =​1577667572​&​from​=​singlemessage​&​isappinstalled​=​0 accessed 28 February 2020. 47 Guancha, ‘“Big Data” Has Become a Highlight for Institutional Reform at Provincial Level: Eight Provinces Have Established Big Data Bureau’ (5 November

Merger control in China’s digital economy

93

the Central All-Round Comprehensive Reform Committee deliberated on and approved the Scheme on the Establishment of Beijing and Guangzhou Internet Courts in July 2018. As for data required by the internet courts when hearing cases, retrieval of and access to data shall be in line with legal conditions, scope and procedure.48 During the 9th internet annual conference held in January 2019, Huawei, China Telecom, China Unicom, China Mobile, Toutiao, Didi and other undertakings proactively and voluntarily declared to join the Self-Discipline Convention on Collection and Application of Network Data and User Personal Information, so as to create a sound, credible and safe network ecosystem together with the authorities and all walks of life.49 4.3.2.4 Prediction of the AMB’s future enforcement Whether the SAMR would also assess data-driven concentration from the perspective of privacy besides input foreclosure is also a concern against the serious backdrop above. The SAMR, which the AMB is affiliated with, is a comprehensive market supervision department, also consisting of Price Supervision and Anti-Unfair Competition Bureau, Cyber Transaction Supervision and Management Bureau, Enforcement and Inspect Bureau and others. Considering the current institutional arrangement, many elements would affect application of merger control rules, such as how the AMB would consider privacy, how it would coordinate with other internal departments of the SAMR, as well as how the SAMR would coordinate with the Office of the Central Cyberspace Affairs Commission in specific cases, with increasing importance of privacy protection in the digital economy. 4.3.3

Competition in Innovation

The digital economy is closely connected with innovation. It is predicted that the digital economy will constitute the main body of innovation competition to be analyzed by global anti-monopoly authorities. In accordance with Article 27 of the AML, elements which shall be considered when reviewing a concentration include, but are not limited to, its influence over market entry and

2018), available at https://​www​.guancha​.cn/​society/​2018​_11​_05​_478375​.shtml accessed 28 February 2020. 48 The Supreme People’s Court, ‘Provisions on Several Issues about Internet Courts’ Hearing Cases’ [2018], available at https://​www​.chinacourt​.org/​article/​detail/​ 2018/​09/​id/​3489797​.shtml accessed 28 February 2020. 49 Science.china.com, ‘Huawei Declares to Join the Self-Discipline Convention on Collection and Application of Network Data and User Personal Information’ (9 January 2019), available at http://​science​.china​.com​.cn/​2019​-01/​09/​content​_40637829​ .htm accessed 28 February 2020.

94

Challenges to assumptions in competition law

technical progress, which has laid the legal foundation for consideration of innovation. Article 8 of the Interim Provisions on Assessment of the Impact of Concentrations on Competition promulgated by MOFCOM in 2011 further makes it clear what influence a concentration could have on technical progress, both positively and negatively.50 What is worth mentioning, Article 1 of the SAMR Draft Amendment to the AML introduced ‘encouragement of innovation’ to the legislative purposes of the AML. It not only reflects China’s support for innovative economies from the perspective of competition law, but has also triggered more heated discussion as regards how the SAMR should review concentrations in the digital economy, especially against the backdrop of the ‘tolerant and prudent’ principle51 proposed by the Chinese government to regulate nascent industries. The 18th Communist Party Congress put forward the ‘Innovation Driven Development Strategy’ at the end of 2012, acting as the strategic support for improving overall national strength. More specifically, scientific and technological innovation is the strategic support for improving social productivity and comprehensive national strength, which shall be put in the key position of national development. When delivering the government report on 22 May 2020 during the 3rd Meeting of the 13th NPC, Premier LI Keqiang introduced that a batch of major innovative achievements had been made, and that business startups and innovation continued to surge nationwide.52 In practice, MOFCOM has reviewed concentrations from the perspective of innovation. In 2017, MOFCOM conditionally cleared a merger between Dow and Du Pont. MOFCOM concluded that the merger would eliminate or restrict competition in the markets of selective herbicides for rice and pesticides for rice, the competition analysis of which has covered innovation. MOFCOM observed that the transaction would bring side effects to technical progress in the market of selective herbicides for rice. More specifically, before the concentration, both Dow and Du Pont were significant innovative powers 50 MOFCOM, ‘Interim Provisions on Assessment of the Impact of Concentrations on Competition’ [2011], available at http://​www​.mofcom​.gov​.cn/​aarticle/​b/​c/​201109/​ 20110907723440​.html accessed 28 February 2020. 51 For example, Chinese Central Government, ‘Government Working Report – Delivered by Keqiang Li, Premier of the State Council, at the 5th Session of the 12th National People’s Congress on 5 March 2017’, ‘We will formulate regulatory rules for nascent industries in keeping with the principle of encouraging innovation and conducting regulation in a tolerant and prudent way, and guide and promote the healthy development of emerging industries’ [2017], available at http://​www​.gov​.cn/​guowuyuan/​ 2017​-03/​16/​content​_5177940​.htm accessed 28 February 2020. 52 Chinadaily.com, ‘The Government Working Report (Bilingual)’ (2020), available at https://​language​.chinadaily​.com​.cn/​a/​202006/​01/​WS5ed4​6379a310a8​ b241159ce5​.html accessed 3 January 2020.

Merger control in China’s digital economy

95

in the market and competed fiercely in research and development (‘R&D’), pouring huge amounts of money, being equipped with strong innovative power and having rich product reserves. The merger would eliminate the basis for competition. After the concentration, both parties might lack momentum to do R&D, decrease investment in the current parallel innovation field (products having the same targets), and delay the launch of new products, which would drag technical progress in the market.53 In 2018, MOFCOM conditionally cleared Bayer’s acquisition of Monsanto, and concluded that the concentration would eliminate or restrict competition in the global digital agricultural market. It also concerned assessment of innovation. Before the transaction, both Monsanto and Bayer were essential innovative powers in the market of digital agricultural market, investing hugely in R&D; while after the transaction, Bayer might probably lessen investment in innovation, which would have a detrimental impact on technical progress. Furthermore, Bayer might also raise technical thresholds to block market innovation.54 In February 2020, the SAMR declared to conditionally clear Danaher Corporation (‘Danaher’)’s acquisition of the medical life sciences and biopharmaceutical businesses of General Electric Company (‘Target Business’).55 As for the hollow fiber tangential flow filter, both the global and Chinese markets were highly concentrated.56 There were mainly two competitors on the global market pro-concentration, while the Target Business’s market share was as high as 45–50 per cent. Even if Danaher did not produce hollow fiber tangential flow filter pro-concentration, it was doing R&D into relevant products, bearing prominently innovative attributes (confidential information) in comparison to the products on the then market. The SAMR was concerned that after-concentration, the merged entity’s incentive to invest in and commercialize innovative substitutes would be discouraged. Accordingly, the introduction

53 MOFCOM, Decision of 2 May 2017, Public Announcement No. 25 [2017], available at http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201705/​20170502568075​.shtml accessed 28 February 2020. 54 MOFCOM, Decision of 13 March 2018, Public Announcement No. 31 [2018], available at http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201803/​20180302719123​.shtml accessed 28 February 2020. 55 SAMR, Anti-Monopoly Review Announcement on the Conditional Clearance of Danaher Corporation’s Acquisition of the Medical Life Sciences and Biopharmaceutical Businesses of General Electric Company [2020], available at http://​www​.samr​.gov​.cn/​ fldj/​tzgg/​ftjpz/​202002/​t20200228​_312297​.html accessed 1 March 2020. 56 HHI for global market was 4,118, while that for Chinese market was 3,752.

96

Challenges to assumptions in competition law

of innovative products could be postponed, which would exert side effects on market competition and technical advancement.57 How to value innovation, including how to judge the possibility to innovate, innovation motivation, as well as costs and benefits of innovation, are challenges being confronted by various anti-monopoly agencies, which still needs theoretical research and enforcement exploration. Since the relationship between a concentration and innovation is not clear in every case, the Chinese anti-monopoly agency should be really cautious when assessing concentrations from the perspective of innovation. 4.3.4

Leveraging Effect

It is quite common for an undertaking with market power to influence adjacent market(s) in the digital economy. Through a concentration, an undertaking could weaken competition in the adjacent market(s) by making use of its market power in another market, which is usually called ‘leveraging effect’. Under the influence of undertaking platformalization, network flow coming first and trans-sector competition, the condition for application of leveraging theory widely questioned before is also changing. ‘The Commercial Use of Consumer Data: Report on the CMA’s Call for Information’ published by the Competition & Market Authority in 2015 pointed out that: ‘Respondents raised concerns about the potential for consumer data to be used to generate or exacerbate market power in a single market, or being used as a source of power that could be leveraged into a related market.’58 The ‘Challenges for Competition Policy in a Digitalized Economy’ released by the European Parliament in 2015 also emphasized the effect certain conducts might have on adjacent markets through leverage.59 The Chinese anti-monopoly agency has applied leverage theories of harm in the past. In the blocked Coca-Cola’s acquisition of Huiyuan, MOFCOM concluded that Coca-Cola had the ability to leverage its dominant power in the carbonate beverage market to the juice beverage market, so as to restrict 57 One of the remedies is that Danaher/the merged entity promised to continue with the R&D of the Emily program (confidential business) for two years since the deal was settled. 58 Competition and Market Authority, The Commercial Use of Consumer Data – Report on the CMA’s Call for Information (2015) 9, available at https://​assets​ .publishing​.service​.gov​.uk/​government/​uploads/​system/​uploads/​attachment​_data/​file/​ 435817/​The​_commercial​_use​_of​_consumer​_data​.pdf accessed 28 February 2020. 59 European Parliament, Challenges for Competition Policy in a Digitalized Economy (2015) 31–33, 61–62, available at http://​www​.europarl​.europa​.eu/​RegData/​ etudes/​STUD/​2015/​542235/​IPOL​_STU(2015)542235​_EN​.pdf accessed 28 February 2020.

Merger control in China’s digital economy

97

or eliminate competition in the latter and would finally harm legitimate rights and interests of consumers. Furthermore, after the concentration, Coca-Cola’s controlling power over the juice beverage market would be significantly enhanced through making use of two famous fruit juice brands, ‘Mei Zhiyuan’ and ‘Huiyuan’, together with its dominant power in the carbonate beverage market and leveraging effect. It would remarkably increase the obstacle to enter the juice beverage market.60 Besides, in Walmart’s acquisition of 33.60 per cent of shares in Newheight Holding, considering that Walmart had a mature storage and delivery system, wide supply channels and high brand awareness in China’s brick-and-mortar retailing market, MOFCOM was concerned that it had the ability to leverage its advantage in the market of brick-and-mortar retailing to Yi Shi Duo No. 1 Shop’s online retailing businesses, which would substantially increase the merged entity’s competitiveness in the market of online retailing. MOFCOM also extended the review into effects the acquisition might have on the Chinese value-added telecommunications market. Investigation showed that the new entity could enter the Chinese value-added telecommunications market through Yi Shi Duo No. 1 Shop and would be capable of rapidly expanding its businesses on the basis of its comprehensive competition advantage combining brick-and-mortar retailing and online retailing, so as to obtain dominance in the market of value-added telecommunications and substantially bigger bargaining power over online platform users. It might eliminate or restrict competition in the Chinese value-added telecommunications market.61 In practice, internet titans have engaged in large scale trans-sector acquisitions. The possibility for application of leverage theory to the following anti-monopoly review of non-horizontal concentrations should not be excluded, considering MOFCOM’s opening attitude towards it in Coca-Cola’s acquisition of Huiyuan and Walmart’s acquisition of 33.60 per cent of shares in Newheight Holding.

60 MOFCOM, Decision of 18 March 2009, Public Announcement No. 22 [2009], available at http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​200903/​20090306108494​.shtml accessed 28 February 2020. 61 MOFCOM, Decision of 14 August 2012, Public Announcement No. 49 [2012], available at http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201303/​20130300058730​.shtml accessed 28 February 2020.

98

Challenges to assumptions in competition law

4.4 REMEDIES 4.4.1

Choice of Remedies

Remedies62 consist of structural remedies and behavioral remedies. While the former includes divestment of certain business units, behavioral remedies regulate the future behavior of the merged entity. This might consist of commitments by the merging parties not to engage in certain conduct or contractual arrangements such as compulsory licensing or access to intellectual property. They require ongoing supervision by the competition agency. In contrast to structural remedies, behavioral remedies do not provide a clean break because the implementation and supervision requirements will unavoidably result in the establishment of specific, continuous connections between the merged entity and other market operators. Generally speaking, competition agencies prefer structural remedies over behavioral remedies, as structural remedies are regarded as being more effective in avoiding the potential negative impacts of mergers.63 With respect to digital economy, how the anti-monopoly agency would choose between structural remedies and behavioral remedies is well worth contemplation. Since the digital economy is characterized by network effects and multi-sided market, what influences may they have on choice between remedies? Will behavioral remedies be more valued in the digital economy, just as the practice in China before? Even if in the digital economy, behavioral remedies are still confronted with dilemmas, such as difficulty in design and supervision, as well as over remedies brought by the authorities’ continuous intervention of the market. In this regard, how to apply behavioral remedies is still well worth deeper research. Furthermore, the blockchain technology could also be made use of to support the arrangement of remedies, such as to solve the problems of supervision enforcement.64

62 Please note that in China, ‘restrictive condition’ is more commonly used than ‘remedy’ or ‘commitment’. For more details, please find Article 29 of the AML, etc. Wei Han, ‘Merger Remedies in China: Past, Present, and Future’ [2013] (2) CPI Antitrust Chronicle February, 2. 63 Alison Jones and Brenda Sufrin, EU Competition Law – Text, Cases and Materials (Oxford University Press, 2016), 1195–1200. 64 OECD, Blockchain Technology and Competition Policy – Issues Paper by the Secretariat (2019) 8–9, available at https://​one​.oecd​.org/​document/​DAF/​COMP/​ WD(2018)47/​en/​pdf accessed 28 February 2020.

Merger control in China’s digital economy

4.4.2

99

Would China Continue to Prefer Behavioral Remedies in the Digital Economy?

In China, Article 29 of the AML stipulates that mergers may be approved with remedies. Since the entry into force of the AML in 2008, the Chinese competition authority has conditionally cleared 50 mergers. Remedies already imposed include both structural and behavioral ones. Although the AML has been in effect only for more than ten years, the widespread application of behavioral remedies in merger review has been heatedly discussed among competition law scholars,65 particularly the rather unique hold-separate remedy.66 The use of this unique condition could partly be explained by the specific Chinese context and market environment. In addition, changes in the markets, especially the digital market, require constant adaptation, which is a challenge faced by each competition agency in the world. As such, the function and use of experimental and innovative remedies is advised to not be arbitrarily denied. In the following anti-monopoly review of concentrations in the digital economy, the SAMR would remain open to behavioral remedies, and the possibility for issuing creative remedies circling around data, algorithm, privacy and innovation cannot be excluded. Naming openness of data as an example, Article 3 of the Provisions on Imposing Remedies on the Concentration of Undertakings (Trial Implementation)67 explicitly stipulated, ‘requiring the undertakings participating in a concentration to make available their respective networks, platforms and other infrastructure, license key technologies (including patents, proprietary technologies or other IPRs), terminate exclusive agreements’. From the perspective of China’s merger control practice, there is no case directly connected with data (input) foreclosure. Nevertheless, MOFCOM did express concern of input foreclosure in precedents, while behavioral remedies, such as openness, have been applied several times. The possibility for SAMR to recognize a certain data set held by the undertaking concerned as input indispensable for competition and require it to open access to its data

65 For example, Ariel Ezrachi and Wei Han, ‘Merger Remedies – the Chinese Experience’ (2015) 3(1), Journal of Antitrust Enforcement, i74, available at https://​ academic​.oup​.com/​antitrust/​article/​3/​suppl​_1/​i69/​225289 accessed 28 February 2020. 66 Ibid., i80–i82, i86–i87. Han Wei, Yin Ranran and Zheng Xiong, ‘Standalone Hold Separate Orders as Remedies in Chinese Merger Control’, (2019) in Wang Xiaoye, The Pioneer of Competition Law in China – Liber Amicorum, Concurrences (Institute of Competition Law, 2019), 27–42. 67 MOFCOM, ‘Provisions on Imposing Remedies on the Concentration of Undertakings (for Trial Implementation’ [2014], available at http://​www​.mofcom​.gov​ .cn/​article/​b/​c/​201412/​20141200835207​.shtml accessed 28 February 2020.

100

Challenges to assumptions in competition law

set to third parties after confirming competition concerns cannot be excluded. Nevertheless, the risk of conflicts between opening access to certain data sets and personal information protection provisions in other laws and regulations, such as the PIPL to be enacted, Cyber Security Law, E-Commerce Law and Personality Right Section of the Civil Code, cannot be ignored.68 It is quite normal for various conflicts to come up during behavioral remedy implementation. Nevertheless, the Chinese competition agency does not acknowledge arbitrating disputes in this regard, especially those of openness to indispensable input. Accordingly, introduction of arbitration mechanism could be another goal for the Chinese merger control system.

4.5

PROSPECTS OF THE SAMR’S ENFORCEMENT

4.5.1

Increasing Attention Has Been Paid to Monopolistic Issues in the Digital Economy

During a meeting held on 11 December 2020, the Political Bureau of the Central Committee of the CPC proposed to strengthen anti-monopoly (enforcement) and to avoid disorderly expansion of capital for the first time.69 It shows determination of the top in terms of anti-monopoly law enforcement. The SAMR has also showed great interest in regulating the conduct of digital companies. Sources have it that the SAMR has initiated the investigation into the ‘either-or’ provision compelled by Alibaba.70 Besides the three administrative punishment decisions introduced above, the SAMR declared that it is investigating other mergers concerning VIE, such as the merger between Huya and Douyu, the top two Chinese online game live broadcasting platforms.71 On 10 November 2020, the SAMR released the draft of Anti-Monopoly Guiding Opinions in the Platform Economy to solicit public opinions by 30 November 2020. The guiding opinions target all the monopolistic conducts undertaken by internet platforms, fully taking attributes of the digital economy 68 Wei Han and Yajie Gao, ‘Promote Openness or Strengthen Protection? Application of Law to Data Competition in China’ [2018] CPI Antitrust Chronicle, available at https://​papers​.ssrn​.com/​sol3/​papers​.cfm​?abstract​_id​=​3240598 accessed 28 February 2020. 69 Sohu, ‘The Political Bureau of the Central Committee of the CPC Requires Strengthening Anti-Monopoly and Avoiding Disorderly Expansion of Capital for the First Time’ (2020), available at https://​www​.sohu​.com/​a/​437749652​_115362 accessed 3 January 2021. 70 Chao Yu, ‘Strengthening Anti-Monopoly Is for Better Development’ (people. com.cn, 25 December 2020), available at http://​paper​.people​.com​.cn/​rmrb/​html/​2020​ -12/​25/​nw​.D110000renmrb​_20201225​_3​-07​.htm accessed 3 January 2021. 71 SAMR (n. 17).

Merger control in China’s digital economy

101

into account.72 It is only the start of China’s anti-monopoly enforcement in the digital economy. From all perspectives, it is predicted that the SAMR would take much more proactive actions than ever before when it comes to review of digital mergers. The Chinese theoretical circle has paid growing attention to monopolistic issues in the digital economy. Especially since 2018, topics of data and algorithm could be found in more and more seminars, such as the ‘Latest developments in international anti-monopoly law’ seminar held in November 201973 and the seminar of Fair Competition and Anti-Monopoly in the Digital Economy held by the Development and Research Center of the SAMR in Beijing in December 2018. During the later seminar, heated discussions were held about regulation and governance of trans-sector competition in the internet era, how to coordinate the relationship between Chinese competition policy and high technology industrial policy in the digital economy and other concrete issues, such as data monopoly. Mr. Zhenguo Wu, Director General of the AMB of the SAMR, also attended the seminar.74 It is worth noting that the SAMR is still investigating Didi’s acquisition of Uber. During a press conference held by the State Council Information Office, PRC in November 2018, in response to questions from a journalist, Mr. Zhenguo Wu introduced that, The SAMR is investigating the acquisition in accordance with anti-monopoly laws and regulations. Online car hailing is a nascent social phenomenon, no matter it is in China, the European Union or the United States, which is different from traditional industries. Competition in this market is not only complicated, but also versatile, which has drawn widespread attention all around the society. We are doing research into regular pattern and characteristics of internet competition, comprehensively analyzing and assessing effects the acquisition has on market competition and sector development and would seriously punish monopolistic conducts which would harm consumer rights and interests. Acting as the anti-monopoly enforcement agency, the SAMR pays high attention to the competition problems in the innovative sectors, and would regulate internet and other nascent sectors in accordance with their respective innovative development, new regulatory methods, as well as sticking to tolerant and prudent principles, so as to maintain fair competition in the market, 72 SAMR, ‘Notice of the SAMR to Solicit Public Opinions on the Draft of Anti-Monopoly Guiding Opinions in the Platform Economy’ (2020), available at http://​ www​.samr​.gov​.cn/​hd/​zjdc/​202011/​t20201109​_323234​.html accessed 3 January 2021. 73 CIIAI, ‘Latest Developments in International Anti-Monopoly Law’ (20–21 November 2019), available at http://​ciiai​.cupl​.edu​.cn/​zh/​node/​1450 accessed 28 February 2020. 74 Gov.cn, ‘Roundtable on Fair Competition and Anti-Monopoly in the Digital Economy Seminar Held by the SAMR in Beijing’ (26 December 2018), available at http://​www​.gov​.cn/​xinwen/​2018​-12/​26/​content​_5352249​.htm accessed 28 February 2020.

102

Challenges to assumptions in competition law

and provide a loose and tolerant environment for the emerging form and innovative business model of internet.[75] The market competition mechanism shall play its due role so as to strengthen innovation momentum, while regulatory system over the internet industry should also be reinforced. The SAMR would coordinate with other departments properly in order to strengthen market supervision, prohibit industrial monopoly and clear market entry barriers, and protect legitimate rights and interests of consumers and public interests.76

4.5.2

Insufficient Enforcement Human Resources

There are four divisions affiliated with the AMB which are in charge of merger control, namely Law Enforcement Supervision Division, and Merger Control Divisions 1, 2 and 3.77 There are only fewer than 20 case handlers in total. In comparison, even if three divisions78 are authorized to investigate a monopoly agreement, abuse of dominance and administrative monopoly respectively, market supervision departments at provincial level are empowered to investigate certain types of the three monopolistic conducts above, as stipulated by the Notice of the SAMR on the Authorization of Anti-Monopoly Enforcement Power released in December 2018;79 while the SAMR is the one and only agency entitled to review concentrations. In fact, however, the number of noti75 The Antitrust Source, ‘Interview with Zhenguo Wu, Director General of China’s State Administration for Market Regulation (SAMR)’ (2018), available at https://​www​ .americanbar​.org/​content/​dam/​aba/​publishing/​antitrust​_source/​2018​-2019/​atsource ​ -december2018/​dec18​_wu​_intrvw​_12​_17f​.pdf accessed 28 February 2020, ‘we should strengthen competition supervision in the area of newly emerging markets, adhere to the principles of tolerance and prudence, properly handle the relationship between innovation stimulation and fair competition maintenance, follow the market rules, and avoid casual and excessive intervention, so as to create a fair competition market environment for the sustainable and healthy growth of the newly emerging markets’ (page 5); ‘to protect, encourage, and promote innovation. An era of innovation-driven economy requires us to pay more attention to the innovation environment and digital economy’ (page 9). 76 Xinhua.net, ‘The SAMR Is Now Investigating Didi’s Acquisition of Uber’ (16 November 2018), available at http://​www​.xinhuanet​.com/​fortune/​2018​-11/​16/​c​ _129995829​.htm accessed 28 February 2020. 77 The Antitrust Source (n. 75) 2. More specific, the Merger Control Divisions are ‘in charge of reviewing merges of different industries’; while functions and duties of the Law Enforcement Supervision Division include ‘investigating illegally implemented mergers and supervising the enforcement of merger remedies’. 78 Ibid. They are Monopoly Agreement Investigation Division, Abuse of Dominance Investigation Division and Administrative Monopoly Investigation Division. 79 Other legal bases include Article 10(2) of the AML, ‘The authority for enforcement of the Anti-monopoly Law under the State Council may, in light of the need of work, empower the appropriate departments of the people’s governments of provinces, autonomous regions or municipalities directly under the Central Government to take

Merger control in China’s digital economy

103

fied concentrations is much higher than that of other monopolistic cases. Since the VIE structure is no longer an obstacle, notification of concentrations in the digital economy would become normality. It would make the insufficient human resources in enforcement even worse. We believe that the SAMR could consider taking the following actions to solve the problem. First and foremost, good use should be made of the local enforcement force. Since the key to institutional reform in recent years is streamlining organs and decreasing enforcers, it is not workable to enlarge the enforcement team of the SAMR in the short term. Against this background, it would be proper to make better use of local enforcement. More specifically, the SAMR could explore empowering market supervision departments from developed regions in which concentrations are really active, such as Beijing, Shanghai, Guangdong, Zhejiang and Jiangsu, to review concentrations on the basis of ‘location of notifying undertaking’ and other standards. Secondly, revise substantive merger control rules. The SAMR is advised to pay special attention to data, privacy and innovation on the basis of deep research into the digital economy and try to ameliorate relevant rules.80 For example, the Interim Provisions on Assessment of the Impact of Concentrations on Competition could be complemented with special competition elements for the digital economy. Besides introduction of more clear rules about how to judge controlling power, notification of joint ventures and others will also facilitate merger control in the digital economy. As for controlling power, rules on how to judge change of control for the purpose of data integration should be provided. With respect to joint ventures, MOFCOM treated all joint ventures in the same way, no matter if they are fully functional or non-fully functional, as long as the notification threshold has been passed.81 After the institutional integration, clear guidelines should also be given to coordination between the three Merger Control Divisions and the Monopoly Agreement Investigation Division. Thirdly, internal knowledge management should be strengthened. Good knowledge management could strengthen internal communication, facilitate knowledge transfer, cut redundant work, enhance enforcers ability, and increase working efficiency. If combined with suggestions for empowering

charge of relevant enforcement of the Anti-monopoly Law in accordance with the provisions of this Law’. 80 Interim Provisions on Reviewing Concentrations of Business Operators (Draft for Public Opinions Solicitation) were released by the SAMR in January 2020. Unfortunately, attributes of the digital economy were not reflected. Available at http://​www​.samr​.gov​.cn/​fldj/​tzgg/​zqyj/​202001/​t20200107​_310322​.html accessed 28 February 2020. 81 Ariel Ezrachi and Wei Han (n. 65).

104

Challenges to assumptions in competition law

local authorities to review concentrations, it would be necessary to train local enforcers with support of better knowledge management. Furthermore, an enforcement database could be established, while results of various exchange training programs and external commissioned research projects should be better transformed to application, so as to improve the quality of enforcement. Fourthly, more support should be provided by technical experts. Besides external legal and economic experts, technical experts also play an irreplaceable role in the anti-monopoly enforcement in the digital economy. From the long-term perspective, not only external technical experts would be needed, a permanent internal independent technical support team is also advised to be set up.82 Last but not least, international exchange in anti-monopoly enforcement in the digital economy should continue to be intensified. China keeps frequent communications with main jurisdictions, such as the regularly held BRICS International Competition Policy Conference83 and the China-EU Competition Policy Week.84 Naming the China-EU Competition Policy Week as an example, it has extended from the traditional anti-monopoly issues to emerging sectors. During the 19th Competition Policy Week held in October 2019, application of competition law to the digital economy was discussed, covering topics of transaction-value-based notification threshold, big data, algorithms, interim measures, innovation policy, etc.85 Ever since the reform of the Chinese competition agency, it is necessary to further enhance international communications through bilateral cooperation agreements, especially in the area of the digital economy. Besides, in order to strengthen cooperation with competition authorities from other jurisdictions, the SAMR could also consider joining the International Competition Network.

4.6 CONCLUSION The digital economy has infiltrated almost every aspect of the Chinese economy and has infused it with impetus. In comparison to the vibrant M&As 82 Wei Han, Yajie Gao and Ai Deng, ‘Algorithmic Price Discrimination on Online Platforms and Antitrust Enforcement in China’s Digital Economy’ (2018) the antitrust source, 9, available at https://​www​.americanbar​.org/​content/​dam/​aba/​publishing/​ antitrust​_source/​2018​-2019/​atsource​-august2018/​aug18​_wei​_8​_16f​.pdf accessed 28 February 2020. 83 VI BRICS International Competition Conference 2019, available at https://​brics​ -icc​-2019​.org/​en/​ accessed 28 February 2020. 84 Available at https://​co​mpetitionc​ooperation​.eu/​accessed 28 February 2020. 85 Competition Cooperation – EU-ASIA Partnership, ‘19th EU-China Competition Week’ (2019), available at https://​co​mpetitionc​ooperation​.eu/​19th​-eu​-china​ -competition​-week/​accessed 28 February 2020.

Merger control in China’s digital economy

105

in the digital economy, only few official investigations have been initiated by MOFCOM. In recent years, increasing concentrations have attracted public attention to competition concerns, and people from all walks of life deeply care about SAMR’s attitudes in this regard. The SAMR has made it clear that VIE is no longer the obstacle for merger notification through three administrative punishment decisions and is also reviewing other digital mergers. If the SAMR Draft Amendment to the AML is deliberated and approved by the SC of the NPC, the SAMR per se would legally obtain stronger competence to have more potentially problematic concentrations in the digital economy under its review. As for competition assessment in certain cases, special attention should be paid to openness to data, privacy protection, innovation competition and leveraging effect, on the basis of characteristics of the digital economy and Chinese anti-monopoly legislation and practice. With respect to remedies, how to choose between structural and behavioral ones is still well worth attention. From the anti-monopoly enforcement in the more than last decade, the possibility for the Chinese competition authority to keep open to behavioral remedies should not be excluded. Only effective anti-monopoly enforcement can help ensure adequate competition in China’s digital economy so as to provide rich soil for more excellent undertakings, like Alibaba and Tencent. Maintaining a market which could constantly bring up undertakings with high quality is what the value of anti-monopoly law is, which could benefit the ultimate consumers in the end. After all, only effective competition can truly protect consumer rights and interests.

5. The adoption of specialised competition tribunals in Latin American countries: transplants and commonalities among them Claudia O’Kane INTRODUCTION One of the main objectives of this chapter is to offer a theoretical framework of transplants of institutions or lessons aiming to guide countries, particularly Latin American countries given the commonalities shared among them, searching for practical solutions to problems similar to those experienced in Mexico and which provoked the adoption of a specialised competition tribunal; about how to transfer the lessons provided by this country. In transferring such lessons it is crucial that policymakers within governments understand the analysis of the transplantation phenomenon. The review of relevant literature evidences that the transplantation of legal systems, institutions or lessons is a complex matter.1 Observing what has been done elsewhere, and how, offers a good comparison opportunity for countries looking to diagnose the strengths and weaknesses of their own institutions, policies, or laws under enquiry, and to quickly transfer the adopted lessons that have been previously tested elsewhere.2 Nevertheless, such adaptation requires the theoretical insights detailed in Section 5.1 of this chapter if failures are to be avoided. Lessons from other countries also need to be aligned with prevailing informal institutions (Section 5.1.1). Since the analysis of transplantation is complex, and its comprehension crucial, the first aim of Section 5.1 is to provide a general conceptual frame1 Juan Couyoumdjian, ‘Are Institutional Transplants Viable? An Examination in Light of the Proposals by Jeremy Bentham’ (2012) 8 Journal of Institutional Economics 494. 2 Richard Rose, Learning from Comparative Public Policy: A Practical Guide (Routledge. 2004) 3, 6.

106

The adoption of specialised competition tribunals in Latin American countries

107

work of the most common definitions of the terms necessary for an understanding of the transplantation phenomenon, and the most relevant theories that underpin its advance. The focus centres on the study of the following concepts: (1) institutions; (2) transplants; (3) institutional change; (4) culture; and (5) design of institutional transplants. In relation to institutions, Section 5.1 identifies institutions as restraints humanly devised to bring order within society, whether as formal or informal. Equally, the study of the judiciary as institution reveals the importance of having independent and effective judiciaries, and brings attention to how dependent, inefficient, poor quality, inconsistent and ineffective judiciaries impact the legitimacy of such a branch. This section also presents a general framework for legal transplants, whose examination confirms that the debates around the correlation between law and society are abundant and remain unsettled, and that the theories around transplant phenomena are plenty, complex and profoundly divided. Section 5.1 also helps highlight that, despite the unsettled nature of the debate, different motivations promote the vast use of transplants, particularly among developing countries. A typology based on such motivations has been designed (Section 5.1.2). Regardless of the motivation, it appears that the success or failure of transplants rests on the compatibility or not between the prevailing informal institutions of the recipient, such as social or cultural factors, and the new transplanted institution, such as a specialised tribunal. With a view to providing a better understanding of the likelihood of acceptance or rejection of a transplanted formal institution, Section 5.1 also illustrates how and why institutions change. From this, an analysis of different theories is offered. The New Institutional Economics doctrine clarifies that informal institutions are drivers of change, which allow the embedding of formal institutions, and that formal and informal institutions change at different paces. In addition, and with a view to exploring in more detail the notion of informal institutions, Section 5.1 provides an analysis of the four main streams of research that explain how the acquirement of culture occurs. The evolutionary anthropology theory exemplifies how culture is an intrinsic part of the individual and their development. Therefore, neglecting the study of culture when analysing transplantation of institutions or lessons would be deficient. In relation to the design of transplants, Section 5.1 contrasts the various alternatives and shows that the ones most likely to succeed are those that recognise the importance of considering local needs, local knowledge, and local prevailing informal institutions. The analysis in Section 5.1 proceeds as follows. First it explores the meaning of institutions. Second, it analyses the advancement of the theoretical framework of transplants. Third, it presents the different theories of institutional change. Fourth, it studies the definition of culture and its effect on institutions.

108

Challenges to assumptions in competition law

Fifth, it describes the different designs of transplants, and provides some recommendations to avoid failure. Finally, it provides concluding remarks.

5.1 TRANSPLANTS There has been considerable research into and interest in the concepts of institutions, and transplants, as well as their relevance in policy development.3 The large numbers of studies seeking to explain these concepts have in fact impinged their understanding, due to the ambiguity and inconsistency observed in the use of terms, and in the explanatory process of the existing theories.4 To address this challenge, this section aims to describe and summarise the most common concepts, arguments and theories in an attempt to gain a good understanding of the transplantation phenomenon, since the subject matter of this research pertains to the possible transplantation of specialised competition tribunals ‒ an example of a formal institution. 5.1.1 Institutions The literature review provides the most common definition of institutions, that given by North, according to which institutions are understood as the ‘rules of the game in a society, or, more formally, are the humanly devised constraints that shape human interaction’,5 represented also as economic and political agents interacting under a shared system of encouragements6 or as ‘stable patterns in social interactions’.7 Institutions are also defined as ‘a set of rules that structure social interactions in particular ways’,8 instruments used to create steady expectations of the behaviour of others by imposing constraints,9

3 Geoffrey Hodgson, ‘What Are Institutions’ (2006) XL Journal of Economic Issues 1, 1. 4 Christopher Kingston and Gonzalo Caballero, ‘Comparing Theories of Institutional Change’ (2009) 5:2 Journal of Institutional Economics, 151–152; Hodgson (n. 3) 1. 5 Douglass North, Institutions, Institutional Change and Economic Performance (Cambridge University Press, 1990) 3. 6 Couyoumdjian (n. 1) 491. 7 Martin De Jong, Konstantinos Lalenis and Virginie Mamadouh, ‘An Introduction to Institutional Transplantation’, in Martin De Jong, Konstantinos Lalenis and Virginie Mamadouh (eds), The Theory and Practice of Institutional Transplantation Experiences with the Transfer of Policy Institutions (Kluwer Academic Publishers, 2002) 3. 8 Jack Knight, Institutions and Social Conflict (Cambridge University Press, 1992) 2. 9 Geoffrey Hodgson (n. 3) 2.

The adoption of specialised competition tribunals in Latin American countries

109

simply as the outcomes of human relations and desires,10 or ‘as durable systems of established and embedded social rules that structure social interactions’.11 Institutions have also been considered as ‘a complex of status-role relationships, which is concerned with a particular area of activity within any specified social system’.12 Equally, it has been indicated that individuals create institutions as an external tool in an attempt to give structure and order to their environment.13 The above human-created institutions are defined as either formal or informal. Formal institutions are positive, explicit or written down, enforced by a third party,14 and usually observed for example in the form of constitutions and laws.15 Informal institutions are not written down, are self-enforced, and exemplified as social norms or conventions derived from moral codes, ideologies or culture.16 From the cluster of definitions presented earlier it could be said that institutions are essentially humanly devised constraints, whether formal or informal, which aim to bring order to social interactions. Considering that the focus of this chapter rests on the adoption of specialised competition tribunals, the next step is therefore to study the judiciary as an institution. a The judiciary as formal institution Of the three main institutional powers within the state, the judiciary has been seen as the weakest or perhaps the most neglected; in the words of Hamilton, the executive holds the sword, the legislature commands the purse, while the

Ibid. 8. Ibid. 13. 12 Howard Kaplan, ‘The Concept of Institution: A Review, Evaluation, and Suggested Research Procedure’ (1960) 39 Social Forces 2, 179. 13 Douglas North and Arthur Denzau, ‘Shared Mental Models: Ideologies and Institutions’ (1994) 47 Kyklos 4. 14 North (n. 3) 4; Kingston and Caballero (n 4) 154; Nils Goldschmidt, ‘A Cultural Approach to Economics’ (2006) Intereconomics 177; Gerard Roland, ‘Understanding Institutional Change: Fast-Moving and Slow-Moving Institutions’ (2004) 38 Studies in Comparative International Development 4, 111; Martin De Jong and Suzan Stoter ‘Institutional Transplantation and the Rule of Law: How This Interdisciplinary Method Can Enhance the Legitimacy of International Organisations’ (2009) 2 Erasmus Law Review 03, 318; Martin De Jong, ‘The Pitfalls of Family Resemblance: Why Transferring Planning Institutions between “Similar Countries” is Delicate Business’ (2004) 12 European Planning Studies 7, 1057–1058. 15 Couyoumdjian (n. 1) 491. 16 Kingston and Caballero (n. 4) 158; De Jong and Stoter (n. 14) 318; Roland (n. 14) 111; De Jong (n. 14) 1057–1058; Couyoumdjian (n. 1) 491; Goldschmidt (n. 14) 177. 10 11

110

Challenges to assumptions in competition law

judiciary does not have influence over the sword or the purse.17 Research has also focused on the legislative and the executive, neglecting the study of law and courts, particularly in Latin America.18 Despite this apparent disadvantage, the judiciary is in fact the formal institution that limits the power of the executive and legislative under the ‘checks and balances’ role, by independently enforcing the law, and reviewing that law and policymaking comply with the constitution, so becoming the final arbiter of the law.19 In the field of human rights, the judiciary constitutes one of the most important institutions, whose function consists in protecting and promoting said rights.20 Regarding the interaction between public authorities and individuals, the judiciary protects the latter from arbitrary state action, and ensures their civic and political liberties.21 In relation to the rule of law, it has been widely agreed among scholars that an independent and effective judiciary is a vital factor for its application.22 Economic freedom and economic growth have been equally correlated with independent and efficient judiciaries, under the assumption that the protection of property rights via enforced agreements leads to more developed credit markets, as this reduces the uncertainty of rights not being upheld and increases the willingness to invest more.23 The notion of an independent judiciary refers broadly to the ability to perform, detached from any external political influence from the executive or

17 Alexander Hamilton, John Madison and John Jay, The Federalist Papers (Signet, 2003) 464; Ralf Dahrendorf, ‘A Confusion of Powers: Politics and the Rule of Law’ (1977) 40 The Modern Law Review 1, 8. 18 Joseph Staats, Shaun Bowler and Jonathan Hiskey, ‘Measuring Judicial Performance in Latin America’ (2005) 47 Latin American Politics and Society, University of Miami 4, 78. 19 Rafael La Porta, Florencio Lopez-de-Silanes, Cristian Pop-Eleches and Andrei Shleifer, ‘Judicial Checks and Balances’ (2004) 112 Journal of Political Economy, 2–4. 20 Winston Langley, Encyclopaedia of Human Rights Issues since 1945 (Greenwood Publishing Group, 1999) 174–175. 21 Staats et al. (n. 18) 78. 22 Kenneth Dam, The Rule of Law and Economic Development (Brookings Institution Press, 2007) 93; Staats et al. (n. 18) 78. 23 La Porta et al. (n. 19) 3; Dam (n. 22) 21, 93–95; Sean Dougherty, ‘Boosting Growth and Reducing Informality in Mexico’ (2015) OECD Economics Department Working Papers No. 1188, 22–23 https://​www​.oecd​-ilibrary​.org/​economics/​boosting​ -growth​-and​-reducing​-informality​-in​-mexico​_5js4w28dnn28​-en accessed 12 July 2018; Giuliana Palumbo, Giulia Giupponi, Luca Nunziata and Juan Mora-Sanguinetti, ‘The Economics of Civil Justice: New Cross-Country Data and Empirics’ (2013) OECD, Economics Department Working Papers No. 1060, 2 https://​www​.oecd​-ilibrary​ .org/​economics/​the​-economics​-of​-civil​-justice​_5k41w04ds6kf​-en accessed 13 July 2018.

The adoption of specialised competition tribunals in Latin American countries

111

the legislative, and particularly to the ability of every judge to make autonomous decisions.24 The relevant literature claims that independent courts are essential to effectively review competition decisions and to adjudicate competition law disputes between private parties.25 For this reason, it is essential to observe the following structural factors: (1) selection and promotion based on merit rather than on political criteria;26 (2) good salaries and prohibition to reduce them;27 (3) life tenure with the only possibility of being removed for cause;28 (4) fair transfers and promotions according to pre-established rules;29 and (5) impartial assignment of cases.30 Judicial effectiveness broadly refers to enforcement, under the premise that good laws cannot substitute a judiciary system incapable of making them enforceable, or that good laws can turn into bad ones if they are not applicable, and even that any effort to improve substantive laws is meaningless if they are not enforced.31 This effectiveness constitutes one among several indicators used to measure the performance of the judiciary, where factors such as efficiency, quality and uniformity have been commonly included.32 In general terms, efficiency refers to the time to disposition of a case,33 quality responds to the knowledge of the law and its correct application,34 and uniformity means producing similar decisions in factually similar disputes. This section has reviewed the general notions of independence and effectiveness, and Section 5.1.2 explores where the legitimacy of the judiciary rests. As the judiciary lacks the ‘sword’ or the ‘purse’, its legitimacy resides in the levels of acceptance from the citizens.35 Thus, when a judiciary is independent, effective, efficient, makes accurate and uniform decisions, positively influences the general perceptions of the public and translates them into acceptance, then the institution is provided with legitimacy.

Staats et al. (n. 18) 79–80. OECD, ‘Judicial Perspectives on Competition Law’ (2017) 3. 26 Dam (n. 22) 115. 27 Ibid. 113–114. 28 Ibid. 114. 29 Randall Peerenboom, ‘Toward a Methodology for Successful Legal Transplants’ (2013) 1 The Chinese Journal of Comparative Law 1, 16. 30 Ibid. 31 Dam (n. 22) 93–95. 32 Staats et al. (n. 18) 79–80. 33 Dam (n. 22) 101. 34 Staats et al. (n. 18) 94. 35 John Yoo, ‘In Defense of the Court’s Legitimacy’ (2001) 68 The University of Chicago Law Review, 781. 24 25

112

Challenges to assumptions in competition law

Research has shown that citizens of countries with well-performing judiciaries express greater confidence in those institutions,36 which is not the case if the judiciary is not independent or effective. If the judiciary system is slow, uncertain, costly, ineffective, biased, or if decisions are contradictory or of poor quality, citizens will bring their cases with scepticism or will abstain from bringing them. Equally, if the powers of the executive and legislative are limitless, the respect for human, civic and political rights will be compromised, the implementation of the rule of law will be undermined, and the fostering of economic development will be hindered. After providing a basic definition of the concept of institutions, and a general analysis of the judiciary as institution, Section 5.1.2 explores the meaning of transplants of institutions in order to set out the factors of effective performance. This discussion is important, as it will allow conclusions to be drawn with regard to the transplantation of specialised courts, as an institution. 5.1.2

Theoretical Framework

The examination of the relevant literature shows that the debate around how law and society correlate is unsettled. The same can be said of the transplant phenomena. With respect to effective legal transplants, the relevant theories are often contradictory and complex. Despite the unsettledness, different motivations promote the widespread use of transplants, particularly among developing countries. One example that illustrates this scenario refers to developing countries adopting or modernising their competition regimes, as well as to creating or strengthening their competition authorities during the 1990s, in pursuit of financial aids.37 The success or failure of transplants is not easy to predict, but it appears that when problems arise it is essentially when social and cultural factors of the recipient country are abstracted from the analysis, or when similarities and differences are not evaluated, making the chosen case unsuitable. Transplant is a word that has been introduced to the legal world to describe the ‘phenomenon of borrowing, copying, or mirroring the laws or rules or institutions of one society for use in another’.38 As with laws, rules or institutions, lessons can also be transferred from one place to another. Incessantly, countries have been searching for experiences from elsewhere in the hope 36 Ryan Salzman and Adam Ramsey, ‘Judging the Judiciary: Understanding Public Confidence in Latin American Courts’ (2013) University of Miami, 88. 37 Michal Gal, ‘The Ecology of Antitrust’ in Competition, Competitiveness and Development: Lessons from Developing Countries (UNCTAD, 2004) 23. 38 Roger Cotterrell, ‘Is there a Theory of Legal Transplants?’ in David Nelken and Johannes Feest (eds), Adapting Legal Cultures (Hart Publishing, 2001) 73.

The adoption of specialised competition tribunals in Latin American countries

113

of finding solutions to their own problems by adopting what has worked in parallel environments.39 The advisability of transplants has been the source of endless debate. In Montesquieu’s view it was implausible that the law of one country might fit into the legal system of another,40 while Bentham advocated for the transfer of one-size-fits-all,41 the same position demonstrated by Watson, who considered the extent of Roman law as the best way to exemplify how institutions can easily be adopted,42 under the assumption that ‘legal rules move easily and are accepted into the system without too great difficulty … even when the rules come from a very different kind of system’.43 This was further developed by theories claiming that law succeeds irrespective of the society in which it operates, or that law and rules are portable and autonomous, and can therefore be transplanted.44 Legrand, who estimated that the law is the product of culture-specific determinations, rejected Watson’s proposition, claiming that once the law is moved to a different host culture the meaning changes immediately, making it impossible to transplant.45 The same rejection has been shown by Cotterrell, who has expressed that it is not possible to find logic in legal transplants as the concept itself is unclear, complex, and the variables to measure success or failure too numerous and often undefined, but most importantly because communities show different social relations that the law cannot harmonise or integrate entirely.46 Nelken stated that in order to avoid rejection of a legal transplant, it was essential to appreciate the social and political context in which law was shaped.47 Middle-way approaches have been considered, where the study of certain conditions has been contemplated to foresee transferability, first 39 Jerneja Penca, ‘Transnational Legal Transplants and Legitimacy: The Example of “Clean” and “Green” Development Mechanisms’ (2016) 36 Legal Studies 4, 706. 40 Marina Kurkchiyan, ‘What to Expect from Institutional Transplants? An Experience of Setting Up Media Self-regulation in Russia and Bosnia’ (2012) International Journal of Law in Context, 115–116. 41 Couyoumdjian (n. 1) 497; Juan Couyoumdjian, ‘An Expert at Work: Revisiting Jeremy Bentham’s Proposals on Codification’ (2008) 61 Kyklos 4, 505. 42 Kurkchiyan (n. 40) 117. 43 Alan Watson, Legal Transplants: An Approach to Comparative Law (University of Georgia Press, 1993). 44 David Nelken, ‘Toward a Sociology of Legal Adaptation’, in David Nelken and Johannes Feest (eds), Adapting Legal Cultures, (Hart Publishing, 2001) 7–8. 45 Pierre Legrand, ‘The Impossibility of “Legal Transplants”’ (1997) 4 Maastricht Journal of European and Comparative Law 2, 114, 117. 46 Roger Cotterrell and Austin Sarat, Law, Culture and Society: Legal Ideas in the Mirror of Social Theory (Ashgate Publishing Limited, 2006) 8, 108, 116. 47 Nelken (n. 44) 10.

114

Challenges to assumptions in competition law

looking just at the similarities, then focusing on the divergences, and then evaluating both.48 Later, a functionality theory was introduced, according to which the origin of the transplant is unimportant as long as it can serve well to address the same social need, underlining the importance of choosing a suitable case study.49 Likewise, with the aim of explaining transplant processes, a typology was developed based on the reasons that motivate them. First, the Cost-Saving Transplant, whose main driver is to save money and time when finding a solution to a new problem.50 Second, the Externally Dictated Transplant, more common in developing countries and intended to satisfy foreign entities pursuing trade agreements, financial aid or political autonomy.51 Third, the Entrepreneurial Transplant, where the motivation resides in the opportunity to gain expertise by adopting a foreign system, as well as to obtain political or economic benefits.52 Fourth, the Legitimacy-Generating Transplant, which occurs due to the prestige of foreign models, common in developing countries suffering from weak institutions where the missing authority is sought via the act of importing such models.53 The use of transplants is common among developing countries, where their rejection has also been common.54 Whether the transplantation is voluntary or forced, if it is cost-saving or builds legitimacy, it has been claimed that the foundation of the problem is that the incompatibility between imported formal institutions and the prevailing informal institutions in the receiving country is what hinders the levels of acceptance.55 It is considered that local conditions

48 Helen Xanthaki, ‘On Transferability of Legislative Solutions: The Functionality Test’, in Constantin Stefanou and Helen Xanthaki (eds), Drafting Legislation – A Modern Approach (Ashgate Publishing Limited, 2008) 2; Helen Xanthaki, ‘Legal Transplants in Legislation: Defusing the Trap’ (2008) 57 The International and Comparative Law Quarterly 3, 661. 49 Xanthaki, ‘On Transferability’ (n. 48) 3; Xanthaki, ‘Legal Transplants’ (n. 48) 673. 50 Jonathan Miller, ‘A Typology of Legal Transplants: Using Sociology, Legal History and Argentine Examples to Explain the Transplant Process’ (2003) 51 The American Journal of Comparative Law, 845–846. 51 Ibid. 847–849. 52 Ibid. 849–851. 53 Ibid. 854–857. 54 De Jong et al. (n. 7) 278–279. 55 Nils Goldschmidt and Joachim Zweynert, ‘The Two Transitions in Central and Eastern Europe as Processes of Institutional Transplantation’ (2006) XL Journal of Economic Issues 4, 906.

The adoption of specialised competition tribunals in Latin American countries

115

play a more important role during the transplant process than the conditions of the donor.56 When examining the different theories around transplants, the approach according to which transplants are unthinkable loses validity when we consider that an effective way for countries to find solutions to their needs is by learning from the lessons offered by their counterparts, and by adapting them. In the same way, the one-size-fits-all approach, where the assumption is that blueprints work in every case, is idealistic considering that every country has a unique set of socio-economic conditions, and without a critical consideration about what is needed and how it is needed, the chances of success are reduced. The middle-way approaches seem to be more realistic, as the analysis of similarities or divergences, or both, as well as the examination of the suitability of the case study, is essential to make the replication feasible. Given that one of the purposes of this chapter is to offer some guidance about the transferring of lessons, the first thing that countries need to do is to establish whether Mexico is a suitable case study to satisfy their needs, the second is to determine their similarities or divergences, and the third is to adapt the lessons appropriately. Such adaptation requires compatibility between the prevailing informal institutions, such as moral norms, customs and culture, and the new adopted formal institution, in this case a specialised competition tribunal. Particularly in the catching-up period, the adaptation of a foreign formal institution or a foreign lesson tends to be a rapid process, whereas informal institutions change at a slow pace, resulting in an incompatibility between them.57 For this reason, an understanding of how institutional change happens is essential, as Section 5.1.3 now shows. 5.1.3

Institutional Change

This section shows that ignoring prevailing informal institutions endangers the levels of acceptance of transplanted formal institutions, in this case a specialised competition tribunal. At the same time, it explores the pace at which new formal transplanted institutions take root and why it is important to marry this with the pace at which prevailing informal institutions change. For this purpose, this section starts by briefly exploring the theories whose emphasis was on rational and collective choices (formal institutions) relegating informal rules (informal institutions) as drivers of change, moving to evolutionary ones 56 Miller (n. 50) 845; Pistor Berkowitz and J. Richard, ‘Economic Development, Legality, and the Transplant Effect’ (2003) 47 European Economic Review 1, 165–168, 174, 180. 57 Joachim Zweynert, ‘Economic Culture and Transition’ (Springer, 2006) 41(4) Intereconomics, 182.

116

Challenges to assumptions in competition law

where collective choices were inconceivable, and ending with a new movement where collective choices and informal rules are combined as drivers of change. There follows a summary of the most important clusters of research in this area: (a) Collective choice; (b) Evolutionary theories; and (c) New Institutional Economics. a Collective choice theories Collective choice theories suggest that the rationality of individuals, organisations, communities or the state promotes a change of rules aspiring to obtain their own benefit.58 Some theories assign different roles to political actors, the judiciary, or third parties as drivers of institutional change.59 Collective choice theorists have considered the informal rules that can be collectively adopted, such as ethical codes or moral norms, as drivers of change. Informal norms, which cannot be rationally and collectively evaluated, and whose evolvement is spontaneous and decentralised, such as social norms and conventions (moral, ideologies, culture), have been ignored by such theorists.60 This collective choice approach rightly holds that individuals or communities promote institutional change but it fails to incorporate social conditions such as culture as promoters of it. The theory assumes that culture will also be ignored when the transplant of a formal institution is contemplated. As a result, the compatibility gap between this informal institution and the new formal entity will be considerable, risking the levels of acceptance of the latter. For the purpose of this chapter, the compatibility gap suggests that if the specific social conditions of the countries contemplating the adoption of a specialised competition tribunal are not considered, then the risks of rejection of the incorporation of such a formal institution will be higher. Evolutionary theories b The core of this cluster of theories centres on the Darwinian principles of variation, selection, and inheritance, as follows: where the change appears unintentionally or deliberately but in any case uncoordinated ‒ variation; where the successful institutions remain and the unsuccessful disappear – selection; and where the replication derives from learning, imitation, and experimentation – inheritance.61

Kingston and Caballero (n. 4) 155–157. Ibid. 157–158. 60 Ibid. 158–160. 61 Ibid. 58 59

The adoption of specialised competition tribunals in Latin American countries

117

Evolutionary theories include informal rules as initiators of change but they are not seen as a product of collective action or political process.62 The most important contribution of this theory is that informal rules are contemplated as initiators of change, but the theory does not explain why and how change happens from an uncoordinated process. It fails also to elaborate on why unsuccessful institutions remain, despite the apparent selection process, failing to justify the necessity of transplants if naturally unsuccessful institutions disappear, as in this case, an inefficient generalist court. Most recently, a new approach has emerged, the New Institutional Economics, rooted in the interaction between formal rules (usually purposefully designed according to collective choice theories) and informal rules (usually spontaneously developed according to the evolutionary theories). This theory is explored in more detail next. c New Institutional Economics According to Williamson, exogenously informal rules determine to what extent formal rules are embedded, according to four levels of institutions and the speed at which their rules change. Thus, it takes centuries or millennia for culture and norms to change (highest level – informal institutions), decades or centuries to change constitutions, laws, and property rights (second level – formal rules), years to change contractual relations (third level – institutions of governance), and a short time to change prices and quantities in individual contracts (lowest level).63 Subsequently, North advanced the recognition that informal rules are the core of institutional change, which coexist and evolve simultaneously with formal rules. He recognised that formal rules are the result of a political process (collective choice theories) fundamentally stimulated by informal rules, whose evolution responds to a process of cultural transmission (evolutionary theories).64 In relation to how quick institutions change, North’s view is that institutional change is the product of an enlargement of small changes; he saw change as gradual,65 and path-dependent, as institutions evolve at the same time as mental models.66 Finally, Roland has argued that formal institutions change quickly and deliberately as products of a centralised political process called ‘fast-moving’, while informal institutions change slowly, responding to a continuous, evolu-

Ibid. 167. Ibid. 167–168. 64 Ibid. 168. 65 Ibid. 66 North and Denzau (n. 13) 22. 62 63

118

Challenges to assumptions in competition law

tionary and decentralised process that is ‘slow-moving’.67 Institutional change according to Roland is the result of an upheaval of informal institutions which speeds up the change of formal institutions.68 The New Institutional Economics doctrine serves the purpose of this chapter by providing a valuable analysis that helps explain why it is important to find compatibility between the prevailing informal institutions of the recipient country and the new transferred formal institutions. The first important input is the recognition of informal institutions as drivers of change. The second is the identification of informal institutions as facilitators of embedment of formal rules. The third is the acknowledgement of the pace at which informal and formal institutions change ‒ the former very slowly, and the latter rather fast. The last input is that informal institutions are more likely to endure than formal institutions. From these deductions, it is evident why ignoring prevailing informal institutions endangers the levels of acceptance of transplanted formal institutions, in this case a specialised competition tribunal, as well as why it is important that the pace at which new formal transplanted institutions change encompasses the pace of prevailing informal ones. Section 5.1.4 studies in more detail the concept of culture in an attempt to provide a better understanding of informal rules as facilitators of embedment. 5.1.4 Culture This section examines the concept of culture in order to specifically show how intrinsic this factor is to human behaviour and to appreciate why different cultures respond differently to the same incentive, in this case, the adoption of a specialised competition tribunal, which in some cases leads to acceptance and in others rejection. The study of the relevant literature suggests that culture plays an important role in the process of accepting imported formal institutions and presents the link between the various theories and the adoption of institutions. The concept of culture has been the object of much academic study, with the assumption that the study of culture is indispensable for understanding human behaviour.69 It has been argued that culture represents those customary

Roland (n. 14) 110. Kingston and Caballero (n. 4) 168–169. 69 Adam Jr Gifford, Reviewing: Robert Boyd and Peter Richerson, ‘Not by Genes Alone: How Culture Transformed Human Evolution’ (2008) Springer Science+Business Media, 193. 67 68

The adoption of specialised competition tribunals in Latin American countries

119

beliefs and values that people acquire.70 In explaining how the acquisition of these beliefs and values occurs, four main streams of research are relevant: (1) Standard Social Science Model (SSSM); (2) Evolutionary Psychology (EP); (3) cultural evolution; and (4) evolutionary anthropology. According to the Standard Social Science Model, culture and social behaviour are a product of learning with no genetic contribution, while for the evolutionary psychologist the brain has a large number of innate modules stimulated by specific social and environmental circumstances, which result in human behaviour.71 F. A. Hayek, one of the main exponents of cultural evolution, contends that culture is not genetically transmitted nor rationally designed but is a process whereby culture and human reasoning develop concurrently.72 Successively, evolutionary anthropology’s main contribution has been the proposition that there is an ontogenetic relation between biological and cultural evolution, where new forms of social cognition trigger new forms of cultural learning that in turn bring new cumulative cultural evolution, so social learning is the critical part of cultural development.73 This social learning is the result of an imitative process, rather than the emulative one used by some animals, that allows individual development at a social and cognitive level, in which individuals learn how to act and think. This social environment has been referred to as culture, and its individual embedment as ‘enculturation’.74 Regarding the insights of evolutionary anthropology, it appears that culture is an intrinsic part of the individual, whose function is to shape their behaviour within society from generation to generation, in which the society is a reflection of the subject. Hence, as Becker has affirmed ‘Individuals have less control over their culture than over other social capital … culture is largely a “given” to individuals throughout their lifetimes’.75 North has equally stated that: ‘culture, human evolution and economic development are closely interrelated’.76 The analysis of these theoretical insights offers different explanations for how culture is generated and diffused. While the explanations are diverse, the concurrent theme is that such a process is complex, deep-rooted, imposed, and

70 Ibid.; Luigi Guiso, Paola Sapienza and Luigi Zingales, ‘Does Culture Affect Economic Outcomes?’ (2006) 20 Journal of Economic Perspectives 2, 23. 71 Gifford (n. 69) 194–195. 72 Friedrich Hayek, Law, Legislations and Liberty. A New Statement of the Liberal Principles of Justice and Political Economy (Routledge, 1982) 71. 73 Goldschmidt (n. 14) 180. 74 Ibid. 75 Guiso et al. (n. 70) 24. 76 Goldschmidt (n. 14) 179.

120

Challenges to assumptions in competition law

makes individuals distinct. Accordingly, Richerson and Boyd have claimed that individuals from different cultures respond to the same incentive in different ways.77 In the context of the value of adopting a specialised competition tribunal, it follows that the fundamental deduction is that culture should not be neglected when envisioning such adoption, and is important in terms of aligning the imported formal institution with the prevailing informal institutions to avoid rejection. After considering culture as a factor of successful transplantation, the next section explores the most common designs of institutional transplants. 5.1.5

Design of Institutional Transplants

Two contrasting perspectives have been identified in the design of how transplantation will occur. The first is known as the ‘actors pulling in’ argument, where the main attention is given to the political actors who will shape the conditions of the new institutions according to their preferences, and where the analysis of compatibility is less important.78 The second is known as the ‘goodness of the fit’ argument, where the precept is that only institutions compatible with the culture of the country to which they are being transferred can be transplanted.79 In addition, an important stream of research has explored three different types of transplantation design and has accordingly constructed a ‘stickiness’ theory referring to degrees of success. The first type is foreign introduced exogenous (FEX), where the design is constructed and imposed by outsiders. The second is indigenously introduced endogenous (IEN), where the institutions emerge spontaneously grounded in metis. The third is the indigenously introduced exogenous (IEX), where the institutions are indigenous but externally introduced by a formal‒local authority.80 Metis refers to local knowledge, gained through experience and practice, which is not written down (locals’ norms or customs).81 According to this approach, metis is the glue that gives institutions stickiness, so the highest level of success is more likely to be obtained when metis is incorporated into

Gifford (n. 69) 193–194. De Jong et al. (n. 7) 184. 79 Zweynert (n. 57) 184. 80 Peter Boettke, Christopher Coyne and Peter Leeson, ‘Institutional Stickiness and the New Development Economics’ (2008) 67 American Journal of Economics and Sociology, 2, 332–335. 81 Ibid. 338. 77 78

The adoption of specialised competition tribunals in Latin American countries

121

the design of the new institutions. In other words, neglecting the analysis of metis creates incongruities between new institutions and the local conditions.82 Similarly, another distinction has been made between two types of design. The ‘blueprint’ where the design consists of transplanting institutions from more advanced economies regardless of the local conditions of the receiving country, and the ‘tacitness’, in which tacit local knowledge is fundamental for the development of institutions, being envisioned at a local level only.83 It has been implied that these designs are unworkable separately, it being necessary for them to converge via harmonisation of local institutions with those from abroad.84 As the previous section has revealed, culture is a unique intrinsic factor of human development and social interaction and means that different societies respond differently to the same incentive. For this reason, in the context of designing courts that review the decisions of competition authorities, culture should not be ignored. Instead, designs which consider local conditions, regardless of their labelling (metis, tacitness or indigenous), should be appraised when evaluating the viability of the transference of lessons related to the adoption of specialised competition tribunals in order to ensure a well-functioning judiciary. Thus, the possibility to emulate, whether as an internal intention to catch-up with development or as an external imposition aimed at obtaining assistance, deserves meticulous scrutiny to avoid unexpected outcomes such as resistance, rejection, or conflicts between prevailing informal institutions and transported competition tribunals. The following recommendations are essential with a view to avoiding such pitfalls: • The advance analysis of the compatibility between imported competition tribunals and local informal institutions.85 • The consideration of multiple court models instead of just one.86 • The involvement of domestic policy actors with the new institutional transplant.87 • The recognition of personal, financial, and physical opportunities or limitations, as well as the knowledge of the specific regulations of the field.88

Ibid. 338–340, 344; Couyoumdjian (n. 1) 493. Dani Rodrik, ‘Institutions for High-Quality Growth: What They Are and How to Acquire Them’ (2000) National Bureau of Economic Research 7540, 15. 84 Ibid. 16. 85 De Jong and Stoter (n. 14) 321–322; Roland (n. 14) 120. 86 Ibid. 322. 87 De Jong and Stoter (n. 14) 1066. 88 Ibid. 1065. 82 83

122

Challenges to assumptions in competition law

In addition, Rose has recommended the following: (1) clear and objective purposes; (2) a single goal; and (3) a simple design.89

CONCLUSIONS It has been observed that institutions have been created to bring order to society. In this way, the judiciary is an important institution designed to bring order at a different level, whose acceptance and legitimacy depend on its effectiveness. It has also been observed that the transplantation phenomenon is a complex one whose advisability and effectiveness remain debatable, but that despite the unsettledness, the use of transplants is very common, particularly among developing countries. The success or failure of transplants is difficult to predict, although it has been affirmed that the crux of the problem emerges when there is no compatibility between the prevailing informal institutions and the new formal transplanted institutions. Seeking to understand why it is important to make formal and prevailing informal institutions compatible, it was necessary to understand how institutions change and at what pace. It was found that informal institutions are generators of change, that they constitute an essential factor that helps embed formal institutions, and that formal institutions may change quickly, while institutional change remains slow. The study of the concept of culture was relevant to the underlying chapter by showing that if this factor were not included when contemplating the adoption of a specialised competition tribunal, the risks of rejection would be higher. In terms of design of transplants, it has been seen that informal local institutions give stickiness to formal institutions, that it is more likely that a formal institution compatible with the culture of the recipient can be transplanted, and that, if close attention is not paid to local needs and efforts are not made to harmonise intended formal institutions with prevailing informal ones, the levels of legitimacy and effectiveness will be compromised, which could lead to the emergence of what has been referred to as ‘institutional traps’.90 The purpose of this chapter has been to present the theoretical framework of transplants of institutions and so offer an overview of its complexity, and to grasp the core components that need to be deliberated to avoid rejection of a transplant, an analysis that was required to understand the complexity of the process, to anticipate possible difficulties, and to establish some recommendations that would allow the best outcomes when emulating lessons.

Rose (n. 2) 10. Goldschmidt and Zweynert (n. 55) 907.

89 90

The adoption of specialised competition tribunals in Latin American countries

123

If countries do intend to emulate the type of judiciary adopted in Mexico, it is indispensable to evaluate thoroughly local needs, prevailing informal institutions, and local opportunities as well as local limitations, in order to encourage participation from local political actors and to ensure good levels of acceptance.

6. Where influence lies in the International Competition Network Christopher Townley1 6.1 INTRODUCTION It ‘is an important and growing’2 phenomenon that firms’ manufacturing and distribution chains often affect several countries. The same applies to any anti-competitive arrangements that they adopt. Over 120 legal systems now have competition rules. This number exploded in the 1990s, accompanied by many competition agencies (National Competition Authorities (NCAs)).3 A confusing array of substantive and procedural competition rules ensued. This competition soup can raise costs for those operating in several countries (and their consumers). It can also generate democratic and public interest benefits, as each jurisdiction tries to protect the values it holds dear.4 Some have sought influence through extra-territorial application of their competition rules. Another approach is to develop international rules. Sometimes through bilateral discussions, such as between the USA and the EU;5 or consider the (unsuccessful) attempts to develop international competition rules.6 Many international organisations also occupy the competition 1 My thanks to ASCOLA for their wonderful 2019 conference and encouraging this publication. Thanks too to Mattia Guidi and Mariana Tavares. 2 John Fingleton, ‘Competition Agencies and Global Markets: The Challenges Ahead’, in Paul Lugard (ed.), The International Competition Network at Ten: Origins, Accomplishments and Aspirations (Intersentia 2011), 174–175. 3 Jacint Jordana, David Levi-Faur and Xavier Fernández-i-Marín, ‘The Global Diffusion of Regulatory Agencies Channels of Transfer and Stages of Diffusion’ [2011] 44 Comparative Political Studies 1343. 4 Christopher Townley, A Framework for European Competition Law: Co-ordinated Diversity (Hart Publishing 2018), chapter 1. 5 Bruno Zanettin, Cooperation between Antitrust Agencies at the International Level (Hart Publishing 2002). 6 Merit Janow and James Rill, ‘The Origins of the ICN’, in Paul Lugard (ed.), The International Competition Network at Ten: Origins, Accomplishments and Aspirations (Intersentia 2011), 26.

124

Where influence lies in the International Competition Network

125

arena (such as the Organisation for Economic Co-operation and Development (OECD) and the United Nations Conference on Trade and Development (UNCTAD)). However, some claim they both had limitations (the OECD did not include many countries with competition laws; and UNCTAD’s work can be too vague and pitched at the lowest common denominator).7 These (and other) initiatives encouraged dialogue between NCAs, but this did not always remove tensions; think of the aborted mergers of GE/ Honeywell or Boeing/ McDonnell Douglas. Although infrequent, such tensions could be high profile. In 1997, the US Attorney General established the International Competition Policy Advisory Committee (ICPAC), to examine ‘multijurisdictional merger review; the interface of trade and competition issues; and future directions in enforcement co-operation between US antitrust authorities and their counterparts around the world’.8 ICPAC delivered its report in February 2000. It sought to provide an environment conducive to the expansion of international trade, to be tolerant to different nations, and to enhance world welfare. It also sought to remove barriers to market access stemming from anti-competitive private trade restrictions. To do this, it advised pushing states to apply their competition rules in a non-discriminatory way, to ignore non-competition factors in merger review (with limited exceptions), and to establish independent NCAs. In order to better pursue these aims, ICPAC suggested that the USA, amongst other things, consider creating a global competition network where: government officials, private firms and non-governmental organisations could consult on antitrust matters. ICPAC recommended [aiming for]… ‘greater convergence of competition law and analysis, common understanding, and common culture.’9

The ICPAC Report was influential. In February 2001 the International Bar Association organised a meeting for many of the world’s senior competition officials and practitioners, to support the creation of ‘a new organisation directed exclusively at international antitrust enforcement’.10 In October 2001, NCAs ‘from 14 jurisdictions launched the International Competition Network

7 Hugh Hollman and William Kovacic, ‘The International Competition Network: its Past, Current And Future Role’, in Paul Lugard (ed.), The International Competition Network at Ten: Origins, Accomplishments and Aspirations (Intersentia 2011) , 68–69. 8 ICPAC, Final Report (2000), Executive Summary, 1. 9 http://​i​nternation​alcompetit​ionnetwork​.org/​about/​history​.aspx accessed 23 March 2020. Also, ICPAC, Final Report (n. 8) Executive Summary, 23, 29. 10 http://​i​nternation​alcompetit​ionnetwork​.org/​about/​history​.aspx accessed 17 December 2020.

126

Challenges to assumptions in competition law

(ICN) as an international forum of enforcement authorities exclusively devoted to competition issues’.11 Some see policy networks, like the ICN, as a good way of dealing with transnational problems, in ways that supplement normal national democratic institutions.12 They can enhance states’ ability to influence what happens beyond their borders. Networks are often portrayed neutrally, but they are political.13 They facilitate and constrain actors and outcomes; creating winners and losers, depending upon how they are designed and how they are used. When we think of ICN influence, two processes are relevant. First, some might try to influence ICN outputs. Secondly, ICN outputs might influence the regulatory processes of its members, and others. There is a wider question about the legitimacy of these two processes too. This chapter describes the first process. Section 6.2 briefly introduces the ICN. Section 6.3 describes who is present in the ICN. Section 6.3.1. notes that NCAs, not states, are members and asks whether it matters. Section 6.3.2 discusses the NCAs that are present in the ICN (and some that are not) and asks why this is. Section 6.3.3 asks similar questions about non-governmental actors (NGAs). Section 6.4 analyses how some NCAs and NGAs seem to have disproportionate influence over the ICN’s agenda and work product, in an organisation where all NCAs formally have an equal vote and NGAs do not have a vote at all. Section 6.5 concludes. This chapter is part of a larger book project that also looks at the second process (ICN influence on states’ regulatory processes) and considers the ICN’s efficiency, effectiveness and legitimacy claims.14 The ICN’s formally neutral structures, on closer examination, reveal powerful mechanisms which facilitate strong NCAs (from a narrow group of principally, rich, ‘Western’

11 ICN, ICN Statement of Achievements 2001–2013 (2013), 3. The jurisdictions involved were: Australia, Canada, European Union, France, Germany, Israel, Italy, Japan, Korea, Mexico, South Africa, United Kingdom, United States, and Zambia. Half are in North America or the European Union. However, there are other jurisdictions, some with established competition systems, and some newer regimes. 12 R. A. W. Rhodes, Understanding Governance: Policy Networks, Governance, Reflexivity and Accountability (Open University Press 1997) 9. 13 Ibid, 11, 160. In relation to the European Competition Network, Wilks and Sturm call networks political structures. Schaub and Tesauro protest: Claus Dieter Ehlermann and Laraine Laudati (eds), European Competition Law Annual 1997: Objectives of Competition Policy (Hart 1998) 15, 34–35, 77–79, and 172–173. 14 Christopher Townley, Mariana Tavares and Mattia Guidi, The Law and Politics of Global Competition: Influence and Legitimacy in the International Competition Network (Oxford University Press Forthcoming, 2021).

Where influence lies in the International Competition Network

127

states)15 and NGAs from these same states, to influence the weak; and grant ‘technical competition experts’ influence over general state interests. This, largely undiscussed area, has major implications for the ICN. It also has major implications for those participating (or those currently considering participating) there.

6.2

THE ICN: ESTABLISHMENT, MISSION, STRUCTURE AND HOW IT WORKS

By 2013 the ICN had grown to 128 NCAs from 111 jurisdictions.16 It is the most extensive NCA network worldwide, including many influential NCAs from the USA and the EU. NCAs accede by letter to the ICN ‘Mission and Activities’.17 All NCAs enjoy the same membership rights and privileges.18 NGAs, such as major corporates, law firms and academics, also play a role (although NGAs cannot be ICN members). International organisations, such as UNCTAD and the OECD, are often involved in ICN work too. The ICN focuses exclusively on competition law and policy, both substance and procedure.19 It describes its mission as follows (note the similarity to the ICPAC Report, above): to advocate the adoption of superior standards and procedures in competition policy around the world, formulate proposals for procedural and substantive convergence, and seek to facilitate effective international cooperation to the benefit of member agencies, consumers and economies worldwide.20

These aims have not formally changed much over time. The two key ICN ‘institutions’ are the Steering Group and the Working Groups. The Steering Group’s powers include: selection of ICN Chair, and three vice chairs, from amongst its members; approving new NCAs as ICN members; identifying subjects of potential interest to ICN members and then establishing Working Groups in areas of need, selecting Working Group Chairs, and reviewing and approving Working Group work; approving draft recommended practices from the Working Groups (once the Steering Group approves a recommended practice it is sent to the ICN Annual Conference for approval by all NCAs); inviting (and approving) international organisations to 15 Wendy Ng, ‘Changing Global Dynamics and International Competition Law: Considering China’s Potential Impact’ (14th ASCOLA Conference) 5. 16 ICN, ICN Statement of Achievements 2001–2013, 3. 17 ICN, ICN Operational Framework (2012), 1. 18 Ibid, 2. 19 ICN, ICN Factsheet and Key Messages (2009), 1. 20 ICN, ICN Statement of Achievements 2001–2013, 3.

128

Challenges to assumptions in competition law

contribute to ICN activity. The Steering Group also deals with ICN finances and controls the hosting of ICN events.21 Working groups also do a lot of ICN work. The current substantive working groups are: advocacy, agency effectiveness, cartels, mergers and unilateral conduct.22 Working Groups’ powers include: determining the nature and ambit of their work, within the Steering Group’s mandate; drafting work plans to be discussed and agreed at the Steering Group; inviting others (including, NGAs) to support their work; allocating leadership tasks and drafting (to NCAs or NGAs) for their work; drafting and approving recommended practices in their work plans (subject to final approval by the Steering Group and then Annual Conference); arranging teleconferences and workshops in their areas.23 In terms of how the ICN works and how it says that it will come to ‘decisions’, the ICN is a: project-oriented, consensus-based, informal network of antitrust agencies from developed and developing countries that will address antitrust enforcement and policy issues of common interest and formulate proposals for procedural and substantive convergence through a results-oriented agenda and structure.24

The ICN relies on the goodwill of NCAs (and NGAs), and their staff. There is an annual conference for all NCAs (and some NGAs). The annual conference and some workshops are held in person. To keep costs down, most other meetings are held either through telephone conferences, or online. The ICN does not have a rule-making function. It makes recommended practices (general suggestions on optimal enforcement of competition rules) 21 ICN, ICN Operational Framework (2012), 2–3; http://​www​.i​nternation​ alcompetit​ionnetwork​.org/​uploads/​library/​doc608​.pdf, 6; http://​www​.i​nternation​ alcompetit​ionnetwork​.org/​uploads/​library/​doc994​.pdf, 2; http://​www​.i​nternation​ 1; http://​www​.i​nternation​ alcompetit​ionnetwork​.org/​uploads/​library/​doc995​.pdf, alcompetit​ionnetwork​.org/​uploads/​library/​doc996​.pdf, 2; http://​www​.i​nternation​ alcompetit​ionnetwork​.org/​uploads/​library/​doc998​.pdf, 2; http://​www​.i​nternation​ alcompetit​ionnetwork​.org/​uploads/​library/​doc1001​.pdf. Other ICN bodies deal with things like planning the annual conference. Please note that since the original research work the ICN seems to have fundamentally redesigned its website, changing the page links and removing many documents that were originally available. Where possible, I have replaced old links with new ones. However, where I have been unable to find the new links for documents, I have left the old ones as this can help readers locate documents in cached sites. 22 Others have closed: antitrust enforcement in regulated sectors (2003–2005), and telecommunications (2005–2006). There are also Operational Working Groups, for example, dealing with membership. 23 ICN, ICN Operational Framework (2012), 5. 24 ICN, Memorandum on the Establishment and Operation of the International Competition Network (2002), 1.

Where influence lies in the International Competition Network

129

where agreed through consensus by all NCAs in the ICN Annual Conference. Even then, individual NCAs: ‘decide whether and how to implement’25 them.

6.3

WHO HAS INFLUENCE IN THE ICN (AND WHY THEY WANT IT)?

Political theorists underline policy networks’ influence through developing new ideas or legitimising certain courses of conduct. Schmidt and Radaelli say that networks are particularly influential where their discourse ‘contains cognitive arguments that demonstrate the policy programme’s relevance, applicability and coherence: and normative arguments that resonate with long-standing or newly emerging values’.26 Wilks applies this to competition networks, saying that the: programme of activities ‘intended to foster the creation of a common competition culture’… matches the discourse concept rather neatly. It is primarily coordinative, it embraces a cognitive framework drawing on competition law and economics, and it has a strong normative thrust in pursuit of the perceived benefits of competition.27

The composition of those present in the ICN (and those who are not) is important in terms of who has (the potential for) influence there. States, or at least their NCAs, are key actors. Section 6.3.1 asks whether it matters that the ICN is an organisation of competition authorities rather than states. Then, Section 6.3.2 looks at which NCAs are present in the ICN (and which are not) and ask why this might be. At this point, the ICN may look like an organisation made up of a balanced group of NCAs with near global reach. However, NCAs are not the only actors that are keen to be involved. In fact, the ICN has always been open to a group of private actors too. In the ICN’s own words: While it is primarily a network of competition authorities, the ICN has always sought to reach out to the entire competition community and take on additional Ibid, 1. Vivien Schmidt and Claudio Radaelli, ‘Policy Change and Discourse in Europe: Conceptual and Methodological Issues’ [2006] 27 West European Politics, 188. 27 Stephen Wilks, ‘Agencies, Networks, Discourses and the Trajectory of the European Competition Enforcement’ [2007] 3 European Competition Journal 437, 430 (discussing proposals for a European Competition Network; similar points apply to the ICN). Similarly, Sonja Wälti, ‘Multi-Level Environmental Governance’, in Henrik Enderlein, Sonja Wälti and Michael Zürn (eds), Handbook on Multi-Level Governance (Edward Elgar Publishing 2010) 416; Frans van Waarden and Michaela Drahos, ‘Courts and (Epistemic) Communities in the Convergence of Competition Policies’ [2002] 9 Journal of European Public Policy 913, 933. 25 26

130

Challenges to assumptions in competition law

expert advice through the engagement of Non-Governmental Advisors (‘NGAs’) … NGAs are private practitioners, representatives of non-governmental international organizations, members of industry and consumer groups, and academics, who volunteer to contribute to the ICN’s work.28

So, Section 6.3.3 charts the involvement of NGAs, breaking this down by the types of actor, the geographic origin of these actors, and their area of expertise. Many NGAs are multinational firms (and their advisors). They often hail from rich, ‘Western’ states and favour competition rules founded on a narrow, neo-liberal, paradigm. Like NCAs, NGAs claim to be acting in the public interest. However, this group of actors has the ability and incentive to influence ICN work products in their own self-interest. 6.3.1 States Under the ICN’s operational framework, only a ‘national or multi-national competition agency’ can be an ICN member.29 This is unlike some other international organisations, such as the OECD, for example, where the states themselves are members, even when competition law and policy is being discussed. According to the ICN: Government officials and members of the antitrust bar recognized that the best way to promote sound and effective antitrust enforcement in the wake of economic globalization was through the establishment of a network of competition authorities and international competition specialists.30

Some believe that restricting the ICN to a network of NCAs and NGAs, excluding lawmakers and policy-makers, is what made the ICN so effective in its first ten years.31 NCAs may have more common interests and want to focus on competition policy. The more similar those involved are (in terms of interests, outlook and sense of problem) the more quickly and easily they are

ICN, NGA Toolkit (2012), 1. ICN, ICN Operational Framework (2012), 1. 30 https://​www​.i​nternation​alcompetit​ionnetwork​.org/​about/​ last accessed 24 March 2020. 31 Pieter Kalbfleisch, ‘From Ditchley Park to the Hague: A Journey Worthwhile’, in Paul Lugard (ed.), The International Competition Network at Ten: Origins, Accomplishments and Aspirations (Intersentia 2011) 4. 28 29

Where influence lies in the International Competition Network

131

likely to find common ground. Von Finckenstein (founding Chairman of the ICN) explains: The ICN formula works because: participants are involved as experts, not as representatives of governments; discussions focus on the production of optimal practices without the immediate concern for implementation; we are also able to avoid defensive posturing which leads to a lowest common denominator approach; public and private sectors, and member agencies from developing and developed economies, participate on an equal footing.32

Populating the ICN with competition experts from the NCA and NGAs, puts ‘expert argument’ front and centre. It may expedite discussions, where permissions do not have to be sought elsewhere.33 The composition of ICN membership is likely to change both the way that the ICN works and its outputs. There are two reasons why those that established the ICN might not have wanted to involve states. First, discussions between NCAs present differences as ‘technical’, rather than political. This is likely to enhance the power of ‘technically advanced’ NCAs (principally the ICN’s founding members). The ICN provides these competition wolves with opportunities to advance their industrial policy goals through soft extra-territorial advocacy of the rules they favour, by dressing them up as technically advanced solutions which benefit all. In addition, excluding states allows these powerful NCAs to more easily co-opt the ICN’s aims, agenda and recommendations for their own ends. Let us take these points in turn. Scholars have long debated under which conditions soft law (‘non-binding’ rules, such as ICN recommended practices) emerges. States often prefer one set of standards over another. There is an incentive to persuade other states to adopt their standards, because this entails no cost of adapting national regulation, and it allows that country’s firms to expand their activities in other countries at lower costs. For these reasons, powerful actors will use soft law to promote the adoption of their own standards at the expense of others, and they will comply with international soft law only as long as its prescriptions are already in line with their interests.34 How actors seek to ‘persuade’ others to change their standards is likely to be different in a network of NCAs, rather than states. Trans-governmental networks (like the ICN) involve domestic officials from different countries, specialising in a certain area. The growing importance of such networks was Konrad von Finckenstein, Introductory Remarks (2003) 3–4. Hollman and Kovacic, ‘The International Competition Network: Its Past, Current and Future Role’ (n. 7) 75. 34 Abraham Newman and Elliot Posner, Voluntary Disruptions: International Soft Law, Finance, and Power (Oxford University Press 2018), 20. 32 33

132

Challenges to assumptions in competition law

reported in the 1970s. By the 1990s, they were seen as different from normal international relations between states. Speaking about transnational networks more generally, Slaughter and Hale explain: They allow domestic officials to interact with their foreign counterparts directly, without much supervision by foreign offices or senior executive branch officials, and feature ‘loosely structured, peer-to-peer ties developed through frequent interaction rather than formal negotiation’35

A key reason for this difference is that states normally negotiate through political bargaining (or force), whereas the ‘domestic officials’ in these networks focus upon ‘expert argument’. NCAs that are particularly well endowed with the relevant expertise are likely to benefit. There are two broad types of changes that experts might seek in order to achieve the kinds of competition rules that they want (in their own or other states). There may be a need for legislative change to alter divergent national competition rules themselves. Such changes are not normally within the gift of an NCA (although the ICN helps NCAs push for local legislative change too36). However, NCAs can do a lot without involving states. Competition rules tend to be widely drafted, leaving lots of room for different interpretations. This creates opportunities for some NCAs to push for alignment by others. This likely empowers those seen as technically strong (whatever that means in this context) because they can push less technically well-resourced NCAs into following them, even when ‘technical changes’ are highly distributional.37 Differences are debated in terms of who has the ‘best technical rules’, rather than being seen as an occasion for bargaining between winners and losers. Often the question of who wins or loses (either at the private or state level) under a certain arrangement is completely ignored. So, the focus on ‘objective’ expertise expands the space for soft law interventions from a narrow set of coordination games to a much wider set of situations. The kind of ‘expert argument’ that is relevant, and how much this replaces political bargaining is linked to network members’ perceptions of their own roles. Those setting up the ICN would likely select a focus where they are 35 Kal Raustiala, ‘The Architecture of International Co-operation: Transgovernmental Networks and the Future of International Law’ [2002] 43 Virginia Journal of International Law 1, 5; Anne-Marie Slaughter and Thomas Hale, ‘Transgovernmental Networks and Multi-Level Governance’, in Henrik Enderlein, Sonja Wälti and Michael Zürn (eds), Handbook on Multi-Level Governance (Edward Elgar Publishing 2010) 358. 36 See discussion around n. 68. 37 Townley, A Framework for European Competition Law: Co-ordinated Diversity (n. 4), Chapter 3.

Where influence lies in the International Competition Network

133

particularly strong. According to Weber, when actors come together to do technical tasks, a feeling of ‘belonging together’ can emerge and communal relationships can develop.38 Identities then can change, through mutual orientation around a common project (due to similar concerns, interests or aspirations).39 Network members develop a kind of interdependence. Poorer NCAs likely start to identify with the aims of their colleagues from richer NCAs. This helps them to fit in but can make it hard to challenge the status quo in any fundamental way.40 By way of example, involving national experts in EU committees can increase the role of national interests. However, Ossege reports different outcomes where network actors’ primary self-perception is as experts with accountability to their ‘profession’ (or epistemic community), rather than being engaged in political bargaining for their state.41 In such contexts, scientists have used their specialist knowledge, to marginalise national politicians (favouring ideas from their epistemic community).42 Formally all NCAs may be equal, but the ICN provides these more technically competent ‘competition wolves’ with new opportunities to advance their industrial policy goals through soft extra-territorial advocacy of the rules they favour. In an arena which they dominate they can dress up their preferences as the most technically advanced solutions. The ICPAC Report even hints at this. When concluding on the ‘need’ for a new forum it notes that the Asia-Pacific Economic Cooperation forum (APEC) is ‘built on this recognition that “peer” pressure is capable of advancing some liberalization and harmonization of practices even without binding legal instruments’. ICPAC envisioned the ICN (as it was to become) as a body where ‘nations can usefully explore areas of

Max Weber, Economy and Society (University of California Press 1978) 41. Marie-Laure Djelic and Sigrid Quack, ‘Transnational Communities and Governance’ in Marie-Laure Djelic and Sigrid Quack (eds), Transnational Communities: Shaping Global Economic Governance (Cambridge University Press 2012), 12–13. 40 Policy networks prevent changes either on principle, or where new ideas threaten their members’ interests: Schmidt and Radaelli, ‘Policy Change and Discourse in Europe: Conceptual and Methodological Issues’ (n. 26), 203. 41 Christoph Ossege, European Regulatory Agencies in EU Decision-Making (Palgrave Macmillan 2016), 34–35 (analysing EU committees, in medicines, food safety and chemicals, but his conclusions seem relevant here). Similarly, Djelic and Quack, ‘Transnational Communities and Governance’(n. 39), 20. 42 Ossege, European Regulatory Agencies in EU Decision-Making (n. 41), 11, 40–43. Dani Rodrik, The Globalisation Paradox: Democracy and the Future of the World Economy (W W Norton & Company, Inc. 2011), 36, reports similar things from central bankers in the 1800s. 38 39

134

Challenges to assumptions in competition law

cooperation in the field of competition policy and facilitate further convergence and harmonization’.43 Section 6.3.2 demonstrates that the most advanced ‘technical’ NCAs have disproportionate influence on what is discussed in the ICN. Then, when the discussions start, they can focus attention on the issues and solutions that they consider important. At this point, there is no discussion (in the ICN) about who benefits (states or private actors); the focus is on whether these rules are the most efficient or not. As time goes on, it may be that even those that do not initially embrace such a shift, become socialised into embracing it. This is particularly likely where all NCAs (and their employees) feel part of the competition community and that they benefit from their ICN membership (Sections 6.3.2 and 6.4.3). The second reason why the NCAs that established the ICN might not want to have states involved is that it consolidates their control of ICN inputs, agenda and recommendations. There are two dimensions to this. In the OECD and UNCTAD, competition officials are often accompanied by state officials from outside the NCA (and must often circulate national contributions to each international organisation more widely within their state (and competition officials are often acting as wider state officials)). This emphasis on wider state interests impinges upon the freedom of competition officials to intervene in ways that they might like. This even applies to officials from the EU, the USA and Germany. Setting up the ICN is a way for NCAs to avoid wider control from their own states. In addition, in fora such as the OECD, NCAs’ control over the focus and content of discussions is also challenged by a powerful participation by academics and wider state officials. By shutting out/ better controlling other state actors in the ICN, NCAs can better focus the debate on matters of interest to them, even in ways that their states might dispute. For example, consider NCA responses to ICN questionnaires (inputs). Sometimes, NCAs seem to be promoting a vision of how they think things should be, rather than how they are. In 2006, the ICN’s Unilateral Conduct Working Group (UCWG) asked NCAs and NGAs questions about the aims of their unilateral conduct laws, the assessment of dominance/ substantial market power, and state-created monopolies.44 The report that followed noted that DG COMP (the EU NCA) spoke of three EU competition goals: consumer welfare, market integration and ensuring an effective competitive process as a means to other goals, and as a goal itself.45 Yet, the EU courts regularly rely ICPAC, Final Report (n. 8) 283. ICN – Unilateral Conduct Working Group, Unilateral Conduct Working Group Questionnaire (2006). 45 ICN – Unilateral Conduct Working Group, Report on the Objectives of Unilateral Conduct Laws, Assessment of Dominance/ Substantial Market Power, and State-Crated Monopolies (2007), Annex A. 43 44

Where influence lies in the International Competition Network

135

on wider public policy goals.46 Removing state scrutiny and control from the ICN, gives NCAs more leeway to create (unchecked) stories. Building on such distortions, ICN recommended practices might advocate ‘best practices’ which are highly re-distributive and not quite as widespread as they are made to appear. NCAs come under subtle pressure to ‘conform and modernise’ in line with ‘advanced’ competition systems (or risk being left behind). Competition provisions are so widely written, that it is often possible to change fundamentally the meaning of the provisions by changing one’s interpretation, without any need for legislative change. As a consequence, those within the network often have it within their gift to change fundamentally the meaning of the rules that they enforce, without seeking outside approval. I said that there are two dimensions to why NCAs might want to shut states out of the ICN. I highlighted the benefits (for NCAs) of reducing control (by their own states). In addition, framing competition discussions as technical, rather than negotiations between states, implies that the choice of ‘best practices’ is a value neutral exercise. Yet, the selection of best practices involves choices. The most ‘technically competent’ NCAs are better able to push others to accept their preferences. All of this is much easier if discussions are seen as ‘right or wrong’ rather than invoking bargaining, where everyone should be seen to benefit. Increasing actor homogeneity within the ICN also makes it easier for participants to agree best practices (it is also harder for them to rebel, as they have more to lose). Some studies suggest that the primary allegiance of many NCAs (and their employees) is changing. According to Gerber, NCAs in the EU (the EU also has a policy and enforcement network for NCAs) are now often guided by what they think is best for the EU, rather than specifically their own states ‘because they identify their own interests and/or the interests of the institutions in which they function with the “success” of European integration’.47 Similar affects seem likely in the ICN too. Section 6.4 discusses how NCAs can incentivise each other to change the way that they think and act. 6.3.2

National Competition Authorities (NCAs)

We saw ICN co-operation disarmingly aimed at proposing superior standards and procedures in competition policy; proposing substantive and procedural 46 Christopher Townley, Article 81 EC and Public Policy (Hart 2009), Part B. The UK’s NCA report was similarly partial: Christopher Townley, ‘The Goals of Chapter I of the UK’s Competition Act 1998’, in Piet Eeckhout and Takis Tridimas (eds), Yearbook of European Law, volume 29 (Clarendon Press 2010). 47 David Gerber, ‘Modernising European Competition Law: A Developmental Perspective’ [2001] 22 ECLR 122, 125.

136

Challenges to assumptions in competition law

convergence; and aiding international co-operation.48 Typically, two potential beneficiaries of ICN work are mentioned, NCAs and consumers. For example, Fingleton (a former ICN head) says that the ICN must continue to develop its governance model ‘to ensure that its work addresses the needs of competition agencies and serves the interests of consumers in each member jurisdiction’.49 Some statements point to other beneficiaries too.50 The ICN justifies its value to NCAs by highlighting the dissemination of antitrust experience (enforcement policies, analytical approaches and investigative techniques), and exchanges of recommended practices on topical issues, that it facilitates. This aids convergence between NCAs in their interpretation and application of competition rules (insofar as this is permissible in their national orders). Even where this is not possible, it leads to a better understanding of other NCAs’ actions, which facilitates better co-operation between them. The ICN is quick to note that it is of value to both established and newer NCAs. It explains that the ICN provides: a forum for newer agencies to draw on the experiences and support of more established agencies, for example as they seek to advocate a competition culture in their jurisdiction, where there may face vested local interest groups.51

The achievement of some of the aforementioned benefits can clash with achieving others. There may, for example, be a clash between benefits for different NCAs, between firms and customers, or between benefits for customer groups in different states. We are not told what hierarchy should be followed in such cases. This is likely to be affected by who has influence within the ICN. The ICN has built an incredibly wide membership of NCAs. However, mere presence in the ICN does not generate influence; activity is needed. So, let us examine activity within the ICN Steering Group and Working Groups. Influence in these two groups is less widespread. Only NCAs can be proper members of the ICN Steering Group (the same is true of the Working Groups). Many NCAs and NGAs see membership of the Steering Group a key path to ICN influence. The ICN has process and substantive rules on selecting Steering Group members.52 Outgoing Steering Group members recommend new Steering Group members. The annual conference

See reference at n. 28. John Fingleton, The International Competition Network: Planning for the Second Decade (2010), 8. 50 ICN, A Statement of Mission and Achievements up until May 2006 (2006), 1 (reducing unnecessary costs and uncertainty for firms). 51 ICN, ICN Factsheet and Key Messages (2009), 7; see also 3. 52 ICN, ICN Operational Framework (2012), 3. 48 49

Where influence lies in the International Competition Network

137

confirms this selection, by consensus, often with little discussion. In terms of substantive membership rules, all elected Steering Group members should meet the following criteria: (a) they must be able to make a significant resource commitment to the ICN, mindful of their relative size; and (b) they must have made consistent and effective ICN participation, which may include, but is not limited to: Working Group co-chair experience; leadership of an ICN time-specific or on-going project; or a substantial contribution to the overall work of the ICN. In addition, the Steering Group composition should be representative, in terms of ‘reflecting the geographic and economic diversity of ICN members’. One can understand the logic here. But, once again, these rules embed a bias towards those that have been in the ICN longer, from bigger NCAs, with more time and money to invest. Upon first impression, geographic diversity is largely respected in the Steering Group. Since the first Steering Group, over half of the NCAs were from Europe and North America, whose competition agencies make up just over half of the ICN membership (about 51 per cent).53 Asia (20 per cent) has three or four seats; Africa (16–17 per cent) one; South America tends to have two or three seats (9 per cent); and Australia is almost always present (Oceania makes up for approximately 3 per cent of members). ICN rules also demand ‘economic diversity’. The lack of economic diversity is hard to justify. African, Asian and South American economies are under-represented. Richer, more technically advanced, NCAs (those that started the ICN in the first place) dominate. Steering Group members have been rotationally the same 20 or so jurisdictions. This is most likely the result of both resource availability and an affinity bias.54 Furthermore, as the years pass, their level of experience in leading the ICN increases, making it harder for others to challenge the Steering Group’s composition. The lack of diversity in the Steering Group carries the risk that the only (or main) issues discussed will be the ones of importance to the richest countries, given the Steering Group’s power over the ICN agenda. Perhaps this is the point? The third category of membership where influence can arise (after general ICN membership and Steering Group membership), is in Working Groups. Any NCA can join any Working Group, so there is not an issue about Working

53 People from 24 different NCAs were involved in the Steering Group between 2006 and 2015. Thirteen from the EU and North America. The remainder were from new jurisdictions (Barbados, Brazil, Mexico, Morocco, South Africa, South Korea and Turkey) and others (Australia, Japan, Russia and Singapore). 54 The affinity bias (an unconscious bias that we all have to a greater or lesser degree) makes it likely that out-going actors recommend those like them (from similar states with similar backgrounds); Kathleen Nalty, ‘Strategies for Confronting Unconscious Bias’ [2016] 45 The Colorado Lawyer 45, 46.

138

Table 6.1

Challenges to assumptions in competition law

Chairs of the Unilateral Conduct Working Group from 2006 to present

2006–2009

2009–2012

2013–2018

2018–present

Germany

Germany

Turkey

Australia

USA

USA

UK

Italy

Sweden

South Africa

Mexico

Group membership. Particular influence comes through being a Working Group co-Chair. The Steering Group appoints them (think of affinity bias again). By way of example, Table 6.1 shows the Chairs of the UCWG from when it was established in 2006 until today: Once again, we see the power of the EU and North America with seven chairs (out of 11, 63 per cent) compared to the rest of the world: Australia, Mexico, South Africa and Turkey all had one chair. This is unrepresentative in terms of its make-up.55 To be representative at least geographically, half of the seats should be given to countries outside of North America and Europe (they got 36 per cent). This lack of diversity in the UCWG co-chairs means that (in its appointment of these officers) the Steering Group fails on both the geographical and economic fronts, to select members ‘reflecting the geographic and economic diversity of ICN members’, contrary to the ICN Operational Framework.56 Influence matters because it is likely to affect who benefits (most) from ICN activity. The unanimity requirement implies that (absent individual or collective mistake) all ICN action must be a pareto improvement for all NCAs. That does not mean that collective action maximises public goods either for the whole world, or even for ICN members. As regards global welfare, networks are political institutions. They benefit their members. Those outside of the

55 ‘US and European participants tend to play the central roles in the working groups and annual meetings’, David Gerber, Global Competition: Law, Markets and Globalisation (Oxford University Press 2010), 116. Similarly, Mariana Tavares, ‘The ICN: One of the Many International Priorities’, in Paul Lugard (ed.), The International Competition Network at Ten: Origins, Accomplishments and Aspirations (Intersentia 2011), 346. 56 ICN, ICN Operational Framework (2012), 6.

Where influence lies in the International Competition Network

139

network may suffer, even if they have relevant interests.57 As regards maximising the welfare of ICN members as a whole, Olson shows that: unless the number of individuals in a group is quite small [unlike the ICN], or unless there is coercion or some other special device to make individuals act in their common interest, rational, self-interested individuals will not act to achieve their common or group interests.58

So, while all ICN members must benefit, welfare is unlikely to be maximised (although Section 6.4.3 argues that the epistemic community may affect this to some degree59). Nor is it likely that the benefits generated by the ICN are equally divided across all ICN members (states). Particular influence comes through membership of the Steering Group and as co-chairs of Working Groups. NCAs from the EU and USA might focus upon four main advantages when assessing the ICN’s costs and benefits, as a first step on the road to convergence around their own standards for competition law and policy.60 First, there are several benefits for the EU and US NCAs themselves; if their own rules are accepted, regulators do not have to learn new rules. Secondly, convergence should benefit EU and US firms, as these firms (and their advisors) do not need to learn/ pay for competition advice on new competition systems.61 This is especially so as the EU/US competition rules should help to open up new markets to EU/ US firms. Thirdly, the ICN may be a relatively cheap route to increasing global convergence. To organise a coalition every time there is a major incident, could be more expensive (and less effective).62 Finally, there are benefits to NCAs’ employees. There is a widespread intellectual belief, amongst the epistemic community of ‘competition people’, that the neo-liberal paradigm that many propagate is positive

57 L. Schaap and M. van Twist, ‘The Dynamics of Closedness in Networks’, in Walter Kickert, Erik-Hans Klijn and Joop Koppenjan (eds), Managing Complex Networks: Strategies for the Public Sector (Sage 1997), 72. 58 Mancur Olson, The Logic of Collective Action (Harvard University Press 1965), 2. Emphasis original. 59 Olson did not think that epistemic communities would normally have much influence in larger groups; ibid, 60–65, but there is reason to believe that their power is stronger here. 60 Raustiala, ‘The Architecture of International Co-operation’ (n. 34), 41. 61 ICPAC, Final Report (n. 8) Executive Summary, 21, see also 22–25. Similarly, ICN Unilateral Conduct Working Group, State-Created Monopolies Analysis Pursuant to Unilateral Conduct Laws (2006), 1; Hollman and Kovacic, ‘‘The International Competition Network: Its Past, Current and Future Role’ (n. 7), 55–56. 62 Arthur Stein, ‘Neoliberal Institutionalism’, in Christian Reus-Smit and Duncan Snidal (eds), The Oxford Handbook of International Relations (Oxford University Press 2008), 209.

140

Challenges to assumptions in competition law

for states large and small (rightly or wrongly). This adds an important, almost evangelical, dimension to support the work of these individuals. There are some costs that counterbalance these benefits for the EU and the USA. First, there is the cost of setting up and supporting the ICN itself. Secondly, there are the costs of participating in ICN activities. This can involve a lot of resources. For example, the US Federal Trade Commission currently has the equivalent of several permanent employees dealing with ICN-related matters.63 Technical assistance and training of other NCAs presumably requires a diversion of funds from other activities, such as competition enforcement at home. So, there is a potential cost in terms of their own consumers, which needs to be considered. Many of these costs are, of course, double-edged swords, as they are what gives the EU and US influence within the ICN, see Section 6.4. There are times when EU/ USA competition systems do not align. The ICN is not a good forum for resolving conflicts.64 This is because unanimity is required for the adoption of new policies. The unanimity rule was a wise choice by the EU/US. It protects them from ICN rules that they themselves disagree with. A comprehensive grouping of most of the world’s NCAs might be an ideal place to thrash out complex problems. Yet, the ICN is not normally used for this. Its main role seems to be to ‘encourage’ others to adopt the existing policies of some older NCAs (particularly the EU and USA). While it is theoretically conceivable that poorer NCAs might suggest practices that ICN members unanimously agree to incorporate into ICN recommended practices, the problems and lack of experience of these NCAs limit the chances of this. So, the benefits that new jurisdictions seek from their ICN membership are unlikely to be the same as those that the EU and USA pursue.65 In fact, the ICN can assist with all three fundamental problems that new jurisdictions face. First, facing down local vested interest groups. In part, this is about persuading states of the benefit of a strong competition regime. Such a regime also benefits the EU and the USA if local competition rules are applied inline with their thinking. This is because it should help to open up

63 https://​www​.ftc​.gov/​policy/​international/​international​-competition last accessed 24 March 2020. 64 Examples of conflicts are evident. For example, think of the disagreements between the US and the EU (principally Germany) on unilateral effects and digital markets. Also, Harry First and Spencer Weber Waller, ‘Antitrust’s Democracy Deficit’ [2013] New York University School of Law, Centre for Law, Economics and Organisation: Law and Economics Research Paper Series Working Paper No. 13-05. 65 There are not just two groups (rich and poor) of NCAs. There is a continuum (some NCAs may be in one group on certain issues, but not others). I limit discussion to two groups to add clarity to the analysis.

Where influence lies in the International Competition Network

141

new markets to entry from EU and US firms. The ICN can support an NCA in many ways which boost its (and its staff) profile. One example is to hold ICN events in that country.66 This emphasises competition’s importance and hints to the NCA’s government that the international community may take a hostile act against the NCA negatively. The ICN can also support new jurisdictions in their struggle to get more resources and better powers. Both can help the NCA to be more active in its role (helping the EU/US if the NCA is acting in line with their views on competition). For example, the ICN might write reports arguing that effective enforcement can only be achieved under certain conditions.67 These might say that NCAs need: sufficient funds to be able to employ well-trained staff and actively enforce the competition rules; certain powers, such as an ability to offer leniency sometimes; or it could be an institutional issue, like the amount of independence that the competition community considers optimal for an NCA. These reports, written by NCAs, help them all back home, as they generally ask for more money and power. Finally, new NCAs might get access to training through ICN participation. Competition enforcement is complex where a lot of the know-how is hidden. So, training is valuable. Despite the costs, providing training can benefit the EU and USA (and they have done a lot of this68), because it allows them to focus the minds of staff in the new NCA on goals, techniques and processes that the EU and USA approve of. Poorer NCAs certainly appreciate assistance. It develops their staff’s techniques (and competence), boosts their self-esteem and initiates them into the competition epistemic community. Once this happens, their views and preferences are likely to further align with those from the EU and USA. In this way, new NCAs form alliances with their more influential peers to ‘punch above their weight’. But the nature of this alliance should not be forgotten. Despite the emphasis on equality in the ICN, this is not an alliance of equals. New NCAs know (or suspect) that valuable assistance like this only comes to the compliant (NCAs that support, and implement, the policy preferences of the EU and the USA). To stand effectively against influential NCAs

66 An example was when El Salvador’s (then, recently created) NCA organised an ICN Cartel workshop in El Salvador. This meant that it could show its government that it has its peers’ respect and trust. 67 The ICN has assisted several NCA members (such as Brazil, India, Jersey and Peru) when they sought external support for domestic reforms. NCAs want such work to be done more often, ICN Merger Working Group, ICN Recommended Practices for Merger Notification and Review Procedures: Member Self-Assessment: Report on 2016 Survey Results (2017), 17–18. 68 Tavares, ‘The ICN: One of the Many International Priorities’ (n. 55), 335.

142

Challenges to assumptions in competition law

on a ‘good practice’ is another matter. You need to be from a reasonably strong and large economy, with strong leadership that is prepared to be isolated. This is rare.69 New NCAs normally have neither the knowledge nor the confidence to take such a stance. They may also fear that they could not afford the price that this strategy involves. 6.3.3

Non-Governmental Actors (NGAs)

Global governance is a key hunting ground for multinationals. Individual firms may have little individual influence, but together they are a force to be reckoned with.70 Firms seek influence by openly contesting the creation or application of norms, but this involves risks. So, firms often prefer to operate in the shadows. They may try to confine ‘decision-making to relatively “safe” issues [that is ones which are comparatively innocuous to them]’.71 Firms also try to influence the media, think tanks, etc. to foster beneficial beliefs in them (and their audiences), about what is possible/desirable.72 In order to assess how easily firms may be able to exercise influence over regulators in a particular field, we need to consider institutional and economic characteristics. Independence shields NCAs from political protection and makes capture of NCAs particularly valuable. NCAs’ prevalent ideology is also important. US multinationals have been extremely successful at changing this in the area of competition policy, placing great emphasis on consumer welfare arguments (the long-term version benefits firms) and excluding ‘non-economic’ policy concerns (ensuring more consistent regulatory outputs). Both of these shifts benefit multinational firms at the expense of smaller firms and others (such as the environment, employment, consumer interests, etc.). As for economic characteristics, Zingales highlights firms’ ability to: make implicit promises of future career opportunities;73 lobby the 69 One example of this is when the South African Competition Tribunal, in the UCWG discussions on competition law goals, said that there are other concerns, such as social integration, that need to be considered in competition law. But the South Africa Competition Tribunal is from a major economy and its charismatic president (David Lewis) felt able (and willing) to row against the mainstream. 70 Stephen Wilks, The Political Power of the Business Corporation (Edward Elgar 2013), 5 and 43–47. 71 Peter Bachrach and Morton Baratz, ‘Two Faces of Power’ [1962] 56 The American Political Science Review 947, 948. 72 Paul Pierson, Goodbye to Pluralism? Studying Power in Contemporary American Politics (Presentation at the Wildavsky Forum for Public Policy, Goldman School of Public Policy, April 2015), 7. 73 One example may be Citigroup’s work to change a US law which sought to sever ties between commercial and investment banking, Luigi Zingales, ‘Towards a Political Theory of the Firm’ [2017] 31 Journal of Economic Perspectives 113, 122.

Where influence lies in the International Competition Network

143

legislator;74 (threaten to) sue unfavourable legislation;75 lobby the regulator to change its policy or enforcement practices;76 and hide relevant information.77 It also assists firms to align their self-interest with a ‘public interest’ (for example, national security); and to control their image through advertising, etc. This raises a challenge for studies like this one. If one cannot study what one cannot see then how can we investigate NGA power (I have already pointed out several times how opaque ICN processes are)? The starting point must be the insight that those with power do not only use it to achieve specific, individual advantages for themselves. They also use it to change the ‘rules of the game’, to stack the deck in their favour for the future. Stigler says as ‘a rule, regulation is acquired by the industry and is designed and operated primarily for its benefit [at the expense of the wider public]’.78 So, it is no surprise that major EU and US multinationals, and the lawyers that represent them, seem to have disproportionate voice in the ICN, in fact. NGAs appear to be principally major multinational firms (and their representatives), such as:79 Alcan, Barclays Bank, British Telecom, Charles River Associates, Compaq Computer, General Electric, Goldman Sachs, Intesa Sanpaolo, Google, Microsoft, NERA, Philips, Rio Tinto, Shell, Sumitomo, the US Chamber of Commerce, Vodafone, and Walmart. These firms are identifiable because they contributed to ICN discussions as panel speakers. The lack of publicly available attendance sheets for ICN events, and a lack of information

74 Pierson, Goodbye to Pluralism? (n. 72), regarding the healthcare industry; Thomas Philippon, The Great Reversal: How America Gave up on Free Markets (The Belknap Press 2019), on US competition law. 75 Zingales, ‘Towards a Political Theory of the Firm’ (n. 73), 124 discusses the US SEC’s decision to drop its rule preventing institutional investors from nominating their own representatives to corporate boards, in the face of litigation from the Business Roundtable. 76 According to Thomas Lambert, ‘Lobbying on Regulatory Enforcement Actions: Evidence from US Commercial and Savings Banks’ [2019] 65 Management Science 2545, regulators are 45% less likely to initiate enforcement actions against lobbying banks. 77 Zingales, ‘Towards a Political Theory of the Firm’ (n. 73), 124, reports that DuPont delayed liability for contaminating the water supply near one of its US factories for over 30 years by hiding information. 78 George Stigler, ‘The Theory of Economic Regulation’ [1971] 2 Bell Journal of Economics 3, 3. 79 See, ICN Merger Review Working Group, 2002–2003 SubGroup Work Plans (2002), 6–11; NGA Submission, NGA Report on Implementation of ICN Recommended Practices for Merger Notification Procedures (2003); ICN, The Future of the ICN in its Second Decade – FINAL REPORT (2016), 27; http://​www​.icn2016​.sg/​sites/​live​ .icn2016​.site​.gsi​.sg/​files/​ICN​_2016​_con​_booklet​_Program​.pdf.

144

Challenges to assumptions in competition law

about precisely who does what in many ICN activities, makes it hard to know which NGAs are active and what they do. In addition, such firms are nested in a supportive environment of service providers. Being parasitic upon them, these service providers are also dependent upon their clients’ continuing success in order to, in their turn, thrive. This could well alter their incentives. Wilks says, rather than reinforcing ‘a legal philosophy that controls and limits corporate activity (which German law firms tend to do) … [major UK and US law firms] have lubricated corporate strategy and profit seeking’.80 Competition policy and enforcement are highly re-distributive. So, ICN discussions and outputs are also theoretically of interest to small firms, consumers, employees and their representatives, such as consumer associations, trade unions (and relevant law firms). Yet, consumer groups (and their law firms) are ‘under-represented’.81 There are also very few small firms actively involved in ICN work. The ICN has challenged economists to get involved.82 This initiative has not been very successful. There are not many other kinds of service providers (such as: business schools, management consultants and accountants) in attendance either. There is also a lack of political scientists, sociologists, and international relations scholars, who all have something to say about the distributional interests and outcomes of the current rules. Behavioural economists and psychologists would also be interesting additions, as consumers are often not as rational as many competition authorities assume. When one looks to see where NGAs are based, many are from Europe and North America. We are told that many NGAs are ‘from younger jurisdictions such as Armenia, Pakistan, Vietnam and others’.83 ICN opacity makes this hard to check. In any event, active influence, not presence, is most valuable. The ICN admits that NGAs from Europe and North America are normally the most active participants.84 The ICN and NGAs believe that NGAs’ participation is important.85 The ICN thinks that firms act for the benefit of all: ‘independence and integrity are Wilks, The Political Power of the Business Corporation (n. 70), 82. Eleanor Fox, ‘Linked-In: Antitrust and the Virtues of a Virtual Network’, in Paul Lugard (ed.), The International Competition Network at Ten: Origins, Accomplishments and Aspirations (Intersentia 2011), 125. 82 Martin Mandorff, ‘Engaging Economists in the ICN: Uniting under a Common Language’ [2012] Competition Policy International , 1–2. 83 Boge (2005), Closing Speech 4th Annual ICN Conference, 6–8 June 2005, 2. 84 ICN, The Future of the ICN in its Second Decade – FINAL REPORT (2016), 14–15. 85 Fostering links between NCAs/NGAs is a key ICN aim, ICN Cartel Working Group, 2010–2011 Work Plan (2010), 4. For multinationals, Ronald Stern, ‘The Role of 80 81

Where influence lies in the International Competition Network

145

key features of the NGA role. NGAs work towards the advancement of the ICN’s general interest to the benefit of all stakeholders.’86 NGAs say similar things: The goals of NGAs, and indeed all competition law advisers, are in principle aligned with those of the agencies in that everyone in the competition community wants to promote and encourage competition law and culture. We are all aiming to effect justice by creating and maintaining good systems.87

The idea that a group of largely major Western multinational firms (and their representatives) would act independently, for the benefit of society in general, is, at best, naïve.88 It seems relevant that a group of (largely Western, rich) NCAs (many without express political authority) came together to make ‘best practice’ rules at the US’ initiative (after ICPAC argued that this would support US industrial policy). NCAs did this supported by major, Western multinationals (and their advisors). From the start, NGAs were then invited into this policy network (to help design the rules which regulate them). NCAs often give three reasons for supporting NGAs’ ICN presence. In their view, NGAs provide: more resources, advice, and increased legitimacy. NCAs often complain of a lack of resources. So, if NGAs can provide ideas, basic research or secretarial assistance, this could save NCAs money, while increasing ICN outputs. Section 6.4 explains why such access is particularly beneficial for NGAs. Secondly, NCAs are keen to get business advice (technical expertise and different perspectives).89 For example, the ICN notes that competition enforcement may impose unnecessary costs.90 NGAs can highlight where such costs arise and suggest ways of reducing them. This ties in with ICPAC’s original industrial policy aims for the ICN, discussed in the Introduction (Section 6.1). One need only highlight here firms’ ability to hide the ICN in Fostering Convergence: An NGA’s Perspective’, in Paul Lugard (ed.), The International Competition Network at Ten: Origins, Accomplishments and Aspirations (Intersentia 2011). For law firms, Calvin Goldman, Robert Kwinter and Navin Joneja, ‘A Perspective on the Role and Contribution of NGAs at the ICN’, in Paul Lugard (ed.), The International Competition Network at Ten: Origins, Accomplishments and Aspirations (Intersentia 2011), 383. 86 ICN, NGA Toolkit (2012), 3. Similarly, ICN Advocacy Working Group, Summary of ICN Teleseminar on Building Agency/ Bar Relationships (2009), 2. 87 Dave Anderson, ‘Reflections on the ICN and its NGAs: Advocacy and Implementation’, in Paul Lugard (ed.), The International Competition Network at Ten: Origins, Accomplishments and Aspirations (Intersentia 2011), 279–280. 88 Wouter Wils, ‘Independence of Competition Authorities: The Example of the EU and its Member States’ [2019] 42 World Competition 149, note 76. 89 ICN, NGA Toolkit (2012), 2–5; ICN Steering Group, Draft Minutes (2011), 3. 90 ICN, ICN Factsheet and Key Messages (2009), 3.

146

Challenges to assumptions in competition law

relevant information (in their discussions with, one might call it lobbying of, NCAs).91 Thirdly, the ICN thinks that more diversity in the actors present in the ICN ‘enhances the legitimacy and relevance of the ICN, of the work it produces, of the standards it aims to offer’.92 Legitimacy ‘through inclusiveness’ is problematic. If all relevant perspectives are neither equally present, nor equally resourced (as happens in the ICN) then policy-making is unlikely to be aligned with the public interest. Wils says it is almost inevitable that ‘among the different NGAs, those representing (directly or indirectly) the views and interests of big business will end up being most vocal’.93 In order to study these effects empirically, one needs to examine (over time) the evolution of policy options, examine who favours certain outcomes, what enabled the outcome that won to win (context, etc.), and the long-term effects of these outcomes on preferences and political resources. An historical perspective is vital here because ‘power is something that develops over time and simultaneously becomes less visible as it does so’.94 This sort of analysis of the ICN is extremely difficult, given the lack of transparency. Having said that, Jenny reports a discussion in the ICN Annual Meeting in 2004. There, some NCAs from developing countries wanted the ICN to start work on recommended practices on abuse of dominance. This idea was flatly rejected. Some suspect that: the refusal of the ICN to take up the issue of abuse of dominance was due to the reluctance of business-related NGAs who did not want to see competition authorities encouraged to go after large multinational firms using abuse of dominance or monopolisation provisions.95

It was another two years before the Steering Group created the UCWG. Even if the NGAs lost the battle in the end, their ability to disrupt ICN work likely brought considerable extra profits to those taking illicit advantage of their market power (and their advisors). Many actors with relevant interests are not present in the ICN (or not as present). There are many potential reasons for this, even if competition policy and enforcement seriously affects them. First, they may not know of the ICN as a forum. Secondly, they may not comprehend the competition rules, or they Wilks, The Political Power of the Business Corporation (n. 70), 109. ICN, NGA Toolkit (2012), 5. 93 Wils, ‘Independence of Competition Authorities’ (n. 88), note 76. 94 Pierson, Goodbye to Pluralism? (n. 72), 15–16. 95 Frederic Jenny, ‘The International Competition Network and the OECD Competition Committee: Differences, Similarities and Complementarities’ in Paul Lugard (ed.), The International Competition Network at Ten: Origins, Accomplishments and Aspirations (Intersentia 2011), 101. 91 92

Where influence lies in the International Competition Network

147

may feel alienated by ‘competition speak’ and the relatively closed community enforcing the rules. Thirdly, they may not have the resources to attend the ICN (or feel that their resources could be better spent elsewhere). This could be due to a collective action problem. As sellers (as a group in a market) are normally much more concentrated than buyers (as a group), then sellers are often much more motivated to fight over any surplus than buyers. This is because the benefit to each individual seller could potentially be more than for each buyer. Fourth, perhaps these other actors have asked to be involved in the ICN, but the NCA that they approached refused. As we have seen, NCAs are the gatekeepers for NGA participation. In competition policy, independent regulators are often tasked with maintaining profitable firms in managed markets. Wilks says: ‘In these fields citizens have become consumers; as voters, like elected politicians, they are self-righteously kept at arm’s length.’96 There is no public data on whether any applicants were refused and, if so, why.

6.4

HOW DOES THIS INFLUENCE ARISE?

The ICN operates by consensus. Before a recommended practice is officially adopted, it must be approved at the ICN Annual Conference. Some see this as an effective brake on more powerful NCAs and NGAs from imposing their will. However, states use their power when they create international institutions and when they operate within them. They use this power to achieve their ends and to constrain the power of others acting there.97 Political scientists are wary of solely focusing on formal decision-making mechanisms. In fact, this section argues that powerful NCAs and NGAs invoke several mechanisms, intentionally or not, to pressure weaker NCAs into accepting ICN processes and outputs, even when they might not be in their best interests. Section 6.4.1 examines epistemic communities; Section 6.4.2 discusses powerful NCAs; then, Section 6.4.3 investigates powerful NGAs. 6.4.1

The Epistemic Community and the Technical Nature of Rules

In complex areas like competition policy and enforcement, as van Waarden and Drahos emphasise, ideas are vital, given the uncertainty about what is best and how best to achieve it.98 There is a lot of room for influence. Competition Wilks, The Political Power of the Business Corporation (n. 70), 255. Stephen Krasner, ‘Structural Causes and Regime Consequences: Regimes as Intervening Variables’ [1982] 36 International Organisation 185. 98 Ideas are ‘important in complex policies with uncertainty concerning eventual outcomes. Since the effects of competition policy are difficult to measure, ideas matter … There are various economic schools of thought, with different concepts of compe96 97

148

Challenges to assumptions in competition law

rules are often ambiguously drafted (‘abuse of dominance’, ‘restraint of trade’), leaving wide discretion in their application. So, we need to understand experts’ impact on NCA decision-making. In addition, many NCAs enjoy significant independence from government influence, making them particularly powerful decision-makers (subject to appeal). So, many competition rules leave a lot of room for interpretation; and interpretation is influenced by ‘expert opinion’. Some see the ICN as a neutral forum, especially as states are not involved.99 However, this is not the case. ‘Expert opinion’ (such as, economic analysis in the competition rules should not be tainted by public policy concerns100) always has a political result, even if those advocating it do not know it (or even aim for it). Experts are: political actors whose preferences set the ‘technical’ agenda and define the range of possible outcomes in a decision-making process. Expert knowledge is itself always political because it is always acquired in a particular social context, and reflects the political-economic structures and social relations which generated and which reproduce that context. Each expert uses his or her own theoretical approach to produce technical solutions … As Robert Cox puts it, ‘theory is always for someone and for some purpose‘ (Cox, 1981: 128101). In other words, technical solutions are never purely technical; they always have a political purpose even if those propagating the solutions are not fully aware of it.102

One way of identifying the political element in an ‘expert opinion’ is to focus on the consequences of expertise. Most experts avoid this. This approach: requires an assessment, however crude or incomplete, of consequences – who wins and who loses? Who decides and who submits to the decisions of others? Only by identifying the stakes of expert action can we understand its politics.103

The EU and USA will focus on the merits of their rules, emphasising liberal, pro-privatisation, open markets. This focus has consequences for who wins

tition and different ideas concerning competition policy’, van Waarden and Drahos, ‘Courts and (Epistemic) Communities in the Convergence of Competition Policies’ (n. 27), 932. Similarly, William Kovacic, ‘The Modern Evolution of US Competition Policy Enforcement Norms’ [2003] 71 Antitrust Law Journal 377. 99 Tavares, ‘The ICN: One of the Many International Priorities’ (n. 55), 344–345. 100 Mandorff, ‘Engaging Economists in the ICN’ (n. 82), 1. 101 Robert Cox, ‘Social Forces, States and World Orders’ [1981] 10 Millennium Journal of International Studies 126. 102 James Perry and Andreas Nölke, ‘The Political Economy of International Accounting Standards’ [2006] 13 Review of International Political Economy 559, 578. 103 David Kennedy, ‘Challenging Expert Rule: The Politics of Global Governance’ [2005] 27 Sydney Law Review 5, 17.

Where influence lies in the International Competition Network

149

and who loses. It is not a value neutral position.104 For example, they might say that if the NCA in a developing economy were to use competition policy to open up its markets to EU and US firms then this would result in more competition and lower prices (and more choice) for local consumers. However, there are likely to be other winners and losers too, which are not discussed so openly. One set of winners is often firms from the EU and the USA. We need go no further than the ICPAC Report for an admission of this industrial policy aim.105 The flipside of this is that probable losers are local firms (and their employees). Some might consider this an acceptable price to pay for the extra consumer surplus. However, it is one thing to fully debate the consequences of following the EU/ US competition paradigm, including the potential winners and losers, and another thing just to blindly follow the ‘technically best policy’. Competition policy is highly re-distributive. By excluding states from the ICN and encouraging independent NCAs, the EU/ US paradigm also seeks to erase this re-distributive dimension from competition discourse. Different expert opinions will have different re-distributive consequences and so choosing between them (deciding to prioritise some experts over others) is re-distributive too. An exclusive focus on consumer welfare makes the interests of employees and small- and medium-sized enterprises (SMEs) irrelevant in competition law, for example. Section 6.3.1 argued that it benefits rich, Western states and NGAs to frame competition policy and enforcement as technical disciplines prioritising economic expertise. It is hard for poor NCAs, especially those recently invested with competition powers, to influence arguments about which experts to favour. They often have less relevant expertise and may be outgunned by experts from richer, more established NCAs. There are some clues that rich NCAs envisaged a role for newer NCAs which focuses on learning rather than seeking influence.106 Epistemic communities are networks of professionals ‘with recognised expertise and competence in a particular domain and an authoritative claim to policy relevant knowledge within that domain or issue-area’.107 They can

Wilks, The Political Power of the Business Corporation (n. 70), 122-8. ‘US support of new competition policy regimes also creates an opportunity for the United States to share its perspective. This can be important as a means of influencing the legal environment in which US exporters and businesses operate’, ICPAC, Final Report (n. 8) Executive Summary, 31. 106 The ‘ICN’s value to the developing countries cannot solely be judged by their participation in the annual conferences and working groups. They can also benefit from the ICN’s guiding principles and recommended practices’, Boge (2005), Speech on the Opening of the 4th ICN Annual Conference from 6–8 June 2005 in Bonn, 2. 107 Peter Haas, ‘Introduction: Epistemic Communities and International Policy Co-ordination’ [1992] 46 International Organisation 1. 104 105

150

Challenges to assumptions in competition law

be really influential. The competition epistemic community108 includes competition academics, employees (of NCAs and NGAs), and other competition practitioners and officials. It is widely accepted that the epistemic community influences what is considered to be relevant expertise in competition policy and enforcement.109 Expert opinion is informed by the epistemic community’s own broader world-view. It is not objective. While epistemic communities: provide consensual knowledge, they do not necessarily generate truth. The epistemological impossibility of confirming access to reality means that the group responsible or articulating the dimensions of reality has great social and political influence.110

Prioritising one epistemic community over another produces winners and losers. Once again, this is often done through the intellectual leadership of rich, Western academics, competition practitioners and NCAs. It is likely that the expertise chosen reflects the preferences of these actors (not others). Joining the leading epistemic community can benefit individuals111 and NCAs.112 Influence and prestige can be reward in themselves. Influence can also enhance future employment opportunities (Section 6.4.3). Yet, members of the competition epistemic community are not necessarily motivated by self-interest. They often genuinely think that their ideas benefit the economy and consumers;113 reinforcing these beliefs in each other. Years of learning help to embed this perspective.114 Nevertheless, these political effects often occur even when experts are not motivated by selfish ends. 108 I say ‘the competition epistemic community’. However, several communities may compete for influence; Djelic and Quack, ‘Transnational Communities and Governance’ (n. 39), 25–26. 109 Gerber, ‘Modernising European Competition Law’ (n. 47), 126. 110 Haas, ‘Introduction: Epistemic Communities and International Policy Co-ordination’ (n. 107), 23. Similarly, Sol Picciotto, ‘Networks in International Economic Integration: Fragmented States and the Dilemmas of Neo-liberalism’ [1996–1997] 17 Northwestern Journal of International Law and Business 1014, 1037. 111 van Waarden and Drahos, ‘Courts and (Epistemic) Communities in the Convergence of Competition Policies’ (n. 27), 929. More widely, Imelda Maher, ‘Competition Law in the International Domain: Networks as a New Form of Governance’ [2002] 29 Journal of Law and Society 1, 117, and references. 112 In fact, NCAs try to bolster their independence (and ultimately their power), through the ICN, see the discussion around n. 99. This is not surprising. The ICN explicitly aims to ‘facilitate effective international co-operation to the benefit of member agencies’, see n. 25. 113 van Waarden and Drahos, ‘Courts and (Epistemic) Communities in the Convergence of Competition Policies’ (n. 27), 913–934. 114 Mark Pollack, ‘Principal-Agent Analysis and International Delegation: Red Herrings, Theoretical Classifications and Empirical Disputes’ [2007] Bruges Political Research Papers, 7.

Where influence lies in the International Competition Network

151

These choices by the powerful in the ICN matter. The pressure on NCAs to conform with the epistemic community’s view can be immense, the cost of exclusion can be so high.115 But a further important effect is the epistemic community’s ability to socialise the opinions of those from poorer NCAs so that they conform to those of their colleagues from richer NCAs. This could turn officials from these poorer NCAs (and the NCAs themselves) into EU/ US competition cuckoos, embedding these rich world views (and their consequences) into the heart of poorer states. 6.4.2

National Competition Authorities (NCAs)

This section discusses various mechanisms that powerful NCAs can use to get other NCAs to do what they want (and thus capture more ICN benefit for themselves). I examine: political and economic power, a relative lack of resources, and offers of technical assistance. Whether or not the influence generated through these mechanisms is sought, there are likely effects on those not exercising these techniques or blessed with the advantages of more powerful NCAs. 6.4.2.1 Political and economic power ICN membership gives powerful NCAs regular opportunities to persuade others to follow their interpretation of the competition rules. Regular meetings and their control over the ICN’s intellectual agenda, may make this a cheap and effective way for them to exert pressure. Dressing up pressure as international ‘best practice’ also makes it less confrontational (and more effective). If new NCAs do not respond, powerful NCAs can raise the stakes. This might mean less training, or resources, or direct coercion (either directly or through more bilateral, rather than multi-lateral, rule-making). 6.4.2.2 A relative lack of resources There is a lot of ICN activity. This has many benefits. However, to actively participate in all activities (answering questionnaires, commenting on draft recommendations, etc.), and attend all meetings (including physical attendance at the Annual Meeting), costs a lot. Many poorer NCAs cannot afford this time or money (participation involves substantial actual and opportunity costs).116 115 Haas, ‘Introduction: Epistemic Communities and International Policy Co-ordination’ (n. 107), 20. On EU competition law, van Waarden and Drahos, ‘Courts and (Epistemic) Communities in the Convergence of Competition Policies’ (n. 27), 929; Townley, A Framework for European Competition Law: Co-ordinated Diversity (n. 4), 154–155. 116 ICN, The Future of the ICN in its second decade, Draft Final Report (2016), 12.

152

Challenges to assumptions in competition law

This is even worse for NCAs seeking to constructively engage with the ICN.117 So, poorer NCAs may not be heard. This makes it harder for them to protect their own interests when they clash with those of richer NCAs and NGAs. The burden is even greater for Chairs of a Working Group or those in the Steering Group. Fox says that less well-resourced NCAs may find this too expensive. So, these NCAs ‘have less opportunity to participate in setting the agenda and to write first drafts of recommendations’.118 So, less well-resourced NCAs are able to influence policy debates. Perhaps this is the point? The topic has been debated within the ICN itself. For example, it has been suggested that the number of workshops per year should be reduced to allow each NCA to attend as many as possible. Inclusiveness could be undermined if NCAs were forced to miss many of these important events.119 The Steering Group did not formally reduce the number of workshops. Rather, it focuses on preventing conflicts between them, and encourages Working Groups to produce only one workshop per year.120 If the ICN wants to keep poorer NCAs involved, or preferably even increase their involvement, then it needs to help them to do so.121 Good examples of this sort of inclusion were twinning of experienced and inexperienced NCAs in the leadership of every working group. Similarly, the USA, Korea and Japan hosted induction sessions for new NCAs in the ICN explaining what is on offer, helping them to make informed choices about where to devote their energies.122 Another idea was to have a small fee for attending the annual conference, as long as this did not put off developing countries.123 If one is really seeking to help poorer NCAs actively engage with the ICN (without coming under undue influence from others), then it may be sensible to adopt practices that allow similar NCAs from states with similar features to pool resources. It may also help to consider regional arrangements for assistance. Many NCAs are in favour of having regional platforms to address regional interests. This reduces travel time, which is not a major factor in the virtual ICN. However,

117 Not only poor NCAs suffer. Swedish NCA staff make similar points, Dan Sjöblom and Monica Widegren, ‘ICN Membership: Opportunities and Challenges for a Competition Authority’, in Paul Lugard (ed.), The International Competition Network at Ten: Origins, Accomplishments and Aspirations (Intersentia 2011), 355. 118 Fox, ‘Linked-In: Antitrust and the Virtues of a Virtual Network’ (n. 81), 126. 119 http://​www​.i​nternation​alcompetit​ionnetwork​.org/​uploads/​library/​doc998​.pdf, 2. 120 Ibid, 2. 121 Operational Framework Working Group, Report to the Steering Group (2003), 4. 122 David Lewis, ‘Some Reflections on the ICN’, in Paul Lugard (ed.), The International Competition Network at Ten: Origins, Accomplishments and Aspirations (Intersentia 2011), 208. 123 Group, Report to the Steering Group, 7.

Where influence lies in the International Competition Network

153

neighbours often have similar problems, culture and even languages.124 There is still a risk of undue influence, even here. It is no surprise that the EU and USA have not supported such initiatives, which would give them less influence. 6.4.2.3 Offers of technical assistance Richer NCAs and NGAs can also exercise influence over poorer, inexperienced NCAs through offers of technical assistance. Such support may involve sending employees to the poorer NCA to provide general training. It can also mean helping NCAs in specific cases. The US ICPAC Report itself explains why NCAs from the (EU or the) USA might want to use technical assistance to influence poorer NCAs: Technical assistance programs may provide the United States with a voice to support the adoption of sound competition principles and promote the rule of law, especially by transition economies. Further, in light of the proliferation of new antitrust authorities, technical assistance can be used to convey practical experience and advice to emerging antitrust regimes, as well as guidance on the formulation of domestic competition policies that make sense in the globalized economy. US support of new competition policy regimes also creates an opportunity for the United States to share its perspective. This can be important as a means of influencing the legal environment in which U.S. exporters and businesses operate [emphasis added].125

Providing technical assistance can be expensive. That it appears to be a major act of benevolence only increases its power as a mechanism for influence. Nonetheless, richer NCAs are often eager to do it. In part, this is in an effort to increase ICN participation from younger and smaller NCAs.126 The ICN increases in value, especially for richer NCAs and NGAs, the more NCAs participate. This is because wider participation expands the scope of the influence of the rich to more jurisdictions. There are three ways in which the providers of technical assistance programmes can influence. First, it is an additional opportunity to increase pressure from the network and the epistemic community. We have already discussed this above and so we will not go into further details here. Secondly, technical assistance programmes provide a wonderful opportunity literally to ‘get inside’ the poorer NCA, engage with their staff and influence them (including in the ideological underpinnings of the type of competition policy and enforcement to apply). The nature of an education programme implies that inexperienced actors are receiving the support. This knowledge gap makes ICN, The ICN’s Vision for its Second Decade (2012), 19–22. ICPAC, Final Report (n. 8) Executive Summary, 31. 126 ICN, The Future of the ICN in its Second Decade, Draft Final Report (2016), 20. 124 125

154

Challenges to assumptions in competition law

it much harder for the poorer NCAs’ employees to understand (or to even be aware of) the ideological choices that have been made in the competition policy and enforcement advice that they are given, let alone to challenge this advice. Finally, providers of technical support can influence poorer NCAs (and their employees) through a (often implicit) threat to withdraw such support if poorer NCAs become too critical elsewhere. This probably impacts upon the freedom, actual or imagined, of these poorer NCAs to criticise or disagree (on issues such as not approving draft recommended practices, for example). 6.4.3

Non-Governmental Actors (NGAs)

Many of the NGAs’ influence mechanisms mirror what we said about rich NCAs. We have already talked about the power of epistemic communities and offering technical assistance in the ICN. Our focus was on NCAs, but similar observations apply to NGAs. Here I discuss three issues: presence is something; getting into the debate early (similar observations apply to NCAs); and revolving doors. 6.4.3.1 Presence is something Originally, the power of corporations was held in check by elected governments. Wilks argues that this liberal democratic equilibrium has broken down. If we want to see whether (and how) NGAs (particularly major multinationals) exercise power within the ICN we have to check their ability ‘to participate in political decision-making in respect of public policies’: in other words, to assess their impact on regulatory activities that define public priorities and distribute welfare.127 The ICN relies more heavily on NGAs than other international organisations, such as the OECD, UNCTAD and the WTO.128 This ‘privatisation’ of the regulatory space129 within the ICN gives firms (and their regulatory intermediaries, such as law firms130) direct access to competition policy debates and Wilks, The Political Power of the Business Corporation (n. 70), 16–17. Hollman and Kovacic, ‘The International Competition Network: Its Past, Current and Future Role’ (n. 7), 76; Wils, ‘Independence of Competition Authorities (n. 88), note 76. 129 Some say using private actors can be more efficient and innovative; Michael Trebilcock and Edward Iacobucci, ‘Privitisation and Accountability’ [2003] 116 Harvard Law Review 1422. Others fear for regulatory systems’ public values, Paul Verkuil, Outsourcing Sovereignty: Why Privatisation of Government Functions Threatens Democracy and What We Can Do About It (Cambridge University Press 2007). 130 Regulatory intermediaries (such as lawyers) often work between private regulatees and the regulators; Kenneth Auld, David Levi-Faur and Duncan Snidal, 127 128

Where influence lies in the International Competition Network

155

chances to exert pressure and influence over policy-makers and enforcers. The imprecise wording in many competition rules, means the distinction between policy and enforcement often collapses. When it comes to indirect types of influence (such as agenda setting), the ICN has explicitly highlighted NGAs’ role in identifying topics for discussion and work. The ICN specifically says that NGAs help to identify topics for discussion and work and may be consulted when Working Groups plan their future work so that they can offer suggestions towards the ICN’s substantive agenda.131 We have already seen how NGAs argued against establishing a UCWG in the ICN. This contributed to nearly two years of delay in the creation of this Working Group.132 The ICN has also highlighted several other roles which give NGAs the opportunity to exert direct pressure towards their desired outcomes. NGAs contribute to the drafting and reviewing of ICN: recommended practices, case handling manuals, databases and toolkits. Section 6.4.3.2 explains why drafting these documents is a significant coup for the major multinationals seeking specific outcomes. NGAs also conduct research and prepare background papers133 (one can readily see how this might help them to get what they want). Fox gives a concrete example of NGAs (and rich NCAs) applying pressure in relation to the content of rules. At issue was how to assess market power (the discussion took place in the UCWG). The South African NCA wanted to use a market share proxy for market power. It argued that most developing countries do not have the resources for more complex assessments and market share works relatively well as a rule of thumb. However, many major multinationals that were present in that Working Group (as NGAs) and some developed NCAs disagreed. They said that assessing market power is highly complex and market shares mean little in and of themselves. Before any finding of dominance (or equivalent), they said that lots of economic evidence was needed, just focusing on market shares would be over inclusive. Regardless of the merits of this discussion, the method that NGAs favoured would be much more expensive and likely to reduce the amount of such cases that poorer NCAs undertake. Interestingly, the South African NCA bowed to pressure and agreed to the recommended practice. If NCAs are so free, why

‘Theorizing Regulatory Intermediaries: The RIT Model’ [2017] 670 The ANNALS of the American Academy of Political and Social Science 14. 131 ICN, NGA Toolkit (2012), 3. Similarly, ICN, ICN Factsheet and Key Messages (2009), 4. 132 David Lewis, Enforcing Competition Rules in South Africa: Thieves at the Dinner Table (Edward Elgar 2013), 241. 133 Hollman and Kovacic, ‘The International Competition Network: Its Past, Current and Future Role’ (n. 7), 59.

156

Challenges to assumptions in competition law

didn’t it refuse to condone a practice that it disagreed with (unless persuaded on the merits, of course)? 6.4.3.2 Getting in early Structural power (economic or political) and an ability to disseminate self-serving ideas may count for little without the ability to influence the debate from an early stage. Lall highlights two mechanisms which, together, may explain the distributional consequences of regulatory processes for complex, technical matters (such as competition policy and enforcement).134 The two mechanisms are: First, they affect which stakeholders have access to regulatory officials’ informal social networks (ISNs) – networks that play a crucial role in determining who claims first-mover position and, thus, sets the agenda … Second, they give rise to distinctive increasing returns processes that serve to reinforce and entrench proposals made at an early stage of rulemaking, guaranteeing that first movers exert an enduring influence over the content of international standards.135

The idea here is that regulatory standards are adopted over several phases. They start with agenda setting, then planning, then a discussion paper, consultation, and approval. The technical complexity of the issue affects those who have access to ISNs and thus who may get a first-mover advantage. Those involved earlier tend to enjoy more influence than stakeholders arriving later. The discussion here explains, in part, why being on the Steering Group is (ever more) important, as well as chairing Working Groups, rather than merely participating in the ICN as most NCAs do. However, Lall focused on the benefits to private actors from ‘getting in early’. Here, I discuss his two mechanisms (access to informal social networks and increasing returns processes) to see if they might explain how NGAs could manipulate the ICN to support their desired distributional consequences.136 Might a first-mover advantage benefit NGAs? Do those involved earlier enjoy more influence? NGAs are involved in many ICN tasks, including suggesting topics for discussion, drafting and reviewing recommended practices, etc. Knowledgeable NGAs can see problems in the existing rules and suggest ways of solving them,

134 Ranjit Lall, ‘Timing as a Source of Regulatory Influence: A Technical Elite Network Analysis of Global Finance’ [2014] Regulation and Governance 1, 6, Lall’s focus is financial regulation, but competition is similar in important respects. He originally focused on NGAs, but it is also relevant for NCAs. 135 Ibid, 2. See also, Elliot Posner, ‘Sequence as Explanation: The International Politics of Accounting Standards’ [2010] 17 Review of International Political Economy 639, 648. 136 Lall, ‘Timing as a Source of Regulatory Influence’ (n. 134), 6.

Where influence lies in the International Competition Network

157

which reflect their regulatory preferences (or ignore gaps where it is not in their interest to point them out). NGAs with ‘timely and high-quality information’ about the regulatory agenda, may suggest a new Working Group or the scoping of a project (or an idea within a project), before others are even aware that the issue is being discussed. This helps them to set or lead the agenda. In addition, because of their specialist knowledge and skills, these NGAs are more likely to be delegated (trusted with) key roles, such as drafting relevant sections of a recommended practice or doing feasibility studies. NGAs from the EU and the USA are likely to propose rules that already apply to them (and they are familiar with), making them more knowledgeable, likely leaders in this area (and better able to win arguments, should any arise). One NGA has even suggested that NGAs be given expedited access to potential reforms, so that they can comment early.137 This desire fits with Lall’s theory of the benefits for business of getting in early to the decision-making process. Yet, it is not all NGAs that will have such ready (and influential) access to the ICN. Only a few have really close personal relationships with NCA officials, regularly interacting with them. Not many NGAs have sufficient experience of international markets and enough technical knowledge and vision to think of how to improve the rules (and add to NCA knowledge), and so this pool of influential NGAs is limited. Those with influence often work in (or for) large multinationals who have a lot of experience with competition rules from around the world. Some NGAs’ expertise can exceed that of NCAs. Lall notes, ‘regulators often depend on the expertise provided by ISN members to develop rules that make operational sense’.138 NCAs often ask senior NGAs for input (linking expertise and early involvement, which explains why NGAs consider both so valuable). So, even if consumers and consumer associations became NGAs, their lack of technical experience is likely to limit their influence. Lall suggests that the ICN should be wary of delegating crucial regulatory functions to NGAs, to protect the independence of officials. So, NCAs must have resources to attract sufficient expertise from a wide range of disciplines. If delegation cannot be avoided ‘regimes should divide functions among several stakeholders to prevent narrow interests from monopolizing regulatory capabilities’.139 However, in the ICN, they are not currently taking this advice. The current narrow range of interests amongst the NGAs does not look set to change. Furthermore, existing NGAs are getting more and more influence. The

Anderson, ‘Reflections on the ICN and its NGAs’ (n. 87), 282–283. Lall, ‘Timing as a Source of Regulatory Influence’ (n. 134), 5. Emphasis original. 139 Ibid, 19. 137 138

158

Challenges to assumptions in competition law

Steering Group said that NGAs’ role should further increase, including early access to the agenda: A stronger/greater involvement of NGAs in all sessions was supported and may be considered in the early planning stages among the Annual Conference Planning Committee members.140

Solving this problem may help with another issue too. In part, the NCAs are too busy, and this relates to how many Working Groups and other meetings they are going to. If NGAs are given less of a role, this will create more work for the NCAs. As we have seen, this is a problem for many of them. However, if the solution were to create fewer Working Groups and for them to have less activity, we could provide more time for the NCAs to get good solutions. We would also reduce the multinationals’ influence. 6.4.3.3 Revolving doors Revolving door policies mean that NCA employees are encouraged to work in firms. Similarly, firms’ employees are encouraged to spend time working in NCAs. These policies, now common in many places, especially in the EU and the USA, help to consolidate an elite with collective goals, sharing the fruits of power. Importantly, this further extends rich, Western multinationals’ influence and power.141 Two processes are at work. First, all good employees working in NCAs know that, when they leave the NCA, there is the potential to work for a firm. Some may be motivated to moderate their behaviour while working in the NCA in order to make these potential, future benefits that they might receive from NGAs (be they law firms, businesses, etc.) more likely. This might incentivise them (whether this is intentional or not) to seek out and show deference to NGAs’ arguments, in their NCA work. Given their origin, these arguments likely involve (potential) impacts of competition rules on NGAs. They are unlikely to emphasise others’ interests unless they coincide with these multinationals’ interests too. The second process arises due to the perceived lack of funds that many NCAs say they face. In order to ameliorate this, and because they see it as good experience for their employees, many major multinationals and law firms encourage their employees to go on secondment to NCAs. This is a way of subsidising NCAs. Obviously, when on secondment these NGA employees are in a prime position to protect and pursue their employers’ interests. Spending

http://​www​.i​nternation​alcompetit​ionnetwork​.org/​uploads/​library/​doc997​.pdf, 2. Wilks, The Political Power of the Business Corporation (n. 70), 78.

140 141

Where influence lies in the International Competition Network

159

time in an NCA also helps these people to make connections with NCA employees to facilitate the first process, see above. Lall says that it is important for entities such as the ICN to ‘prevent informal social links between officials and regulated interests from becoming too strong’. This demands more distance between NCAs and NGAs. One way of achieving this is to slow, or stop, the revolving door between NCAs and NGAs, or at least ensure that NCAs recruit more widely. NCAs might also limit the number of candidates that they hire from NGAs.142 Currently, NCAs seem to have a desire to hire good candidates cheaply, and they can do this, in part, because of the revolving door. If we are to inhibit this, we might also need more financing from government, to make working in an NCA competitive with NGA salaries.

6.5 CONCLUSION We saw the ICN advocating ‘the adoption of superior standards and procedures in competition policy around the world … [formulating] proposals for procedural and substantive convergence, and … [seeking] to facilitate effective international cooperation to the benefit of member agencies, consumers and economies worldwide’. We are told that, in the ICN ‘public and private sectors, and member agencies from developing and developed economies, participate on an equal footing’.143 Formally, NCAs do participate on an equal footing. However, rich, Western NCAs and multinationals have disproportionate influence in the ICN, in fact. This matters. Competition rules are often widely drafted, often giving NCAs lots of interpretative discretion. The ICN dresses up its ‘Superior standards and procedures’ as value neutral. They are not. Framing discussions over the interpretation of competition rules as technical (rather than political), arrived at by a network of ‘NCA specialists’ rather than government officials, is clever. It disguises the re-distributive impact of ICN outputs (and undermines states’ ability to counteract this). The unanimity requirement formally protects all NCAs from undue influence. In reality, it principally protects rich, Western NCAs, because, as we saw, they have many tools that they can employ to influence other NCAs. Welcoming NGAs into the ICN’s heart also gives them (disproportionately rich, Western firms) the chance to pick and frame ICN debates.

Lall, ‘Timing as a Source of Regulatory Influence’ (n. 134), 19. See quotes around nn. 21 and 33.

142 143

160

Challenges to assumptions in competition law

Some have disproportionate influence in the ICN. This matters. So, ‘The Law and Politics of Global Competition: Influence and Legitimacy in the International Competition Network’ tries to improve upon this status quo.

7. Cartels’ little helpers: a comparative study of the case law regarding the facilitators of collusion in Europe, United States and South America Andrés Calderón 7.1 INTRODUCTION Cartels become more sophisticated as antitrust agencies evolve. With the purpose of mitigating the risks of detection or identification of its members, cartels look for newer ways to hide their collusive agreements. One way is by employing the help of a third-party that could organize, manage, and monitor the cartel and its rules. More importantly, these third-party agents can help cartel members fly under the radar of antitrust agencies. Cartels facilitators have recently become an important topic in international competition law literature, as a result of new case law studying the role of these facilitators. The prominent rise of facilitators is probably due to United States v Apple (hereinafter, the e-books case). However, American courts had already addressed the role of facilitators some years ago. In the e-books case, the United States Department of Justice filed a complaint against Apple and the five largest editorials,1 which had given Apple the right to publish their books in an electronic format under an agency model. Per this new contractual arrangement, the publishers would set the retail prices of the books sold through Apple, paying the company a commission of 30 percent. Additionally, Apple required the publishers to establish a most-favored-nation (MFN) clause, by which they were obligated to make the selling price of the e-books on the Apple platform the lowest on the market. The publishers also agreed to pressure the incumbent Amazon into adopting the new agency

1 Complaint. United States v Apple Inc., 952 F. Supp. 2d 638 (S.D.N.Y. 2013); aff’d 791 F.3d 290 (2d Cir. 2015).

161

162

Challenges to assumptions in competition law

model like the one with Apple, and to accept the minimum price fixed by the publishers.2 According to the district court, the agency model and the MFN clause were part of a conspiracy designed to increase the prices of e-books and eliminate price competition at the retail level.3 The decision was upheld by the Second Circuit Court of Appeals, which emphasized Apple’s responsibility.4 Across the Atlantic, other jurisdictions such as the United Kingdom have also developed a consistent case law regarding cartel facilitators. Replica Kits,5 Hasbro6 and the Dairy Case7 all involved arrangements between competitors that used their common suppliers as intermediaries to line up their retail prices. Retailers would provide sensitive commercial information to their suppliers knowing that it would be shared with their competitors. Then the retailers would complain to the common supplier, asking for the supplier’s mediation when a competitor would lower its prices. The landmark case (AC Treuhand) for the European Union presents a few differences insofar as the facilitator in AC Treuhand was not a common supplier for retailers who wished to coordinate their prices.8 Several competitors in the thermal stabilizers market retained the services of AC Treuhand, a Swiss consultancy firm, specifically for collusive purposes and for avoiding detection. The firm was in charge of collecting sensitive business data from the cartel members, analyzing it, processing it, and sharing it with the cartel members. AC Treuhand was also in charge of moderating and solving controversies between the market competitors. The European Commission imposed an individual fine on AC Treuhand of €348,000 for its role as facilitator. While in South American countries, there are fewer cases involving cartel facilitators, the legal discussion has not differed much from their European and North American counterparts. First at issue is whether competition law applies only to the market participants directly involved in the anticompetitive practice

2 Brent Fisse, Facilitating Practices, Vertical Restraints and Most Favoured Customers (2016) accessed 11 March 2020 at https://​www​.brentfisse​.com/​images/​ Fisse​_​-​_Facilitating​_practices​_most​_favoured​_customer​_clauses​_220516​.pdf. 3 United States v Apple, Inc., 952 F. Supp. 2d 638 (S.D.N.Y. 2013). 4 United States v Apple Inc., 791 F. 3d 290 (2d Cir. 2015). 5 See Decision of the Office of Fair Trading No. CA98/06/2003, ‘Price-fixing of Replica Football Kit’; Decision of the Competition Appeal Tribunal, Case 1021 and 1022/1/1/03 Allsports Limited and JJB Sports plc v Office of Fair Trading [2004] CAT 17. 6 See Decision of the Office Fair Trading No. CA98/18/2002, ‘Agreements between Hasbro UK Ltd and distributors fixing the price of Hasbro toys and games’. 7 See Decision of the Office Fair Trading No. CA 98/03/2011, ‘Dairy retail price initiatives’. 8 Case C-194/14 P, AC-Treuhand AG v Commission, 2015  EU:​C:​2015:​717.

Cartels’ little helpers

163

or also to the other agents that may collaborate with the planning, execution or concealing of a competition violation. Secondly, can a state entity or governmental official qualify as a cartel facilitator and therefore be held responsible for violating competition law? This chapter aims to take a look at the South American experience dealing with both topics of discussion regarding cartels facilitators. We will review the local case law and policy of the most active competition agencies. Section 7.2 is dedicated to Peru, arguably, the South American country with the most active and open discussion of the subject. Section 7.3 focuses on Chile. Section 7.4 and 7.5 address the Colombian and Argentinian practices, respectively. Section 7.6 looks at the differences and some common ground in the South American region. Finally, we conclude with some remarks on the aspects where South American antitrust enforcement could take the lead, especially, in advancing the expansion of competition law scope to include public officials who promote or collaborate with anticompetitive practices.

7.2 PERU 7.2.1

Peruvian Legislation

In Peru, competition law dates back to 1991 with Legislative Decree No. 701, which was replaced in 2008 by Legislative Decree No. 1034, the Law for the Repression of Anticompetitive Conducts. Since its origins, Peruvian competition law has been applied to any ‘economic agent’ or any entity who undertakes an economic activity. Importantly, this law explicitly includes directors and managers of companies that commit anticompetitive behavior, provided that they have participated in the planning or execution of such practices. Likewise, associations and guilds are also included in the scope of the law, provided that their members take part in the anticompetitive economic activity. In fact, both the directors of a company and trade or professional associations have specific fines set out in the law depending on the violation (up to 100 taxation units for representatives and up to 1,000 taxation units for associations).9 In 2015, Legislative Decree No. 1205 was approved, incorporating a modification of the law’s scope to explicitly include the figure of the facilitator as a possible recipient of penalties for their part in carrying out anticompetitive

9 Articles 43.2 and 43.3, Legislative Decree No. 1034, (25 June 2008), Diario Oficial ‘El Peruano’.

164

Challenges to assumptions in competition law

practices.10 Accordingly, the Explanatory Preamble to Legislative Decree No. 1205 stated the need to expressly incorporate a cartel’s facilitators within the scope of application of the Law due to the important role they played in the formation and survival of a cartel.11 The Law also contemplates the inclusion of officials, directors, and public servants who facilitate anticompetitive practices, if their participation does not occur in the exercise of their regular duties. Although this provision has yet to be applied, this could mean an important advance in the fight against the cartels, which are often encouraged by irregular governmental actions and state pressures. 7.2.2

Peruvian Case Law

The Commission for the Defense of Free Competition (CLC) of INDECOPI (the Peruvian competition agency) is the only Latin American authority that has explicitly addressed the role of cartel facilitators, interpreting them as distinct subjects aside from the directors of a company or from trade or professional associations, and separate from the companies participating in an anticompetitive practice happening in the market. The first time the CLC studied the role of the agents that, without participating in the market where a collusive agreement took place, were instrumental to the effectiveness of a cartel, was the well-known Pharmacies Case. The Technical Secretariat of the CLC, in the context of several dawn raids, found emails exchanged between the laboratories and pharmacies. Dozens of these communications revealed a particular collusive scheme among the pharmacy chains. Pharmacies came into contact with laboratories or distributors in order to compare and control the prices of their competitors. The emails contained complaints with requests for the matching of prices, complaints about price discounts or commercial promotions from their competitors, communications about upcoming price increases, coordination of agreements and ‘gentlemen’s agreements’, among others. The laboratories or the distributors were then in charge of communicating with the pharmacy chains in order to transfer the grievances from their peers. This way, the pharmacies’ representatives avoided communicating with each other. It was a three-sided coordination.12

Legislative Decree No. 1205 (23 September 2015), Diario Oficial ‘El Peruano’. Presidencia del Consejo de Ministros. ‘Exposición de Motivos del Decreto Legislativo 1205’, Sistema Peruano de Información Jurídica, 8–9. 12 Decision of the Commission for the Defense of Free Competition No. 078-2016/ CLC-INDECOPI [309–310]. 10 11

Cartels’ little helpers

165

Here, the laboratories not only participated as intermediaries for communications, but also sought to avoid direct contact between the competitors; thereby mitigating the risk of exposing the collusive practice. The CLC decided that the infringing practice was essentially horizontal in nature, and that the role played by the laboratories, then, was to facilitate the cartel’s activities and to prevent its detection.13 The CLC fined five pharmacy chains for horizontal collusion, but not the laboratories. Although it recognized their participation in carrying out the collusive practices as facilitators, the CLC understood that the laboratories were not liable under the competition law in force at that time.14 Then, the CLC issued its decision on the Hemodialysis Case, which involved 33 dialysis clinics that engaged in a collusive agreement aimed at increasing the referential value in the public biddings to contract hemodialysis services. The state entity Essalud, a health insurance and health care provider, called for the tenders for private clinics, which could provide hemodialysis services to a number of its patients. In order to determine the reference value that Essalud would have to pay, the governmental entity requested quotes from different private suppliers, so that, based on the various proposals, it could calculate a probable cost in the market. The CLC detected that there was an Association of Private Hemodialysis Centers, which served as a meeting and agreements platform for hemodialysis clinics. This Association would set out the instructions to be followed by all the clinics before Essalud, such as submitting identical quotes to Essalud, in order to artificially increase the price, or not participating in the tenders that did not meet the agreed referential value, as a way to put pressure on Essalud. Additionally, the cartel of clinics relied upon the advice of Mr. Raúl Barrios, the Association’s lawyer, who participated in the planning and administration of the cartel, including designing the cost structure used to determine the price agreed upon. Furthermore, as a way to exert pressure on Essalud, he proposed to file a number of complaints and observations with the sole purpose of having Essalud accept their concerted prices.15 However, similar to the Pharmacies Case, the lawyer Raúl Barrios was not punished for his part as a facilitator for the cartel. The CLC considered that at the time the actions took place, a rule allowing the penalization of facilitators was not in force.16

15 16 13 14

Ibid. at 104. Ibid. at 107. Ibid. at 496–497. Ibid. at 498, 500.

166

Challenges to assumptions in competition law

In addition to cartels facilitators, in Peru, it is undisputed that competition law applies to guilds and professional associations. INDECOPI has punished professional bars including those of pharmaceutical chemists, lawyers, notaries, engineers, and trade associations of transport, cattle breeders, and bakers, among others.17

7.3 CHILE 7.3.1

Chilean Legislation

Article 3 of the Chilean competition law, Legislative Decree No. 211, prohibits every act or convention that may, individually or collectively, restrict or hinder free competition. Article 3 subsections a, b, c and d include an open list of anticompetitive practices. Chilean Law (Article 3) does not state a clear difference between the types of persons to which the prohibitions of the Law apply.18 Article 26 of the aforementioned law states that the Tribunal for the Defense of Free Competition may impose fines to any person that participated in the anticompetitive practice.19 The Law is so broadly drafted that it has been interpreted to include any party that participates in an anticompetitive conduct. Likewise, Article 62 of Legislative Decree No. 211 – introduced in 2016 by Law No. 20.945, to reinstate criminal penalties for hard-core cartels – states that a criminal punishment may be imposed not only to a participant in the anticompetitive practice but also to the person organizing the anticompetitive action.20

17 Instituto Nacional de Defensa de la Competencia y de la Protección de la Propiedad Intelectual, Guía sobre asociaciones gremiales y libre competencia, (8 August 2019) accessed 11 March 2020 at https://​www​.indecopi​.gob​.pe/​ documents/​1902049/​3761587/​Gu​%C3​%ADa+​de+​Asociaciones+G ​ remiales+​y+​Libre+​ Competencia​.pdf/​38048db5​-e2fb​-7ba0​-f889​-609f864f7d07. 18 ‘Whoever executes or celebrates, individually or collectively, any action or convention that prevents, restricts or hinders free competition, or that may produce those effects, will be punished with the sanctions set forth in article 26 of this law …’. Legislative Decree No. 211, 7 March 2005, Diario Oficial [D.O.] (Chile). 19 ‘Fines can be imposed to the corresponding legal person, directors, administrators and every person that may have participated in the [anticompetitive] act.’ Legislative Decree No. 211, 7 March 2005, Diario Oficial [DO.O.] (Chile). 20 ‘Article 62. – Whoever enters into or orders to enter into, executes or organizes an agreement between two or more competitors, to fix sales prices or buying prices in one or more markets; to limit the production or supply; to divide, allocate or assign market quotas or areas; or to affect the result of public biddings conducted by public companies, private companies in charge of public utilities, or public entities, will be

Cartels’ little helpers

167

The wording of Article 62 makes it possible to interpret that a person or company may be liable for establishing an anticompetitive agreement, even if said person is not a part of that agreement. This third-party may be a firm, a person, or an association. This interpretation opens the door to punish a facilitator of an anticompetitive practice (with administrative fines or with imprisonment, if applicable). 7.3.2

Chilean Case Law

In 2014, the Tribunal for the Defense of Free Competition (TDLC) ruled on a very exemplary case. There, the Chilean National Economic Prosecutor’s Office (Fiscalía Nacional Económica, FNE) conducted an investigation into three poultry companies (Agrosuper, Aritzia, Don Pollo) and the poultry association (Asociación de Productores Avícolas de Chile, APA21). As a result of the investigation, all of them were punished pursuant to Article 3 of Legislative Decree No. 211. The poultry companies had exchanged sensitive commercial data22 on sales prices and sales volumes to collectively forecast the future demand and to agree on referential prices for poultry meat. They also colluded to limit the production of poultry meat in the Chilean market, by defining coordinately a predetermined level of production and sales quota for each company.23 The TDLC also found that the APA was a permanent coordinator, facilitating the cartel’s existence24 by promoting and monitoring compliance with the agreement.25 The TDLC went as far as ordering the dissolution of the APA.26 While the TDLC did not make a distinction between the articles of the Law applicable to the cartel members and to the trade association, it did emphasize the role played by the APA in the collusive scheme.27

punished …’. Legislative Decree No. 211, 7 March 2005, Diario Oficial [DO.O.] (Chile). 21 Decision of the Tribunal de Defensa de la Libre Competencia No. 139/2014, 25 September 2014 [1]. 22 Ibid. at 150. 23 Ibid. at 163. 24 Ibid. at 180. 25 Ibid. at 178. 26 Ibid. at 183. 27 ‘[A] trade association may have a facilitating role in opening markets and new businesses for its members or affiliates. However, if its purposes are perverted, this may pose a danger for free competition, inasmuch as it helps the coordination of the market conducts of its associates and, by doing so, it changes the competitive process.’ Ibid. at 157.

168

Challenges to assumptions in competition law

In 2015, the Chilean Tribunal decided another important case involving the Association of Obstetric Gynecologists of Ñuble (AGGOÑ) and 25 of its members. All of them were punished pursuant to Article 3 of the Legislative Decree No. 211. The health professionals had jointly decided to terminate their respective contracts with the provisional health institutions in Ñuble (Isapres), and agreed to the new minimum prices to be charged to the Isapres’ patients. The TDLC concluded that the Association worked as a vehicle to materialize the anticompetitive agreement,28 and that the Association could be held responsible separately from its members.29 Thus far, the Tribunal has yet to punish a cartel facilitator other than a trade or professional association. However, in 2014, the Tribunal took into consideration the role that some companies played in a collusive agreement without being part of the conspiracy. In this investigation, the FNE found that four intercity passenger transport companies (Tur Bus, Pullman, Cometa, Romani), colluded to block their competitors from access to bus terminals.30 The members of the conspiracy coordinated actions that tended to block their competitors’ access to bus terminals in the cities of Antofagasta, Coquimbo, La Serena and Valparaíso,31 thereby delaying or preventing entry of new competitors to the market of interurban transportation services to and from those cities. To achieve this, the investigated parties used their affiliated companies32 to help them secure the available spaces at the bus terminals, even if they were not going to use them. Nevertheless, the Tribunal did not punish those affiliated companies. Arguably, the most important case dealing with cartels facilitators in Chile was decided in 2019. The TDLC fined three supermarkets (Cencosud, SMU and Walmart) for price-fixing in the retail prices of fresh chicken meat,33 between the years 2008 and 2011, with the help of their wholesale suppliers. According to the collusive agreement, the retail price could not be lower than the wholesale price plus taxes.34 More specifically, the supermarkets had to follow the wholesale prices established by the suppliers (Ariztia y

28 Decision of the Tribunal de Defensa de la Libre Competencia No. 145/2015, 1 April 2015 [35]. 29 Ibid. at 36. 30 Decision of the Tribunal de Defensa de la Libre Competencia No. 134/2014, 30 January 2014 [76]. 31 Ibid. at 50. 32 Ibid. at 42. 33 Decision of the Tribunal de Defensa de la Libre Competencia No. 167/2019, 28 February 2019. 34 Ibid. at 35.

Cartels’ little helpers

169

Agrosuper).35 Compliance with this rule was conditioned to the abidance by their collusive counterparts.36 The wholesale suppliers not only fixed the chicken resale price, but they also had the responsibility of monitoring supermarkets from any departure from the agreement. The supermarket chains only communicated via their common suppliers, showing their concern when a competitor did not follow the agreement.37 According to the TDLC, this was a ‘hub and spoke’ type of practice.38, 39 The three investigated supermarkets argued that they did not have a horizontal agreement, but individual vertical agreements imposed by the chicken meat wholesalers.40 While the supermarket chains did not effectively coordinate between themselves but with the suppliers, the TDLC argued that these vertical agreements could become part of a horizontal collusive arrangement if the competitors were conscious that the same rules applicable to them were to be followed by their competitors.41 Indeed, taking into consideration the emails exchange between a supermarket and its suppliers,42 the Tribunal confirmed that the agreement also involved the rest of supermarket chains, inasmuch as the supermarkets would require compliance with the agreed upon retail price.43 Furthermore, the Tribunal confirmed this interpretation with the testimonies from the executives of Cencosud and a former Ariztia employee (one of the wholesale suppliers). Those testimonies reaffirmed that the supermarkets requested the wholesalers’ intervention so that the latter would make the other

Ibid. at 35. Ibid. at 98. 37 Ibid. at 100. 38 The FNE filed two expert opinions supporting its theory that this case consisted in a horizontal anticompetitive agreement: Okeoghene Odudu, Indirect information exchange. Report for FNE regarding the Praga investigation. Filed by the FNE in the Case ‘FNE’s complaint against Cencosud and others’ (Case No. 304-16), 30 November 2017. Spencer Weber Waller, Final Expert Opinion and Declaration of Professor Spencer Weber Waller Regarding Horizontal Price Fixing Investigation in the Retail Sale of Poultry in Supermarkets, 23 December 2015. Filed by the FNE in the Case ‘FNE’s complaint against Cencosud and others’ (Case No. 304-16), 30 November 2017. 39 Decision of the Tribunal de Defensa de la Libre Competencia No. 167/2019, 28 February 2019 [37]. 40 Ibid. at 33. 41 Ibid. at 39. 42 Ibid. at 88. 43 Ibid. at 70. 35 36

170

Challenges to assumptions in competition law

supermarkets comply with the same rules (retail prices), or punish them with a penalty in case of non-compliance.44 Finally, even though the chicken meat wholesalers fulfilled an essential role in the supermarkets’ anticompetitive scheme, they were not punished by the Tribunal, since the FNE did not include them in its indictment.45

7.4 COLOMBIA 7.4.1

Colombian Law

Colombia is one of the few jurisdictions where there has been an express recognition of the role of ‘facilitator’ in the law. Competition law in Colombia dates back to 1959, with Law No. 155, which was later reinforced by Decree No. 2153 in 1992 and Law No. 1340 in 2009. The competition authority is an administrative agency named Superintendence of Industry and Commerce (SIC). Article 4, subsection 15, of Decree No. 2153 establishes the economic penalties associated with violating the law. Subsection 16 further includes specific fines for natural persons. Due to this distinction, subsection 15 has been interpreted to be solely applicable to businesses (normally, legal persons) and subsection 16 to be applicable to their representatives (natural persons). Subsection 16 was amended in 2009, to include the possibility to impose fines ‘to any person that collaborates, facilitates, authorizes, executes or tolerates conduct[s] that violate[s] the norms for the protection of competition …’.46 For more clarity, in 2015, the Government enacted Decree No. 1523, introducing specific regulations for the Leniency Program. The Leniency Program was introduced through Law No. 1340, Article 14. Decree No. 1523 contains a number of definitions, including one for ‘economic agents’ and a separate one for ‘facilitators’.47 Decree No. 1523 expressly states that both the economic agents and the facilitators are eligible to apply for the Leniency Program.

Ibid. at 59. Ibid. at 37. 46 Law No. 1340 of 2009, 21 July 2009, Diario Oficial [D.O.]. 47 ‘2. Market agent: Every person that conducts an economic activity that affects or may affect its development, notwithstanding its form or legal nature, whatever the activity or economic sector may be. 3. Facilitator. Any person that collaborates, facilitates, authorizes or tolerates conducts that may consist in restrictive commercial practices, according to article 2 of Law No. 1340 of 2009, and the norms that complement or modify it.’ 44 45

Cartels’ little helpers

171

This distinction makes clear that according to the Colombian legislators, the facilitator of an anticompetitive practice is a separate entity from the economic agent that engages in an anticompetitive conduct; and that both individuals may be considered violators of Colombian competition law. 7.4.2

Colombian Case Law

As previously described, the facilitator is a different person from the undertaking that engages in an anticompetitive practice. Nevertheless, Colombian laws (Decree No. 2153, Article 4, subsection 16) do not distinguish between a person working within a firm and a person outside the firm hired for the specific purpose of helping a cartel. In other words, a facilitator could potentially be an administrator or manager of the anticompetitive agent, but it could also be a completely different person, such as a consultant, an advisor, or a provider of services. Thus far, the SIC has used Article 4, subsection 16, mainly to punish directors and administrators responsible for the anticompetitive conduct executed by the company they worked for. For instance, in 2015, the SIC punished two companies, Melquiades and Fundesoe, for infringing Article 47 of the Decree No. 2153 (horizontal collusion). The two companies filed the same bid in a public tender conducted by the Colombian Family Welfare Institute for the supply of nutritional supplements. It was later revealed that spouses Olga Florez and Jose Camacho controlled both companies, and simulated competition between the two companies to increase their chances of winning the tender. The SIC not only fined the two undertakings but it also punished Florez and Camacho, citing Article 4, subsection 16. While interpreting said part of the Law, the SIC argued that it was not necessary for there to be a direct relationship between the economic agent and the person who participated or influenced in the anticompetitive conduct.48 The SIC admitted then that an external agent (i.e., a facilitator) may be held responsible for infringing the competition law even if said agent is not a participant in the market where the collusion takes place. The competition agency also had the opportunity to address the question of whether a public official may be found responsible for a competition law infringement and it answered it affirmatively. This happened in the case against the Mayor of Bogota and other public officials of the Bogota Aqueduct, Sewerage and Cleanness Company of Bogota (EAB), the

48 Decision of the Superintendencia de Industria y Comercio No. 91235, 24 November 2015 [19].

172

Challenges to assumptions in competition law

Special Administrative Unit for Public Services (UAESP) and Bogota Water Company, all of whom were punished by the SIC. The SIC concluded that all of the investigated parties took action to monopolize the market for sewerage and cleanliness in the city of Bogota and to exclude potential competitors.49 Furthermore, the SIC considered that the public officials took part in this anticompetitive scheme and were held responsible as representatives for the public entities involved.50 In a more recent case, the SIC took one step further, investigating public officials not in their role as representatives of businesses or entities but as independent facilitators of cartel of private companies. The SIC initiated an investigation against Reingegas, Aribuk and SEG 3A Gas de Colombia, all firms dedicated to the installation of domestic natural gas. These companies were selected by the Mayor, Mr. Rodrigo Guarín Lemeses, to be in charge of the construction and installation of the internal pipelines for the supply of domestic natural gas in the city of La Mesa. After being selected, in a meeting arranged by the office of the Mayor, the three companies along with the Secretary of Public Works, Mr. Javier Pérez Rojas, signed an agreement, which included several conditions such as abiding by a price list and punishing any party who deviated from the agreement.51 The Colombian Competition Authority punished the Mayor due to its ‘determinant role in the celebration and execution of the price fixing agreement’, inasmuch as he ‘(i) propitiated an scenario contrary to free competition by choosing and preferring Reingegas, SEG 3A and Aribuk; (ii) induced consumers to hire those companies; (iii) actively participated in the anticompetitive agreement; and (iv) altered the dynamics of free competition by directing demand towards those three (3) agents’.52 Both public officials, Mr. Guarín and Mr. Pérez, were fined for collaborating with the cartel, according to Decree No. 2153, Article 4, subsection 16. While the individuals argued that the scope of the competition law did not reach them, the SIC considered that ‘free competition norms do not differentiate the legal condition of the investigated persons, instead, the only criterion is whether the investigated parties perform an economic activity or have an impact in an economic activity, or not’.53

49 Decision of the Superintendencia de Industria y Comercio No. 25036, 21 April 2014 [26]. 50 Ibid. at 121–146. 51 Decision of the Superintendencia de Industria y Comercio No. 91153, 14 December 2018 [32–33]. 52 Ibid. at 36. 53 Ibid. at 66.

Cartels’ little helpers

173

Consequently, according to Colombian case law: (1) Decree No. 2153, Article 4, subsection 16 is applicable to any individual that participates or collaborates with a cartel, either as a representative of an undertaking or either as an external agent; and, (2) a cartel facilitator may be a private party or a public official. On a related matter, trade associations and professional associations have undoubtedly been considered within the scope of application of the Colombian competition law. The SIC has steadily interpreted that these associations can infringe Article 46 of the Decree No. 2153 (as amended by the Law No. 1340).54,55

7.5 ARGENTINA 7.5.1

Argentinian Legislation

According to Argentinian competition law (Law No. 27442, of 2018), only economic agents can infringe the law, notwithstanding whether they are natural or legal persons, private or public entities. The only condition, according to Article 4, is that they perform economic activities with an effect on the national market. Articles 57 and 58 describe the different liability rules applicable to legal and natural persons. Article 57 states that companies are held responsible for the acts performed by natural persons on their behalf. Article 58 declares a shared liability between legal persons and its directors, managers, administrators, and legal representatives when their actions or omissions helped, encouraged or allowed the infraction. There are no specific provisions for infringement facilitators in the law, and a literal interpretation may lead some to think that only market competitors or their representatives could be held responsible for an antitrust violation. 7.5.2

Argentinian Case Law

The Argentinian antitrust authority (formerly, the National Commission for the Defense of Competition, or CNDC; now, the Tribunal for the Defense of

Decree 2153/1992, 31 December 1992, Diario Oficial [D.O.] (Colom.). Superintendencia de Industria y Comercio, Cartilla sobre la aplicación de las normas de competencia frente a las asociaciones de empresas y asociaciones o colegios de profesionales (21 September 2019) accessed 11 March 2020 at http://​www​.sic​.gov​.co/​si tes/default/files/files/cartilla_sobre_asociaciones_de_empresas_y_asociaciones_de_ profesionales2.pdf. 54 55

174

Challenges to assumptions in competition law

Competition) has had the chance to study the role played by different agents that could qualify in the definition of anticompetitive practices facilitators. For instance, an interesting case solved in 2002 involved two television channels (TRISA and TSCA) who held the exclusive right to broadcast local football matches, and three pay-television operators (VCC, Multicanal and Cablevision). As a result of the CNDC’s investigation, the competition authority concluded that TRISA and TSCA had engaged in a minimum resale price scheme with the pay-tv providers, as the latter had charged a fixed price to its clients for the broadcasting of the encrypted signal containing the football matches. While TRISA and TSCA were held as mainly responsible for the anticompetitive conduct, the authority also fined the pay-tv operators as ‘secondary participants’.56 The CNDC supported its ‘secondary liability’ interpretation from Article 46 of the Criminal Code, given that the pay-tv operators’ joint acceptance of the commercial terms provided by TRISA and TSCA worked as a ‘facilitating’ mechanism for the resale price scheme, and diminished the risk of negotiating conditions for each cable company individually. Nevertheless, the fact that the CNDC decided to attribute the ‘facilitating’ role to the cable companies and not to the TV channels is quite odd. If the CNDC would have interpreted this commercial arrangement as a horizontal collusion between the pay-tv providers, then the TV channels would have been categorized as helpers or ‘facilitators’ in a similar way as the United States Department of Justice did on the e-Books case. Another important case is the one filed by the CNDC against five valve manufacturers (Vaspia, Foragen, Tidar, Errepol, and Metalúrgica VG) and the company Axle S.A. The latter was appointed as the commercial agent for all five manufacturers. Axle was in charge of dealing and answering business requests from companies demanding valves. Therefore, Axle would set out the prices, commercial conditions, and decide which manufacturer would sell the valves for each buyer. The CNDC found evidence of the manufacturers’ agreement and its exclusive dealer. The investigated practice had already ceased by the time the CNDC had issued its final ruling. Nevertheless, the authority suggested fines for Tidar, Errepol and Axle.57

56 Decision of the Comisión Nacional de Defensa de la Competencia No. 28. Case No. 064-002331/1999, [10]. 57 Comisión Nacional de Defensa de la Competencia. Annual Report 1997 [25].

Cartels’ little helpers

175

While the CNDC did not specifically classify Axle as ‘facilitator’, its actions fit perfectly into that role. Axle was neither a competitor nor a competitor’s worker. It was a separate company, helping members of a collusive agreement. Trade associations and guilds have also been fined for antitrust violations in Argentina, even if the law does not explicitly refer to these types of organizations. In 2005, a case in point was decided by the CNDC involving several cement producers.58 The CNDC fined the companies Loma Negra, Cemento San Martin, Juan Minetti, Corcemar, Cementos Avellaneda, Cementos el Gigante, and Petroquimica Comodoro, for being part of a cartel to fix the price of cement and to allocate the market shares in different areas of the country. The guild of cement producers (Asociación de Fabricantes de Cementos Portland, or AFCP) was also fined for making its statistical system available to the cartel. The AFCP collected highly sensitive business information from each company (including sales and levels of production data) in the Portland cement market, and shared it with all the manufacturers on a weekly basis. This helped the cartel to monitor compliance with the agreement. The trade association also hired an external auditor to confirm the figures provided by the companies, and to estimate the amount each firm had to contribute to the association. Thus, the CNDC found that the AFCP had built an instrument to support and control the market and product allocation scheme organized in the cement market facilitating the tacit coordination between manufacturers.59 In 1986, the CNDC ruled on another relevant case, where several navy drivers’ unions and sand production companies were punished. The sand producers had reached a collusive agreement to limit its production and artificially raise its sales price. The sand companies paid the unions to ensure that their crews would report and oversee the amount of sand production that was agreed upon with the cartel members. The CNDC ruled that the agreement between the sand producers and the unions was part of a concerted action to limit the supply of sand and to increase 58 Opinion of the Comisión Nacional de Defensa de la Competencia No. 513. 25 July 2005. CNDC v Loma Negra, Cemento San Martin, Juan Minetti, Corcemar, Cementos Avellaneda, Cementos el Gigante, Petroquimica Comodoro and Asociación de Fabricantes de Cemento Portland (AFCP). 59 ‘In conclusion, a concerted exchange of information between competitors, with the same degree of detail, speed and disaggregation as the one implemented through the AFCP’s statistical system. This not only represented an instrument for control and support of the quota and market allocation agreement to a national level, but it also represented a mechanism that facilitates the tacit coordination between cement producers, with a potential damage to the economic general interest (articles 1 and 41, subsection b, of the Law No. 22.262.’ Opinion of the Comisión Nacional de Defensa de la Competencia No. 513.

176

Challenges to assumptions in competition law

its price. Hence, the CNDC fined both the sand companies and the unions for infringing the Argentinian competition law.60 Similar to other jurisdictions, Argentina’s competition authority recently published a ‘Guide on Defense of Competition for Associations, Business Guilds and Professionals Associations and Bars’ (December 2018), expressly indicating that these organizations are subject to the application of competition law.61 The CNDC has previously investigated and punished these types of organizations in various markets such as health clinics, private security and surveillance, cement, and pharmacies among others.

7.6

MOVING FORWARD

South American competition agencies have shown acuteness in identifying the facilitator’s role in anticompetitive schemes. Each of the South American countries mentioned in this chapter have consistently ruled that trade associations and professional bars are within the scope of competition law. While some countries have specific legal rules (e.g., Peru) or guidelines (Colombia, Chile and Argentina) to address these organizations, they have not been uniformly labeled as ‘facilitators’ (See Box 1, Table 7A.1 in Appendix). Considering the impact that trade associations have on their members, they have been on the receiving end of fines and even dissolution orders when competition authorities have discovered the anticompetitive measures those associations have adopted in the market. The situation becomes more complex when the agent that collaborates with a cartel has no affiliation with market participants. If we follow the categorization made in Europe (e.g., AC Treuhand, Replica Kits) and the United States (e.g., Toys ‘ R’ Us, Apple), the facilitator’s role is reserved for the separate economic agents that do not participate or have influence in the market where the anticompetitive practice takes place.

60 ‘In return for their cooperation, unions received maintenance for all ships without taking into account level of production that they generated.’ Opinion of the Comisión Nacional de Defensa de la Competencia Case No. 70.332/84, 29 September 1986 [2]. 61 Comisión Nacional de Defensa de la Competencia, Guía sobre defensa de la competencia para asociaciones y cámaras empresariales y colegios y asociaciones profesionales (10 December 2018) accessed 11 March 2020 at https://​www​.argentina​ .gob​.ar/​sites/​default/​files/​guia​_camaras​_y​_asociaciones​_empresariales​_10​-12​-2018​_0​ .pdf.

Cartels’ little helpers

177

In a broad way, all of the South American antitrust agencies studied in this chapter have dealt with cases of cartel’s facilitators, but with different approaches (See Box 2, Table 7A.1 in Appendix). For instance, neither Argentine legislation nor case law directly addresses the discussion of whether an agent qualifies as a cartel facilitator, even though some cases like the TRISA and TCSA case were appropriate to elaborate on the matter. In contrast, Peru expressly uses the term ‘facilitator’ to discuss the role played by an advisor (Hemodialysis case) and by a group of suppliers (Pharmacies case) as intermediaries in a collusive strategy. And while the Peruvian antitrust agency did not punish those facilitators, it recognized their roles and called for an amendment to the Repression of Anticompetitive Conducts Act expressly including facilitators in its subjective scope of application. Similarly, in Chile, the TDLC fined a cartel of supermarkets that received the intermediation of their poultry suppliers (Supermarkets case), but did not punish the wholesalers as facilitators inasmuch as the prosecutor (the FNE) did not include them in its accusation. Colombia presents a very particular case, because its legislation has an express legal recognition for the term ‘facilitator’. However, the SIC has not directly addressed the distinction between a person who facilitates the execution of an anticompetitive practice as part of a market participant (as a representative) and an outside person who collaborates with the cartel’s activities (as a facilitator). Nevertheless, this has not been an obstacle for the SIC to punish a public official who propitiated a price-fixing agreement between private companies (La Mesa case). This first comparative look at how South American antitrust authorities deal with cartels facilitators leaves us with some preliminary conclusions. First, all competition authorities under study (Argentina, Colombia, Chile and Peru) have investigated cases where cartels facilitators have played an integral role as part of a collusive scheme. Second, some competition authorities whose law did not have an express reference to ‘facilitators’ (Argentina and Chile) have not considered they were restricted from prosecuting or even sanctioning a cartel facilitator (Argentina). Only the Peruvian competition agency expressly recognized the role played by a cartel facilitator and still opted not to prosecute them. Instead, the Peruvian competition agency proposed an amendment to the law expressly considering cartels facilitators as within its subjective scope of application. Argentina, on the other hand, appears to be following EU and US case law by considering that a cartel facilitator violates competition law regardless of the lack of express mandate for its role in the law. Meanwhile, Colombian competition law does have an express but still broad reference to facilitators, and has considered that the scope of the law

178

Challenges to assumptions in competition law

reached even to public officials but has not punished a case of a private cartel facilitator yet (See Boxes 3, 4 and 5, Table 7A.1 in Appendix). This comprehensive approach has been a subject of criticism, specially, by some European authors, who have questioned the unpredictability of the European Commission’s ruling in AC Treuhand. Rick Busscher, Martin Herz, and Hans Vedder, for instance, believe that the decision to penalize a cartel facilitator must be made on the part of the legislator and not the European Court of Justice.62 Peruvian competition agency, INDECOPI, would concur with this reasoning, and that may be why they withdrew from punishing cartel facilitators before the law was amended to expressly include them. We disagree with the criticism about ‘uncertainty’. If the collusion involved a hard-core cartel such as price-fixing, market allocation or bid rigging, it would be quite difficult to imagine a scenario where an intermediary would not know that they are facilitating a clear violation of the law. Secondary or accomplice liability rules exist and are commonly applied in different areas of the law (e.g., Criminal Law, Consumer Protection, Intellectual Property). Unsurprisingly, this type of responsibility could also apply to third parties in competition law who consciously intermediate in a collusive arrangement. Our third conclusion regards the possibility of applying competition law to public officials or governmental entities who engage as cartel facilitators. Peru is the only country that has expressly addressed this issue in its legislation. Since its 2015 amendment, Peru’s competition law has contemplated the possibility of penalizing a public official as an infringer of the law for facilitating a collusive agreement. Taking a step further, the Colombian Competition Authority has gone as far as penalizing a public official for helping the execution of a private cartel in the La Mesa case (See Box 6, Table 7A.1 in Appendix). This measure could be the most important advancement in the South American approach to cartel facilitators. While this issue has not been covered by South American literature in competition law, the international discussion has been very controversial. Commenting on Spanish case law, for instance, Francisco Marcos – one of the academics that has studied in depth the application of competition law to public officials – has criticized extending competition law to public powers not acting as market participants.63 62 Rick Busscher et al., The Shortest Competition Judgment Ever: AC- Treuhand II, European Law Blog, accessed 11 March 2020 at https://​europeanlawblog​.eu/​2015/​11/​ 09/​the​-shortest​-competition​-judgment​-ever​-ac​-treuhand​-ii/​. 63 ‘It is true – as it is well put by the CNC in its decisions – that [the] Public Administration should respect the complete legal system, and so it is that any public entity that promotes, fosters, or favors the realization of anticompetitive illicit conduct will be acting against the constitutional mandate, and its acting is reprehensible,

Cartels’ little helpers

179

Julia Ortega adds on to that criticism; considering it absurd to impose fines on the Public Administration for an antitrust infringement since the monetary burden will ultimately be supported by citizens.64 We disagree with Marcos and Ortega. Public officials could very well be deterred not only through monetary fines but also through individual administrative penalties. Even though competition law was originally designed to punish market participants, there is no real reason to avoid extending its application to public officials when they have a clear mandate to respect competition and not to engage in illegal activities which may corrupt the competitive system. We believe this is the case when a public servant acts as a facilitator for a hard-core cartel. Eleanor Fox and Deborah Healy have analyzed the application of competition law to public officials in a very comprehensive comparative study, where they concluded that it was beneficial for antitrust law to expressly cover state officers, inasmuch as a proper understanding of the anticompetitive practice might need to involve them.65 Interestingly, Fox and Healy’s study shows the growing international trend of competition laws prohibiting government restraints to competition within most Central and Eastern European nations; where the state officials were prone to help cartels.66 Their study also shows that interviewees consider that ‘[competition] laws should apply to state officials who join and facilitate illegal private conspiracies and bid-rigging rings’, that is, to public officials acting as cartel facilitators.67

but that does not make its conduct an infringement of the Free Competition Law.’ Francisco Marcos, La sanción a las administraciones públicas cuando restringen la competencia o promueven conductas anticompetitivas sin actuar como operadores económicos en el mercado, Working Papers IE-Law School AJ8-237, (21 September 2017) accessed 11 March 2020 at https://​papers​.ssrn​.com/​sol3/​Delivery​.cfm/​SSRN​_I D3040026_code242325.pdf?abstractid=3040026&mirid=1. 64 Julia Ortega, Se puede sancionar a la Administración en el ámbito de la Defensa de la Competencia, Almacén Derecho, accessed 11 March 2020 at https://​ almacendederecho​ . org/​ c uando​ - se​ - puede​ - sancionar​ - la​ - administracion​ - ambito​ - la​ -defensa​-la​-competencia/​(last update 31 January 2017). 65 ‘The benefits of antitrust coverage of complicit state officers can be significant and coverage of the official natural. It can be natural because (and if) the officer is part of a unified conspiracy and is a necessary link to its success. The narrative of the conspiracy may importantly involve him or her … Moreover, the threat of antitrust liability may provide the decisive incentive to officials not to join illegal conspiracies and grease their wheels. The fact that a defendant is not a participant in the relevant market does not preclude liability.’ Eleanor M. Fox and Deborah Healy, When the State Harms Competition – The Role for Competition Law, 79 Antitrust Law Journal 769, 805 (2014). 66 Ibid. at 773. 67 Ibid. at 812.

180

Challenges to assumptions in competition law

In this context, the South American expansion of the competition law’s scope to apply to public officials facilitating the functioning of a cartel, whether express or implicit, might not be that surprising. Instead, we could be witnessing the consolidation of a doctrine, where cartels and their helpers are punished, no matter the identity of the person.

7.7 CONCLUSION The South American experience shows that the cartel facilitator’s role is a popular topic in the region and that the advancements within the case law regarding this issue are not so different than in Europe and the United States. Controversies over the necessity of an express reference to cartels facilitators in the law and the possibility of punishing public officials who promote horizontal collusion have also arisen in South American countries like Peru and Colombia. Populism and setbacks from a past of centralized-state interventions in the economy are two factors that continuously haunt South American market competition. Public officials are always tempted to support anticompetitive collusion and intervene in the market if it meant extracting personal benefits. Unsurprisingly, competition agencies have since fought against both private and public restraints to competition and have started to expand the scope of application of their respective competition laws to cover and punish cartel facilitators regardless of their (private or public) identity. Fortunately, while cartels are becoming more sophisticated, antitrust authorities ‘have leveled the playing field through their new legislation and prosecution techniques’.68

68 Andrés Calderón, El otro ‘colaborador’: Una mirada comparada sobre los facilitadores de los cárteles en el Derecho de la Competencia, 73 Themis 113, 128 (2018).

Cartels’ little helpers

APPENDIX Table 7A.1

Facilitators’ approaches by South American antitrust agencies

Box 1: Associations and guilds covered by competition law Country

Result

Argentina

Yes

Chile

Yes

Colombia

Yes

Peru

Yes

Box 2: Cartel cases involving a facilitator Country

Result

Argentina

Yes

Chile

Yes

Colombia

Yes

Peru

Yes

Box 3: Express recognition of ‘facilitator’ in the law Country

Result

Argentina

No

Chile

No

Colombia

Yes

Peru

Yes

Box 4: Express or implicit recognition of ‘facilitator’ in case law Country

Result

Argentina

Yes (Trisa and TSCA; Axle)

Chile

Yes (Chicken meat; Supermarkets)

Colombia

Yes (La Mesa)

Peru

Yes (Pharmacies and Hemodialysis)

Box 5: Fines applied to a cartel facilitator Country

Result

Argentina

Yes (Trisa and TSCA; Axle)*

Chile

No

Colombia

Yes (La Mesa)

Peru

No**

Box 6: Public officials or entities as cartel facilitators covered by competition law or case law Country

Result

Argentina

To be determined

181

182

Challenges to assumptions in competition law

Chile

To be determined

Colombia

Yes

Peru

Yes

Notes:

* Under secondary liability. ** Due to the absence of an express legal provision.

PART III

New institutional challenges

8. Public interest considerations in European merger control regimes Oliver Budzinski and Annika Stöhr1 8.1 INTRODUCTION Protecting and safeguarding market competition as the principal coordination mechanism of supply and demand serves the public interest. Coordinating economic activities through competitive markets provides a superior solution to the fundamental problem of scarcity.2 Consequently, market competition generates more social welfare than any alternative coordination mechanism. Notwithstanding, there may be additional public interests that potentially stand in conflict with market competition. Some of them may require additional policy tools adjusting market results according to these goals, like for instance in the case of income redistribution through taxes and social security services. 1 We would like to thank the participants of the 14th ASCOLA Conference, ‘Challenges to the Assumptions at the Basis of Competition Law’ (Aix-en-Provence, June 2019) and the 2019 Annual Conference of the German Law and Economics Association (Hanover, July 2019) for valuable comments on an earlier version of the chapter. 2 The theories of market failure do not stand in contrast to this general wisdom. The term market ‘failure’ is somewhat misleading since it ‘only’ implies that markets do not achieve the perfect level of social welfare all alone. Intelligent institutions and smart interventions may improve social welfare in these cases. Notwithstanding, the market does not ‘fail’; it still produces superior results to alternative coordination mechanisms. Note also that only ingeniously designed interventions may improve welfare here; real-world interventions, distorted by imperfect political and administrative processes and influenced by lobbyism, may actually perform worse than the imperfect market. Thus, the theories of market failure are rather theories of market imperfections and do not represent cases where less competition serves the public interest better than more competition. This is even true for the extreme case of public goods. The rare cases where users of a non-rival good cannot be excluded from consumption because of inherent goods characteristics imply the necessity for the public to enforce payment of users in order to avoid widespread free rider problems – but notwithstanding a competitive market still serves welfare better than a monopoly or bureaucratic administration.

184

Public interest considerations in European merger control regimes

185

In these cases, having effectively competitive markets as a starting point is still beneficial for public interest. In rare cases, however, a social goal may be better achieved through the natural opposite of competition, i.e. through market power. We will call such goals non-market public interests. If there is a trade-off between effective competition and non-market public interest, then a democratic society must be allowed to value the conflicting goal higher than the welfare effects of effective competition (in this special area), given the decision process is made under awareness of the welfare costs of pursuing the conflicting public interest. Therefore, there may be scope for correcting purely competition-based decisions on non-market public interest grounds. Note, however, that in the real world, vested interests may hide behind ostensible public interests in order to protect or create non-competitive rents – at the expense of society. In an imperfect world, the identification problem of real public interest versus ostensible public interest must be considered, given the imperfections of political and administrative processes and the power of lobbyism. Merger control represents a particular relevant field of competition policy where these types of public interest considerations play a role. Competition authorities analyse mergers and acquisitions with respect to their effects on competition. The goal of merger control is to protect competition through blocking combinations of companies that substantially or significantly lessen competition in the affected markets. However, in exceptional cases, non-market public interest considerations may justify to allow for anticompetitive concentration if two conditions are fulfilled: (1) the non-market public interests outweigh the public interest of competitive markets; and (2) the anticompetitive merger is instrumental for achieving the respective non-market public interest goal. While the public interest-based clearance of an anticompetitive concentration is the dominant case, it is also conceivable that a pro-competitive merger is challenged on non-market public interest grounds. Here, antitrust authorities would allow the merger because of a lack of significant anticompetitive effects and/or considerable efficiency advantages, whereas non-market public interest considerations advocate blocking the merger. The topic of public interest considerations in merger control recently received renewed attention in the context of the European Commission’s decision to block the merger between Siemens and Alstom. Politicians from the home countries of the two companies eventually published a ‘Franco-German Manifesto for a European industrial policy fit for the 21st Century’.3 Here the

3 ‘A Franco-German Manifesto for a European Industrial Policy Fit for the 21st Century’ (Bundesministerium für Wirtschaft und Energie and Ministère de L’Économie et des Finances) https://​www​.bmwi​.de/​Redaktion/​DE/​Downloads/​F/​franco​-german​

186

Challenges to assumptions in competition law

German and French Ministers for Economic Affairs suggest an appellation right for the European Council, so that it can override merger prohibitions by the European Commission on the grounds of public interest considerations. This proposal would extend an already long list of public interest considerations and exceptions in European merger control regimes. These include, e.g., securing jobs, supplying goods such as energy and medical care, or preserving the free press. However, some of these public interest considerations clearly do not match our definition, as they do not trade-off with effective competition. This contribution aims to give an overview on public interest considerations in European merger control regimes and takes a deeper look into German merger control, where the Federal Minister of Economic Affairs may allow a merger deemed to be anticompetitive by the Federal Cartel Office (FCO) on public interest grounds (so-called ministerial approval; ‘Ministererlaubnis’). We answer the research questions (1) whether the public interests stipulated in the respective competition laws are non-market in nature; and (2) if, in anecdotally chosen cases, the respective permitted anticompetitive mergers were instrumental to achieve the stated non-market public interest goals. For the German case study, we (3) provide an overview whether the postulated public interest goals have ex-post been promoted by the past cases of ministerial approvals.

8.2

PUBLIC INTEREST CONSIDERATIONS IN MERGER CONTROL

In the European Union, merger control takes place on the European level (based upon EU law with the European Commission as the competition authority) as well as on the member state level (based upon national law enforced by national competition authorities). Additionally, member states can also directly apply European law.4 Under strict conditions, member states have the possibility to consider public interests when applying EU law by the provisions of Article 21 EUMR. Public security, plurality of the media and prudential rules are seen as prima facie legitimate to constitute exceptions for the otherwise exclusive consideration of competitive aspects, whereas the European

-manifesto ​ - for ​ - a ​ - european ​ - industrial ​ - policy ​ . pdf ​ ? ​ _ ​ _ blob ​ = ​ p ublicationFile ​ & ​ v ​ = ​ 2 accessed 31 January 2020. 4 See for questions of competence allocation and delineation: Oliver Budzinski and Arndt Christiansen, ‘Competence Allocation in EU Competition Policy as an Interest-Driven Process’ (2005) 25(3) JPP 313; Oliver Budzinski, ‘An Economic Perspective on the Jurisdictional Reform of the European Merger Control System’ (2006) 2(1) ECJ 119; Oliver Budzinski, The Governance of Global Competition: Competence Allocation in International Competition Policy (Edward Elgar 2018).

Public interest considerations in European merger control regimes

187

Commission may decide case-by-case if other public interest aspects can be applied ex-ante in a given merger case.5 The different member states of the EU design their merger control regimes individually and can therefore decide whether they include public interest considerations or not. If they do, they can shape their regulations differently, e.g. in terms of sectors covered, consequences of a successful application of public interest considerations in a merger case, or the decisive body that clears or prohibits a merger based on public interests. Table 8.1 gives an overview of the EU member states that actually include public interest considerations into their merger control legislation. Shown are all states in which the respective public interest aspects are included in competition law – besides that, there are several states that include public interest considerations on single case level, without explicit legal basis. That makes a deeper investigation of the instruments and the reasons behind the usage difficult and, in tendency, inaccurate. Thus, they are not included in Table 8.1 and will not be analysed in this chapter. For deeper analysis, we choose four European countries where public interest considerations are included in the respective competition law. We describe public interest aspects in the Austrian, German, French and UK6 merger control. To ensure comparability of our analysis, we examine the different merger control regimes under the same comparative values: (1) we first check if the respective regime stipulates truly non-market public interests following our definition. Secondly (2), we analyse the instruments with regard to their transparency and the legal certainty for all parties involved. In a third and fourth step, we (3) look into the frequency of the actual usage of the instruments, and (4) evaluate anecdotally if the decisions actually promoted public interest. In this step, we analyse if the anticompetitive merger was instrumental to reach the respective non-market public interest goals. The last step (5) is a final evaluation of the respective instrument in terms of asking if it is, in its current form, suitable to include important non-market public interest aspects into merger control decisions. 8.2.1

Public Interest Considerations in Austrian Merger Control

In Austrian merger legislation, public interests must be taken into account regarding media mergers. Besides anticompetitive effects, media mergers can also have negative effects in terms of decreasing media plurality and increas5 OECD, Public Interest Considerations in Merger Control – Summaries of Contributions (DAF/COMP/WP3/WD(2016)30, 2016) 6. 6 At the time of writing, the UK was still a member of the EU and has an interesting way of including public interests into their merger control. Therefore, it is used here for comparison.

188

Table 8.1

Challenges to assumptions in competition law

Public interest considerations in EU merger control regimes

Country

Public interest considerations

Austria

Merger in media sector is to be blocked if it is likely to decrease media plurality (§13 KartG) ‘Federal Cartel Prosecutor’ to represent public interests in important cases or industries (§75 KartG)

Finland

Electricity market has special thresholds: total volume of electricity distributed at 400V or number of electricity plants belonging to the merging parties amounts to over 25% of the national electricity grid – then, intervention in merger does not require that a significant impediment to effective competition can be demonstrated (Section 25 of the Finnish Competition Act)

France

Minister for Economic Affairs has power to evoke a case for reasons of general interest: industrial development, competitiveness of undertakings relative to international competition, or creation/maintenance of jobs (Article L. 430 – 7 – 1 Code de Commerce)

Germany

Minister of Economic Affairs can clear blocked merger based on public interest considerations: competitiveness of undertakings relative to international competition, macroeconomic advantages, outstanding interest of the general public (§42 GWB)

Hungary

Merger that helps to preserve jobs or to assure security of supply can be claimed to be of strategic importance on the national level (Article 24/A of the Competition Act)

Ireland

Special rules for mergers in the media sector to preserve media pluralism (section 23 of the Competition Act)

Italy

Italian Competition Authority (ICA) can exceptionally allow an otherwise prohibited merger if relevant national economic interests are involved (Article 25 (1) of Law No. 287/90) After clearance by the ICA, Italian Prime Minister can prohibit the acquisition of an Italian company by a foreign firm if in the foreign country Italian undertakings are discriminated against, especially in connection with the ability of Italian firms to acquire local companies (Article 25 (2) of Law No. 287/90) ICA can clear an anticompetitive merger in the banking sector to protect the economic stability of one or more parties involved in the merger (Article 25 (5-bis) of Law No. 287/90)

Netherlands

Minister of Economic Affairs can clear a blocked merger based on not specified public interest considerations (Article 47 of the Dutch Competition Act)

Portugal

Regulatory Authority for the Media can intervene to protect freedom and pluralism of the media (Article 4B of the Portuguese Television Act) Council of Ministers can intervene based on fundamental strategic interests of the national economy (Article 41 of the 2014 PCA’s By-laws)

Public interest considerations in European merger control regimes

Country

Public interest considerations

Spain

Council of Ministers can clear a prohibited merger due to public interests (Article

189

10 of the Competition Act 15/2007): defence and national security, protection of public security or public health, free movement of goods and services within the national territory, environment protection, promotion of technological research and development, guarantee of adequate maintenance of the objectives of sectorial regulation United Kingdom

Secretary of State can intervene if he considers public interest aspects to be present: national security, media plurality, stability of the UK financial system (section 58 of the Enterprise Act)

ing media bias. If a merger between media companies is likely to decrease media plurality, it must be prohibited (§13 KartG). Media plurality is defined by the Austrian competition law as ‘a variety of independent media companies, which are not related to each other within the meaning of §7 KartG and by which a media coverage of different opinions is ensured’ (§13 (2) KartG). Hence, media plurality is not only about a wide variety of titles; instead, independent coverage must be ensured through a respective ownership structure.7 Therefore, higher concentration in this market can be seen as a threat to media plurality – showing that protection of competition and public interest considerations do not trade-off here. Besides the application of public interests in media mergers, there is another particularity in Austrian merger control: the so-called ‘Bundeskartellanwalt’ (Federal Cartel Prosecutor, FCP).8 His main task is ‘the representation of public interests in matters of competition law at the Higher Regional Court of Vienna as a cartel court’ (§75 KartG). What constitutes public interests in the context of this law, however, is not defined.9 An evaluation of the two particularities of the Austrian merger control regime regarding our five criteria yields the following insights: 1. Non-market public interests Media plurality, minimization of media bias, and independent media coverage could be seen as non-market public interest reasons for otherwise anticompetitive mergers if market-powerful media companies experience fewer incentives 7 ‘Medienfusionskontrolle’ (Bundeskanzleramt) https://​www​.bundeskanzleramt​ .gv​ . at/​ a genda/​ m edienrecht/ ​ r undfunk ​ - und ​ - medienrecht​ - in​ - osterreich/​ m edien fusionskontrolle.html accessed 31 January 2020. 8 Beirat für Wirtschafts- und Sozialfragen, Zukunft der Wettbewerbspolitik in Österreich (No. 84, 2010). 9 ‘Der Bundeskartellanwalt’ (Bundesministerium Justiz) https://​www​.justiz​.gv​ .at/​home/​home/​justiz/​justizbehoerden/​bundeskartellanwalt​~36c​.de​.html accessed 31 January 2020.

190

Challenges to assumptions in competition law

to bias news than media firms facing intensive competition.10 However, under empirically realistic conditions like imperfect information, information asymmetries, and the presence of politically motivated media entrepreneurs, effective competition among media companies serves these goals better than narrow oligopolies or a powerful monopoly.11 Furthermore, the definition of media plurality used by the Austrian competition authorities makes media plurality dependent on the ownership structure and, thus, on the competitive conditions on the market. For these reasons, media plurality cannot be seen as a non-market public interest. The FCP is a further instance that is able to include public interests into merger proceedings and can also intervene in merger cases on grounds of purely competition-related aspects, for instance, market delineation.12 The public interests the prosecutor should represent according to the law are not defined in detail but seem to mostly not coincide with our definition of truly non-market aspects that cannot be reached (better) through competition. 2. Transparency The regulation of media mergers included in Austrian competition law and the proceedings in such cases seem to be relatively transparent, whereas the tasks of the FCP are less clear. This may be because the law does not define the public interests it shall represent. Historically, the FCP was set up to give politicians the opportunity to intervene in antitrust proceedings.13 This way, the institution is potentially prone to lobbyistic influence. Due to the considerable scope for interpretation of public interest and the potentially occurring political influence, it depends on the individual acting of the FCP whether it weakens or strengthens the protection of competition. 10 Sendhil Mullainathan and Andrei Shleifer, ‘The Market for News’ (2005) 95(4) AER 1031. 11 Ronald H. Coase, ‘The Market for Goods and the Market for Ideas’ (1974) 64(2) AER 384; Matthew Gentzkow and Jesse M. Shapiro, ‘Media Bias and Reputation’ (2006) 114(2) JPE 280; Matthew Gentzkow and Jesse M. Shapiro, ‘Competition and Truth in the Market for News’ (2008) 22(2) JEP 133; Simon P. Anderson and John E. McLaren, ‘Media Mergers and Bias with Rational Consumers’ (2012) 10(4) JEEA 831. 12 Inter alia, Alfred Mair and Gustav Stifter, Republik Österreich: Bundeskartellanwalt – Tätigkeitsbericht 2015, https://​docplayer​.org/​55738771​ -Taetigkeitsbericht​ -2015​ .html accessed 6 January 2021; Alfred Mair and Gustav Stifter, Republik Österreich: Bundeskartellanwalt – Tätigkeitsbericht 2016, file:///​ C:/​Users/​anst3458/​AppData/​Local/​Temp/​t​%C3​%A4tigkeitsbericht​_bkanw​_2016​.pdf accessed 6 January 2021; Alfred Mair, Gustav Stifter and Heinz L. Majer, Republik Österreich: Bundeskartellanwalt – Tätigkeitsbericht 2017, file:///​C:/​Users/​anst3458/​ AppData/​Local/​Temp/​t​%C3​%A4tigkeitsbericht​_bkanw​_2017​_final​.pdf accessed 6 January 2021. 13 Beirat für Wirtschafts- und Sozialfragen (n. 8).

Public interest considerations in European merger control regimes

191

According to his activity reports of the last years, the current FCP seems to interpret his tasks mostly as the protection of consumer welfare – which is a public interest in general, but, at the same time, is also the goal of competition policy itself. Therefore, it represents no additional non-market public interest. The activity reports show a variety of interferences of the FCP in several merger cases. However, we did not find any case where the FCP intervened because of truly non-market aspects. Altogether, the institution FCP lacks transparency, not just in terms of the reasons to intervene, but also in terms of actual tasks, and increases legal uncertainty. 3. Usage As the activity reports show, the FCP intervenes in several cases per year – mostly with the aim of improving the respective market delineation and, therefore, preventing or containing anticompetitive effects of mergers. In 2017, he was also active in one merger in the TV sector, where he interfered to implement comprehensive behavioural remedies to especially ensure the diversity of news reporting.14 Besides this combination of the two public interest intervention possibilities in one case, real non-market public interest cases are rare. 4. Success/instrumentality There are positive effects achieved through the FCP – mostly regarding the protection of competition. Due to the rare intervention on non-market public interest grounds, potential effects in this regard are difficult to assess. Nevertheless, the separate rules for media mergers seem to cause the desired positive effects on overall welfare in terms of the public interest in media plurality. After the 2017 Pro7 Sat1 Puls 4/ATV merger, comprehensive news media coverage could be preserved through behavioural remedies.15 Nevertheless, since it required post-merger remedies to safeguard media plurality that existed pre-merger, it is difficult to understand why a prohibition of the merger – protecting both plurality and competition – would not have been the superior solution. Obviously, intensity of competition and public interest achievement were not a trade-off here. Overall evaluation 5. The overall assessment of the public interest considerations included in Austrian competition policy displays room for reform. Whereas the incentives of special rules regarding mergers in the media sector are understandable,

Ibid. Mair, Stifter and Majer (n. 12).

14 15

192

Challenges to assumptions in competition law

the objective could be achieved with a stricter competition policy, as there is no apparent trade-off between the public interest competitive markets and the public interest media plurality. Additionally, the institution FCP is non-transparent. A clearer definition of tasks and especially relevant public interest aspects that the FCP shall preserve should be provided in order to improve transparency, legal certainty, the achievement of targets and, above all, to effectively represent non-market public interests. The arbitrariness of the public interest interpretation seems to be the most relevant problem with the current institutional settings, allocating considerable interpretation scope to and, thus, sensitively relying on the individual FCP. Thus, a narrow definition of non-market public interests according to our definition (see Section 8.1) is imperative. 8.2.2

Public Interest Considerations in French Merger Control

The French Minister of Economic Affairs can overrule a decision of the competition authority (Autorité de la Concurrence, FCA) if there are reasons that are linked to industrial development, business competitiveness in terms of international competition, and/or creation and preservation of employment. There are two possible starting points for the Minister to intervene in a FCA-merger case on public interest grounds: after the FCAs clearing of a Phase I merger review, the Minister can ask the FCA to open a Phase II in-depth investigation. The second possibility occurs after the FCAs decision in a Phase II investigation; the Minister can then overrule this decision because of the mentioned public interest aspects. Therefore, he has the power to prohibit a pro-competitive merger or to allow an anticompetitive merger on public interest grounds (Article L. 430 – 7 – 1 Code de Commerce). The analysis of the French merger control regime in terms of public interest considerations shows the following results: 1. Non-market public interests The three mentioned potential public interest reasons do not trade-off with market competition and, thus, do not match our definition of non-market public interests. All of them are linked to macroeconomic effects, which can be reached better through competitive markets than through market-power-creating mergers. They do not stand in conflict with the protection of competition as the general goal of merger control and competition policy. In particular, the preservation of employment is highly unlikely to be achieved by market concentration since mergers lead to the reaping of synergy effects, which implies

Public interest considerations in European merger control regimes

193

a reduction of workforce16 and, under certain circumstances, an increase of involuntary unemployment.17 Furthermore, mergers increase market power on labour markets, leading to worse conditions for employees.18 Additionally, international competitiveness is dynamically harmed if domestic competition intensity decreases because of anticompetitive concentration. In the course of time, firms who do not have to compete on their home markets become less fit for international competition. 2. Transparency The public interests are explicitly stated in the law; however, their formulation remains rather broad. This may lead to uncertainty for merging companies in terms of the applicability of the reasons to their specific case. On the other hand, this broad formulation and missing specific definition or delimitation of e.g. positive effects on the industrial development may also lead to exploitation by the companies. Such open formulations are prone to lobbyism and could be utilized by the companies to maximize their individual profits under the guise of public interests. Nevertheless, the specifications for the procedure increase transparency and legal certainty for all parties potentially involved. 3. Usage The Minister’s right to re-examine a merger decision of the FCA was implemented in French merger control in 2008 with the advent of the FCA. Since then, it was used by the respective Minister only once in the 2018 Agripole/

16 Inter alia, Anju Seth, ‘Sources of Value Creation in Acquisitions: An Empirical Investigation’ (1990) 11(6) SMJ 431; Martin Conyon, Sourafel Girma, Steve Thompson and Peter W. Wright, ‘The Impact of Mergers and Acquisitions on Company Employment in the United Kingdom’ (2002) 46(1) EER 31; Klaus Gugler and B. Burcin Yurtoglu, ‘The Effects of Mergers on Company Employment in the USA and Europe’ (2004) 22(4) IJIO 481; Tarun K. Mukherjee, Halil Kiymaz and H. Kent Baker, ‘Merger Motives and Target Valuation: A Survey of Evidence from CFO’s’ (2004) 14(2) JAF 7; David N. Margolis, ‘Should Employment Authorities Worry about Mergers and Acquisitions’ (2006) 5(2) Port Econ J 167; Anca D. Chirita, ‘The Impact of Economic Efficiency on Employment: A Case Study of Mergers and Acquisitions’ (2016) 82nd International Atlantic Economic Society Conference Washington D.C., USA, 13–16 October 2016. 17 Oliver Budzinski and Jürgen-Peter Kretschmer, ‘Horizontal Mergers, Involuntary Unemployment, and Welfare’ (2016) 21(3) JER 297. 18 David Autor, David Dorn, Lawrence F. Katz, Christina Patterson and John Van Reenen, ‘Concentrating on the Fall of the Labor Share’ (2017) 107(5) AER 180; Suresh Naidu, Eric A. Posner and E. Glen Weyl, ‘Antitrust Remedies for Labor Market Power’ (2018) 132(2) HLR 537.

194

Challenges to assumptions in competition law

Cofigeo case,19 where the Minister used his power to clear the merger without remedies (that were imposed by the FCA, who cleared the merger under conditions) for reasons of job security.20 4. Success/instrumentality In the mentioned case, the remedies drawn by the FCA raised concerns by the Minister regarding the potential impact of the remedies on labour and industrial development. Therefore, he induced a ‘Phase III’ investigation after which he overruled the FCA’s decision by allowing the merger to take place without any remedies.21 Whether, in fact, these remedies would have had a negative impact on the job situation or the industrial development of the firm and/or the French overall economy remains unclear. However, mergers in general tend to have negative effects on the number of jobs in the merged company, due to synergy effects, as well as on labour standards. Another questionable point is if the development of one single company should be put on one level with the development of the overall economy. It is at least controversial if the job situation and economic development of a single firm can really be of general public interest. Overall, there is no discernible conflict between the objective of preserving jobs and the objective of protecting competition. Therefore, the anticompetitive merger (now without remedies) is not likely to help achieve the goal of job preservation. 5. Overall evaluation The fact that the ‘evocation’ power of the French Minister of Economic Affairs is not based on clearly defined public interest reasons and the rare usage of the instrument overall indicate that there is in fact no real need for it in the merger control regime – at least in the current form.

19 ‘Acquisition of William Saurin’ (Autorité de la Concurrence Press Release, 14 June 2018) https://​www​.aut​oritedelac​oncurrence​.fr/​en/​communiques​-de​-presse/​14​ -june​-2018​-acquisition​-william​-saurin accessed 31 January 2020. 20 ‘The Minister of Economy Intervenes to Authorize the Takeover of William Saurin by Cofigeo Without Divestment of Assets’ (Dentons Competition Newsletter August 2018, 27 August 2018) https://​www​.dentons​.com/​en/​pdf​-pages/​ge​neratelist​ ingpagepdf​ ? isPdf​ = ​ t rue​ & ​ I temId​ = ​ u bZ​ 3 5SKD3ZNdh​ v CRXetaOe/​ o ​ A 8iULfZgg6​ 5fQXJVce7pTClbiwHOIw​=​=​accessed 31 January 2020. 21 Christina Renner and Petra Kupka, ‘Developments in National Competition Laws (April 1, 2018 – June 30, 2018) – France’ (2018) 68(9) WuW 459.

Public interest considerations in European merger control regimes

8.2.3

195

Public Interest Considerations in UK Merger Control

There are several possibilities to raise public interest concerns to the UK government: (a) the Secretary of State for Business, Energy and Industrial Strategy can intervene on grounds of national security and financial stability of the UK financial system; and (b) the Secretary of State for Digital, Culture, Media and Sport can intervene on grounds of media quality, plurality, and standards. It is also possible for the respective Secretary of State to intervene because of additional public interest considerations. These additional considerations can be implemented into section 58 of the 2002 Enterprise Act after the two Houses of Parliament approve the order to do so. In so-called special public interest cases, (c) mergers that do not meet the common thresholds for merger investigations may still be investigated solely on public interest grounds. These special rulings apply to mergers involving government contractors holding confidential information regarding the country’s defence and certain mergers in the media sector.22 The proceedings of a public interest merger case are not strictly divided into a ‘competition-analysis part’ and a ‘further public interests part’. If the above-mentioned public interest considerations are potentially relevant for the case, the respective Secretary of State may issue a public interest considerations notice (PIIN). After this issue, the Competition and Markets Authority (CMA) examines the case in terms of competition and non-competition aspects at the same time, including third party views (Phase 1). If there were in fact important public interest aspects found in Phase 1, the CMA initiates Phase 2, where additional reports are provided to the Secretary of State. The Secretary eventually makes the decision whether or not he permits or allows the merger based on public interest considerations.23 The analysis results based on our five questions are as follows:

22 Alex Chisholm and Nelson Jung, ‘The Public Interest and Competition-Based Scrutiny of Mergers: Lessons from the Evolution of Merger Control in the United Kingdom’ (2014) October 2014 (1) CPI Antitrust Chronicle; Maurits Dolmans, Nicholas Levy, Romano Subiotto QC, Paul Gilbert and David R. Little, ‘UK Introduces New Thresholds for National Security Mergers’ (Cleary Gottlieb Alert Memorandum, 22 March 2018) https://​www​.clearygottlieb​.com/​-/​media/​files/​alert​-memos​-2018/​uk​ -introduces​-new​-thresholds​-for​-national​-security​-mergers​-pdf​.pdf accessed 31 January 2020; Federico Mor, ‘Contested Mergers and Takeovers’ (2018) House of Commons Library Briefing Paper No. 5374. 23 Competition and Markets Authority, Mergers: Guidance on the CMA’s Jurisdiction and Procedure (2014).

196

Challenges to assumptions in competition law

1. Non-market public interests The public interest aspects potentially justifying interventions of politics are mostly not non-market and may be achieved (better) through competition. Only a potential threat to national security could justify an intervention of the Government on grounds of public interests. Financial stability represents a more controversial case. 2. Transparency The rules regarding public interest implementation in this respective merger control are transparent and the proceedings are predefined. Nevertheless, due to the unclear potential reasons and the possibility to include new public interest reasons at any time, the overall assessment of the transparency and legal certainty through the procedure is rather negative. The possibility of exploitation of the instrument by companies is existent. 3. Usage The possibility of including public interest considerations into UK merger control cases was used regularly in the past. There were three cases of media mergers, and one merger case has been assessed for its potential impact on financial stability. This merger was also the reason to include the public interest of stability of the UK financial system in law during the 2008 financial crisis.24 The public interest of national security was used in six cases, showing the important role of this reasoning in UK merger control. Thus, the merger rules (including the above-mentioned thresholds for special public interest cases) regarding the government’s power to scrutinize mergers and takeovers that may raise national security concerns have been updated in 2018. These new rules especially apply to tech firms developing military technology and shall keep the countries’ openness to trade and investment, while not risking national security.25 4. Success/instrumentality All the merger cases in the financial and the media sector raised controversial discussions on their actual positive effects on public interests afterwards.26

24 Louise Smith, ‘The Lloyds-TSB and HBOS Merger: Competition Issues’ (2008) House of Commons Library Briefing Paper No. SN04907. 25 ‘New Merger and Takeover Rules Come into Force’ (UK Government Press Release, 11 June 2018) https://​www​.gov​.uk/​government/​news/​new​-merger​-and​ -takeover​-rules​-come​-into​-force accessed 31 January 2020. 26 Cosmo Graham, ‘Public Interest Mergers’ (2013) 9(2) ECJ 383; ‘What Are the Issues in Fox’s Sky Deal?’ (BBC News, 23 January 2018) https://​www​.bbc​.com/​news/​ business​-40434381 accessed 31 January 2020.

Public interest considerations in European merger control regimes

197

As mentioned above, the issues are either controversial or do not represent non-market public interests according to our definition. Therefore, we conclude that in fact non-market public interests were not protected in these cases. Without an in-depth ex-post evaluation of the other cases in the defence sector and their results, a clear statement on the actual protection of public interests in these cases is difficult to derive. Nevertheless, in these cases actual public interests, which potentially stood in conflict with the goal of protecting market competition, were taken into account and could at least potentially have outweighed the public interest of safeguarding competition. 5. Overall evaluation Since there is only one actual public interest aspect included in UK merger control – national security – the overall assessment of the instrument in its current form is rather negative. As in the other merger control regimes, the reasons justifying political intervention in merger cases should be revised in terms of including only public interests that cannot be achieved through competition. The development of the UK merger rules concerning cases that potentially hold a threat to national security can be evaluated as positive.

8.3

THE ‘MINISTERERLAUBNIS’ IN GERMAN MERGER CONTROL

As a case study for an ex-post analysis of the actual effects of public interest considerations in merger control, we utilize the example of German merger control and the so-called ministerial approval or ‘Ministererlaubnis’. In German merger control, competition and market analysis are meant to be strictly separated from further (e.g. public interest) aspects. The German FCO (Bundeskartellamt) prohibits a merger if it would significantly impede effective competition (§36 (1) GWB) – this decision is made solely based on competition aspects. After the FCO’s decision, the companies involved can apply for ministerial approval. This approval can be granted if, in the individual case under consideration, the competition restrictions resulting from the merger, which were determined by the FCO, are outweighed by (a) macroeconomic advantages, (b) an overwhelming public interest, and/or (c) the strengthening of the international competitiveness of the companies. Prior to the decision, an opinion from the Monopolies Commission (MC)27 must be obtained, stating 27 The MC is an independent advisory body that provides statutory and special opinions on competition policy, competition law and regulation issues and advises the government and other legislatures in these areas. See: ‘Mission’ (Monopolkommission) https:// ​ w ww ​ . monopolkommission ​ . de/ ​ e n/ ​ m onopolies ​ - commission/ ​ m ission ​ . html accessed 31 January 2020.

198

Challenges to assumptions in competition law

whether the reasons for the exception are reasonable and likely to be achieved through the anticompetitive merger. The Federal Minister of Economic Affairs decides upon the approval decision and he may deviate from the MC’s advice. An approval must not be granted, however, if the anticompetitive effects of the concentration in question would jeopardize the overall market economy (§42 (1) GWB). In order to safeguard and support the public interest reasons, remedies may be combined with the permission (§42 (2) GWB). Overall, the decision of the Federal Minister of Economic Affairs is solely based on the three mentioned aspects and explicitly not a repeal of the decision of the FCO. Since its implementation in the early 1970s, 23 applications for a ministerial approval were submitted to the Federal Ministry of Economic Affairs. Ten of them were granted (with or without conditions), six were rejected, and seven were withdrawn by the companies. Applications were made by companies from different sectors, like energy, media, engineering, and hospitals.28 An overall remarkable aspect is that the Federal Minister decided against the explicit recommendation of the MC in some of the approval cases – but in none of the rejection cases. The members of the MC recommended a ministerial approval only four times, whereas the Federal Minister decided more than twice as many cases in favour of the companies. This may indicate how the instrument can be (ab)used to enforce vested interests.

28 ‘Übersicht über die bisherigen Anträge auf Ministererlaubnis nach § 24 Abs. 3/§42 GWB’ (Bundesministerium für Wirtschaft und Energie, 6 September 2019) https://​www​.bmwi​.de/​Redaktion/​DE/​Downloads/​Wettbewerbspolitik/​antraege​-auf​ -ministererlaubnis​.pdf​?​_​_blob​=​publicationFile​&​v​=​9 accessed 31 January 2020.

Public interest considerations in European merger control regimes

199

The pros and especially the cons of this instrument have been subject to controversial academic discussion.29 Stöhr and Budzinski30 add to this discussion by analysing the actual effects of nine cases where the ministerial approval was granted and by evaluating ex-post whether the goals for the approvals were achieved. Table 8.2 presents an overview of these approved cases. Table 8.2 shows that the reasons leading to a ministerial approval are various and not always strictly non-market. According to our definition of public interests, only the case of Uniklinik Greifswald/KKH Wolgast is based on aspects that cannot be achieved through competition. Other cases only partly include such aspects – for example, the quality of jobs in EDEKA/ Kaiser’s Tengelmann or the security of know-how regarding working teams in Thyssen/Hüller-Hille. The overview also shows that some ‘groups’ of reasons are more successful in terms of receiving an approval than others. Interestingly, these are mostly reasons that are formulated in a rather vague way and, thus, tend to be challenging in ex-post analysis. This may be due to the characteristics of public interest considerations in general or deliberately designed in order to make accountability for decisions more difficult. Nevertheless, the ex-post effects analysis shows mostly negative results, i.e. the ministerial approvals rarely achieved their own goals. This is particularly true for public interest considerations mixing in economic- and competition-related aspects, whereas the picture looks somewhat more positive for cases of truly non-market public interest reasoning. Table 8.3 gives an overview of the cases and states if the expected public interest effects of the anticompetitive merger actually occurred and if, in fact, they were accomplished through the respective merger. The most recent 29 Inter alia, Hans O. Lenel, ‘Zum Teerfarbenurteil und zur sogenannten Fusionskontrolle’ (1972) 23 ORDO 307; Michael Krakowski, ‘Gemeinwohlvorteile durch Großfusion?’ (1989) 69(8) Wirtschaftsdienst 368; Steffen J. Roth and Michael Voigtländer, ‘Die Ministererlaubnis für den Zusammenschluss von Unternehmen – ein Konflikt mit der Wettbewerbsordnung’ (2002) 51(2) Zeitschrift für Wirtschaftspolitik 231; Florian Bien, ‘Die Berücksichtigung nichtwettbewerblicher Aspekte in der Fusionskontrolle – Gibt es Alternativen zur Ministererlaubnis?’ (2016) 4(10) NZKart 445; Rupprecht Podszun (2016), ‘Die Ministererlaubnis – Einbruch der Politik ins Recht der Wirtschaft’ (2016) 9 NJW 617; Rupprecht Podszun, ‘Die 9. Novelle des Gesetzes gegen Wettbewerbsbeschränkungen (GWB) – Stellungnahme als Sachverständiger im Wirtschaftsausschuss des Deutschen Bundestages zur Vorbereitung der Anhörung am 23.1.2017’ (2017) accessed at https://​www​.bundestag​.de/​resource/​blob/​489168/​ eff​ffe1ad50da​2f28f43442​b2d8be7c1/​podszun​-data​.pdf 18 December 2020; Oliver Budzinski and Annika Stöhr, ‘Die Ministererlaubnis als Element der deutschen Wettbewerbsordnung: eine theoretische und empirische Analyse’ (2018) 69 ORDO 216. 30 Annika Stöhr and Oliver Budzinski, ‘Gemeinwohl durch Marktmacht? – Eine Ex-Post-Analyse der Ministererlaubnis-Fälle’ (2019) 10 WuW 508.

200

Table 8.2

Challenges to assumptions in competition law

Cases of granted ministerial approvali

Case (Year) – sector

Recommendation MC

Reason for approval

VEBA/Gelsenberg (1974)

Do not approve

Securing energy supply

– Energy Babcock/Artos (1976) –

(approved) Do not approve

Mechanical engineering Thyssen/Hüller-Hille (1977) –

Workplace security (approved under conditions)

Approve under conditions

Mechanical engineering

Workplace security for technically high-skilled teams (securing know-how) (approved under conditions)

VEBA/BP (1978/79) – Energy

Do not approve

IBH/WIBAU (1981) –

Approve

Securing energy supply (approved under conditions)

Mechanical engineering Daimler-Benz/MBB (1989)

Workplace security (approved)

Approve under conditions

– Aerospace

Privatization of Airbus and prevention of market monopolization in favour of Boeing (approved under conditions)

E​.ON/​Ruhrgas (2002) – Energy

Do not approve

Uniklinik Greifswald/KKH

Approve

Securing low-cost energy supply (approved under conditions)

Wolgast (2007/08) – Hospital

Securing status as university clinic and research in the field of community medicine (approved)

EDEKA/Kaiser’s Tengelmann

Do not approve

(2015/16) – Food retail

Workplace security in terms of quality and quantity (approved under conditions)

Miba/Zollern (2019) – Mechanical Engineering

Do not approve

Know-how and innovation potential for energy system transformation, maintaining a competitive SME sector (approved under conditions)

Note: i BMWi, ‘Erlaubnis eines Zusammenschlusses zur Sicherung der Energieversorgung, Verfügung des Bundeswirtschaftsministers vom 1. Februar 1974’ (1974) 24 WuW 343; BMWi, ‘Erlaubnis eines Zusammenschlusses unter Auflagen, Verfügung des Bundeswirtschaftsministers vom 17. Oktober 1976’ (1976) 26 WuW 659; BMWi, ‘Teilerlaubnis eines Zusammenschlusses mit Entflechtungsauflage, Verfügung des Bundeswirtschaftsministers vom 1. August 1977’ (1977) 27 WuW 663; BMWi, ‘Erlaubnis eines Zusammenschlusses zur Absicherung der Energieversorgung, Verfügung des Bundeswirtschaftsministers vom 5. März 1979’ (1979) 29 WuW 499; BMWi, ‘Ministererlaubnis wegen Verbesserung der Möglichkeiten des Marktzutritts auf Auslandsmärkten und Erhaltung von Arbeitsplätzen, Verfügung des Bundeswirtschaftsministers vom 9. Dezember 1981’ (1981) 31 WuW 163; BMWi, ‘Ministererlaubnis eines Zusammenschlusses mit erheblichen Auswirkungen auf Märkte für Rüstungsgüter, Verfügung des Bundeswirtschaftsministers vom 6. September 1989’ (1989) 39 WuW 947; BMWi, ‘Ministererlaubnis unter Auflagen, Verfügung des Bundeswirtschaftsministers vom 5. August 2002’ (2002) 52 WuW 751; BMWi,

Public interest considerations in European merger control regimes

201

‘Wissenschaftliche Bedeutung als Gemeinwohlgrund, Verfügung des Bundeswirtschaftsministers vom 17. April 2008’ (2009) 59 WuW 683; BMWi, Verfügung in dem Verwaltungsverfahren Edeka/Kaiser’s Tengelmann, Gesch.-Z.: I B 2 – 22 08 50/01 (2016); BMWi, Verfügung in dem Verwaltungsverfahren Miba/Zollern, Gesch.-Z.: I B 2 – 20302/14–02 (2019).

Table 8.3

Effects of ministerial-approved mergersi

Case

Reason

Effect occurred?

Caused by merger?

Energy Cases

Security of energy supply

Yes

No

Reasonably priced energy

No

-

supply Babcock/Artos

Workplace security

No

-

Thyssen/Hüller-Hille

Workplace security for

Rather Yes

Rather Yes

technically high-skilled teams (securing know-how) IBH/WIBAU

Workplace security

No

-

Daimler-Benz/MBB

Privatization

Rather No

-

Decrease federal subsidies

Rather No

-

Increasing competitiveness

Yes

Yes

Uniklinik Greifswald/

Securing status as university

Yes

Yes

KKH Wolgast

clinic Yes

Yes

Yes

No

No

-

Securing community medicine research EDEKA/Kaiser’s

Workplace security in terms

Tengelmann

of quality Workplace security in terms of quantity

Note: i Stöhr and Budzinski (n. 30).

case, Miba/Zollern, is excluded because an analysis of the effects cannot be made due to the short time since the decision. We combine the three cases in the energy sector, where nearly the same reasoning was used in the single cases and only one particular market is relevant for the ex-post analysis. Security of jobs in terms of quantity is often used as a public interest concern. However, general empirical evidence shows that mergers lead to overall fewer jobs and negative labour market effects (see Section 8.2.2), whereas competition has a positive impact on the development of employment. Our analysis shows that this is also true for the respective ministerial approval cases – there are slight indications for positive effects of mergers on workplace security in only one case. In the case Edeka/Kaiser’s Tengelmann, the (very specifically defined) ‘quality’ of jobs was formally kept but only through special conditions coming with the approval. The merger itself has had negative effects on the overall job situation – jobs were already cut shortly

202

Challenges to assumptions in competition law

after the ministerial approval. The security of reasonable-priced energy supply was used as a potential positive effect of the three merger cases in the energy sector. By comparing energy price developments in different European countries and checking for potential blackouts, we conclude that there were no such blackouts and the development of energy pricing in Germany was not particularly positive compared to neighbouring countries. Only the hospital case can be seen as a success in terms of reaching public interest goals through an anticompetitive merger.31 In all other cases, the occurrence of the positive effects that were the reasons for the approvals cannot be verified ex-post. Our case study shows that even in an established public interest legislation, where competition-related and non-competition aspects are strictly separated and the instrument is implemented in competition law, positive public interest effects occur only in a few cases and in even fewer cases are the result of the approved anticompetitive merger. Therefore, we conclude that the instrument of a ministerial approval in its current form does not represent a beneficial part of the German competition policy – and should definitely not be extended to the European level. Our ex-post analysis shows that the instrument does not fulfil its duties – in most cases, the predicted positive public interest effects were not capable of outweighing the negative effects on competition. Thus, the instrument must undergo significant reforms (if it is not abolished altogether). There are several possibilities for these reforms: e.g. changing the possible reasons for an approval to strictly non-market public interests, or changing the decision body.32

8.4

CONCLUDING REMARKS: ECONOMIC JUSTIFICATION OF PUBLIC INTEREST CONSIDERATIONS IN MERGER CONTROL?

Table 8.4 gives an overview of the results of our analysis. The overall evaluation of the differently designed instruments in all four jurisdictions tends to be rather negative – especially regarding the public interests that may justify an intervention.

31 This specific hospital merger was necessary to preserve the status of university hospital, and, therefore, the education of young doctors in the region as well as the existence of the unique research area ‘community medicine’. In Germany, there are specific regulations and requirements for university hospitals: in order to ensure the quality of the education of young doctors, university hospitals must meet minimum thresholds of case numbers of specific treatments (in particular types of surgery). Only due to the merger (increasing the number of beds and treatments), the university hospital of Greifswald could maintain meeting these thresholds. 32 Budzinski and Stöhr (n. 29).

203

Public interest considerations in European merger control regimes

Table 8.4

Analysis results for exemplary merger control regimes French Merger

Austrian Merger

UK Merger

German Merger

Control

Control

Control

Control

No

Partly

Partly

Partly

Transparency of

Proceedings – yes

Institution of

Proceedings – yes

Proceedings – yes

procedure

Reasons – no

Federal Cartel

Reasons – no

Reasons – no

Non-market public interests?

Prosecutor – no Regulation of media mergers – yes Usage of instrument

Once

Rare

Regularly

Rare

Success/

Consequences

Rather no

Partly

Partly

instrumentality of

are not yet

instrument to reach

foreseeable, but Rather negative

Rather negative

Rather negative

public interest goals

rather no

Overall evaluation of

Rather negative

instrument

All the current concepts suffer from definitions of public interest that are too broad and not focused on non-market public interests, i.e. such public interests that are better achieved through market power than through effective competition. This offers scope for injecting vested interests of politicians, companies and other powerful lobbies into the decision process. At the end of the day, this harms social welfare without serving any real public interest. Consequently, the answer to our first research question, whether the public interests stipulated in the respective competition laws are non-market in nature, is clearly: no, since none of the regimes limits itself to non-market public interests. Further deficiencies include that none of the analysed regimes and cases include checks whether the market power resulting from the permitted anticompetitive mergers is instrumental to achieve the stated public interest goals. Given these results, it is not surprising that the empirical record of past public interest interventions into merger control is mixed at best from an ex-post perspective, which represents the answer to our second research question. For Germany, the desired public interest effects were (probably) achieved only in a minority of cases and in even fewer cases, the market power of the allowed anticompetitive merger was instrumental in achieving the desired public interests. There is no indication that the records for other regimes are likely to look any better. Thus, altogether, the public interest considerations in the current regimes are deficient and contribute neither to welfare nor to the interests of the public. Calls for extending provisions like these either in scope or to the European level

204

Challenges to assumptions in competition law

cannot be supported by academic analysis. Such plans must be expected to be detrimental to social welfare without serving any real public interest. Quite in contrast to the calls for an extension, a beneficial reform must narrow and focus the scope for public interest considerations and provide institutional checks and balances against abuses. Primarily, only non-market public interest can provide a ground for overriding competition-focused merger control decisions. In order to safeguard this, we propose a four-step test:33 1. Identification of a social goal that is demonstrably in the interest of the general public. 2. Evidence of a trade-off between this public goal and effective competition. 3. Proof that this goal is better achieved by market power than by the protection of the competition process. 4. Balancing the welfare costs of the considerable weakening of effective competition with the benefits of better fulfilment of the social goal, including proof that no alternative, ‘milder’ instruments exist that would enable an equivalent achievement of the social goal. The enforcement of this test must be given into the hands of an independent decision-making body like a law court. In order to maintain transparency and protect the competition authority from political pressure, it should be strictly separated from the competition policy analysis and decision. If a competent authority wants to get a merger control decision overruled on public interest grounds, it needs to seek permission of an anticompetitive merger or prohibition of a pro-competitive merger in front of the court. The burden of proof with respect to all four steps must be allocated to the party seeking the overturning of the merger control decision on public interest grounds.34 In the aftermath of the prohibited Siemens/Alstom merger, the latest controversy around public interest considerations in EU merger control focuses on promoting National or European Champions. German and French politicians are calling for ‘a more ambitious European industrial strategy with clear objectives for 2030’35 including changes in European merger guidelines and better possibilities to subsidize European companies to promote their international competitiveness.36 This implies that there would be a public interest for generating and/or supporting European Champions and that this represents a trade-off with effective domestic competition. The underlying

Ibid. Ibid. 35 Bundesministerium für Wirtschaft und Energie & Ministère de L’Économie et des Finances (n. 3) 1. 36 Ibid. 4. 33 34

Public interest considerations in European merger control regimes

205

logic would be that companies with domestic market power can use the supracompetitive rents from domestic markets (which are yielded at the expense of customers and society) to increase their international competitiveness and, thus, reap higher rents from foreign markets that overcompensate the domestic welfare losses. However, there is vast theoretical and empirical evidence that companies protected from domestic competition tend to concentrate on exploiting their market power to the detriment of consumers and society as well as on lobbying for and extending non-competitive pensions and privileges (rent-seeking instead of competitiveness) – ultimately also to the detriment of their international competitiveness.37 Thus, there is no non-market public interest in protecting single industries or companies from (national or international) competition.

37 Inter alia, Frédéric Jenny and Damien Neven, ‘Competitio Policy in the Aftermath of the Siemens/Alstom Prohibition: An Agenda for the New Commission’ (2019) 2-2019 Concurrences 2; Massimo Motta et al., ‘Open Letter: More, Not Less, Competition Is Needed in Europe’ (CPI, 10 February 2019) https://​www​.​competitio​ npolicyint​ernational​.com/​more​-not​-less​-competition​-is​-needed​-in​-europe/​ accessed 31 January 2020; Jeromin Zettelmeyer, ‘The Return of Economic Nationalism in Germany’ (2019) PIIE Policy Brief 19-4 https://​ www​ .piie​ .com/​ publications/​ policy​ -briefs/​return​-economic​-nationalism​-germany accessed 31 January 2020.

9. Formalism, fairness and freedom of contract: abuse of dominance in the UK courts and the business ‘Achilles heel’ of objective justification1 Barry J. Rodger 9.1 INTRODUCTION This chapter will undertake a detailed review of the UK courts’ jurisprudence in relation to the application of the abuse of dominance prohibition under Article 102 TFEU and the Chapter II prohibition of the Competition Act 1998. The outline hypothesis is that one can observe a change in understanding of the ‘abuse’ concept by the judiciary recently away from a more formalistic freedom of contract/property based approach grounded on laissez-faire free market based ideology towards jurisprudence which is willing to engage with and be more critical of businesses’ market strategies and behaviour. First, the chapter will review some central themes in the background theoretical and jurisprudential debate on general approaches to market regulation and the specific scope and role for the abuse prohibition in competition law, and consider the interplay between fairness and freedom of contract in that context. Second, the chapter will undertake a detailed review of the UK abuse of dominance case-law followed by a close analysis of the key issues which arise thereunder, notably the courts’ approach to objective justification for commercially rational behaviour under the prohibition. In particular it will be suggested that more recent judicial rulings on the determination of what constitutes abusive behaviour, for instance in Purple Parking, Arriva, Socrates and Achilles,2 have indicated a more critical approach to purported justifications for market-based business decisions.

1 Many thanks to Liam MacLean, CMA for his helpful comments on an earlier draft though responsibility for all errors and omissions remain mine alone. 2 See discussion further infra.

206

Formalism, fairness and freedom of contract

9.2

BACKGROUND DEBATE

9.2.1

Markets, Fairness and Formalism

207

The discussion in this chapter is partly inspired by the ‘controversial’ work by Ayal in Fairness in Antitrust: Protecting the Strong from the Weak,3 where he argues that monopolists have not received a fair deal and for greater protection of the property and contract rights of companies with market power, relying on (what Ayal views as) basic tenets of society such as freedom of contract and property rights to justify certain business behaviour. This argument reflects a belief in markets, and their self-correcting nature, as opposed to a more critical and interventionist approach to markets advocated by many other antitrust scholars.4 An appreciation of this broader context in which ‘monopoly’ laws function was also evident in Stucke’s seminal article, ‘Reconsidering Competition and the Goals of Competition Law’,5 wherein he posed the question whether post-financial crisis one could have the same trust or faith in the functioning of markets. In that context he has questioned what is meant by the term ‘competition’ and stressed that the term is normative, with different conceptions of what we perceive as an appropriate degree of competition, noting for instance the pejorative use of the term ‘cut-throat’ competition.6 Ayal’s contribution to the debate in Fairness in Antitrust also allows us to reflect in this chapter on judicial approaches to the abuse doctrine, in relation to the UK courts at least, in the wider context of jurisprudential theory. His thesis is redolent of legal formalism and a laissez-faire world view based on the freedom of the individual to strike private bargains.7 As Duxbury notes,8 legal formalism focuses on private interests; views state regulation as undesirable and unnecessary and asserts that private rights should be governed solely by contractual rights and duties: an approach epitomised by the Supreme Court in Lochner v New York.9 This Lochner style laissez-faire approach was subsequently subjected to considerable criticism focused on the need for

3 Idi Ayal, Fairness in Antitrust: Protecting the Strong from the Weak (Hart Publishing, 2010). 4 See for instance, Tim Wu, The Curse of Bigness, Antitrust in a New Gilded Age (Columbia Global Reports, 2018). 5 (2011) 81 Mississippi Law Journal 107. 6 See Maurice Stucke, ‘ Is Competition Always Good?’ (2013) 1(1) JAE 162–197. 7 Neil Duxbury, Patterns of American jurisprudence (OUP, 1995), in particular, Chapter 1 ‘The Challenge of Formalism’. See also e.g. Patrick Selim Atiyah, The Rise and Fall of Freedom of Contract (Oxford Clarendon Press, 1979) at pp. 226–231. 8 Duxbury (n. 7) Chapter 1. 9 1905, 198 US 45.

208

Challenges to assumptions in competition law

real economic freedom, based on equality of bargaining power.10 Giocoli11 recounts the interpretation of laissez-faire constitutionalism as driven by social Darwinists who ‘identified the only function of the State as the protection of property and enforcement of contracts’. Accordingly laissez-faire liberalism was tied up with individualism and freedom of contract and non-interference with property rights and their value.12 In affirming and applying the core constitutional principles of economic freedom and liberty of contract, laissez-faire constitutionalist judges did nothing more than ensuring that state action supported, rather than threatened, those natural rights of property and exchange that were fundamental to both classical liberalism and classical political economy.13

In contrast with the debate on formalism in a specific EU competition law application context,14 formal rationality or legal formalism signifies the exclusion of non-legal considerations from law-making and adjudication: ‘formalism implies a narrow ... approach to legal control – the use of clearly defined, highly administrable rules, an emphasis on uniformity, consistency and predictability, on the legal form of transactions and relationships and on literal interpretation … according to rule’.15 This jurisprudential background has also influenced the antitrust debate where academics concerned by the prominence of the Chicago School of Economics promoted alternative theories of antitrust based on non-efficiency goals and a policy of legislative activism.16 The relevance of the debate is epitomised by Posner’s conservative ideology – based on wealth maximisa-

10 Duxbury (n. 7) Chapter 5, ‘Economics in Law’, for example at p. 324: ‘Lawyers such as Robert Hale, Karl Llewellyn, John Dawson and John Dalzell endeavoured to show that the economic freedom by which the Supreme Court swore was not really freedom at all. That it was merely freedom to engage in economic coercion.’ 11 Nicola Giocoli, ‘Classical Competition and Freedom of Contract in American Laissez Faire Constitutionalism’ 10 June 2014, available at http://​papers​.ssrn​.com/​sol3/​ papers​.cfm​?abstract​_id​=​2448419. 12 See also John Stuart Mill, Principles of Political Economy (Longmans, Green & Co, 1848 [1909]) II ii 3. 13 Giocoli (n. 11) at 48. 14 Wouter Wils, ‘The Judgment of the EU General Court in Intel and the So-called More Economic Approach to Abuse of Dominance’ (2014) 37(4) World Competition 405–434. 15 Doreen McBarnet and Christopher Whelan, ‘The Elusive Spirit of the Law; Formalism and the Struggle for Legal Control’ (1991) 54 MLR 848. 16 See for example Robert Pitofski, ‘The Political Content of Antitrust’ [1979] 127 University of Pennsylvania Law Review 1051–1075.

Formalism, fairness and freedom of contract

209

tion – which essentially upholds the ‘legitimacy of existing power relations’.17 Sunstein has for instance stressed that the law – even of the liberal formalistic legal tradition – has an inevitable role in shaping preferences, i.e. preferences for autonomy and private property.18 For example, in the context of EU competition law, the Irish freezer exclusivity case saga highlighted the inherent paradox in competition law.19 The Irish High Court found no convincing reason why the defendant (HB) should unilaterally abandon the widespread practice of freezer exclusivity and noted that: ‘It may well be that strategies adopted by a firm in a dominant position would prevent competitors from eroding its market power: that is what they are intended to do.’20 This view may partly be a manifestation of the historical formal legal tradition in emphasising the sanctity of property rights.21 Indeed, one of the reasons that competition law was introduced by statute in the UK was the perceived laissez-faire attitude by the judiciary to market conduct which, while rational for the businesses undertaking it, restricted, distorted, or eliminated competition.22 More recently, there has been a resurgence in antitrust scholarship in the US, branded ‘hipster’ or ‘neo-Brandeisian’ antitrust, which has generally argued for a more effective and aggressive antitrust enforcement particularly against ‘tech giants’ in the digital economy.23

17 Duxbury (n. 7) at 402. See Richard Posner, Economic Analysis of Law (2nd edn., Little Brown, 1977). 18 See Cass Robert Sunstein, After the Rights Revolution: Reconceiving the Regulatory State (Harvard University Press, 1993) and Free Markets and Social Justice (OUP, 1997). 19 See Barry Rodger, and Angus MacCulloch, Competition Law and Policy in the EU and United Kingdom (5th edn., Routledge, 2014) at p. 12. 20 H.B. Ice-cream Limited, Plaintiff v Masterfoods Limited trading as Mars Ireland, High Court [1990] 2 IR 463. See Ken Murphy, ‘Irish Court Rules on Freezer Exclusivity in the Ice Cream Market’ [1992] 6 ECLR 270. See also Barry Rodger, ‘The Big Chill for National Courts: Reflections on Market Foreclosure and Freezer Exclusivity under Article 81’ (2004) 11 Irish Journal of European Law 77–116. 21 See the discussion on the legal formalism tradition in American jurisprudence in Duxbury (n. 7) particularly Chapter 1. 22 See the brief discussion in Rodger and MacCulloch (n. 19) Chapter 1, pp. 21–23. 23 See for instance Lina M. Khan, ‘Amazon’s Antitrust Paradox’ (2016) 126 Yale LJ and ‘The New Brandeis Movement: America’s Monopoly Debate’ (2018) Journal of European Competition Law and Practice 131; John Newman, ‘Reactionary Antitrust’ (2019) No. 4 Concurrences Revue; Sandra Marco Colino, ‘The Antitrust “F” Word: Fairness Considerations in Competition Law’ (2019) Journal of Business Law 329.

210

9.2.2

Challenges to assumptions in competition law

Article 102 TFEU: Rationale, Uncertain Scope and the Concept of ‘Normal Competition’

This backdrop clearly has implications for the proper interpretation of and scope of the relevant laws for dealing with market power abuses, in this context Article 102 TFEU and the equivalent UK provision, Chapter II of the Competition Act 1998. It should be noted that the domestic UK prohibition, contained in section 18 of the Competition Act (‘the Chapter II prohibition’) is identical in formulation to Article 102 (save for the effect on inter-state trade requirement).24 Furthermore section 60 of the 1998 Act25 required the domestic prohibition(s) to be interpreted consistently with the EU prohibitions in Articles 101 and 102 and in particular ensuring consistency with the principles laid down by the Court of Justice of the European Union (CJEU). Following UK withdrawal from the EU, section 60 has been replaced by section 60A which requires consistency only in relation to pre-Brexit CJEU rulings, and even provides flexibility to depart in circumstances where appropriate from those interpreative rulings.26 It is uncertain yet at this early stage to what extent, if any, the domestic UK courts are likely to adopt a divergent interpretation in any specific contexts of the abuse concept. Accordingly, in order to analyse the UK courts’ application of the abuse of dominance provision, it is important at the outset to try and identify the parameters of the Article 102 abuse prohibition. The starting point is to reaffirm Ibanez-Colomo’s statement that there remains ‘considerable uncertainty about the exact boundaries of Article 102’.27 He notes the core problem with the notion of abuse in that ‘multiple substantive standards co-exist, overlapping and contradicting one another. As a result it is difficult to anticipate how a particular case will be decided.’28 Two further interesting and recurrent themes are revisited by Ibanez-Colomo – the ‘pigeon-holing’ or formal approach to the treatment of certain types of behav-

24 Roberto Nazzini, ‘A Welfare-Based Competition Policy under Structuralist Constraints: Abuse of Dominance and OFT Practice’, in Barry J. Rodger (ed.), Ten Years of UK Competition Law Reform (Dundee University Press, 2010), pp. 97–137. 25 Ben Rayment, ‘The Consistency Principle: s60 of the Competition Ac 1998’, in Barry J. Rodger (ed.), Ten Years of UK Competition Law Reform (Dundee University Press, 2010). 26 Post-Brexit, section 60 is likely to be revised to remove the requirement for consistency with CJEU rulings. As introduced by the Competition Amendment (etc.) (EU Exit) Regulations 2019, SI 2019/93. 27 See Pablo Ibanez-Colomo, ‘The Law on Abuses of Dominance and the System of Judicial Remedies’ (2013) 32(1) Yearbook of European Law 389–431 at 389. 28 Ibid. at 395. Although it can also be argued that the concept of abuse is by necessity sufficiently flexible to respond to changes in markets and new forms of anti-competitive market behaviour.

Formalism, fairness and freedom of contract

211

iour as abusive and the elusive nature of the scope of the ‘abuse’ concept which is predicated on conduct which departs from normal competition. How far the methods of the dominant undertaking have to differ ‘from those governing normal competition’ is indeed a controversial question.29 The European Court has arguably taken a broad pro-competition approach, condemning methods of competition which in practice could be exclusionary but which many critics would argue were ‘normal’ in competitive situations. Consequently, there have been ongoing debates regarding case-law developments under Article 102 relating to abuse, for instance in relation to Intel, and the appropriate legal rules for dealing with exclusionary discounting practices.30 These again reflect broader uncertainty about the goals and objectives of competition law.31 In defending the General Court’s 2014 Intel judgment, Wils rejected an EU approach to competition based on Chicago School tenets which would neglect the process values of undistorted competition including the right to compete on the merits and equality of opportunity.32 The debate has continued and gained prominence more recently in relation to the Commission’s infringement findings against Google, notably in Google Search (Shopping) in 2017,33 currently on appeal before the General Court.34 The argument based on the central importance of the principle of freedom to compete as a distinct goal of EU competition law, primarily relevant in

See Rodger and MacCulloch (n. 19), pp. 110–111. Intel Corp v Commission (Case C-286/09) General Court [2014] 5 CMLR 9; and (Case C-413/14 P) ECJ [2017] 5 CMLR 18. See for example Pablo Ibanez-Colomo, ‘Intel and Article 102 TFEU Case Law: Making Sense of a Perpetual Controversy’, LSE Law, Society and Economy Working Papers 29/2014. See also, R Grasso, ‘Conditional Pricing in Europe and the CJEU’s Ruling in Intel: What’s New?’ (2018) 4(2) CLPD 25–41 and Nicolas Petit, ‘Intel and the Rule of Reason in Abuse of Dominance Cases’ (2018) 43(5) EL Rev 728–750. 31 For an alternative view on the importance of this debate, see http://​ chillingcompetition​.com/​2015/​09/​17/​whats​-with​-this​-obsession​-with​-the​-objectives​-of​ -eu​-competition​-law/​accessed 3 January 2021. 32 See Wils (n. 14). See also Amartya Sen ‘Markets and A achievements and Limitations of the Market Mechanism in Promoting Individual freedoms’ (1993) 45 Oxford Economic Papers 519, at 536. 33 https://​ec​.europa​.eu/​competition/​elojade/​isef/​case​_details​.cfm​?proc​_code​=​1​ _39740 accessed 3 January 2021. 34 See e.g. Carsten Koenig, ‘Form, Effects or Both? – The More Economic Approach and the European Commission’s Decision in Google Search’ (2019) 44(5) EL Rev 680–693, and E. Iacobucci and F. Ducci, ‘The Google Search Case in Europe: Tying and the Single Monopoly Profit Theorem in Two-sided Markets’ (2019) 47(1) EJLE 215–242. See also Pablo Ibanez Colomo, ‘Indispensability and Abuse of Dominance: From Commercial Solvents to Slovak Telekom and Google Shopping’ forthcoming. 29 30

212

Challenges to assumptions in competition law

relation to Article 102,35 is often linked with a reliance on the Ordoliberal School36 – whereby the competitive process itself has a normative value.37 This principle does not rely on economic effects or require proof of exclusion and appears to fit better with court based adjudication, albeit it provides more of a general principle than an ‘operational standard for adjudication’.38 As Lianos has stressed, in the absence of a clear understanding of the term ‘competition’39 the teleological method of EU law judicial law-making facilitated a more goal-oriented legal interpretation of the competition rules.40 This leaves the key issue of how we construct, interpret and apply rules to distinguish between normal competition on the merits and abusive behaviour. This will inevitably be dependent on the level of trust in the free play of market forces, and consequent views on the role and objectives of competition policy and how these should apply to market power behaviour. In this context, there has been sustained academic discussion and critique of the underlying rationale of the abuse prohibition in the EU in recent years.41 Akman reaffirms that Article 102 TFEU and the rather abstract term of ‘abuse’ is quite flexible but rejects the notion that the provision was driven by an ordoliberal approach, which would place more emphasis on protecting the economic freedom of

35 See Liza Lovdahl Gormsen, A Principled Approach to Abuse of Dominance (CUP, 2010); cf. Pinar Akman, ‘The Role of “Freedom” in EU Competition Law’ (2014) 34(2) Legal Studies 183–213. 36 See generally, Ioannis Lianos, ‘Some Reflections on the Goals of EU Competition Law’, in Ioannis Lianos and Damien Geradin (eds), Handbook on European Competition Law, Substantive Aspects (Edward Elgar, 2013) Chapter 1. See also Liza Lovdahl Gormsen, ‘The Concept between Economic Freedom and Consumer Welfare in the Modernisation of Article 82 EC’ (2007) 3(2) European Competition Journal 329–344. 37 Lianos (n. 36) p. 56. See Heinz Grosskettler, ‘On Designing an Economic Order: The Contribution of the Freiburg School’, in Donald A. Walker (ed.), Perspectives on the History of Economic Thought, Volume II, Twentieth Century Economic Thought, Selected Papers from the History of Economics Society Conference 1987 (Edward Elgar, 1987). 38 Lianos (n. 36) at 53. 39 See for instance Maurice Stucke (n. 6). 40 Lianos (n. 36) at 69. 41 See Pablo Ibanez-Colomo, The Shaping of EU Competition Law (CUP, 2018), Part II Ch. 4 ‘The Notion of Abuse within the Meaning of Article 102 TFEU’; Ekaterina Rousseva, Rethinking Exclusionary Abuses in EU Competition Law (Hart Publishing, 2010); Roberto Nazzini, Foundations of European Union Competition Law: The Objective and Principles of Article 102 (OUP, 2011); Pinar Akman, The Concept of Abuse in EU Competition Law: Law and Economic Approaches (Hart Publishing, 2012). See also Pinar Akman and Luke Garrod ‘When are Excessive Prices Unfair?’ (2011) 7(2) Journal of Competition Law and Economics 403, 426. See also Akman (n. 35).

Formalism, fairness and freedom of contract

213

market participants. This debate on the dividing line between acceptable and prohibited competitive conduct is certainly not restricted to the EU, and there has been, of course, considerable literature comparing and contrasting the approaches to market power in the EU and the US.42 As Schweizer notes: ‘Like US antitrust law, EU law struggles with formulating an adequate test for distinguishing between pro-competitive conduct, or “competition on the merits”, on the one hand, and anti-competitive, exclusionary conduct on the other….’43 The EU abuse case-law has been categorised broadly into two main strands: exclusionary and exploitative abusive behaviour.44 Jones and Lovdahl-Gormsen provide an excellent overview of the former in this passage: in seeking to identify unlawful exclusionary behaviour, the EU courts have not, in contrast to their US counterparts, been unduly concerned about their interpretation of Article 102 giving rise to false positives and chilling, rather than fostering, competition … the judgments seem to prefer a construction of Article 102 that will ensure the process of competition and rivalry between firms is preserved and protected. The anti-competitive effects of a dominant firm’s conduct thus tends to be assumed where it is capable of excluding equally efficient competitors or is liable to remove or restrict a buyer’s freedom to choose its sources of supply, to bar competitors from access to the market, to apply dissimilar conditions to equivalent transactions with other trading parties or to strengthen the dominant position by distorting competition. Critics complain that the EU rules in this area may, in consequence, be over-inclusive.45

Gal’s account of exploitative abuses chimes with the themes in this chapter,46 noting the excessive prices prohibition as a ‘microcosm’ of competition law

42 See, for example, Donald Baker, ‘An Enduring Antitrust Divide across the Atlantic over Whether to Incarcerate Conspirators and When to Restrain Abusive Monopolists’ [2009] 5(1) European Competition Journal 145–200; Eleanor Fox, ‘Monopolization and Abuse of Dominance: Why Europe is Different’ (2014) 59(1) The Antitrust Bulletin 129–152. See also Heike Schweitzer ‘Parallels and Differences in the Attitudes towards Single-Firm Conduct: What Are the Reasons? The History, Interpretation and Underlying Principles of Sec 2 Sherman Act and Art.82 EC’ EUI Working papers Law 2007/32. 43 Schweitzer (n. 42) at 38. 44 See for instance Rodger and MacCulloch (n. 19), Chapter 4; see Pablo Ibanez-Colomo (n. 41) Chapter 4. 45 See Alison Jones and Liza Lovdahl Gormsen, ‘Abuse of Dominance: Exclusionary Pricing Abuses’, Chapter 10 in Ioannis Lianos and Damien Geradin (eds), Handbook on European Competition Law, Substantive Aspects (Edward Elgar, 2013), p. 472. Emphasis original. 46 See Michal Gal, ‘Abuse of Dominance – Exploitative Abuses’, Chapter 9 in Ioannis Lianos and Damien Geradin (eds), Handbook on European Competition Law, Substantive Aspects (Edward Elgar, 2013).

214

Challenges to assumptions in competition law

in seeking to balance ‘social Darwinism with … social justice’.47 The central paradox is how to deal with market processes which lead to and result in ‘earned monopoly rewards’,48 although Gal also noted that the prohibition ‘can’t be explained by a simplistic property or freedom of trade argument’.49 Such a formalistic approach to freedom of contract and property rights would, she explained, effectively annul the competition laws.50 9.2.3

Fairness and Freedom of Contract

The broad notion of fairness in contractual relations and its potential application as an underlying principle in the application of Article 102, has been explored by Akman in The Concept of Abuse in EU Competition Law: Law and Economic Approaches.51 While the notion of fairness underlies the abuse prohibition to some extent, its role is unclear. Although Akman suggests a tension with ordoliberalism52 (based on freedom), they may be reconciled to the extent that, in contrast with Ayal’s approach, fairness is concerned with market participants other than the dominant undertaking. A central question posed by Akman for Article 102 in the context of both the jurisprudential debate here and the case-law analysis, is the extent to which fairness may impede the concept of freedom of contract. In this context, one may consider fairness at both the vertical and horizontal levels, although there is some overlap in the case-law.53 It should be noted that Akman concludes that a fairness based approach to the abuse prohibition would be too vague, a threat to freedom of contract and would produce unwanted uncertainty for businesses in B2B relations.54 Nonetheless, in parallel with developments in US antitrust scholarship, there is increasing support for a broader approach to EU compe-

Ibid. at p. 385. Ibid. at p. 397. See also Rodger and MacCulloch (n. 19), Chapter 1. 49 Gal (n. 46) at p. 398. 50 Ibid. 51 Pinar Akman (2012) (n. 41), Chapter 4 ‘Fairness’. 52 Ibid. at 151 et seq. 53 See e.g. Purple Parking discussed further infra. Indeed the focus in the abuse prohibition on inequality of bargaining power involving a dominant party, as Akman highlights, may reflect the idea of procedural fairness which underpins the role of fairness in contemporary contract law. See Akman (2012) (n. 41) at 159–167. See also for instance the potential scope for separate provision on abuse of a superior bargaining position: see for example Masako Wakui and Thomas K Cheng, ‘Regulating Abuse of Superior Bargaining Position under the Japanese Competition Law: An Anomaly or Necessity?’ (2015) 3 JAE 302–333. 54 Akman (2012) (n. 41) at 182. 47 48

Formalism, fairness and freedom of contract

215

tition law, and the abuse prohibition in particular, which would incorporate fairness considerations.55 These broad themes of the uncertain scope of the prohibition, the potential for intervention in market-ordered outcomes, and possible tensions with respect to formal contract and property rights that may result from fairness based approaches, provide a canvas for viewing the case-law developments in the UK, and particularly the hypothesis that more recent case-law evidences a broader scope for the application of the abuse doctrine.

9.3

UK COURTS’ ABUSE CASE-LAW

9.3.1 Introduction This is the key section of the chapter, in which we will provide an overview of the case-law in the UK courts, essentially the courts of England and Wales, and the Competition Appeal Tribunal (CAT).56 This will not be a comprehensive coverage of all case-law since accession by the UK to the EU in 1974.57 It should be stressed that the chapter will focus on case-law where there has been judicial consideration of the scope of the abuse doctrine involving the application of the abuse of dominance prohibition under either/both EU law and UK law: Articles 102 TFEU and the Competition Act 1998 Chapter II prohibition respectively, bearing in mind the equivalent nature of the abuse prohibition under both. Accordingly there will be little (or no) emphasis and analysis of case-law involving other key aspects of the prohibition, for instance in relation to determination of issues regarding what constitutes an ‘undertaking’,58 the test for the relevant market, whether or not there is dominance or an effect on 55 Colino (n. 23). See also Damien Gerard, ‘Fairness in EU Competition Policy: Significance and Implications’ (2018) 9 Journal of European Competition Law and Practice 219. 56 In Scotland there has been very limited case-law and/or judicial consideration of the scope of the abuse doctrine. See for instance, Argyll Group plc v The Distillers Company plc, 4 April, 1986, 1987 SLT 514; Millar & Bryce Limited v Keeper of the Registers of Scotland, 1997 SLT 1000, 2 May 1997; Arriva Scotland West Ltd v Glasgow Airport Ltd [2011] CSOH 69. 57 See Barry Rodger ‘Competition Law Litigation in the UK Courts: A Study of All Cases to 2004 – Part I’ [2006] ECLR 241; ‘Competition Law Litigation in the UK Courts: A Study of All Cases to 2004 – Part II’ [2006] ECLR 279; ‘Competition Law Litigation in the UK Courts: A Study of All Cases to 2004 – Part III’ [2006] ECLR 341; ‘Competition Law Litigation in the UK Courts: A Study of All Cases 2005–2008– Parts I and II’ [2009] 2 Global Competition Litigation Review 9–114 and 136–147; and ‘Competition Law Litigation in the UK Courts: A Study of All Cases 2009–2012’ [2013] 6(2) GCLR 55–67. 58 See for instance UKRS Training Ltd v NSAR Ltd [2017] CAT 14.

216

Challenges to assumptions in competition law

inter-state trade (for Article 102 TFEU) or an effect on trade within the UK (for the Chapter II prohibition). These are important issues, pre-requisites for the application of the prohibition, but they fall outside the focus of this chapter which concentrates on what type of behaviour constitutes an ‘abuse’ and in particular how this may limit or impact upon the commercial competitive strategy of dominant undertakings. We will also focus on the period since the introduction of the UK competition law Chapter 2 prohibition, which mirrors the Article 102 prohibition and came into force as of 1 March 2000.59 This is not because we will be focusing only on the domestic abuse provision, which is in identical terms to Article 102 and, at least until UK withdrawal from the EU, required to be interpreted consistently with the CJEU interpretation of that provision,60 but because there have been more substantive final judgments in this period and considerably more judicial discussion of the scope of the abuse doctrine. As highlighted, the UK courts in applying Article 102 and the domestic prohibition, must follow and apply the CJEU jurisprudence,61 but as indicated above, the abuse doctrine and the case-law afford considerable scope for interpretation according to the circumstances of the case, and one’s view of the competitive process and the role for competition law intervention in that process. The focus of the analysis will be on cases where the courts have delivered substantive final judgments on the application of the abuse doctrine, although we will also cover case-law where there have been interesting discussions concerning the scope of the abuse doctrine and its potential application in some (but certainly not all) of the cases involving summary judgments or interim process.62 In this chapter, in addition to case-law before the ordinary 59 See Rodger and MacCulloch (eds), The UK Competition Act: A New Era for UK Competition Law (Hart, 2000); Barry James Rodger (ed.), Ten Years of UK Competition Law Reform (Dundee University Press, 2010). 60 Under section 60 of the Competition Act 1998. See Ben Rayment, ‘The Consistency Principle: Section 60 of the Competition Act 1998’ and Roberto Nazzini, ‘A Welfare Based Competition Policy under Structuralist Constraints: Abuse of Dominance and OFT Practice’ Chapters 4 and 5 in Barry James Rodger (ed.), Ten Years of UK Competition Law Reform (Dundee University Press, 2010). Note the revised section 60A introduced following UK withdrawal from the EU, above at n 26. 61 Section 60 Competition Act 1998. 62 The most straightforward and important case law involves a substantive final judgment on the competition law issue, normally following trial, or proof as it is known in Scots law. Summary judgment cases are where the defendant or the claimant has sought a summary judgment in order to dismiss the action or to strike out the defence. The broad banner of ‘interim process’ covers a range of situations in which judgments are given in a competition law dispute during the procedural phases of the litigation. For a fuller discussion and outline of the distinction between substantive final judgments, summary judgments and interim process judgments, and for discussion of the earlier case-law, see Rodger, Barry, ‘Competition Law Litigation in the UK Courts –

Formalism, fairness and freedom of contract

217

courts (notably the High Court in England and Wales) we shall only be considering the limited abuse prohibition jurisprudence of the CAT in litigation under section 47A of the Competition Act in the context of inter-partes’ commercial disputes as opposed to the CAT’s role and rulings in relation to Article 102/Chapter 2 in the public enforcement context.63 Furthermore it should be stressed that we will not be considering the enforcement practice of the Competition and Markets Authority (CMA) and its predecessor the Office of Fair Trading (OFT) in relation to the abuse prohibitions in this chapter.64 9.3.2

Abuse Case-law Post Competition Act 1998

There has been considerable discussion of all competition case-law before the UK courts over various periods by the current author to 2016,65 but the focus here is on abuse case-law post 1 March 2000, when the 1998 Act Chapter 2 prohibition came into force. Generally it should be stressed that there have been relatively few case-law judgments involving Article 102 since 2000. The clear preponderance of Article 102 cases was in the 1980s and 1990s and these were primarily in cases where Article 102 was used as a euro-defence, for the most part unsuccessfully and primarily in intellectual property-related cases as a defence to some form of IPR infringement action, for instance in Duracell A Study of All Cases to 2004’ [2006] ECLR Parts I, II and III at 241, 279 and 341; and ‘Competition Law Litigation in the UK Courts: A Study of All Cases 2009–2012’ (2013) 6 (2) Global Competition Litigation Review 55–67. 63 For a discussion of the role of the CAT, see David Bailey, ‘The Early Case Law of the Competition Appeal Tribunal’, Chapter 2 in Barry James Rodger (ed.), Ten Years of UK Competition Law Reform (Dundee University Press, 2010). Note that the CAT has a role in private enforcement in follow-on actions, and following the introduction of the Consumer Rights Act 2015 also in stand-alone actions, but in undertaking its role in follow-on actions the CAT is bound by prior infringement decisions (of the Commission or OFT/CMA) and only considers issues of causation and quantum: see Rodger ‘UK Competition Law and Private Litigation’ Chapter 3 in Barry James Rodger (ed.), Ten Years of UK Competition Law Reform (Dundee University Press, 2010). 64 See Barry Rodger, ‘Application of the Domestic and EU Antitrust Prohibitions: An Analysis of the UK Competition Authority’s Enforcement Practice’ (2019) JAE 1–38. 65 See Barry Rodger ‘Competition Law Litigation in the UK Courts: A Study of All Cases to 2004 – Part I’ [2006] ECLR 241; ‘Competition Law Litigation in the UK Courts: A Study of All Cases to 2004 – Part II’ [2006] ECLR 279; ‘Competition Law Litigation in the UK Courts: A Study of All Cases to 2004 – Part III’ [2006] ECLR 341; ‘Competition Law Litigation in the UK Courts: A Study of All Cases 2005–2008– Parts I and II’ [2009] 2 Global Competition Litigation Review 9–114 and 136–147; ‘Competition Law Litigation in the UK Courts: A Study of All Cases 2009–2012’ [2013] 6(2) GCLR 55–67; and Barry Rodger ‘Competition Law Private Enforcement in the UK Courts: Case-Law Developments 2013–2016’ [2017] 10 GCLR 128–144.

218

Challenges to assumptions in competition law

International Inc. v Ever Ready Limited.66 There was little or no discussion in any of the pre-2000 case-law of the precise nature of the abuse concept. Since the 1998 Act, there has been a clear prevalence of Chapter 2 cases.67 There have been a number of private court disputes in this period post 1 March 2000 in which the domestic or EU prohibition has been raised by one of the parties.68 The first judgment on the substance of the Chapter II prohibition was delivered eight months after its introduction, on 2 November 2000, by Lawrence Collins J in the High Court, Chancery Division, in Claritas (UK) Ltd v The Post Office and Postal Preference Service Ltd.69 This judgment was in the context of an action for an interim injunction to restrain the despatch of consumer preference questionnaires using the Royal Mail brand and logo throughout the UK by the Royal Mail. The Post Office owned 30 per cent of Postal Preference Service Ltd (PPS), and Claritas, a company involved in procuring, supplying and analysing consumer information for a fee for businesses, claimed that this amounted to an abuse of the Royal Mail’s dominant position. The key question was whether, at this interim stage, there was a serious issue to be tried. Different strands of European jurisprudence were analysed. First, the possibility of an abuse based on refusal to license IP rights was rejected as Claritas would not be excluded from the market because it could not use the Royal Mail logo.70 Secondly, the application of the essential facilities doctrine was deemed inappropriate as Claritas retained full access to the postal facilities. Finally, in any event, given that the alleged abuse was not in a market in which the Post Office was dominant, Claritas had failed to demonstrate that this situation fell within the scope of the Tetra Pak II requirement for close links between the dominant market and abuse market.71 Although the appli Ch D, 10 March 1989 [1989] RPC 731. Claritas(UK) Limited v The Post Office and Postal Preference Service limited [2001] ECC 12; Network Multimedia Television Ltd and another v Jobserve Ltd [2001] EWCA Civ 2021 [Network Multimedia Television ltd (T/A Silicon.com) v Jobserve Ltd, April 5 2001 Chancery Division (Transcript) 5 April 2001]; Getmapping PLC v Ordnance Survey [2002] EWHC 1089 (Ch) [2002] UKCLR 411; Wireless Group PLC v Radio Joint Audience Research Ltd [2004] EWHC 2925 (Ch) [2005] ECC 99; Seafood Holdings Ltd v My Fish Company Ltd [2017] EWHC 766 (Ch). 68 See Aidan Robertson, ‘Litigating under the Competition Act 1998: The Early Case Law’ [2002] Comp Law 335; Barry Rodger, ‘Competition Law Litigation in the UK Courts – A Study of All Cases to 2004’ [2006] ECLR Parts I, II and III at 241, 279 and 341 and ‘Competition law litigation in the UK courts: a study of all cases 2009–2012’ (2013) 6(2) Global Competition Litigation Review 55–67. 69 [2001] UKCLR 2. This was followed by the DGFT Decision in relation to the same dispute, Consignia plc and Post Preference Service Ltd, Decision CS 98/4/2001, 15 June 2001. 70 See Case T-68/89 Radio Telefis Eireann v Commission [1991] ECR II-485. 71 Case C-333/94P [1996] ECR I-5951. 66 67

Formalism, fairness and freedom of contract

219

cation was dismissed, this early resort to the prohibition in private litigation demonstrated an awareness of the possibilities under the Act for aggrieved competitors. Network Multimedia Television Ltd and another v Jobserve Ltd,72 concerned an appeal as to whether the judge at first instance was wrong to hold that there was a ‘serious issue’ to be tried in an action for injunction for alleged breaches of section 18 of the Competition Act 1998. The appeal was rejected, damages were not an adequate remedy for the claimant, and the balance of convenience favoured the grant of interim relief. The court was not convinced that Oscar Bronner disposed of the issue and the judge was right to hold that there was a seriously arguable case.73 Arguably, the two most significant (interim process and/or summary judgment) cases in the early period following the introduction of the Chapter 2 prohibition both involved Laddie J,74 where he focused on the relationship between commercial decisions taken by a dominant undertaking and the scope for intervention under an abuse based prohibition. In Suretrack Rail Services Limited v Infraco JNP Ltd,75 he stressed that the abuse prohibitions are ‘not designed to allow the courts or the relevant regulatory authorities to interfere on a day by day basis with the running of business. Even big businesses are entitled to win in the market place. Even big businesses are entitled to take rational decisions which are beneficial to them and harmful to competitors or others in the market.’76 He emphasised the point made earlier by him in Getmapping,77 that: an abuse which is incapable of being objectively justified is one which no rational and fair person could justify. A decision which respectable and fair-minded businessmen could disagree with is not thereby made incapable of being objectively justified. There may be many views as to whether a particular course is objectively justified or not. The need to demonstrate that the dominant trader’s actions are

72 21 December 2001 [2001] EWCA Civ 2021 [Network Multimedia Television ltd (T/A Silicon.com) v Jobserve Ltd, April 5 2001 Ch D]. 73 See Barry Rodger, ‘Early Steps to a Mature Competition Law System; Case Law Developments in the First Eighteen Months of the Competition Act 1998’ [2002] 2 ECLR 52–67. See also for example, Intel Corp v Via Technologies Inc. 26–28 November, 20 December 2002 [2003] FSR 33 [Intel Corp v Via Technologies Inc., 11–12 April, 14 June 2002 [2003] FSR 12]. See also Intel Corp v Via Technologies Inc. 26–28 November, 20 December 2002 [2003] FSR 33 [Intel Corp v Via Technologies Inc., 11–12 April, 14 June 2002 [2003] FSR 12]. 74 See Her Majesty’s Stationary Office and another v The Automobile Association Limited and another [2001] ECC 34, which highlighted the difficulties in pleading an abuse case. 75 [2002] EWHC 1316 (Ch). 76 Ibid. para 20. 77 Getmapping plc v Ordnance Survey [2002] EWHC 1089 (Pat).

220

Challenges to assumptions in competition law

incapable of being objectively justified means that a person alleging abuse on that basis has a high hurdle to cross.78

The clear implication is that the bar for intervention on the basis of the abuse prohibition should be set high and that considerable deference should be afforded to rational commercial business decisions by dominant undertakings. It is the contention of this chapter that we may be able to observe a shift in some of the more recent English court judicial pronouncements on the abuse prohibition (and the relationship between commercial legitimacy and the objective justification defence under Article 102 / the Chapter II prohibition) towards an approach whereby courts will engage critically regarding the merits of the purported commercial and other justifications for certain competitive strategies. 9.3.3

Substantive Final Rulings in Abuse Cases

In Hendry and others v The World Professional Billiards and Snooker Association Limited (WPBSA),79 it was held that the rule that players could not enter tournaments organised elsewhere without prior written consent was held to breach sections 2 and 18 of 1998 Act, Articles [101] and [102] and also constituted an unreasonable restraint of trade. The competition law issue failed in various cases after trial: in Arkin v Borchard Lines ltd and others, following a 49 day trial, the claimant was unsuccessful in claims of alleged infringements under Articles [101] and [102] regarding the market for liner services to and from Israel. It was held that as prices never fell below average variable costs and the failure to pursue a policy of undercutting competitors was strong evidence against eliminatory intent and therefore there was no breach of either Articles [101] or [102]; Leeds City Council v Watkins & Whiteley concerned alleged breaches of Chapters 1 and 2 Competition Act 1998 and Articles [101] and [102], which were dismissed due to insufficient and unconvincing evidence regarding the relevant market and dominance; in Ineos Vinyls Ltd v Huntsman Petrochemicals (UK) Ltd80 the claimant manufacturers had failed to establish that the defendant supplier of ethylene had a dominant position and abused it by charging prices and delivery charges that were unfair within the meaning of the EC Treaty Article [102] and the Competition Act 1998 section 18, as they had failed to analyse the costs incurred in producing the product, compare those costs with the price charged, or evaluate the resulting profit.

Suretrack (n. 75) at para 26. 5 October 2001 [2002] ECC 8. 80 [2006] EWHC 1241 (Ch). 78 79

Formalism, fairness and freedom of contract

221

In AttheRaces Ltd v British Horseracing Board, the claimant, which supplied websites, TV channels, and other media relating to British horse racing, alleged that the defendant, which had a central role in the organisation and promoting of British horse racing and which kept a computerised database of pre-race data, including the date and place of the race meeting, name of the race, list of horses entered, etc., had abused a dominant position and thereby infringed both Article [102] and the Chapter 2 prohibition. BHB effectively held a monopoly in the provision of the pre-race data and it was held in the High Court that it had abused its dominant position by excessive, unfair and discriminatory pricing.81 It was held, following the Court’s jurisprudence on refusal to supply, that the pre-race date constituted an ‘essential facility’, essential to ATR’s business, and that a constructive refusal to supply was caught by the prohibitions. Referring to court case-law on excessive and discriminatory pricing, the price was excessive in comparison to the cost to BHB plus a reasonable return, and discriminatory in being markedly higher than the sum normally charged to other broadcasters. On appeal, this ruling was overturned by the Court of Appeal, which was sceptical about Article [102]/Chapter II becoming a general provision for the regulation of prices.82 The Court stated that exceeding cost was a necessary but not a sufficient test for abuse of dominance, there was little evidence that competition on the market was being distorted by BHB’s demands and that the value to ATR of the pre-race data was relevant in determining whether the price was excessive. Furthermore, differential pricing was not necessarily abusive, and prices essentially were dependent on market forces. Accordingly, BHB’s pricing strategy was not abusive and the Court of Appeal clearly advocated a restrained approach to court involvement in claims of excessive or discriminatory pricing. It is interesting to note that all of the five more recent court rulings on the abuse prohibition have concerned a dominant undertaking limiting access to a related or downstream market, and an assessment of arguments raised to objectively justify those commercial decisions. Two of those cases have both involved abusive behaviour by airport operators.83 In Purple Parking Ltd v Heathrow Airport Ltd,84 Mann J in the High Court held that the defendant airport operator had abused its dominant position by changing existing arrangements to effectively exclude ‘meet and greet’ car park operators from the airport terminal forecourts while maintaining its own ‘meet and greet’ operations there. Mann J undertook a comprehensive analysis of the existing

83 84 81 82

[2005] EWHC 3015 (Ch) Ch D. [2007] EWCA Civ 38, CA. See also Arriva Scotland West ltd v Glasgow Airport Ltd [2011] CSOH 69. [2011] EWHC 987 (Ch).

222

Challenges to assumptions in competition law

EU (and UK) jurisprudence on the essential facilities doctrine and margin squeeze abuse. However, it was stressed that the defendant’s ‘pigeon-hole’ approach whereby a claim should fit within the category of essential facilities or fail was not accepted by Mann J: a court is entitled to look at conduct, and ask the overall question of whether there is an abuse by reference to various ways of committing that abuse, and is not forced to find one single appropriate label to that abuse … and apply some test only applicable to that form.85

The claimant’s claim was based on an allegation of the application of dissimilar conditions to equivalent transactions, resulting in anti-competitive behaviour, and this was upheld on the basis of the effect on consumers. It was stressed that commercial justification did not constitute an objective justification for abusive behaviour, and the objective justification arguments based on security and congestion were rejected on the facts; it was clear that the airport operator was motivated by an anti-competitive intent. In Arriva The Shires Ltd v London Luton Airport Operations Ltd,86 ATS had a concession agreement with the operator of Luton Airport (‘Luton Operations’) to carry passengers (over 1 million a year) on the 757 bus service from the airport direct to London Victoria. Following a tender process, the concession was awarded to a rival bus operator. ATS claimed that Luton Operations was dominant in the market for granting rights to operate bus services and that they had abused it in two ways: (1) the tender procedure in awarding the new concession was unfair, and (2) through the abusive terms contained in the award of the new concession to the rival bus operator. The first part of the claim was unsuccessful as the tender process was fair, but the seven year exclusive period awarded to the successful tender bid operator was deemed to be abusive. The analysis and application of the abuse and objective justification in this judgment merit further analysis. It was noted that exclusionary abuses fall into two categories, where the dominant undertaking: (a) competes on the downstream market and is acting to foreclose that market to its own advantage;87 or (b) distorts competition on the upstream market between itself and its competitors by entering contracts with customers to buy exclusively from the dominant undertaking.88 This case clearly did not

Ibid. at para 102. [2014] EWHC 64 (Ch). 87 See Commercial Solvents v Commission [1974] ECR 223. This is also the type of abuse involved in Purple Parking. 88 See for instance Case C-549/10 P Tomra Systems ASA v Commission [2012] 4 CMLR 27. 85 86

Formalism, fairness and freedom of contract

223

fall within the second category but Rose J rejected the defendant’s assertion that to fall within the first category the dominant undertaking had to derive a competitive advantage or commercial benefit through the exclusionary conduct; for instance, by using its dominance in one market to improve its own position in a downstream market.89 Clearly, Luton Operations were not active on the downstream bus services market, but in any event they gained important commercial and financial advantages from the concession, which gave them a percentage of revenue earned by the bus operator. Rose J also held that the grant of an exclusive right to the bus route for a lengthy period of seven years affected competition on that downstream market and was anti-competitive. This reasoning is analogous with European Commission decisions in relation to the grant of media rights to broadcast football matches over an extended period.90 As in Purple Parking, it was stressed that the objective justification defence was not available simply where a business decision was commercially rational; prohibited abusive conduct normally invariably furthers the business interests of the dominant undertaking. Both these cases demonstrate a robust approach to abusive conduct and a restrictive approach to defences based on the objective justification for dominant undertakings’ business decisions.91 The only case in the recent quintology of cases where there was not a finding of an abuse was Streetmap.EU v Google Inc.92 It involved another alleged abuse of a dominant position involving the ‘interaction of competition between online search engines and competition between suppliers of online maps’.93 Basically the claimant contended that the preferential and prominent display of their own online map product restricted competition from competing suppliers of online maps. It was accepted for these purposes that Google held a dominant position in online search engines. The discussion in the case focused on the actual or potential anti-competitive effect of the alleged distortion of competition in online maps. Evidence of actual effect was required and it was not abusive where the effect was on a separate market from the market where the undertaking was dominant and the effect was not appreciable. There was 89 See for example Case T-128/98 Aeroports de Paris v Commission [2000] ECR II-3939, and SEL-Imperial Ltd v The British Standards Institution [2010] EWHC 854 (Ch). 90 See Commission Decision, Case COMP/38173, Joint Selling of the Media Rights to the FA Premier League, 22 March 2006. 91 Nonetheless, in practice claimants have often struggled to satisfy all the elements required for a successful abuse of dominance claim. See for instance: PIK Facilities Ltd v Watson’s Ayr Park Ltd (2005) CSOH 32; Chester City Council v Arriva plc [2007] EWHC 1373 (Ch); and BHB Enterprises Ltd v Victor Chandler International Ltd [2006] EWHC 1074 (Ch). 92 [2016] EWHC 253 (Ch). 93 Ibid. at [4] per Roth J.

224

Challenges to assumptions in competition law

considerable evidence to the effect that the Google Maps Onebox innovation did not have an appreciable effect in taking custom away from the complainant and was not reasonably likely to give effect to anti-competitive foreclosure. Roth J proceeded to make observations in relation to the defendant’s alternative argument based on objective justification. Roth J considered their commercial strategy to be objectively justified in order to secure efficiency gains and there were no other effective proportionate or viable alternatives that they could use to achieve the same legitimate aim.94 There were specific technical arguments to support Google’s revised commercial strategy and there was considerable evidential support, but the validity of a commercial strategy is clearly justiciable and courts will not simply accept a plain argument that there was a commercial decision to objectively justify a strategy. The two most recent final judgments on the determination of what constitutes abusive behaviour share a number of common elements: both rulings by the CAT in private enforcement proceedings under the 1998 Act Chapter II prohibition, and also involving the Chapter I prohibition; both concerning access to markets related to regulatory systems run by quasi-public body undertakings; and both involving failed attempts to use the objective justification defence where an abuse is prima facie established. In Socrates Training Ltd v the Law Society of England and Wales,95 the claimant, a training course provider, brought proceedings against the Law Society in respect of the society’s Conveyancing Quality Scheme. The scheme provided accreditation for solicitors’ firms engaged in residential conveyancing, and incorporated mandatory training in mortgage fraud and anti-money laundering. The claimant argued that the requirement for scheme members to obtain training exclusively from the Law Society was an abuse of a dominant position under the Chapter II prohibition (and also an anti-competitive agreement under the Chapter I prohibition) by tying or bundling applying the principles set out in Streetmap.EU.96 The Law Society’s argument that training was an integral part of the scheme was rejected; the supply of training was a separate market and was different from the supply of accreditation.97 Crucially, the CAT rejected the Law Society argument that no reasonable alternative to the mandatory training could be provided.98

Ibid. at paras 142–177. [2017] CAT 10. 96 Streetmap.EU Ltd v Google Inc. [2016] EWHC 253 (Ch) [2016] Info. TLR 146 [2016] 2 WLUK 347. 97 Microsoft Corp v Commission of the European Communities (T-201/04) [2007] ECR II-3601. 98 See Ordem dos Tecnicos Oficiais de Contas v Autoridade da Concorrencia (C-1/12) [2013] 4 CMLR 20, paras 166–176. 94 95

Formalism, fairness and freedom of contract

225

In Achilles Information Ltd v Network Rail Infrastructure,99 Network Rail’s operation of supplier assurance schemes, whereby potential suppliers seeking to work on its managed railway infrastructure could be approved only by the Railway Industry Supplier Qualification Scheme (RISQS) run by the Rail Safety and Standards Board, constituted (in addition to a breach of the Chapter I prohibition) an abuse of a dominant position contrary to section 18. The defendant was assumed to be dominant in the market for the operation and provision of access to national rail infrastructure and the anti-competitive effects complained of were allegedly in the market for the provision of supplier assurance services in the UK rail industry. It was stressed that in order to constitute an abuse, the conduct complained of must lead to anti-competitive foreclosure on the related market100 and it was not necessary for there to be some commercial benefit to be gained by the dominant undertaking from its conduct before that conduct could be condemned as abusive, as earlier confirmed in Arriva The Shires Ltd.101 the dominant undertaking had to show that the exclusionary effect was counter-balanced or outweighed by advantages that benefited consumers and that the conduct was proportionate. That conduct was not objectively justified (similarly in relation to the breach of Chapter I) by the need to ensure safety on the railway network because it was not impossible to achieve those aims without the RISQS-only rule.102

9.4

SPECIFIC ISSUES IN THE ABUSE CASE-LAW

In this section of the chapter, we will address various specific issues relating to the nature and scope of the abuse prohibition dealt with by the UK case-law; namely: 1. 2. 3. 4.

The competition test inherent in the prohibition; The way in which abuses can be pigeon-holed; Analysis of the objective/subjective nature of the prohibition; Commercial decisions and the scope for the objective justification defence; and 5. The interface between abuse infringements and contract/property rights.

[2019] CAT 20. Streetmap.EU Ltd v Google Inc. [2016] EWHC 253 (Ch). 101 Supra. This finding was upheld on appeal by the Court of Appeal [2020] EWCA Civ 323. 102 Paras 311, 313. This issue was not considered by the Court of Appeal [2020] EWCA Civ 323. 99

100

226

Challenges to assumptions in competition law

9.4.1

The Competition Test

There is very little in-depth analysis of the type of restriction of competition required in the case-law, although aspects of this have been considered by the High Court in Purple Parking and Arriva and by the CAT in Socrates and Achilles respectively.103 The question of whether competition requires to be eliminated was considered in Purple Parking. It was held that the prohibition would be ‘potentially contravened if the effect on the competition was less than an elimination of effective competition, but still had some real and serious effect. Distortion is not synonymous with elimination.’104 True ‘essential facilities’ cases were contrasted on the basis that: If something is essential to competition on the market, then if only the dominant person has access to it then, a fortiori, there cannot be competition. Competition will inevitably be eliminated. If there can be competition then the facility is not essential.105

The AG’s Opinion in Oscar Bronner was ‘not setting out the requirements of liability in an essential facilities situation’ but the references to elimination reflected the facts of particular cases and even in ‘essential facilities’ type abuses, Mann J, doubted ‘if elimination of competition, as opposed to a significant enough distortion, is required’. This indicates a clear scepticism towards an abuse test, even in essential facilities cases, which would require the ‘elimination of competition’ and is in line with the General Court approach in Microsoft.106 The Arriva case, discussed above, re-emphasised that the dominant undertaking need not benefit from any distortion of competition. Accordingly, the exclusive grant of rights in the new concession in that case was considered to be similar in effect to the grant of rights to broadcast football matches, assessed in various earlier cases by the Commission.107 ‘The grant of exclusivity for a long period to a single downstream provider of rights has a distortive effect on competition where competitors cannot enter the downstream market to compete with the undertaking to whom the rights have been granted.’108 The And also the CAT in Burgess and Sons v OFT [2005] CAT 25, at para 312. Para 94. 105 Para 82. 106 See the debate at http://​chillingcompetition​.com/​2015/​03/​30/​some​-thoughts​-on​ -alfonsos​-theory​-of​-article​-102​-tfeu/​ accessed 3 January 2021. 107 Particularly Joint selling of the media rights to the FA Premier league (Case 38173) decision of 22 March 2006 (‘FAPL’). 108 Supra para 106. 103 104

Formalism, fairness and freedom of contract

227

grant of exclusivity dampened competition on the downstream coach services market by preventing other operators from providing direct services to London, and this would impact on customer choice and prices. Again, ‘dampening’ of competition as opposed to a higher threshold of elimination was sufficient to trigger the abuse prohibition. Roth J in Streetmap.EU adopted a more restrictive approach to the issue of the requisite effect on competition, again in a situation where the effects were on a related market (the market for online maps) where the undertaking was not dominant. In that context a de minimis threshold would apply and to constitute an abuse the effect must be appreciable,109 although the Court had earlier held that ‘The impugned conduct must be reasonably likely to harm the competitive structure of the market.’110 Demonstration of actual foreclosure was unnecessary, although an analysis of market developments led Roth J to the conclusion that the alleged abusive behaviour (the introduction of the Maps Onebox) did not have the required appreciable effect on competition and was not reasonably likely to cause anti-competitive foreclosure.111 Nonetheless, while reiterating the test as set out by Roth J in relation to abuses in related markets, the CAT in its subsequent judgment in Achilles established an abuse on the basis that the Network Rail rule (RISQS-only) had ‘the effect, in the interim of weakening competition in the market’.112 Moreover, the CAT explored the classic ‘normal conditions of competition’ mantra for distinguishing acceptable from prohibited types of business behaviour, stating that it would be ‘reluctant to hold that it is consistent with recourse to normal methods of competition under Chapter II for a dominant infrastructure operator to foreclose in-market competition in ancillary service markets in the absence at least of fair competition for the market’.113 As Stucke has stressed, identifying what constitutes acceptable competitive behaviour is a normative judgment.114 Furthermore, in the other recent ‘related markets’ abuse case, Socrates Training Ltd, the CAT noted that it was not ‘inappropriate’ to consider the actual effect of alleged anti-competitive conduct ‘but it is necessary to bear in mind that the degree to which conduct has actually had (or not had) an anti-competitive effect is only evidential. That is because demonstration of a potential effect is sufficient.’115

109 At paras 97–98, relying inter alia on Whish and Bailey, Competition Law (8th edn, OUP, 2015) p. 212. 110 Ibid at para 88. Italics added for emphasis. 111 See in particular at paras 135–139. 112 Supra at para 297. Italics added for emphasis. Upheld on appeal by Court of Appeal [2020] EWCA Civ 323. 113 At para 301. 114 Stucke (n. 6). 115 At para 150.

228

Challenges to assumptions in competition law

The CAT rejected the need for proof of actual foreclosure of a competitor for an abuse to be established. 9.4.2

Pigeon-Holing of Behaviour/Abuses

There is a clear tension in the case-law and academic commentary about the appropriate balance to be maintained between dealing with specific abusive behaviour in its particular context and on the basis of the circumstances of the case, and a legal approach which provides greater certainty and predictability in the application of the law, particularly given that commercial behaviour which constitutes a breach of Article 102/the Chapter II prohibition may be illegal, fined and lead to contracts being unenforceable.116 The modernised focus on effects-based competition law enforcement creates further tension between specific enforcement and a categorisation based approach. The latter can be criticised for being too formalistic and rigid in prohibiting market behaviour without a proper analysis of its competitive effects,117 yet also be used in a pejorative sense where there are perceived limits on the scope of application of the abuse doctrine based on earlier formulaic legal tests set out by the courts, notably the CJEU. It is in the latter sense that the term pigeon-holing has been utilised in the High Court. The scope of the ‘essential facilities’ doctrine and its potential application has been a recurring theme in the case-law in the English courts. For instance in The Wireless Group Plc v Radio Joint Audience Research Ltd,118 albeit a summary judgment case, Lloyd J stressed at para 47: Equally, the categories of conduct which may constitute abuse are not exhausted by the examples considered in cases already decided. I do not propose to elaborate on the existing case-law in relation to essential facilities, which would best be done in a case where facts have been established, or at least are clearer than they can be on a summary application of this kind. I am not prepared to dismiss this claim on the basis that the conduct alleged cannot be an abuse because it does not fall within the essential facilities doctrine.

116 The ongoing relevance of categorisation (Or ‘pigeon-holing’) of abuses is evident in recent work by Pablo Ibanez Colomo, ‘Indispensability and Abuse of Dominance: From Commercial Solvents to Slovak Telekom and Google Shopping’ forthcoming. 117 See Wils (n. 14); Cf. Pablo Ibanez-Colomo, ‘Intel and Article 102 TFEU Case Law: Making Sense of a Perpetual Controversy’, LSE Law, Society and Economy Working Papers 29/2014. 118 16 December 2004 [2004] EWHC 2925 (Ch) [2005] ECC 99.

Formalism, fairness and freedom of contract

229

The rejection of the all-encompassing scope of the essential facilities doctrine119 is also evident in Purple Parking Ltd v Heathrow Airport Ltd.120 Mann J in the High Court held that the defendant airport operator had abused its dominant position by changing existing arrangements to effectively exclude ‘meet and greet’ car park operators from the airport terminal forecourts while maintaining its own ‘meet and greet’ operations there. Mann J undertook a comprehensive analysis of the existing EU (and UK) jurisprudence on the essential facilities doctrine and margin squeeze abuse. However, it was stressed that the defendant’s ‘pigeon-hole’ approach, whereby a claim should fit within the category of essential facilities or fail, was not accepted by Mann J. He recited the CJEU in Deutsche Telekom v European Commission case C-280/08P at para 173: 173. Furthermore, the list of abusive practices contained in Article 82 EC [an equivalent to section 18 for practical purposes] is not exhaustive, so that the practices there are merely examples of abuses of a dominant position. The list of abusive practices contained in that provision does not exhaust the methods of abusing a dominant position …

Lloyd J was critical of counsel for the defendants’ ‘pigeon-holing’ approach:121 79. Mr Brealey’s submissions require one to treat each of those examples, and the ‘essential facilities’ type of abuse, as being individual pigeon-holes into which one must fit a case, and having thus fitted it to fulfil a list of criteria said to be applicable to that pigeon-hole. That is an erroneous approach. The statutory examples, and those developed by subsequent case-law, are ways in which the basic wrong can be committed, but at all times an eye must be kept on the basic wrong itself.

The case did not have to fit into the category of essential facilities or fail, but ‘a court is entitled to look at conduct, and ask the overall question of whether there is an abuse by reference to various ways of committing that abuse, and is not forced to find one single appropriate label to that abuse … and apply some test only applicable to that form’.122 Nonetheless, it is inevitable in the context of private enforcement of competition law based rights that courts in a precedent based system particularly where there is a line of jurisprudence from the European Court, will adopt a categorisation approach that starts with 119 There was a backlash against the ‘rising tide’ of the essential facilities doctrine prior to Oscar Bronner, but ironically, this case is a judicial rejection of the straight-jacketing effect of the perceived limited scope of the essential facilities abuse requirements. 120 [2011] EWHC 987 (Ch). 121 See ibid. para 99. 122 Ibid. at para 102.

230

Challenges to assumptions in competition law

identification of a particular type of abuse as identified in earlier case-law. Accordingly Achilles was categorised as a ‘related markets’ case and Socrates as a ‘tying or bundling’ case. Moreover, as indicated above, the CAT judgment in Socrates demonstrated that even though the effects-based approach has been advocated (even by the Commission) to determine whether anti-competitive behaviour is anti-competitive and prohibited, an abuse can still be established without evidence of an actual foreclosure effect. 9.4.3

Objective/Subjective Test

It is generally considered in the case-law and literature that abuse is an objective concept, requiring neither an intention to harm nor any morally reprehensible dimension. The CJEU gave what has become a benchmark test of abuse in Hoffmann-La Roche: The concept of abuse is an objective concept relating to the behaviour of an undertaking in a dominant position which is such as to influence the structure of a market where, as a result of the very presence of the undertaking in question, the degree of competition is weakened and which, through recourse to methods different from those which condition normal competition in products or services on the basis of the transactions of commercial operators, has the effect of hindering the maintenance of the degree of competition still existing in the market or the growth of that competition.123

However, intent may be of real significance in relation to specific abuses, such as predatory pricing. There is no formal exception incorporated into the text of Article 102 (or the Chapter II prohibition) that corresponds to Article 101(3), but conduct may escape condemnation as abuse if it can be objectively justified. For example, there is considerable room for debate whether particular conduct is a firm’s rational response to meet the challenge of competition or if it is abusive. Accordingly there is a close relationship between this issue and the following section on objective justification and commercial strategies which may be legitimately employed under Article 102. There is very little in the UK private enforcement case-law regarding the objective/subjective divide in the abuse case-law but certainly more in the public enforcement CAT jurisprudence, which has focused more on predatory elimination of rivalry.124 This

Case 85/76 Hoffmann-La Roche v Commission [1979] ECR 461 at 541. See David Bailey, ‘Early Case-Law of the Competition Appeal Tribunal’, in Barry James Rodger (ed.), Ten Years of UK Competition Law Reform (Dundee University Press, 2010), Chapter 2 123 124

Formalism, fairness and freedom of contract

231

has really only been considered in any depth by the English courts in one case, Arkin v Borchard Lines ltd & others,125 where Colman J observed as follows: 303. The question whether particular abnormal conduct of a dominant party has or is likely to have the effect necessary to amount to an abuse because it hinders competition is judged objectively: see Hoffmann-La Roche at paragraph 91 and BPB Industries and British Gypsum [1993] 5 CMLR 32 paragraphs 69 et seq. However, where the conduct is not intrinsically abnormal market conduct, such as price reductions, it may amount to an abuse if its purpose is to hinder competition for example by eliminating a competitor or by other means. 304. Where it is necessary to investigate eliminatory intent, however, the test is clearly subjective and not objective. There is, in particular, nothing in the authorities relied on by the claimant (Musique Diffusion Francaise [1983] ECR 1825, Hoffmann-La Roche, supra, BPB Industries, supra,) to suggest that as a matter of law the fact that the eliminatory consequences were reasonably foreseeable must lead to the conclusion that those consequences were intended. The test is essentially subjective – Did the undertaking have the necessary intention to procure the hinderance of competition? – but intention must be proved by inference from its conduct.

Here, intent is conflated with the issue of the commercial justification or purpose for the alleged anti-competitive business strategy. Although Colman J noted that the ‘the rate war was … unusually fiercely fought’,126 he considered there to be no direct evidence of an intent to eliminate and that the evidence constituted no more than ‘an ordinary operation of a participant in a fiercely competitive market and … is not indicative of exclusionary intent’.127 The combination of need to prove intent (albeit, it should be stressed, not a pre-requisite for a finding of abuse!) together with a benevolent approach to aggressive competition by a dominant company would certainly appear to make proving an abuse infringement in these circumstances problematic and arguably runs counter to the special responsibility doctrine. However in the context of analysis of commercial justifications for behaviour and analysis of the objective justification defence to an abuse infringement, the English courts have clearly taken a more intrusive and critical approach, notably in Purple Parking and Arriva. The only recent judgment in which the object/intent dichotomy was explored was in Streetmap.EU, where Roth J accepted the argument by Google that its strategy was motivated by its intention to improve its offering in the market for general search engines.

[2003] EWHC 687 (Comm); [2004] 2 CLC 242, Colman J. Para 319. 127 Para. 320. 125 126

232

Challenges to assumptions in competition law

9.4.4

Commercial Decisions and Objective Justification

A core aspect of the central theme in this chapter on the extent to which the abuse prohibition may impose limits on the way businesses act in a free market economy is the way in which business justifications are considered and dealt with by the courts – whether in a deferential or critical way. This involves practical application of the abstract terminology adopted by the European court in Hoffman la Roche and Michelin of conduct which departs from ‘normal methods of competition’ or which cannot be characterised as competition on the merits. Although judicial consideration of this issue is primarily undertaken at the stage of assessing whether conduct is abusive, the same underlying issues resonate in the often separate and subsequent consideration of whether conduct which is prima facie abusive can be objectively justified. The doctrine of objective justification was developed by the European Courts to allow undertakings to justify their otherwise infringing conduct, and the 2009 Commission Guidance on its Enforcement Priorities also explicitly recognises that the Commission may not challenge behaviour which is ‘objectively necessary’. This form of business justification was neatly encapsulated by Collins J in Global Knafaim Leasing Limited & CGTSN Limited v The Civil Aviation Authority, BAA Limited v Eurocontrol, NATS (En Route) Plc, Nats (Services) Limited, Secretary of State for Transport, Aviation Working Group:128 There may, if an abuse and an adverse impact on competition is found, be a justification. Such an objective justification can exist if the conduct in issue is proportionate to any legitimate commercial interest or public policy objective which may be identified. That is the position here.129

The issue of how commercial/objective justifications are to be considered by the courts has been revisited in both Purple Parking and Arriva. In the latter Rose J contrasted the approach in two earlier interlocutory decisions in the High Court where the test for objective justification was not particularly stringent. In Getmapping plc v Ordnance Survey130 Laddie J noted that it had not been suggested that the particular commercial ‘choice was irrational or even unreasonable’.131 Subsequently in The Wireless Group v Radio Joint Audience Research Ltd,132 Lloyd J struck out a claim for abuse of dominance, describing the defendant undertaking’s business decision as ‘a rational commercial

130 131 132 128 129

[2010] EWHC 1348 (Admin). Para 81. [2002] EWHC 1089 (Pat). Ibid. at para 55. [2004] EWHC 2925 (Ch).

Formalism, fairness and freedom of contract

233

approach’. Lloyd J referred to the decision of Laddie J in the Getmapping case as requiring an assessment of objective justification by the court which ‘must allow undertakings to take business decisions on normal commercial bases and in a normal way’ – although this simply begs the question as to what constitutes ‘business decisions in a normal way’. That approach was criticised and deemed inappropriate for all abuse cases in Arriva by Rose J as follows (at para 133): 133. …Those decisions were not purporting to decide that conduct will always be objectively justified provided it is not ‘irrational or arbitrary’ or provided that it can be described as a normal business decision. As ATS point out, such an approach would be self-defeating since anti-competitive behaviour usually is entirely rational from a commercial point of view, and abusive conduct usually promotes the business interests of the dominant undertaking engaging in it. That is why we need laws and enforcement agencies to prevent it.

Moreover, in Purple Parking the commercial justifications were viewed by the court as demonstrating an anti-competitive effect and thereby negating any potential reliance on an objective justification. It was stressed that Heathrow Airport’s acts in excluding Purple (and Meteor) from the forecourts were about ‘the picking up of business’,133 that ‘these commercial effects can be seen as part of the rationale behind the first proposed change, that at T5’,134 and ‘that what was ultimately proposed was all about commerce and what the meet and greet operators got out of use of the car park’.135 Accordingly ‘the principal motivation … was to disadvantage the meet and greet operators and to make competition more difficult’136 and there ‘were commercial reasons for this change’.137 Mann J subsequently addressed the related issue of objective justification, noting that this way of looking at objective justification (considering whether the conduct is prima facie abusive, and if so then considering whether it is objectively justified) is apparently required by the authorities, as opposed to treating the factor as one going to the overall inquiry as to abuse. This approach is exemplified, for instance, in Mann J’s assessment of the Terminal I forecourt exclusion, stating that motivations are key138 and the claimant could Para 145. Para 151. 135 Para 152. 136 Para 153. 137 Para. 157. 138 Commercial Solvents v Commission [1974] ECR 223. As noted by Justice Rose in Arriva at 134, he held that it was open to the claimant to show that even if the conduct of the dominant company could be objectively justified, if the undertaking’s motivation was in fact to suppress competition, then the plea of objective justification was not open to it. 133 134

234

Challenges to assumptions in competition law

test whether there is an objective justification by seeing whether the evidence shows that that justification was indeed the basis on which the dominant company acted, and a commercial motivation is insufficient for these purposes: 216. … I find that the real motivation for the change in relation to Terminal 1 was a commercial one. That commercial one was to force the off-airport meet and greet operators off the forecourts and into the car park, in order to increase revenues from the car park and to decrease the effectiveness of those operators in their valet parking operations. There was no real eye on the other factors referred to above. 217. That commercial justification is not a good objective justification for the purposes of the section. The other purported justifications are either non-existent or weak, and do not amount to good objectively justifiable reasons for creating the anti-competitive environment that HAL created (or wishes to create) there, taking them each individually and in aggregate.

Accordingly, in the view of Mann J, objective justification would require some factor extraneous to the simple commercial interests of the dominant undertaking – a view which harks back to the tenor of the Commercial Solvents ruling.139 Moreover, Mann J set a rigorous standard of proof required for an objective justification test,140 in stating that the law ‘requires a high degree of necessity if objective justification is relied on to justify what would otherwise be forbidden anti-competitive conduct’141 and the ‘essence is that there is no justification if there is another solution…’.142 Given that the matter was obiter in Streetmap.EU, in that the court had established that there had been no abuse, the court was also satisfied on the evidence that Google’s conduct had been objectively justified on the basis of a range of technical arguments related to the new-style Maps Onebox, and was not disproportionate. Nonetheless, the high threshold to be satisfied for an objective justification defence to succeed in each case where an abuse is established remains. This has been made clear in the two subsequent and most recent cases, Socrates and Achilles. In Socrates, the CAT stressed that the relevant test required the defendant Law Society to show ‘that the anti-competitive effect is outweighed by efficiencies or advantages achieved as a result, and that this restriction was proportionate i.e. that there was no less restrictive alternative of achieving the purpose allegedly pursued’.143 The CAT assessed the arguments and noted critically: ‘the evidence on practical difficulties as 139 Cases 6, 7/73 Commercial Solvents v Commission, [1974] ECR 223. And see United Brands infra. 140 See Arriva at para 134. See also European Commission in Flughafen Frankfurt/ Main AG (34.801) OJ 1998 L72/30 (‘Flughafen’). 141 Para 234. 142 Para 236. 143 At para 166.

Formalism, fairness and freedom of contract

235

scant and unpersuasive’.144 The Law Society’s case that there was no reasonable alternative to its own mandatory training was rejected.145 In Achilles the objective justification arguments were rejected: first on the basis of ulterior safety purposes as the defendants had failed to establish that those safety purposes could be met by a less restrictive rule146 and had failed to demonstrate that the restrictive rule generated sufficient cost efficiencies to offset the loss of competition.147 Both these recent cases highlight the CAT’s willingness to critically engage with objective justification defences in regulatory contexts on the basis of arguments for quality assurance and safety which must be fully supported and proportionate otherwise they will be rejected. 9.4.5

Abuse, Fairness and Freedom to Contract

This chapter has sought to develop a greater understanding of judicial approaches in the UK courts towards the abuse of dominance concept and the extent to which it may reflect fairness considerations in contrast with notions of the free market and traditional conceptions of contractual and property rights.148 The treatment of refusal to supply is a particularly controversial area; due, in no small part, to competition law’s interaction with other long-standing legal principles in the law of contract and intellectual property. Most legal systems have shied away from insisting that an undertaking be forced to contract when it does not see it to be in its best interests to do so. However, under EU law the basic rule is that a refusal to supply by a dominant undertaking must be objectively justified. However, the potential controversies extend beyond simply the refusal to supply doctrine. As we have already discussed, the domestic and EU abuse prohibitions potentially intrude beyond simple exclusionary anti-rival competitive strategies such as predatory pricing, and despite the Commission’s more limited enforcement priorities, private competition law disputes may involve long-standing and ongoing business relationships and aspects of the contractual or property rights of the parties. A willingness to engage with commercial decisions and agreements was demonstrated in Hendry and others v The World Professional Billiards and Snooker Association Limited (WPBSA).149 Subsequently, albeit in a summary

At para 174 At para 176. 146 See also para 315 involving a broader discussion of the relationship between restrictions on competition and health and safety: ‘restriction must be shown to be indispensable to the achievement of the health and safety purpose’. 147 Para 313. 148 Colino (n. 23). 149 5 October 2001 [2002] ECC 8. 144 145

236

Challenges to assumptions in competition law

judgment, we can witness the tensions between IP rights and a relatively wide view of the role of competition, by Mummery LJ in the Appeal Court in Intel Corp v Via Technologies Inc.150 Despite the jurisprudence in EU law in United Brands and General Motors,151 the Commission has focused its enforcement over the last 30 years on exclusionary anti-competitive behaviour, as evidenced by its 2009 Guidance on Enforcement Priorities. Nonetheless there have been some commercial contract disputes in the English courts concerning fairness issues, albeit these have demonstrated a clear reticence by the courts to unravel prices alleged to be unfair and/or excessive. In Ineos Vinyls Ltd v Huntsman Petrochemicals (UK) Ltd152 the claimant manufacturers had failed to establish that the defendant supplier of ethylene had a dominant position and abused it by charging prices and delivery charges that were unfair within the meaning of Article [102] and the Competition Act 1998 section 18, as they had failed to analyse the costs incurred in producing the product, compare those costs with the price charged, or evaluate the resulting profit. This reluctance to become involved in market pricing mechanisms was more evident in AttheRaces Ltd v British Horseracing Board,153 in which BHB’s pricing strategy was not abusive and the Court of Appeal clearly advocated a restrained approach to court involvement in claims of excessive or discriminatory pricing. Nonetheless, and on a parallel with case-law under the anti-competitive agreements provision in Calor Gas Ltd v Express Fuels (Scotland) Ltd,154 Arriva demonstrates that courts will consider behaviour to be abusive, even if the contract/procurement process or contract terms are not unfair per se, where, primarily as a result of the duration and extent of the exclusivity requirements, competition may be restricted. It was held in Calor Gas Ltd that the twin factors of market power and duration ensured that ‘if a nationwide network of principal dealers is tied to the brand leader for at least five years, this will restrict competition, especially in a mature market’.155 The difference is that Calor Gas concerned a horizontal restriction of competition as a result of the single-branding, whereas Arriva demonstrated that the courts may assess 150 26–28 November, 20 December 2002 [2003] FSR 33 [Intel Corp v Via Technologies Inc., 11–12 April, 14 June 2002 [2003] FSR 12]. 151 Case 27/76 United Brands v Commission, [1978] ECR 207; Case 26/75 General Motors Continental NV v Commission [1975] ECR 1367. 152 [2006] EWHC 1241 (Ch). 153 [2005] EWHC 3015 (Ch) and [2007] EWCA Civ 38. 154 [2008] CSOH 13. 155 At para 35. In relation to severability, it was noted that: ‘the proper approach is to view [the clause] as an integral and non-severable part of the overall anti-competitive aspects of the agreement, and thus it, along with other relevant clauses, is void in terms of Article [101(1)]’.

Formalism, fairness and freedom of contract

237

the contracts of a dominant undertaking where the exclusivity affects parties on a downstream market, and the dominant undertaking is not operating or directly benefiting from that market. The judgment in Purple Parking had obvious implications for the dominant undertaking’s freedom to contract and exercise of its property rights. Furthermore, the limitations on contractual autonomy for a dominant undertaking were also clearly evidenced in both the Socrates and Achilles rulings. The English courts have demonstrated a willingness to engage both with the determination of what constitutes ‘normal methods of competition’ and purported commercial, strategic and other rationale for justifying anti-competitive foreclosure, reflected in the CAT’s criticism of the absence of ‘fair competition for the market’ in Achilles.156

9.5

CONCLUDING REMARKS AND REFLECTIONS

There has been limited case-law on the abuse of dominance prohibition before the courts in the UK, and only more recently has there been more detailed judicial consideration of the substantive scope of the abuse concept. The ruling in Arkin confirmed the difficulties in identifying eliminatory intent where required and the Court of Appeal in AttheRaces certainly advocated a restrained judicial approach to alleged contract based ‘market failures’. Moreover, Laddie J adopted an approach which appeared to limit the potential involvement of courts in adjudicating in disputes involving commercial justifications for business decisions. The more recent case-law alludes to notions of fairness and freedom to compete, and a wider legitimacy for court intervention in contractual rights. It is impossible from the limited case-law and judicial consideration of the abuse doctrine to extrapolate a particular theory or approach of the judiciary to the scope of Article 102 / the Chapter II prohibition and the role of fairness but the more recent case-law certainly suggests a lower level of deference to unfettered markets, the legitimacy of rational business decisions (and their outcomes) when made by dominant undertakings, and the sanctity of the principle of freedom to contract and exercise property rights. These were already vulnerable under Article 102 and the jurisprudence of the European Courts, but one can trace a subtle change in the underlying approach from the earlier pronouncements of Laddie J to the later dicta and outcomes in particular in Purple Parking, Arriva, Socrates and Achilles.157

At para 301, italics added for emphasis. Albeit first instance decisions as opposed to the Court of Appeal ruling in AttheRaces. 156 157

10. Using trade tools to counteract anticompetitive conduct within global value chains: competition chapters in trade agreements Galyna Kostiukevych INTRODUCTION T-shirts, smartphones or Nutella jars are still marked with phrases like ‘Made in China’, ‘Made in Turkey’ or ‘Made in Italy’. But does this actually mean that a smartphone was fully manufactured in one country? In today’s globalized world that is impossible, as nowadays countries mostly trade in intermediate products, not finished ones. Recent studies conducted by the Organisation for Economic Co-operation and Development (OECD) and the European Central Bank (ECB) show that today around 70 percent of international trade relies on Global Value Chains (GVCs).1 Intermediate goods (primary goods, parts and components, and semi-finished products) represent more than half of the world’s manufactured imports, whereas intermediate services represent more than 70 percent of the world’s services imports.2 First things first: how could GVCs be defined? These are arrangements between companies specializing in a particular stage of production and usually situated in different countries, which agree to produce intermediate products to which they add some value at every stage of the production according to their specialization. The last stage consists of producing the finished (final) products. The relationship between trade in intermediate products and deep trade agreements can be described as two-way traffic. On the one hand, the exist1 OECD, ‘Trade Policy Implications of Global Value Chains’ (2020), Trade Policy Brief, https://​issuu​.com/​oecd​.publishing/​docs/​trade​_policy​_implications​_of​_global accessed 30 March 2020. 2 Koen de Backer and Sébastien Miroudot, ‘Mapping Global Value Chains’ (2014) Working Paper No. 1777, European Central Bank, p. 2.

238

Using trade tools to counteract anticompetitive conduct within global value chains

239

ence of trade agreements can foster the creation of value chains. On the other hand, countries with many GVCs can be interested in signing deeper trade agreements with countries where other participants of supply chains are established to facilitate and secure their trade relationship.3 Thus, the depth of a trade agreement increases network trade among the countries which signed such an agreement.4 Deep agreements foster GVCs, because they contribute to the increase of backward and forward linkages within GVCs.5 If a policy area is added to a preferential trade agreement (PTA), this results in increasing forward GVC linkages (imports of intermediate (not finished) products which a country uses for its exports) by 0.48 percent and backward linkages (exports in intermediate products that are re-exported by third countries) by 0.38 percent.6 One should keep in mind though that the impact of trade agreements differs across industries. Closer integration matters more for high capital-intensity industries like automotive parts and information and communication technology than for standardized industries like textile production. The previous studies on behind-the-border measures examined why countries decide to insert those commitments into their trade agreements.7 Their conclusion was that in this way it will be more problematic and costly for the parties to the agreement to deviate from their obligations on behind-the-border policies mentioned in the trade agreement.8 Other studies focused specifically on competition provisions and scrutinized some potential goals (selective enforcement versus enhancing the capacity to detect international anticompetitive behavior) of competition provisions in trade agreements.9 Some other contributions categorized the types of competition provisions and examined 3 Gianluca Orefice and Nadia Rocha, ‘Deep Integration and Production Networks: An Empirical Analysis’ (2014) 37 (1) The World Economy 106. 4 Ibid. p. 107; Michele Ruta, ‘Preferential Trade Agreements and Global Value Chains: Theory, Evidence and Open Questions’ (2017) Policy Research Working Paper No. WPS 8190, World Bank Group, p. 19. 5 Cristina Constantinescu, Aaditya Mattoo, Alen Mulabdic and Michele Ruta, Global Trade Watch 2017: Trade Defies Policy Uncertainty – Will It Last? (World Bank Group, 2018), p. 20. 6 Edith Laget, Alberto Osnago, Nadia Rocha and Michele Ruta, ‘Deep Trade Agreements and Global Value Chains’ (2018) Policy Research Working Paper No. WPS 8491, World Bank Group. 7 Terry M. Moe, ‘Power and Political Institutions’ (2005) 3(2) Perspective on Politics 215; Edward D. Mansfield and Helen V. Milner, Votes, Vetoes, and the Political Economy of International Trade Agreements (Princeton University Press, 2012). 8 Ibid. 9 Anu Bradford and Tim Buthe, ‘Competition Policy and Free Trade: Antitrust Provisions in PTAs’ in Elsig Manfred and Andreas Dur (eds), Trade Cooperation (Cambridge University Press, 2015).

240

Challenges to assumptions in competition law

its implications for multilateral trade cooperation.10 However, these studies never made a link between trade agreements, GVCs and competition provisions. All three phenomena were linked only in recent studies produced by the World Bank.11 These studies were purely economic and therefore are based on some assumptions which should be scrutinized by lawyers. For instance, they assume that competition provisions in trade agreements are enforceable. This chapter aims to fill up this lacuna and provides a legal analysis on whether competition provisions in trade agreements can be used to tackle anticompetitive conduct within GVCs. The chapter is structured as follows. Section 10.1 analyzes the impact of PTAs on the development of GVCs and describes potential foreclosure effects of PTAs on companies which could potentially join GVCs. Section 10.2 addresses theories of harm. Section 10.3 scrutinizes competition provisions within several exemplary PTAs. Section 10.4 looks at how competition provisions could be used to tackle anticompetitive conduct.

10.1

INTERRELATIONS BETWEEN PTAS AND GVCS

Before the rise of GVCs, a country could exploit its comparative advantage. Yet nowadays companies are the driving force of GVCs, which use the comparative advantages of different countries.12 Indeed, GVCs are organized by exploiting the comparative advantage of each company in each country.13 In a modern world a competing unit is a GVC, which goes beyond a country’s border, and competes on the market of finished products. Competition law assumes that if there are many competitors on a relevant market, purchasers can easily switch from one manufacturer to another. I argue that in the context of GVCs, the situation is different. Apart from more ‘traditional’ obstacles to switching (interdependence, need for stability, complica-

10 Simon J. Evenett, ‘What Can We Really Learn from the Competition Provisions of Regional Trade Agreements?’ in Philippe Brusick, Lucian Cernat and Ana Maria Alverez (eds), Competition Provisions in Regional Trade Agreements: How to Assure Development Gains (United Nations, 2005). 11 Michele Ruta, ‘Preferential Trade Agreements and Global Value Chains’ in Global Value Chain Development Report 2017: Measuring and Analysing the Impact of GVCs on Economic Development, International Bank of Reconstruction and Development and World Bank Group, p. 177. 12 Richard Baldwin, The Great Convergence: Information Technology and the New Globalization (Harvard University Press, 2016), p.11. 13 Christophe Degain and Andreas Maurer, ‘Implications of Global Value Chains for Trade Statistics and Trade Policy’ in Yuqing Xing (ed.), Uncovering Value Added in Trade: New Approaches to Analyzing Global Value Chains (Asian Development Bank Institute, 2016), p. 36.

Using trade tools to counteract anticompetitive conduct within global value chains

241

tions with finding the same volume of the product), production within GVCs which are created along the PTAs poses an additional challenge of foreclosure. Before explaining this, let us explore in more detail how PTAs and GVCs are interlinked. Trade liberalization played a significant role in the development of GVCs. Nowadays, however, the process of multilateral trade liberalization has slowed down. Instead, further tariff reductions and other preferences are rather given on the basis of trade agreements between several countries (PTAs). There is consensus about two main advantages of PTAs for the countries and consequently their companies. First, PTAs facilitate market access to foreign countries, which helps achieve the economies of scale.14 And second, companies get the possibility to move their stages of production to other countries and therefore exploit the comparative advantage of different countries.15 According to a study conducted by the World Bank, PTAs and the development of GVCs are related.16 That study, together with some others which followed later,17 concluded that deep trade agreements matter more for the PTAs than purely tariff preferences. As the Most Favorite Nation (MFN) tariffs declined, the difference between MFN rates and preferential rates for the parties to PTAs was not significant. 18 As a result, gains from preferential advantage and losses from preferential disadvantage for global exports were quite low. On these grounds, it was concluded that the main driver for the creation of GVCs nowadays is not the level of tariffs, but the depth of PTAs.19 This does not mean, however, that tariffs are totally insignificant for GVC participation. They matter for both forward and backward participation. If a country’s average manufacturing tariff is decreased by 1 percent, that country’s backward participation share in exports will increase by 0.4 percent.20 Tariffs in

14 Kerry A. Chase, ‘Economic Interests and Regional Trading Arrangements: The Case of NAFTA’ (2003) 57(1) International Organization 137, p. 144. 15 Luis Marcelo Florensa, Laura Márquez-Ramos, Inmaculada Martínez-Zarzoso and María Luisa Recalde, ‘Regional Versus Global Production Networks: Where Does Latin America Stand?’ (2015) 47(37) Applied Economics 3938. 16 Michele Ruta, ‘Preferential Trade Agreements and Global Value Chains: Theory, Evidence and Open Questions’ (2017) Policy Research Working Paper No. WPS 8190, World Bank Group, p. 19. 17 Cristina Constantinescu, Aaditya Mattoo, Alen Mulabdic and Michele Ruta (2018) Global Trade Watch 2017: Trade Defies Policy Uncertainty – Will It Last? (World Bank Group, 2018), p. 20. 18 Ibid. 19 Ibid. 20 World Development Report 2020: Trading for Development in the Age of Global Value Chains, p. 42, box 2.2, figure B2.2.2, World Bank https://​www​.worldbank​.org/​ en/​publication/​wdr2020 accessed 28 March 2020.

242

Challenges to assumptions in competition law

the country of destination are also an important factor for GVC participation, as they ensure better market access for foreign companies. If the average tariff for some sector production is reduced by 1 percent in a destination country, a country’s backward GVC integration will increase by 6 percent in the given sector and its forward GVC participation by 7 percent.21 Naturally, higher tariffs preclude companies from participating in GVCs. Preferential tariffs therefore encourage the GVC participation, but it is not the only crucial factor that should be considered. Among other important elements, rules of origin and behind-the border policies should be taken into account. Rules of origins define how much of the product’s value should be generated in the country to enjoy the preferential tariffs. This chapter does not go into details regarding rules of origins, as this question has already been sufficiently addressed in the literature.22 Instead, I focus on competition rules in the context of GVCs, which fall into the category of behind-the-border policies. PTAs are referred to as deep trade agreements, when in addition to the ‘traditional’ set of provisions on tariffs and market access they also cover a broader scope of issues, such as competition policy or investment. In contrast, shallow trade agreements focus solely on tariffs and other border measures.23 More recently, Horn et al. proposed to distinguish between the WTO+ and WTO-X provisions in trade agreements, depending on whether the issue has been covered by the WTO or not. WTO+ provisions cover the same issue as WTO Agreements, but go further than the WTO commitments. For instance, signatories of a PTA can agree to reduce tariffs between them even further than the minimum level established by the WTO Agreements. WTO-X provisions cover a range of areas which are not addressed by the WTO Agreements, such as competition, environment, government procurement and labor.24 As we can see, shallow trade agreements rather contain WTO+ provisions, whereas deep trade agreements comprise WTO-X provisions. This is important to keep in mind, because the cited studies of the World Bank show that deep trade agreements have a decisive impact on the development of GVCs. Deep PTAs foster backward and forward GVC integration.25 If a policy area is added to Ibid. For more information see: Dylan Geraets, Colleen Carroll and Arnoud R. Willems, ‘Reconciling Rules of Origin and Global Value Chains: The Case for Reform’ (2015) 18(2) Journal of International Economic Law 287. 23 Robert Z. Lawrence, Regionalism, Multilateralism, and Deeper Integration (Brookings Institution, 1996). 24 Henrik Horn, Petros C. Mavroidis and Andre Sapir, ‘Beyond the WTO? An Anatomy of EU and US Preferential Trade Agreements’ (2010) 33(11) The World Economy 1565. 25 Swati Dhingra, Rebecca Freeman and Eleonora Mavroeidi, ‘Beyond Tariff Reductions: What Extra Boost from Trade Agreement Provisions?’ (2018) Discussion Paper No. 1532, Centre for Economic Performance. 21 22

Using trade tools to counteract anticompetitive conduct within global value chains

243

a PTA, this results in the increase of forward GVC linkages (imports of intermediate (not finished) products which a country uses for its exports) by 0.48 percent and the growth of backward linkages (exports in intermediate products that are re-exported by third countries) by 0.38 percent.26 The depth of trade agreements is determined by the number of non-WTO areas which are covered by the agreement. Competition policy is not addressed by the WTO. So, competition provisions are one of the markers which are used to identify the depth of PTAs. Importantly, competition is one of the WTO-X areas which are most often included into PTAs. According to the World Bank study, competition provisions are present in 200 PTAs.27 Based on these findings, one might assume that competition provisions are one of the drivers behind the emergence of the new GVCs. However, the mere presence of competition clauses in PTAs does not stimulate the creation of GVCs. Deep trade agreements include provisions on behind-the-border policies, which in turn can improve the functioning of foreign institutions.28 Those provisions may stimulate institutional reforms or at least send a signal that the country is ready to cooperate on these matters with its partners. Stronger regulatory framework defined in the PTA may demonstrate that a country which commits to improve, for instance, its competition policy or government procurement is trustworthy. As a result, companies from that country become more attractive to cooperate with and be involved in GVCs. As we see, PTAs as a tool of trade liberalization stimulate the development of GVCs. There is, however, a dark side to the relationship between GVCs and PTAs. If there is an agreement between a producer of components and an assembler who are located in the countries which have a PTA, they have preferential terms for commerce. From the perspective of GVCs, this means that dealing with a long-term partner on zero tariffs is more beneficial than dealing with a company located in a non-PTA country, where the tariffs are, for instance, 7–10 percent. This tariff would mean that the price of the component becomes higher and, as a result, less competitive. This consideration can contribute to the decision of a company not to switch to an alternative producer

26 Edith Laget, Alberto Osnago, Nadia Rocha and Michele Ruta, ‘Deep Trade Agreements and Global Value Chains’ (2018) Policy Research Working Paper No. 8491, World Bank Group. 27 Claudia Hoffman, Alberto Osnago and Michele Ruta, ‘Horizontal Depth: A New Database on the Content of Preferential Trade Agreements’ (2017) Policy Research Working Paper No. 7981, World Bank Group, p. 11. 28 World Development Report 2020: Trading for Development in the Age of Global Value Chains, p.  54, World Bank https://​www​.worldbank​.org/​en/​publication/​wdr2020 accessed 28 March 2020.

244

Challenges to assumptions in competition law

if it is located in a country which does not have a PTA with the country where the assembler is located. Besides, PTAs are not only about tariffs: other trade preferences (in particular, rules of origin) also matter while choosing a partner for long-term cooperation. Thus, GVCs can result in hindering competition by creating foreclosures to the competitors who are not coming from a country which has a PTA. If a GVC is created along the PTA, companies coming from other countries would automatically be in a worse position than the companies coming from the country of the PTA. Companies from the non-PTA countries would not enjoy the same treatment as the companies from the PTA countries. For instance, the existence of tariffs (even not very high), less advantageous rules of origins or worse regulation of the behind-the-border policies would create barriers for companies from non-PTA countries to join a GVC. Therefore, PTAs can contribute to enhancing competition risks. Companies coming from a non-PTA country would not be able to compete with companies coming from PTA countries.

10.2

THEORIES OF HARM

Anticompetitive conduct that occurs within GVCs is not properly addressed by competition chapters of PTAs as they are often drafted in current PTAs. Anticompetitive conduct within GVCs can be harmful for several actors: other participants of the GVCs, non-participants who are potential competitors to the participants of the GVC, and consumers. Trade agreements mainly address anticompetitive conduct which impedes market access for foreign producers, but often ignores the anticompetitive conduct which adversely affects consumers. The reason of that lies in the prime goals of competition and trade policies. Trade policy focuses on states’ measures and its objective is to ensure that the interests of individual producers are protected.29 In order to achieve this, countries try to reduce or eliminate state barriers to trade – tariffs and quotas and non-tariff measures. The objectives of competition law enforcement are not clearly defined, because there is no consensus about that among the countries.30 Some clarifications regarding the universal objectives of the competition policy were brought by the Working Group on Interaction between Trade and Competition, which was created under the auspices of the WTO to analyze the interlinks between these two policies in 1997. The task of the group was to 29 Report (1998) of the Working Group on the Interaction between Trade and Competition Policy to the General Council, WT/WGTCT/2, World Trade Organization, p. 27. 30 Ariel Ezrachi, ‘Setting the Scene: The Scope and Limits of “International Competition Law”’ in Ariel Ezrachi (ed.), Research Handbook on International Competition Law (Edward Elgar, 2012) p. 10.

Using trade tools to counteract anticompetitive conduct within global value chains

245

analyze how both policies interact and whether it would be possible to complement the international trade regime with competition provisions. In the first Report (1998) the Working Group agreed with the view that the objective of competition policy is to achieve efficiency and consumer welfare.31 However, the exact meaning of the consumer welfare standards as well as the methods to achieve it can vary.32 Moreover, the welfare standard can be extended to also include protection of the competitive process.33 In any case, competition policy is not focused on the interests of the producers. Despite different interpretations, consumer welfare constitutes a central goal of competition policy. This analysis demonstrates that the objectives of trade and competition policies are not entirely the same: trade policy ensures the interests of producers, whereas competition policy is rather focused on consumers (with the exception of ‘protecting competitors’ policy choices). Moreover, trade policy focuses on state (public) measures, whereas competition policy touches upon the anticompetitive conduct of private entities. A very important common feature though is that both policies aim to ensure market access. This is exactly why trade and competition are seen as complementary and explains why they are called two sides of one coin,34 i.e. they are two important policies to ensure market access. Current competition chapters in trade agreements follow this logic. Competition matters for trade only when anticompetitive conduct hinders trade. This is why some competition provisions had been included in the trade agenda: first, in the cases decided by the WTO and then into PTAs. Competition provisions in current trade agreements partially address situations where market access is impeded, but they do not address the challenges which exist in the context of GVCs. For instance, trade agreements do address import cartels – the situation where companies from one country entered into a price-fixing agreement to keep the foreign competitors out of the market. This is an example of anticompetitive conduct which would be relevant purely from a trade perspective, because it would hinder market access for foreign 31 Report (1998) of the Working Group on the Interaction between Trade and Competition Policy to the General Council, WT/WGTCT/2, World Trade Organization, p. 27. 32 International Competition Network, ‘Competition Enforcement and Consumer Welfare: Setting the Agenda’ (2011), https://​www​.i​nternation​alcompetit​ionnetwork​ .org/​wp​-content/​uploads/​2019/​11/​SP​_CWelfare2011​.pdf accessed 1 March 2020. 33 Case C-501/06,GlaxoSmithKline Services Unlimited v Commission [2010], 4 CMLR 2, para 63: ‘Article 101 TFUE aims to protect not only interests of competitors or of consumers, but also the structure of the market and, in doing so, competition as such’. 34 Merit E. Janow, ‘Trade and Competition Policy’ in Patrick F. J. Macrory, Arthur E. Appleton and Michael G. Plummer (eds), The World Trade Organization: Legal, Economic and Political Analysis, Volume 2 (Springer, 2005) p. 448.

246

Challenges to assumptions in competition law

products despite the efforts of trade liberalization. However, with the development of GVCs, trade relations became more complex. I argue that in the context of GVCs, import cartels are not a matter of prime concern. The core idea why GVCs are created, is to allow companies to specialize in different stages of production. This means that one company decides to cooperate with another foreign company, because there are no domestic producers which participate in a different stage of production. This is a reason to look for foreign partners in the first place. Import cartels mean that domestic producers engage in price-fixing to keep foreign producers out of their domestic markets. Basically, this occurs when they want to prevent competition from foreign companies which obtained access to their domestic market. This scenario is rather unlikely in the context of GVCs. Therefore, competition provisions which deal only with import cartels are designed to address the anticompetitive conduct, which creates only ‘classical’ obstacles for international trade related to market access. Export cartels are a more relevant form of anticompetitive conduct in the context of GVCs. Much attention in the scholarly literature has been devoted to their nature,35 in particular to the domestic effects of export cartels36 and it was sufficiently discussed whether there is a need for global rules.37 It is recognized that export cartels primarily affect markets abroad, but some also argue that export cartels may adversely affect domestic markets. In the light of GVCs, their effects on the domestic market are rather irrelevant. What is relevant for the purpose of this chapter is how export cartels can affect participants of GVCs which are situated abroad or companies that would want to become a part of a GVC. In the GVC setting, an example of export cartels would be a horizontal agreement on one stage of a transformation process. Let us assume that a participant of a GVC enters a price-fixing cartel with other producers of the same components. These components are later sold to the downstream companies within the GVC. As a result, the price for the product transformation becomes higher. A company which participates in the next stage of transformation buys the intermediate product for the higher price, because challenging this conduct would be more complicated. In particular, this would result in a disruption of its manufacturing process, which is costly in itself. The company which buys for a higher price simply passes on the costs to the next company. All 35 Florian Becker, ‘The Case of Export Cartel Exemptions: Between Competition and Protectionism’ (2007) 3(1) Journal of Competition Law and Economics 97. 36 Daniel Sokol, ‘What Do We Really Know about Export Cartels and What is the Appropriate Solution?’ (2008) 4 Journal of Competition Law and Economics 967. 37 Brendan Sweeney, ‘Export Cartels: Is There a Need for Global Rules’ (2007) 10(1) Journal of International Economic Law 87.

Using trade tools to counteract anticompetitive conduct within global value chains

247

subsequent companies do the same. Finally, the finished product arrives to a consumer who has to pay a higher price. In this scenario, the party who is harmed is the consumer. This is a typical case where competition law should be applicable, because it is clearly about consumers’ welfare. However, there are no international competition rules on that matter. In theory, the competition authority of the country where the consumers are affected could investigate the anticompetitive conduct. Yet if an export cartelist was involved in the GVCs, the place where the conduct originates and the place where it affects consumers would be quite far apart. This makes an enforcement procedure even more problematic than in the situation where the cartelists are exporting directly to the country which is affected by that conduct. Export cartels are a very controversial topic in the area of competition law: some countries explicitly exempt export cartels from antimonopoly sanctions (even those where competition laws related to other issues are strict). This exception is explained by the lack of incentives to prosecute such cartels. First, export cartels do not hurt domestic consumers, only consumers abroad. Second, the economy of the country might actually benefit from allowing the companies to create an export cartel. Third, some authors who defend this exception claim that export cartels could and should be challenged in the importing country in order not to create a conflict of jurisdictions between the exporting and the importing country.38 However, this argument does not take into account a situation where the importing country cannot extract necessary evidence from the exporting country, or if the importing country cannot implement sufficient remedies.39 In order to do so, the countries would often need to cooperate with the competition agencies of the country where the conduct originates. But in practice such cooperation does not occur very often.40 Last, although export cartels are often presumed to be harmful, some claim that export cartels have pro-competitive effects, in particular for small- and medium-sized enterprises which can start exporting their products only thanks to such cooperation. Therefore, according to them, export cartels help to boost competition instead of restricting it.41 Indeed, small- and medium-sized 38 Damien Geradin and Einer R. Elhauge, Global Antitrust Law and Economics (3rd edn., Foundation Press, 2018) p. 1101. 39 Anu Bradford, ‘International Antitrust Negotiations and the False Hope of the WTO’ (2007) 48(2) Harvard International Law Journal 383. 40 OECD, Working Party No. 3 on cooperation and enforcement, ‘Limitations and constraints to international enforcement’ DAF/COMP/WP3(2012)8 http://​www​.oecd​ .org/​officialdocuments/​publi​cdisplaydo​cumentpdf/​?cote​=​DAF/​COMP/​WP3(2012)8​&​ docLanguage​=​En accessed 27 February 2020. 41 WTO Working Group on the Interaction between Trade and Competition Policy, Note by the Secretariat: Report on the Meeting of 20–21 February 2003 WT/ WGTCP/M/21 (26 May 2003), para 37. See also Spencer W. Waller, ‘The Failure

248

Challenges to assumptions in competition law

enterprises might become more competitive internationally if they decide to cooperate. However, this argument does not justify a full exemption of export cartels from the scope of competition laws. If an objective is indeed to help small- and medium-sized enterprises, the legislation can provide for a de minimis rule42 that already exists in the context of anticompetitive agreements between the undertakings. This principle means that the behavior of undertakings with small market shares is not capable of significantly hindering trade and/or competition. As export cartels escape the competition enforcement by national authorities, I suggest that trade agreements should address this problem. Although right now competition chapters in PTAs are not focused on export cartels, this has to change. If trade agreements enter into play as a tool to counteract the anticompetitive conduct within GVCs, export cartels should not be overlooked. In the context of GVCs international cartels also represent a major concern.43 International cartels are created by producers who are operating in different countries and who intend to cartelize both domestic and export markets.44 International export cartels consist of exporters from more than one country who are interested only in exports. The most prominent examples of international cartel cases were the Motorola cases45 in the US and Innolux case46 decided by the European Court of Justice. These cases dealt with a cartel on the market of LCD panels. LCD panels are intermediate products for smartphones, laptops and TVs. The issue at stake in those cases was how to punish the cartel members who produced the LCD panels and then sent them to third countries where they were assembled into finished products. Only the finished products (mobile phones etc.) entered the US and EU markets. Because there were no direct sales of the cartelized product to the US and EU market, the extraterritorial application of the antitrust was not straightforward.

of the Export Trading Company Program’ (1992) 17(2) North Caroline Journal of International Law and Commercial Regulation 239, 250. 42 Communication from the Commission, Notice on agreements of minor importance which do not appreciably restrict competition under Article 101(1) of the Treaty on the Functioning of the European Union (De Minimis Notice), C(2014) 4136 final. 43 OECD, WTO, UNCTAD joint report prepared for the G-20 Leaders Summit in Saint Petersburg (Russian Federation), ‘Implications of Global Value Chains for Trade, Investment, Development and Jobs’ (2013) https://​www​.oecd​.org/​trade/​G20​-Global​ -Value​-Chains​-2013​.pdf accessed 20 February 2020, p. 18. 44 Frederic Jenny, ‘Export Cartels in Primary Products: The Potash Case in Perspective’ in Simon J. Evenett and Frederic Jenny (eds), Trade, Competition, and the Pricing of Commodities (Centre for Economic Policy Research, 2012). 45 United States v Hsiung, 758 F. 3d 1074 (9th Cir. 2014); Motorola Mobility LLC v AU Optronics Corp., 775 F 3d 816 (7th Cir. 2014). 46 Case C-231/14 P Innolux v Commission [2015] EU:​C:​2015:​451.

Using trade tools to counteract anticompetitive conduct within global value chains

249

As eliminating the anticompetitive conduct which is relevant in the context of GVCs does not always lie in the realm of policy objectives, one might argue that competition chapters should be used purely to complement the goals of trade liberalization. I argue differently: competition chapters should be considered as an instrument to tackle anticompetitive conduct within the supply chains regardless whether the anticompetitive conduct in question clashes with objectives of trade liberalization or simply harms consumers. Some competition chapters in modern trade agreements also mention specific goals of competition provisions. While some recent EU trade agreements still mention that anticompetitive conduct should be eliminated because it can have adverse effects on trade liberalization,47 other agreements often mention specifically that the goals of competition chapters coincide with the goals of competition policy in general, for instance economic efficiency and consumers’ welfare.48 These goals have recently been included in trade agreements, because there is evidence that anticompetitive practices adversely affect people’s welfare by increasing the prices of input goods, by raising the costs for transportation and by otherwise increasing the prices for final goods and services to consumers.49 Export cartels are a clear example of the practices which harm consumer welfare. If inputs producers can enter an export cartel, it is likely that end consumers will have to pay the price increase, as the other companies in-between would simply pass on the increased costs. The next part of the chapter analyzes which types of competition provisions are common in PTAs and whether they can address the issue at stake.

10.3

COMPETITION PROVISIONS IN TRADE AGREEMENTS: CLASSIFICATION

The OECD was a pioneer in categorizing competition provisions in trade agreements. In 2005, the first part of the Secretariat paper which analyzed 47 regional trade agreements was published and was immediately picked up by 47 Agreement between the EU and Japan for an Economic Partnership http://​ trade​.ec​.europa​.eu/​doclib/​docs/​2018/​august/​tradoc​_157228​.pdf​#page​=​333 accessed 1 March 2020; Comprehensive Economic and Trade Agreement between Canada and the EU (CETA), http://​trade​.ec​.europa​.eu/​doclib/​docs/​2014/​september/​tradoc​_152806​.pdf accessed 1 March 2020. 48 US-Singapore free trade agreement, https://​ustr​.gov/​sites/​default/​files/​uploads/​ agreements/​fta/​singapore/​asset​_upload​_file708​_4036​.pdf accessed 1 March 2020. 49 Robert D. Anderson, William E. Kovacic, Anna Caroline Muller and Nadezhda Sporysheva, ‘Competition Policy, Trade and the Global Economy: Existing WTO Elements, Commitments in Regional Trade Agreements, Current Challenges and Issues for Reflection’ (2018) Staff Working Paper ERSD-2018-12, World Trade Organization, p. 28.

250

Challenges to assumptions in competition law

scholars working on this topic.50 In 2006, the OECD published a follow-up Working Paper which analyzed 86 agreements including the ones notified to the WTO Secretariat from 2001 to 2005.51 The OECD defined the following categories: 1. Measures which are related to adopting, maintaining and applying competition measures; 2. Provisions on coordination and cooperation between national competition enforcement entities, which include an obligation to notify another party about enforcement actions, to exchange the information, to consult about competition policy and enforcement. In addition, this category may include obligations related to negative and positive comity; 3. Provisions related to anticompetitive behavior which may prohibit different anticompetitive practices as anticompetitive agreements (horizontal and vertical), abuse of dominance, state aid (subsidies) and anticompetitive mergers. According to the OECD Working Paper, some trade agreements can state that such anticompetitive agreements or abuse of dominance are incompatible with the trade agreement in question.52 However, many trade agreements rather provide for an obligation of the parties to adopt measures to ban anticompetitive conduct. Some trade agreements can include examples of anticompetitive conduct that should be eliminated;53 4. Provisions on non-discrimination, due process and transparency; 5. Provisions which exclude the application of antidumping laws; 6. Provisions which allow a recourse to trade defense mechanisms in the circumstances where a trading partner breaches its obligations related to competition; 7. Provisions on dispute settlement, which exclude competition provisions from the dispute settlement mechanism, consultations and arbitration. The vast majority of the trade agreements put competition provisions outside the scope of dispute settlement. This means that competition chapters are impossible to enforce by using the dispute settlement mechanism. Some

50 Simon J. Evenett, ’What Can We Really Learn from the Competition Provisions of Regional Trade Agreements?’ in Philippe Brusick, Lucian Cernat and Ana Maria Alverez (eds), Competition Provisions in Regional Trade Agreements: How to Assure Development Gains (United Nations, 2005). 51 Oliver Solano and Andreas Sennekamp, ‘Competition Provisions in Regional Trade Agreements’, OECD Trade Policy Working Paper No. 31 COM/DAF/TD(2005)3/ FINAL, pp. 7–9. 52 Ibid. p. 8. 53 Ibid. p. 8.

Using trade tools to counteract anticompetitive conduct within global value chains

251

trade agreements provide for consultation as a first step in dispute settlement. Arbitration is a very rarely mentioned tool in competition chapters; 8. Provisions on flexibility of commitments (non-reciprocal provisions, exemptions and exceptions); technical assistance and capacity building and transition periods. Although the analysis conducted by the OECD did not include the recent trade agreements, this categorization is relevant for also scrutinizing current trade agreements, as they largely have the same types of provisions mentioned in the Working Paper of the OECD. A recent study by the WTO54 analyzed more trade agreements and concluded that out of the 280 trade agreements which were notified to the WTO, 56 percent contain special chapters or provisions on competition.55 The World Bank in their analysis claims that competition chapters are enforceable. As was mentioned before, their study did not analyze those provisions from a legal perspective. As to the enforceability, approximately 34 percent of trade agreements containing a competition chapter allow enforcers to use dispute settlement mechanisms for competition-related issues. Fifty-one percent of trade agreements with competition chapters provide for consultations instead and 15 percent of the trade agreements totally exclude competition from all types of dispute settlement.56 According to the analysis conducted by the WTO, provisions on adoption or maintenance of competition laws are included in 55 percent of the trade agreements. The vast majority of trade agreements with competition chapters define which anticompetitive conduct should be regulated. Eighty percent of trade agreements mention anticompetitive agreements and abuse of dominant position, whereas mergers are addressed in approximately 23 percent of trade agreements. Cooperation on competition policy is mentioned in 66 percent of 54 Robert D. Anderson, William E. Kovacic, Anna Caroline Muller and Nadezhda Sporysheva, ‘Competition Policy, Trade and the Global Economy: Existing WTO Elements, Commitments in Regional Trade Agreements, Current Challenges and Issues for Reflection’ (2018) Staff Working Paper ERSD-2018-12, World Trade Organization. 55 The data for the WTO study was taken from the WTO Regional Trade Agreements Information System (WTO RTA Database), http://​ rtais​ .wto​ .org/​ UI/​ P​ ublicMaint​ainRTAHome​.aspx accessed 12 February 2020; the list of the agreements analyzed by Anderson, Kovacic, Muller and Sporysheva is available as Appendix Table 1 and 2 to their paper. 56 Robert D. Anderson, William E. Kovacic, Anna Caroline Muller and Nadezhda Sporysheva, ‘Competition Policy, Trade and the Global Economy: Existing WTO Elements, Commitments in Regional Trade Agreements, Current Challenges and Issues for Reflection’ (2018) Staff Working Paper ERSD-2018-12, World Trade Organization, p. 33.

252

Challenges to assumptions in competition law

the trade agreements with competition chapters. Obligations to ensure transparency are present in 49 percent of recent trade agreements, whereas obligations related to non-discrimination and procedural fairness are mentioned in 32 percent. The question I would like to address here is whether the abovementioned types of anticompetitive conduct can help to tackle anticompetitive conduct within GVCs. Of course, to some extent any competition provision is more helpful than a lack thereof. If the country does not have any competition laws, then the obligations to adopt competition policy would already be a step forward. However, in order to address competition problems within GVCs, the mere obligation to adopt competition law would not be enough. This is only a preparatory step and without further obligations to eliminate anticompetitive conduct this obligation in itself does not achieve much. Another type of provisions commonly enshrined in trade agreements are provisions on coordination and cooperation between national competition enforcement entities. They are crucial to make two or more competition authorities work together on one case. These provisions can also be found in cooperation agreements related to anticompetitive practices and the application of competition laws. The European Union has such agreements with several countries,57 for instance with the US, Switzerland, Japan, Canada, South Korea and others. The US also signed several cooperation agreements.58 Without diminishing their importance, it should be noted that in the context of GVCs, these agreements do not properly replicate the design of GVCs. On the contrary, as trade agreements foster GVCs, companies established in the countries which signed a PTA most probably either already cooperate or plan to cooperate within GVCs. Also, trade agreements are concluded with more countries than the cooperation agreements on competition. In addition to that, cooperation agreements do not have an enforceability mechanism, whereas a trade framework is more complex and thus can provide more tools to influence another party. Competition chapters in trade agreements are often explicitly excluded from the dispute settlement and thus are not automatically enforceable in a traditional way. In this section, however, I want to show how trade agreements could become a real tool to tackle anticompetitive conduct. First let us have a look at the studies by the World Bank mentioned earlier in the chapter. These studies assumed that competition chapters in trade agree-

57 A full list of the agreements can be found here: https://​ec​.europa​.eu/​competition/​ international/​bilateral/​accessed 25 February 2020. 58 A full list of countries and agreements is available here: https://​www​.justice​.gov/​ atr/​antitrust​-cooperation​-agreements accessed 25 February 2020.

Using trade tools to counteract anticompetitive conduct within global value chains

253

ments are fully enforceable. As a result, the World Bank decided that competition chapters are crucial for GVCs. After we have categorized the types of provisions in competition chapters, it became clear that there are different types of competition provisions. Many trade agreements do not provide for any enforcement mechanism for competition provisions. On the contrary, they are explicitly excluded from the dispute settlement mechanism available for other obligations under the trade agreement. Therefore, the World Bank’s line of reasoning is probably not entirely correct. Competition provisions in trade agreements are rather signaling to trade partners that the country does have competition laws and authorities or will establish them and overall is willing to cooperate on behind-the-border policies. This signal is of course positive for the companies established in the countries which have the trade agreement, as it will create more trust towards their partners. But the signaling function and real enforceability are two different things. Here I am analyzing how countries could potentially use trade agreements if their companies or consumers suffer from anticompetitive conduct in the context of GVC.

10.4

HOW CAN COMPETITION PROVISIONS IN TRADE AGREEMENTS BE USED?

It is important to clarify first how even bilateral trade agreements that do not cover the whole supply chain from the beginning to the end can be used to tackle anticompetitive conduct within GVCs. Imagine that country A and B have a bilateral agreement between themselves, country A is where producers of a component who enter a cartel are located and country B is where the finished product is sold. Before the final product is sold in country B, it goes for transformation or assembling to countries C, D etc. which are not part of the trade agreement between countries A and B, nor do they have a separate trade agreement either with country A or B. What does matter is that the country where the anticompetitive conduct occurs and the country which is affected by such anticompetitive conduct have a trade agreement with competition provisions. If the producers from other countries only pass on the increased costs, but are not involved in the anticompetitive conduct themselves, they are not a real target of antitrust enforcement. Moreover, these countries are not particularly interested in enforcing competition laws against the suppliers from abroad if there are no effects on their markets. In these circumstances, the lack of a trade agreement with the countries where these in-between producers are situated does not matter for our analysis. If country A and B have a trade agreement and country B is affected, regardless of how many links there are between the cartelized component product and the finished product, the affected country B can demand action from country A to tackle the anticompetitive conduct on its territory on the basis of competition chapters in trade

254

Challenges to assumptions in competition law

agreements. Nothing in the texts of today’s trade agreements precludes the countries from doing that. Of course, a practical issue remains – in order to demand action from country A, country B should trace back the value chain and check exactly which company in which country produces inputs that are later incorporated into the finished product that are sold in country B. But demanding an action does not mean that this action will be taken. And here comes into play a question of enforceability. As was mentioned earlier, if competition chapters are specifically excluded from dispute settlement, it is problematic to enforce them. However, there is a way to increase their enforceability. The issue we have to address more generally first is how to make a country enforce something laid down in a trade agreement, especially if it is not particularly willing to do so and if there is no clear mechanism to apply in these circumstances. To make this happen, a trade partner has to have some leverage that would make the government change its attitude to the enforcement issue. Trade agreements, in contrast to agreements on cooperation in competition matters can actually provide for the leverage. This would be a recourse to trade defense measures. But how exactly would that work? Some ideas have already been introduced a decade ago or more in a very limited number of trade agreements. Unfortunately, these provisions did not become mainstream and, so far, they were not invoked by the parties. The concept seems straightforward: competition chapters are complemented with some provisions allowing to revoke trade concessions if the country fails to eliminate anticompetitive conduct on its territory which has its effects on the territory of another party to the agreement. A recourse to trade defense provisions was inserted in several bilateral Euro-Mediterranean agreements between the EU and Jordan,59 EU-Egypt60 and EU-Morocco.61 These provisions are replicated in the trade agreement between the EFTA countries (Iceland, Liechtenstein, Norway and Switzerland) and Jordan.62 The same 59 Euro-Mediterranean Agreement establishing an Association between the European Communities and their Member States, of the one part, and the Hashemite Kingdom of Jordan, of the other part http://​eeas​.europa​.eu/​archives/​delegations/​jordan/​ documents/​eu​_jordan/​eu​_jordan​_assoc​_agrmt​.pdf accessed 28 February 2020. 60 Euro-Mediterranean Agreement establishing an Association between the European Communities and their Member States, of the one part, and the Arab Republic of Egypt, of the other part https://​www​.consilium​.europa​.eu/​en/​documents​ -publications/​treaties​-agreements/​agreement/​?id​=​2001033 accessed 27 February 2020. 61 Euro-Mediterranean Agreement establishing an association between the European Communities and their Member States, of the one part, and the Kingdom of Morocco, of the other part https://​eur​-lex​.europa​.eu/​legal​-content/​EN/​ALL/​?uri​=​ CELEX:​22000A0318(01) accessed 27 February 2020. 62 Agreement between the EFTA states and the Hashemite Kingdom of Jordan https://​www​.efta​.int/​media/​documents/​legal​-texts/​free​-trade​-relations/​Jordan/​EFTA​ -Jordan​%20Free​%20Trade​%20Agreement​.pdf accessed 28 February 2020.

Using trade tools to counteract anticompetitive conduct within global value chains

255

provisions about the recourse to trade measures are also found in the EU-South Africa trade agreement.63 Regarding the EU-Jordan trade agreement, it is complemented by a special agreement on dispute resolution.64 Competition provisions are not excluded from the scope of dispute settlement. There have been no trade disputes involving competition between the EU and Jordan; therefore unfortunately we lack empirical evidence on how these provisions would work. This, however, does not preclude us from a theoretical analysis of these provisions. The agreements mentioned above all have clauses that are similar to the following: If the Community or Jordan considers that a particular practice is incompatible with the terms of paragraph 1, and is not adequately dealt with under the implementing rules referred to in paragraph 3, or in the absence of such rules, and if such practice causes or threatens to cause serious prejudice to the interest of the other party or material injury to its domestic industry, including its services industry, it may take appropriate measures after consultation within the Association Committee or after 30 working days following referral for such consultation.65

This formulation does not provide for a proper solution, only for a direction that may be considered to tackle the anticompetitive conduct. Various other questions arise. For instance, what kind of measures may the parties apply in case another party violates its obligation to eliminate the anticompetitive conduct? How could ‘an appropriate measure’ be defined? And who would evaluate whether it is appropriate, in particular if the competition chapters are explicitly excluded from the dispute settlement? These questions are quite complex and have to be addressed in a follow-up study.

63 Agreement on Trade, Development and Cooperation between the European Community and its Member States, of the one part, and the Republic of South Africa, of the other part, https://​eur​-lex​.europa​.eu/​LexUriServ/​LexUriServ​.do​?uri​=​OJ:​L:​1999:​ 311:​SOM:​EN:​HTML accessed 28 February 2020. 64 Agreement in the form of a Protocol between the European Union and the Hashemite Kingdom of Jordan establishing a dispute settlement mechanism applicable to disputes under the trade provisions of the Euro-Mediterranean Agreement establishing an Association between the European Communities and their Member States, of the one part, and the Hashemite Kingdom of Jordan, of the other part https://​eur​-lex​ .europa​.eu/​legal​-content/​EN/​TXT/​?uri​=​OJ​%3AJOL​_2011​_177​_R​_0001​_01 accessed 1 March 2020. 65 Article 53.6 of the Euro-Mediterranean Agreement establishing an Association between the European Communities and their Member States, of the one part, and the Hashemite Kingdom of Jordan, of the other part http://​eeas​.europa​.eu/​archives/​ delegations/​jordan/​documents/​eu​_jordan/​eu​_jordan​_assoc​_agrmt​.pdf accessed 1 March 2020.

256

Challenges to assumptions in competition law

Another matter that should be examined is whether trade concessions could be revoked by the parties if this possibility has not been explicitly mentioned by the trade agreement itself. This can potentially be justified by a holistic approach. If a country violates one obligation (regardless whether it is purely a trade obligation or related to behind-the border policies), another country can retaliate by using instruments provided for in the trade agreement. In any case, while considering using this mechanism, we should bear in mind that trade defense should be carefully crafted in order not to harm other producers, since there could be more GVCs than the one where the anticompetitive conduct occurs. The bottom line is that a recourse to trade measures should not harm other producers operating in the country which is planning to withdraw some trade concessions.

CONCLUSION In today’s globalized world the production process is very much interlinked. Countries do not compete with each other; instead of producing a finished product in one country, companies cooperate across the borders. Participants of GVCs need more stability in their commercial relations than producers who do not participate in GVCs. This new reality calls for reconsidering how the anticompetitive conduct with international dimension should be tackled. This chapter aimed to show how GVCs and PTAs are related and how this could potentially be used to tackle the anticompetitive conduct within GVCs. This chapter proposes as an alternative solution to look at an instrument which is not discussed enough in the literature: recourse to trade defense. This is shown as an alternative path, but is not presented as a panacea.

Index Abbott, Aiden F. 38 abuse of dominant position EU law approach 210–15, 230 fairness versus freedom of contract 207–9, 214–15, 235–7 Microsoft II (D.C. Cir.) 23, 25–6, 30, 32 Microsoft v Commission (T-201/04) 16, 26–8, 29–30, 31, 32, 33, 34–5, 35–6 objective justifications 28–32, 35–7, 232–5 in platform-to-business markets 10–11, 209 exclusionary strategies 11–16, 20–21 Google Android (AT.40099) 3–4, 7, 13–14, 19, 20 technological tying see technological tying remedies for 4, 18–20 tying and bundling 24–5 see also technological tying in United Kingdom see under United Kingdom see also dominant position AC-Treuhand AG v Commission (C-194/14 P) (ECJ) 162, 178 Achilles Information Ltd v Network Rail Infrastructure (CAT) 225, 227, 230, 235, 237 agricultural data, effect on market power see Digital Agriculture and market power Akman, Pinar 212–13, 214 Alibaba 81, 88–9, 100, 105 Amazon Inc. 9, 12–13, 16 Android antitrust decision (AT.40099) 3–4, 7, 13–14, 19, 20 anti-fragmentation agreements 4, 13–14, 19

Anti-Monopoly Bureau, China see State Administration for Market Regulation (SAMR), China antitrust abuse of dominant position see abuse of dominant position cartel facilitation enforcement see cartel facilitation enforcement in Latin America EU and US federal law compared 22–3, 207–9, 213 on technological tying see technological tying fairness versus freedom of contract 207–9, 214–15, 235–7 in global value chains see trade and competition interactions in global value chains (GVCs) Apple Inc. 10–11 Apple Inc. v Pepper et al (US Sup Ct) 7–8, 20 Apple/Shazam merger decision (COMP/M.8788) 15–16 United States v Apple Inc. (e-books case) (US 2nd Cir.) 161–2, 174 see also mobile operating systems Areeda, Phillip E. 25 Argentina, cartel facilitation enforcement 173–6, 181–2 see also cartel facilitation enforcement in Latin America Arkin v Borchard Lines Ltd (QBD) 220, 231, 237 Arriva Scotland West Ltd v Glasgow Airport Ltd (CSOH) 222–3, 225, 226–7, 231, 232–3, 236–7 asymmetries in data control 11–13 AttheRaces Ltd v British Horseracing Board (CA) 221, 236, 237 257

258

Challenges to assumptions in competition law

Austria, merger control 85, 187–8, 189–92, 203 Ayal, Idi 207, 214 Baidu 87–8 barriers to entry in Digital Agriculture sector see under Digital Agriculture and market power in platform-to-business markets 2–3, 11–16, 20–21 Barrios, Raúl 165 Bayer/Monsanto merger decisions European Commission (COMP/M.8084) 44–52, 54–5, 62, 63, 70, 72–5 MOFCOM (China) 82, 95 see also Digital Agriculture and market power Becker, Gary S. 119 Bentham, Jeremy 113 big data in agriculture, effect on market power see Digital Agriculture and market power big data management bureaus in China 92–3 biotechnology industry 14, 18 Bork, Robert 37 boundary resources in digital economy 5–7 Boyd, Robert 120 Budzinski, Oliver 21, 199 bundling 24–5 see also technological tying Busscher, Rick 178 buyout strategies 14–16 Caffarra, Cristina 39 Calderón, Andrés 180 Calor Gas Ltd v Express Fuels (Scotland) Ltd (CSOH) 236–7 cartel enforcement in global value chains see trade and competition interactions in global value chains (GVCs) cartel facilitation enforcement in Latin America 161–3, 176–80 Argentina 173–6, 181–2

Chile 166–70, 181–2 Colombia 170–73, 181–2 Peru 163–6, 181–2 Chile, cartel facilitation enforcement 166–70, 181–2 see also cartel facilitation enforcement in Latin America China’s digital economy, merger control 80–83, 104–5 notification of concentrations transaction-value-based threshold for 85–6 turnover-based threshold, limitations of 84–5 VIE (variable interest entity) structure 82, 83–4, 103 remedies 98–100 SAMR enforcement action decisions 100–104 substantive competition assessment data openness 86–9 innovation, competition in 93–6 leveraging effect 96–7 privacy 89–93 Claritas (UK) Ltd v Post Office and Postal Preference Service Ltd (Ch. D) 218–19 Coca-Cola/Huiyuan merger decision (MOFCOM) 96–7 collective choice theories of institutional change 116 collusion with cartels see cartel facilitation enforcement in Latin America Colombia, cartel facilitation enforcement 170–73, 181–2 see also cartel facilitation enforcement in Latin America Competition Act 1998 Chapter II prohibition see United Kingdom, abuse of dominant position enforcement competition chapters in trade agreements see trade and competition interactions in global value chains (GVCs)

Index

competition tribunals in Latin America see transplants of competition tribunals to Latin America cooperation in platform-to-business markets see under platform-to-business markets Cotterrell, Roger 112, 113 culture concept 118–20 Cyber Security Law 2016 (China) 91, 92 Danaher Corp/General Electric Co. merger decision (MOFCOM) 95–6 data control asymmetries in digital economy 11–13, 15–16, 17 data-driven market power see Digital Agriculture and market power data openness in China 86–9 De Jong, Martin 108 developing countries cartel facilitators in see cartel facilitation enforcement in Latin America China, merger control see China’s digital economy, merger control limited influence at ICN 137–8, 140–42, 146–7, 151–3, 155–6, 159–60 see also International Competition Network (ICN) transplants of competition tribunals see transplants of competition tribunals to Latin America Dianping 87–8 Didi/Uber China merger (2018) 81, 82, 85, 101–2 Digital Agriculture and market power 53–5 ag-data, concept and role 41–4, 56–7, 66 Bayer/Monsanto merger decisions European Commission (COMP/M.8084) 44–52, 54–5, 62, 63, 70, 72–5 MOFCOM (China) 82, 95 switching costs barrier to entry 56–62, 76–8

259

business model and distance to data effect 65–8, 75, 78 data combination, importance of 64, 75 data substitutability, lack of 49, 52, 68–70, 75–6 economies of scale variations 71–2, 76 historical data sets, importance of 63, 74–5 positive feedback loops instead of direct network effects 72–3, 74 rivalrous nature of data in practice 70–71, 76 ‘Smart Farming’ costs 64–5, 75 digital economy agricultural data, effect on market power see Digital Agriculture and market power data control asymmetries in 11–13, 15–16, 17 platform-to-business markets see platform-to-business markets technological tying see technological tying dispute settlement under trade agreements 250–51 disruptive innovations 14 dominant position abuse of see abuse of dominant position market power 21, 53–5 agricultural data, effect on see Digital Agriculture and market power leverage of 12, 36, 96–7 in platform-to-business markets 2–3, 17, 21 Dow/Du Pont merger decision (MOFCOM) 94–5 Drahos, Michaela 147–8 Du Pont/Dow merger decision (MOFCOM) 94–5 Duoshan 88 Duxbury, Neil 207 e-books case (United States v Apple Inc. (US 2nd Cir.)) 161–2, 174

260

Challenges to assumptions in competition law

e-commerce livestreaming in China 80–81 economic analysis market economics approach 207–9, 214–15, 235–7 ordoliberal approach 212–13, 214 of tying and bundling practices 37–9 effective competition standard 21 effectiveness principle of EU law 31–2 emerging economies cartel facilitators in see cartel facilitation enforcement in Latin America China, merger control see China’s digital economy, merger control limited influence at ICN 137–8, 140–42, 146–7, 151–3, 155–6, 159–60 see also International Competition Network (ICN) transplants of competition tribunals see transplants of competition tribunals to Latin America enculturation 119 entry barriers in Digital Agriculture sector see under Digital Agriculture and market power in platform-to-business markets 2–3, 11–16, 20–21 epistemic communities, influence at ICN 147–51 see also International Competition Network (ICN) essential facilities doctrine 218, 222, 226, 228–9 EU competition law and policy abuse of dominant position 210–15, 230 Google Android (AT.40099) 3–4, 7, 13–14, 19, 20 Microsoft v Commission (T-201/04) 16, 26–8, 29–30, 31, 32, 33, 34–5, 35–6 UK application of EU law see under United Kingdom

cartel facilitation enforcement 162, 176, 178 China-EU Competition Policy Week 104 competition chapters in trade agreements 249, 252, 254–5 effectiveness principle 31–2 EU influence at ICN 133, 135, 138, 139–41, 148–9, 153–4, 159–60 see also International Competition Network (ICN) goals of 134–5 merger control see public interest considerations in European merger control proportionality principle 28–9, 32 US federal antitrust law compared 22–3, 207–9, 213 on technological tying see technological tying evolutionary psychology and anthropology 119 evolutionary theories of institutional change 116–17 ex post merger control 18 export cartels 246–8, 249 Facebook 10, 14, 55 facilitators of cartels see cartel facilitation enforcement in Latin America fairness considerations in antitrust 207–9, 214–15, 235–7 in merger control 23 farm data, effect on market power see Digital Agriculture and market power Finckenstein, Konrad von 131 Fingleton, John 136 Finland, merger control 188 Fox, Eleanor M. 152, 155, 179 France, merger control 18, 188, 192–4, 203 free trade agreements see trade and competition interactions in global value chains (GVCs) freedom of contract approach to antitrust 207–9, 214–15, 235–7

Index

Gal, Michal 213–14 Gawer, Annabelle 3 General Electric Co./Danaher Corp merger decision (MOFCOM) 95–6 Gerber, David 135 Germany Facebook antitrust decision 55 merger control 17, 85, 188, 197–202, 203 Getmapping Plc v Ordnance Survey (Patents Ct) 232–3 Giocoli, Nicola 208 global value chains see trade and competition interactions in global value chains (GVCs) Google Inc. 14 Google Android (AT.40099) 3–4, 7, 13–14, 19, 20 Google Shopping case 12, 22–3, 211 Streetmap.EU v Google Inc. (Ch. D) 23, 28, 223–4, 227, 231, 234 see also mobile operating systems Guarín Lemeses, Rodrigo 172 Hale, Thomas 132 Hamilton, Alexander 109–10 Hayek, Friedrich 119 Healy, Deborah 179 Henderson, Rebecca 3 Hendry v World Professional Billiards and Snooker Association Ltd (Ch. D) 220, 235 Herz, Martin 178 Hippel, Eric von 6 Hodgson, Geoffrey 108–9 Hoffmann-La Roche & Co v Commission (85/76) (ECJ) 230, 232 Horn, Henrik 242 Hovenkamp, Herbert 11, 25 Huawei Technologies Co. 86–7 Huiyuan/Coca-Cola merger decision (MOFCOM) 96–7 Hungary, merger control 188 Ibanez-Colomo, Pablo 210–11 IBM Corp 25, 32, 36

261

ICN see International Competition Network (ICN) ICPAC (International Competition Policy Advisory Committee) 125, 133–4, 145 import cartels 245–6 industrial economics see economic analysis Ineos Vinyls Ltd v Huntsman Petrochemicals (UK) Ltd (Ch. D) 220, 236 information asymmetries in digital economy 11–13, 15–16, 17 innovation China, competition in innovation 93–6 predatory or disruptive 12, 14 Instagram 10, 14 institutions concept 108–9 institutional change theories 115–18 judiciary as formal institution 109–12 see also transplants of competition tribunals to Latin America Intel Corp 3, 211, 235–6 international cartels 248 International Competition Network (ICN) 104, 124–7 ICN mission, structure and activities 127–9 influence of actors 129–30, 159–60 national competition authorities 135–42, 151–4 non-governmental actors 142–7, 154–9 states 130–35 influence, mechanisms of 147 early stage participation 156–8 expert opinions 147–51 offers of technical assistance 153–4 political and economic power 151–3 revolving door policies 158–9 International Competition Policy Advisory Committee (ICPAC) 125, 133–4, 145 international trade agreements see trade and competition interactions in global value chains (GVCs)

262

Challenges to assumptions in competition law

interoperability Digital Agriculture sector, barriers to 61–2, 77 in platform-to-business markets 4, 6, 9, 12, 13 Ireland, merger control 188 Italy, merger control 188 Jenny, Frederic 146 John Deere Co. 49, 60–61 Jones, Alison 213 judiciary as formal institution 109–12 Katalevsky, Dmitry 61 Katz, Ralph 6 Kennedy, David 148 kill zones 2, 14–15, 16 Knight, Jack 108 Kühn, Kai-Uwe 39 Lafontaine, Francine 38 laissez-faire approach to antitrust 207–9, 214–15, 235–7 Lall, Ranjit 156, 157, 159 Lassal ruling (C-162/09) 31–2 Latin America cartel facilitation enforcement see cartel facilitation enforcement in Latin America transplants of competition tribunals see transplants of competition tribunals to Latin America Lauslahti, Kristian 6–7 Leeds City Council v Watkins & Whiteley (Ch. D) 220 legal culture concept 118–20 legal transplants see under transplants of competition tribunals to Latin America Legrand, Pierre 113 leverage of market power 12, 36, 96–7 Li Keqiang 94 Lianos, Ioannis 61, 212 Liu Qihong 3, 13, 16 Lochner v New York (US Sup Ct) 207–8 lock-ins in Digital Agriculture sector see under Digital Agriculture and market power Lovdahl Gormsen, Liza 213

Maimai 87 Marcos, Francisco 178–9 market economics approach to antitrust 207–9, 214–15, 235–7 market failure theories 184 market power 21, 53–5 agricultural data, effect on see Digital Agriculture and market power leverage of 12, 36, 96–7 see also dominant position Mattila, Juri 6–7 Mavroidis, Petros C. 242 McBarnet, Doreen 208 media plurality concerns in merger control 187–90, 191–2 merger control in China see China’s digital economy, merger control in Digital Agriculture sector see Digital Agriculture and market power in Europe, public interest considerations see public interest considerations in European merger control ex post control 18 fairness considerations 23 in platform-to-business markets 17–18, 68–9 remedies 4, 98–100 takeovers as exclusionary strategies 14–16 in United States 23 Mexico, transplants of competition tribunal system see transplants of competition tribunals to Latin America Microsoft Corp exclusionary strategies 10–12, 15 Microsoft II (D.C. Cir.) 23, 25–6, 30, 32 Microsoft v Commission (T-201/04) 16, 26–8, 29–30, 31, 32, 33, 34–5, 35–6 Ministererlaubnis system (ministerial approval of mergers in Germany) 197–202 mobile operating systems

Index

in China see China’s digital economy, merger control as cooperative ecosystems 5–8 economic dependency concern 9–11 exclusionary strategies by pivotal firms 11–16, 20–21 implications for competition enforcement 16–20, 21 technological tying see technological tying Monsanto/Bayer merger decision European Commission (COMP/M.8084) 44–52, 54–5, 62, 63, 70, 72–5 MOFCOM (China) 82, 95 see also Digital Agriculture and market power Montesquieu 113 Moore, James F. 5 multinational companies, influence at ICN 142–7, 154–9 see also International Competition Network (ICN) national competition authorities influence at ICN 135–42, 151–4 see also International Competition Network (ICN) SAMR see State Administration for Market Regulation (SAMR), China native apps 20 Netherlands, merger control 188 Network Multimedia Television Ltd v Jobserve Ltd (CA) 219 New Institutional Economics 107, 117–18 Newheight Holding/Walmart merger decision (MOFCOM) 97 Nokia Corp 14 Nölke, Andreas 148 non-governmental actors, influence at ICN 142–7, 154–9 see also International Competition Network (ICN)

263

non-market considerations in European merger control see public interest considerations in European merger control North, Douglass 108, 117, 119 nowcasting 2, 12 objective justifications for abuse of dominance 28–32, 35–7, 232–5 OECD (Organisation for Economic Co-operation and Development) 125, 127, 130, 134, 154, 249–51 Olson, Mancur 139 ordoliberalism 212–13, 214 Organisation for Economic Co-operation and Development (OECD) 125, 127, 130, 134, 154, 249–51 Ortega, Julia 179 Ossege, Christoph 133 Pérez Rojas, Javier 172 Perry, James 148 personal data protection in Chinese digital economy 89–93 in Digital Agriculture sector 58, 60 Peru, cartel facilitation enforcement 163–6, 181–2 see also cartel facilitation enforcement in Latin America Picasa 14 platform envelopment strategies 12 platform-to-business markets abuse of dominant position in 10–11, 209 exclusionary strategies 11–16, 20–21 Google Android (AT.40099) 3–4, 7, 13–14, 19, 20 technological tying see technological tying cooperation in 2–8 economic dependency concern 9–11 exclusionary strategies by pivotal firms 11–16, 20–21 implications for competition enforcement 16–20, 21

264

Challenges to assumptions in competition law

dominant position in 2–3, 17, 21 merger control 17–18, 68–9 in China see China’s digital economy, merger control Portugal, merger control 188 Posner, Eric 8, 11 Posner, Richard 208–9 predatory innovations 12 predatory pricing 47, 230 preferential trade agreements (PTAs) and GVCs see trade and competition interactions in global value chains (GVCs) privacy protection in Chinese digital economy 89–93 in Digital Agriculture sector 58, 60 progressive web apps 20 proportionality principle of EU law 28–9, 32 PTAs (preferential trade agreements) and GVCs see trade and competition interactions in global value chains (GVCs) public interest considerations in European merger control 184–7, 202–5 Austria 187–8, 189–92, 203 France 188, 192–4, 203 Germany 188, 197–202, 203 United Kingdom 189, 195–7, 203 Purple Parking Ltd v Heathrow Airport Ltd (Ch. D) 214, 221–2, 223, 226, 229, 231, 233–4, 237 Radaelli, Claudio 129 refusal to licence or supply 218, 221, 235 Regulation 1/2003/EC 30 remedies 4, 18–20, 98–100 Richerson, Peter 120 Roland, Gerard 117–18 Rose, Richard 122 SAMR see State Administration for Market Regulation (SAMR), China Sapir, Andre 242 Schmidt, Vivien 129 Schrepel, Thibault 12

Schweitzer, Heike 213 Seppälä, Timo 6–7 Shazam/Apple merger decision (COMP/M.8788) 15–16 Sina Microblog 87 Slade, Margaret E. 38 Slaughter, Anne-Marie 132 ‘Smart Farming’ 54, 64–5, 75 see also Digital Agriculture and market power Socrates Training Ltd v Law Society of England and Wales (CAT) 224, 227–8, 230, 234–5, 237 South America cartel facilitation enforcement see cartel facilitation enforcement in Latin America transplants of competition tribunals see transplants of competition tribunals to Latin America South Korea, merger control 85 Spain, merger control 189 specialised competition tribunals in Latin America see transplants of competition tribunals to Latin America Standard Social Science Model 119 start-up acquisitions in digital economy 10, 14, 15 State Administration for Market Regulation (SAMR), China 82–3 Anti-Monopoly Law, draft amendments 85–6, 94 enforcement activity 81, 84, 89, 95–6, 105 prospects of 93, 100–104 personal data protection measures 92 see also China’s digital economy, merger control states, influence at ICN 130–35 see also International Competition Network (ICN) Steinbaum, Marshall 21 Stigler, George 143 Stillman, Robert 39 Stöhr, Annika 21, 199 Streetmap.EU v Google Inc. (Ch. D) 23, 28, 223–4, 227, 231, 234 Stucke, Maurice 21, 207, 227

Index

Sundmaeker, Harald 61 Sunstein, Cass 209 Suretrack Rail Services Ltd v Infraco JNP Ltd (Ch. D) 219–20 Sweden, merger control 18 switching costs in Digital Agriculture sector see under Digital Agriculture and market power takeovers as exclusionary strategies 14–16 technological tying 22–3, 39–40 economic analysis 37–9 by Google Inc. 13, 22, 24, 33–7, 223–4, 227, 231, 234 new products, levels of deference to 25–8 objective justifications 28–32, 35–7, 234 tying and bundling, general legal position 24–5 see also abuse of dominant position Tencent 81, 86–7, 88, 105 TFEU Article 102 prohibition see EU competition law and policy, abuse of dominant position TikTok 88 Toshiba ruling (C-17/10) 31 trade and competition interactions in global value chains (GVCs) 238–40 competition provisions in trade agreements enforcement of 253–6 typology 249–53 preferential trade agreements, relationship with GVCs 240–44 theories of harm 244–9 transaction-value-based merger notification threshold 85–6 transplants of competition tribunals to Latin America 106–8, 122–3 conceptual framework culture 118–20 institutions 108–9 judiciary as formal institution 109–12 design of institutional transplants 120–22

265

theoretical framework 112–15 institutional change theories 115–18 typology of transplant processes 114 Treaty on the Functioning of the European Union Article 102 prohibition see EU competition law and policy, abuse of dominant position turnover-based merger notification threshold 84–5 two-sided platforms see platform-to-business markets tying 24–5 see also technological tying Uber China/Didi merger (2018) 81, 82, 85, 101–2 UNCTAD (United Nations Conference on Trade and Development) 125, 127, 134, 154 unfair competition disputes in China 86–8 United Kingdom abuse of dominant position enforcement 206–10, 237 case law overview 215–25 competition test 226–8 EU law approach 210–15 fairness versus freedom of contract 207–9, 214–15, 235–7 objective/commercial justifications 232–5 objective/subjective divide 230–31 pigeon-holing of conduct 228–30 cartel facilitation enforcement 162 merger control 18, 189, 195–7, 203 United Nations Conference on Trade and Development (UNCTAD) 125, 127, 134, 154 United States cartel facilitation enforcement 161–2, 176 competition chapters in trade agreements 252

266

Challenges to assumptions in competition law

federal and EU antitrust law compared 22–3, 207–9, 213 on technological tying see technological tying influence at ICN 139–41, 145, 148–9, 153–4, 159–60 see also International Competition Network (ICN) merger control 23 United States v Apple Inc. (e-books case) (US 2nd Cir.) 161–2, 174 Vaheesan, Sandeep 8, 10–11 variable interest entities (VIEs) in China 82, 83–4, 103 Vedder, Hans 178 Vestager, Margrethe 22–3 Vezzoso, Simonetta 5 VIEs (variable interest entities) in China 82, 83–4, 103 Waarden, Frans von 147–8 Walmart/Newheight Holding merger decision (MOFCOM) 97 Watson, Alan 113 Weber, Max 133

Wen Wen 6 Whelan, Christopher 208 Whinston, Michael 38 Wilks, Stephen 129, 144, 147, 154 Williamson, Oliver E. 117 Wils, Wouter 146, 211 Wireless Group Plc v Radio Joint Audience Research Ltd (Ch. D) 228, 232–3 World Bank analysis of competition chapters in trade agreements 240, 241, 242, 243, 251, 252–3 World Trade Organization (WTO) 154 analysis of competition chapters in trade agreements 251–2 Working Group on Interaction between Trade and Competition 244–5 WTO+ and WTO-X provisions 242–3 see also trade and competition interactions in global value chains (GVCs) Wright, Joshua D. 38 Wu Zhenguo 101–2 Zhu Feng 3, 6, 13, 16 Zingales, Luigi 142–3