Research Handbook on Methods and Models of Competition Law 2020944596, 9781785368653, 9781785368646

Citation preview

RESEARCH HANDBOOK ON METHODS AND MODELS OF COMPETITION LAW

RESEARCH HANDBOOKS IN COMPETITION LAW This highly-topical series addresses some of the most important questions and areas of research in competition law and antitrust. Each volume is designed by a leading expert to appraise the current state of thinking and probe the key questions for future research on a particular topic. The series encompasses some of the most pressing issues as well as the foundational pillars of the field, including: merger control, competition damages, abuse of dominance and cartels, amongst others. Each Research Handbook comprises specially-commissioned chapters from leading academics, and practitioners, as well as those with an emerging reputation and is written with a global readership in mind. Equally useful as reference tools or high-level introductions to specific topics, issues and debates, these Research Handbooks will be used by academic researchers, postgraduate students, practising lawyers, competition authority officials and policy makers. Titles in the series include: Research Handbook on Asian Competition Law Edited by Steven Van Uytsel, Shuya Hayashi and John O. Haley Research Handbook on Methods and Models of Competition Law Edited by Deborah Healey, Michael Jacobs and Rhonda L. Smith

Research Handbook on Methods and Models of Competition Law Edited by

Deborah Healey Professor, Faculty of Law, UNSW, Australia

Michael Jacobs Emeritus Distinguished Research Professor of Law, DePaul University, Chicago, USA

Rhonda L. Smith Senior Lecturer, Department of Economics, University of Melbourne, Australia

RESEARCH HANDBOOKS IN COMPETITION LAW

Cheltenham, UK • Northampton, MA, USA

© The Editors and Contributors Severally 2020

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: 2020944596 This book is available electronically in the Law subject collection http://dx.doi.org/10.4337/9781785368653

03

ISBN 978 1 78536 864 6 (cased) ISBN 978 1 78536 865 3 (eBook)

Contents

List of contributorsvii Acknowledgementsx Table of casesxi Table of legislationxxiii PART I

THE OBJECTS AND ECONOMICS OF COMPETITION LAW

1

Competition law in flux: established and emerging approaches to methodology Deborah Healey and Rhonda L. Smith

2

The ambit of competition law: comments on its goals Deborah Healey

12

3

The cost of progress: hurdles facing antitrust’s economic advance Alan Devlin

38

4

The relevance of economics in US, EU and Australian competition law Geoff Edwards and Jennifer Fish

57

5

The use of economics in competition law enforcement in mainland China and Hong Kong Lin Ping and Yan Yu

PART II

2

89

THE CONTENT OF THE LAW

6

Cartel prohibition and the search for deterrent penalties: the United States, the European Union, Australia and China compared Mark Williams

7

Algorithm-driven collusive conduct Rob Nicholls

8

Vertical agreements under EU competition law: proposals for pushing Article 101 analysis, and the modernization process, to a logical conclusion Miguel de la Mano and Alison Jones

9

Unilateral conduct analysis: focus on harm in multiple guises Rhonda L. Smith and Deborah Healey

116 138

167 204

10 Mergers Rhonda L. Smith

228

11

253

Competition law in Japan, Malaysia and the Philippines: an overview Mel Marquis v

vi  Research handbook on methods and models of competition law 12

Building an efficient system of protection of competition in Serbia on its path to the EU Dragan Penezic and Zoran Soljaga

13

Merger review updates in Latin America Fernando M. Furlan

274 294

PART III PARTICULAR ISSUES 14

The interface between intellectual property rights and competition law: implications for public health in sub-Saharan Africa Mor Bakhoum

312

15

Pay for delay in perspective: the impact of adversarial and inquisitorial legalism on pharmaceutical antitrust enforcement Sven Gallasch

336

16

The Australian approach to third party infrastructure access under Part IIIA of the Competition and Consumer Act 2010 Alice Muhlebach

365

PART IV ENFORCEMENT 17

The EU method of antitrust enforcement Andreas Stephan

391

18

Cartel enforcement: critical reflections from the South African experience Simon Roberts

414

19

Procedure and substance in China’s merger control regime Wang Xiaoye and Adrian Emch

436

PART V

COMPETITION POLICY AND OTHER ISSUES

20

An effective way to keep power in an institutional cage: legislation and regulation of administrative monopoly Xu Shiying

21

Competition advocacy: a broader perspective Wendy Ng and Allan Fels

22

Export cartels in times of populist protectionism: challenges and options for young and small competition agencies Pierre M. Horna and Leni Papa

23

Complementarities and tensions between competition and trade law and policy 507 Hassan Qaqaya

452 471

489

Index527

Contributors

Mor Bakhoum is an assistant professor at the Virtual University of Senegal (UVS), an affiliated research fellow at the Max Planck Institute for Innovation and Competition in Munich, and a faculty member at the Munich Intellectual Property Law Center (MIPLC). Miguel de la Mano is Executive Vice President in the Brussels office of Compass Lexicon and formerly Head of Economic Analysis and Evaluation at DG Internal Market. Alan Devlin is a partner in the Washington, DC, office of Latham & Watkins LLP, and an Adjunct Professor at Georgetown Law. Before rejoining the firm, he was Acting Deputy Director of the FTC’s Bureau of Competition. Geoff Edwards is Vice President of Charles River Associates. He is an economist with extensive experience advising in competition and regulatory proceedings internationally. Adrian Emch is a partner at Hogan Lovells in Beijing. He is also an honorary senior fellow at the University of Melbourne, where he teaches Chinese competition law and policy. Allan Fels is a Professorial Fellow at the University of Melbourne and was formerly the Chair of the Australian Competition and Consumer Commission. Jennifer Fish is a Principal at Charles River Associates. Fernando M. Furlan is a Professor of Economic and Corporate Law at Uniceplac Law and Business Schools in Brasilia. He was formerly President of the Brazilian Competition Authority (CADE), a CADE Commissioner and CADE General Counsel. He was formerly deputy minister for Industry, Foreign Trade and Services for Brazil. Sven Gallasch is a Lecturer at Swinburne Law School, Melbourne. Deborah Healey is Professor and Director of the Herbert Smith Freehills China International Business and Economic Law Centre at UNSW Law, Sydney. Pierre M. Horna is Legal Affairs Officer, Competition and Consumer Policies Branch, Division on International Trade, UNCTAD Secretariat. Michael Jacobs is Emeritus Distinguished Research Professor of Law at the DePaul University College of Law.

vii

viii  Research handbook on methods and models of competition law Alison Jones is a Professor at the Dickson Poon Law School of Law, University College London. Lin Ping is a Professor in the School of Economics, Shandong University, and a Professor in the Department of Economics and Director of the Centre for Competition Policy and Regulation at Lignan University, Hong Kong. Mel Marquis teaches at Monash Law School in Melbourne and (in 2020) Doshisha University in Kyoto. From 2011 to 2019 he was Part-time Professor of Law at the EUI in Florence. Alice Muhlebach is a barrister in Melbourne, Australia. Wendy Ng is a Senior Lecturer at Melbourne Law School, where she is Director of the Competition Law and Economics Network, and an Associate Director (China) of the Asian Law Centre. Rob Nicholls is an Associate Professor in Business Law and the Director of the UNSW Business School Cybersecurity and Data Governance Research Network. Leni Papa is a Consultant, Competition Law and Policy, to the Philippine Competition Commission. Dragan Penezic is Head of Regulatory and External Affairs for Serbia and Montenegro for British American Tobacco (BAT). He was previously Secretary General of the Commission for the Protection of Competition of the Republic of Serbia. Hassan Qaqaya is a Senior Fellow in the Global Competition and Consumer Law Program at Melbourne Law School. He was formerly Head of the UNCTAD Competition and Consumer Policies Branch, Switzerland. Simon Roberts is a Professor of Economics at the University of Johannesburg. Rhonda L. Smith is a Senior Lecturer in the Economics Department at the University of Melbourne. She is a former Commissioner of the Australian Competition and Consumer Commission. Zoran Soljaga is Attorney at Law at Moravčević, Vojnović and Partners in cooperation with Schoenherr. He was previously a Senior Legal Advisor at the Commission for the Protection of Competition in Serbia. Andreas Stephan is Professor of Competition Law and Head of School, School of Law and Centre for Competition Policy, University of East Anglia.

Contributors  ix Mark Williams is the Executive Director of the Asian Competition Forum, founded in 2005. He was formerly Professor of Law at Hong Kong Polytechnic University and at the University of Melbourne Law School. Wang Xiaoye is Distinguished Professor at Shenzhen University, and Professor of Law at the Chinese Academy of Social Sciences. She was advisor for the anti-monopoly legislation in the National People’s Congress of China. Xu Shiying is a Professor and Director of the Competition Law Institute at the East China University of Political Science and Law (ECUPL). She is a member and the standing director of the National Economic Law Research Association of the China Law Society, and a former chairman of the Asia Competition Association. Yan Yu is a partner of RBB Economics based in Melbourne.

Acknowledgements

The editors thank the Centre for Law Markets and Regulation at the University of NSW for its ongoing support for this project.

x

Table of cases

AUSTRALIA ACCC v Flight Centre Travel Group Ltd (2016) 261 CLR 203; [2016] HCA 49  164 ACCC v Metcash Trading [2011] FCA 967; [2001] FCAFC 151  242 ACCC v Pfizer Australia Pty Ltd [2015] FCA 113; (2015) 323 ALR 429  84–5, 215–16 ACCC v Pfizer Australia Pty Ltd [2018] FCAFC 78  85 ACCC v Visy Industry Holding Pte Ltd (No 3) [2007] FCA 1617  133 ACCC v Yazaki Corporation [2018] FCAFC 73  135 Betfair v Racing New South Wales (2012) 249 CLR 217  482–3 Betfair v Western Australia (2008) 234 CLR 418  480 BHP Billiton Iron Ore Pty Ltd v National Competition Council (2008) 236 CLR 145.  380 Boral Besser Masonry Ltd v Australian Competition and Consumer Commission (2003) 215 CLR 374 86 Castlemaine Tooheys Ltd v South Australia [1986] HCA 72; (1990) 169 CLR 436  482–3 Cole v Whitfield (1988) 165 CLR 360  480, 482 Commonwealth Director of Public Prosecutions v Nippon Yusen Kabushiki Kaisha [2017] FCA 876 133 Commonwealth v Bank of New South Wales [1949] UKPC 37; [1950] AC 235; (1949) 79 CLR 497 482 Duke Eastern Gas Pipeline Pty Ltd [2001] ACompT 2; (2001) 162 FLR 1  384 Finemores Transport Pty Ltd v New South Wales [1978] HCA 16; (1978) 139 CLR 338  482 Fortescue Metals Group Ltd, Re [2010] ACompT 2; (2010) 271 ALR 256  366–7 Glencore Coal Pty Ltd, Application by (No 2) (2016) 309 FLR 358  377 Glencore Coal Pty Ltd, Application by [2016] ACompT 6  376–7, 381–3, 387–8 Grannall v Marrickville Margarine Pty Ltd [1955] HCA 5; (1955) 93 CLR 55  482 Hamersley Iron Pty Ltd v National Competition Council [1999] FCA 867; (1999) 164 ALR 203 377, 380 Melway Publishing Pty Ltd v Robert Hicks Pty Ltd [2001] HCA 13; (2001) 205 CLR 1  86–7, 368 NT Power Generation Pty Ltd v Power and Water Authority [2004] HCA 48; (2004) 219 CLR 90. 368 Pilbara Infrastructure Pty Ltd v Australian Competition Tribunal [2012] HCA 36; (2011) 193 FCR 57; (2012) 246 CLR  380–82, 384–8 Port of Newcastle Operations Pty Ltd v Australian Competition Tribunal [2017] FCAFC 124; (2017) FCR 115; 346 ALR 669; (2012) 246 CLR  377, 383 Port of Newcastle Operations Pty Ltd v The Australian Competition Tribunal [2018] HCATrans 55 (23 March 2018)  377–8 Queensland Co-operative Milling Association Ltd (1976) 8 ALR 481; (1976) ATPR 40-012  238 Queensland Wire Industries Pty Ltd v Broken Hill Proprietary Co Ltd [1987] FC A 496; (1987) 17 FCR 211; (1987) 78 ALR 407  365, 368 Queensland Wire Industries Pty Ltd v Broken Hill Pty Co Ltd [1989] HCA 6; (1989) 167 CLR 177; (1989) 83 ALR 577  86–7, 368 Radio 2UE Sydney Pty Ltd v Stereo FM Pty Ltd [1982] FCA 223; (1982) 44 ALR 557; (1982) 62 FLR 437  242 Rail Access Corporation v NSW Minerals Council Ltd [1998] FCA 1266; 158 ALR 323; (1998) 87 FCR 517  378 Rural Press Ltd v Australian Competition and Consumer Commission [2003] HCA 75; (2003) 216 CLR 53  87 7-Eleven Stores, Re (1994) ATPR 41-357  238 Services Sydney Pty Ltd, Application by [2005] ACompT 7; (2005) 227 ALR 140  377 Sydney Airport Corporation Ltd v Australian Competition Tribunal (2006) 155 FCR 124 376–7, 381, 383–4 Sydney International Airport [2000] ACompT 1  381–2, 384, 388

xi

xii  Research handbook on methods and models of competition law Trade Practices Commission v Australian Meat Holdings Pty Ltd [1988] FCA 338; (1988) ATPR 40-876; 83 ALR 299  233 Virgin Blue Airlines Pty Ltd, Re [2005] ACompT 5; (2005) 195 FLR 242; (2006) ATPR 42-092 376,  381, 387–8

BRAZIL AB Inbev (Ambev) case (Administrative Process No 08012.010028/2009-74)  Oxygen and Cement Cartel cases (Administrative Processes No 08012.009888/2003-70 08012.011142/2006-79)  Petrobrás SA case (Administrative Process No 08700.002600/2014-30); 

306 and 306 306

CHILE Falabella/D&S, Resolution No 24 – Case NC 199-07  Quiñenco/Terpel Chile, Resolution No 39 – Case NC 399-11 

301–2 302

CHINA Abbott Laboratories/St Jude Medical (MOFCOM Public Announcement No 88 of 2016)  444 Agrium/Potash Corp of Saskatchewan (MOFCOM Notice No 75 of 2017)  108 Anheuser-Busch InBev/SABMiller (MOFCOM Public Announcement No 38 of 2016)  96, 444 ASIMCO Advanced Semiconductor Engineering/Siliconware Precision Industries (MOFCOM Notice No 81 of 2017)  109 Baxter/Gambro (MOFCOM Notice No 58 of 2013)  97–8 Bayer/Monsanto (MOFCOM Public Announcement No 31 of 2018)  109, 113, 447 Becton Dicskinson/US Bud Company (MOFCOM Notice No 92 of 2017)  109 Beijing Municipal Commission of Housing and Urban-Rural Development (concrete industry, abuse of administrative power restricting competition) 29 December 2016  456–7, 460–61 Beijing Municipal Commission of Housing and Urban-Rural Development (concrete industry, abuse of administrative power restricting competition) Notice on quality control price, Jing Jian Fa [2014] No 24, 24 December 2014  460–61 Beijing Municipal Commission of Transport (online taxi business service management system) Fair Competition Review Jing Jiao Wen [2016] No 216, 21 December 2016  468–70 Beijing Ruibang Yonghe Science & Commerce Co Ltd v Johnson & Johnson (Shanghai) Medical Equipment Co Ltd, Civil Judgment of Shanghai High People’s Court, 2012 Hu Gao Min San Zhong No 63, entered 1 August 2013  125–6 Bengbu Health and Family Planning Commission (abuse of administrative power restricting competition), Bureau of Price Supervision of National Development and Reform Commission, Order No [2015] 2175, 17 August 2015  456–7, 459–60 Broadcom/Brocade Communications Systems (MOFCOM Notice No 46 of 2017)  112 Coca-Cola/Huiyuan (MOFCOM Public Announcement No 22 of 2009)  96, 448–9 Essilor/Luxottica (SAIC Notice of 25 July 2018)  99, 111, 113–14, 449 Gansu Province Road Transport Authority (abuse of administrative power restricting competition), Gansu Province Development and Reform Commission, Order No [2015] 7, 1 December 2015 456–7 GE China/China Shenhua Coal (MOFCOM Notice No 74 of 2011)  95–6 GE/Shenhua (remedy lifting decision) (SAMR Public Announcement No 21 of 2018)  444 General Motors/Delphi (MOFCOM Public Announcement No 76 of 200)  448 Glencore/Xstrata (MOFCOM Public Announcement No 20 of 2013)  444 Google/Motorola Mobility (2012 MOFCOM Public Announcement No 25 of 2012)  448

Table of cases  xiii Guangdong Education Department and Glodon Software Company Limited (abuse of administrative power restricting competition), Civil Administrative Appeal Judgment Guangdong High People’s Court, 228 (2015) Administrative Judgment Higher Court Ultimate No 228, 28 June 2017  456–7,  461–2 Guangdong Education Department and Glodon Software Company Limited (abuse of administrative power restricting competition) Sui Zhong Fa Xing Chu Zi, No 149, the Administrative Litigation Judgment, Guangzhou Intermediate People’s Court  462 Hainan Provincial Price Bureau v Hainan Yutai Scientific Feed Company (December 2018; published 24 June 2019).  92 Hebei Provincial People’s Government Department of Transportation (abuse of administrative power restricting competition) 26 September 2014  456–9 HP/Samsung (MOFCOM Notice No 58 of 2017)  108, 112 Huawei v IDC, Guangdong People’s Court Decision 305 (18 April 2014)  94 KLA/Orbotech (SAMR Public Announcement No 7 of 2019)  448 Linde/Praxair [2018] SAMR Public Announcement, 30 September 2018.  444 Maersk Line/Hamburg South American Shipping (MOFCOM Notice No 77 of 2017)  108 Merck KGaA/AZ Electronic Materials (MOFCOM Public Announcement No 30 of 2014)  449 Mitsubishi Rayon/Lucite International (MOFCOM Public Announcement No 28 of 2009)  444, 448 Nanjing Faershi New Energy Co, Ltd v People’s Government of Jiangning District of Nanjing City, Intermediate People’s Court of Nanjing City, Jiangsu Province (2015) Ning Xing Chu Zi No 16, 18 June 2015  456–7, 462–3 NDRC Administrative Penalty Decision [2015] No 1  214 NDRC Administrative Penalty Decision [2017] No 2  214 Novartis/Alcon (MOFCOM Public Announcement No 53 of 2010)  447 NXP/Freescale (MOFCOM Public Announcement No 64 of 2015)  99, 444 Primearth EV Energy (PEVE)/Toyota China/Keliyuan (MOFCOM Notice of 2 July 2014)  98 Qihoo 360 v Tencent (2013)  100–101, 223–4 Qualcomm case Administrative Penalty Decision of 10 February 2015  93–4 Rainbow v Johnson & Johnson (2013)  92, 100–101 SAIC Administrative Penalty Decision [2016] No 1  214 Seagate/Samsung (MOFCOM Public Announcement No 90 of 2011)  447–8 Shandong Province Department of Transportation (abuse of administrative power restricting competition), National Development and Reform Commission, Order No 501, 9 March 2015  456, 458 Shanghai Municipal Transportation Commission (Tourist Industry in the Huangpu River, abuse of administrative power restricting competition) 29 December 2016  456, 458, 460–61 Shenzhen Municipal Education Bureau (Procurement of school supplies, abuse of administrative power restricting competition)  456–7 Sichuan Province and Zhejiang Province Health and Planning Commissions (centralised drug procurement, abuse of administrative power restricting competition) 2 November 2015  456–7 Tetra Pack Administrative Penalty Decision of 16 November 2016  94–5 Thermo Fisher/Life Technologies (MOFCOM Public Announcement No 11 of 2018)  444–5, 447 Thermo Fisher/Life Technology (MOFCOM Notice No 3 of 2014)  96–7, 100 United Technologies/Goodrich (MOFCOM Public Announcement No 35 of 2012)  96, 446 United Technologies/Rockwell Collins (SAMR Public Announcement of 23 November 2018)  444 Western Digital (MOFCOM Letter No 786 of 2014)  444–5 Western Digital (MOFCOM Letter No 787 of 2014)  444–5 Western Digital/Hitachi (MOFCOM Notice No 8 of 2012)  100 Western Digital/Hitachi (MOFCOM Notice No 9 of 2012)  447–8 Wuwei City Road Transport Department (abuse of administrative power restricting competition), Gansu Province Development and Reform Commission, Order No [2015] 5, 24 November 2015  456–8 Yunnan Province Communications Authority (abuse of administrative power restricting competition) 2 June 2015  456, 458

xiv  Research handbook on methods and models of competition law

EUROPEAN UNION Aalborg Portland v Commission Joined Cases C-204/00P, C-205/00P, C-211/00P, C-213/00P, C-217/00P and C-219/00P [2004] ECR I-123  162 AC-Treuhand AG v European Commission, Case T-27/10, Judgment of 6 February 2014; ECLI:​EU:​T:​ 2014:​59  394 Ahlstrom Osakeyhtio v Commission, Joined Cases G-89/85, G-104/85, G-114/85, G-116/85, G-117/85 and G-125/85 to G-129/85 [1993] ECR I-1307 (Woodpulp II)  161, 393–4 Airtours plc v Commission of the European Communities, Case T-342/99, [2004] ECR II-2585; ECLI:​ EU:​T:​2002:​146  63, 232 AKZO Chemie BV v Commission, Case 62/86 [1991] ECR I-3359  220 Allen and Hanburys Ltd v Generics (UK) Ltd Case C-434/85 [1988] ECR 1245  484 Alliance for Natural Health and Others v Secretary of State for Health and National Assembly for Wales Joined Cases C-154/04 and C-155/04 [2005] ECR I-6451  484 Allianz Hungaria Ciztosito Zrt C-32/11 ECLI:​EU:​C2013:​160  119, 174, 396–7 Amazon e-books COMP/AT40.153, (28 July 2017) C(2017) 2876 final  193 AOK Bundesverband v Ichtyol Gesellschaft Cordes C-264/01[2004] ECR I-2493; [2004] 4 CMLR 22 394 Apple/iTunes COMP/39.154 (18 March 2008)  175 Aragonesa de Publicidad Exterior and Publivia Joined Cases C-1/90 and C-176/90 [1991] ECR I-4151 484 AstraZeneca, Case COMP/A.37.507/F3 (15 June 2005)  319 AstraZeneca v Commission Case C-457/10 P [2012] ECR I-770  319 AstraZeneca v Commission Case T-321/05 [2010] ECR 2010 II-2805  319 Asus COMP/AT.40465 (24 July 2018)  67, 172, 179, 196 BASF/Eurodiol/antochim Case COMP/M.2314 Commission Decision 2002/365/EC [2001] OJ L132/45  249 Bayer AG v Commission Case T-41/96, [2000] ECR II-3383  393–4 Bayer AG v Maschinenfabrik Hennecke GmbH & Heinz Süllhöfer Case C-65/86, [1988] ECR 5249 360 BHP Billiton/Rio Tinto Case COMP/M.4985 [2008] OJ C 312/16  250–51 Blackstone/Acetex Case COMP/M.3625, Commission Decision 2005/839/EC [2005] OJ L312/60  251 Boeing/McDonnell Douglas, Case No IV/M.877 [1997] OJ L 336/16  63 Boel v Commission, Case T-142/89 [1995] ECR II-867  162 British Airways plc v Commission of the European Communities, Case C-95/04 P [2007] ECR I-2331; ECLI:​EU:​C:​2007:​166  73–5, 77 Bronner v Mediaprint Case C-7/97 [1998] ECR I-7791; [1999] 4 CMLR 112  365, 367–8 Campari Cases IV/171, IV/856, IV/172, IV/117, IV/28.173 [1978] OJ L 70/69  169 Campus Oil Ltd v Minister for Industry and Energy Case C-72/83 [1984] ECR 2727  484 Cartonboard Case IV/C/33.833 [1998] ECT II-2099  405 Centrafarm BV and Adriaan de Peijper v Sterling Drug, Inc Case C-15/74 [1974] ECR 1147; ECLI:​EU:​ C:​1974:​114  361 Coditel v Ciné Vog Films (Coditel I) Case 62/79 [1980] ECR 881  173 Commission v Anic Partecipazioni SpA Case C-49/92 [1999] ECR I-4125  393 Commission v Belgium Case C-227/06 [2008] ECR I-46  484 Commission v DEI Case C-553/12 P [2014] ECLI:​EU:​C:​2014:​2083  479–80 Commission v Tetra Laval BV, Case C-12/03 P [2005] ECR I-987; ECLI:​EU:​C:​2005:​87  63 Commission v Volkswagen Case IV/35.733 [1998] OJ L124/60  175 Competition Authority v Beef Industry (BIDS) Case C-209/07 [2008] ECR I-8637; (2009) CMLR 6 66,  69, 172, 362, 396 Competition Authority v Beef Industry Development Society Ltd (BIDS), Case C-2-9/07 [2008] ECR I-9291 [Opinion of Trstenjak]  393, 396 Confederación Española de Empresarios de Estaciones de Servicio (CEES) v Compañia Española de Petróleos SA Case C-217/05 [2006] ECR I-11987  193 Consten and Grundig – Etablissements Consten SA and Grundig-Verkaufs-GmbH v EEC Commission Joined Cases 56 and 58/64, [1966] ECR 299  41, 66–7, 70–71, 172–3, 178, 191, 361 Continental Can Co v Commission, Case 6/72 [1975] ECR 495  73

Table of cases  xv Coöperatieve Vereniging ‘Suiker Unie’ UA v Commission, Joined Cases 40-48/73, 50/73, 54–56/73, 111/73, 113–114/73 [1976] 1 CMLR 295  393 Costa v ENEL Case C-6/64 [1964] ECR 585  480 Coty Germany GmbH v Parfümerie Akzente GmbH, Case C-230/16, ECLI:​EU:​C: 2017:941 57, 71–2,  173–4, 192 Criminal Proceedings against Asjes et al Joined Cases C-209/84 to C-213/84 [1986] ECR 1425  479 Dansk Rørindustri A/S v Commission Case C-189/02 P [2005] ECR I-5425  405 Delimitis v Henninger Bräu Case C-234/89 [1991] ECR I-935  173, 176–7 Denon & Marantz COMP/AT.40469 [2018] OJ C 335/3  67 Deutsche Telekom AG v Commission, Case C-280/08 P [2010] ECR I-9555; ECLI: EU:​C:​2010:​603  77, 80–81 Distillers (Red Label) COMP/30.228 [1983] OJ C245/3  191 Dole Food and Dole Fresh Fruit Europe v Commission Case C-286/13 P, ECLI:​EU:​C:​2015:​184 172, 197 e-books COMP/39.847 [2013] OJ C378/25  157, 193 Electrical and mechanical carbon and graphite products COMP/E-23/38.359 [2004] OJ L125/45  402 Engel v Netherlands [1976] ECHR 3; (1976) 1 EHRR 647  391 Erauw-Jacquéry Sprl v La Hesbignonne Société Coopérative Case 27/87 [1988] ECR 1919  173 European Night Services Case T-374/94 [1998] ECR II-3141  119, 393, 396 Expedia Inc v Autorité de la concurrence Case C-226/11 [2012] ECLI:​EU:​C:​2012:​795  172 Federación Española de Empresas de Tecnología Sanitaria (FENIN) v Commission Case C-205/03 P [2006] ECR I-6295  394 Fédération Française des Sociétés d’Assurance v Ministère de l’Agriculture et de la Pêche, Case C-244/94 [1995] ECR I-4013  394 Football Association Premier League Ltd v QC Leisure Case C-403/08 [2011] ECR I-9083  67 France Télécom SA v Commission Case C-202/07 P [2009] ECR I-214  221 Gas Insulated Switchgear COMP/38.889, [2008] OJ C 266/1  400 Gencor v Commission Case T-102/96 [1999] ECR II-759  498 General Electric/Honeywell v Commission Case COMP/M.2220 Decision 2004/134/EC [2004] OJ C 48/1  63 GlaxoSmithKline Services Unlimited v Commission of the European Communities Joined Cases C-501/06 P, C-513/06 P, C-515/06 P and C-519/06 P [2009] ECR I-9291, ECLI:​EU:​C:​2009:​610  67–8, 170, 172–3, 178, 397 Goodyear Italiana-Euram Case IV/23.013 Decisiosn 75/94/EEC [1975] OJ L 38/10  169 Google Search (Shopping) Case COMP/AT.39740 [2018] OJ C 9/11  207, 217 Graphite Electrodes Case COMP/E-1/36.490 [2002] OJ L100/1  400 Groupement des cartes bancaires (CB) v European Commission, Case C 67/13 P [2014] ECLI:​EU:​C:​ 2014:​2204  57, 64, 68–70, 119, 172, 174, 191, 196–8, 362–3, 397 Guinness/Grand Metropolitan Case COMP IV/M.938 Decision 98/602/EC, [1998] OJ L 288/24  63 Hoffmann-La Roche & Co AG v Commission of the European Communities, Case 85/76 [1979] ECR 461, ECLI:​EU:​C:​1979:​36  72, 74, 77–9, 85–6, 197, 280 Höfner and Elser v Macrotron GmbH Case C-41/90 [1991] ECR I-1979  118–19, 394 Huawei Technologies Co Ltd v ZTE Corp Case C-170/13 [2015] ECR 477  43 Huls v Commission Case C-199/92P [1999] ECR 1-4287  160, 393 ICI v Commission, Case 48/69, (1972) ECR 619  393 ICI v Commissioner (dyestuffs), Case 48/69, [1972] ECR 619  158–61 INEOS/Kerling Case COMP/M.4734 [2007] OJ C 174/11  251 INEOS/Solvay/JV, Case COMP/M.6905 [2014] OJ C 273/11  251 Intel Corp. v European Commission, Case C-413/14 P[2017] ECLI:​EU:​C:​2017:​632 57, 64, 168, 197, 217 Intel Corp v European Commission, Case T -286/09 [2010] ECR I-93; ECLI:​EU:​T:​2014:​547 77–80, 168 Kali + Salz/MdK/Treuhand Case IV/M.308 Commission Decision 94/449/EC [1993] OJ L186/38 249 Keck and Mithouard, Criminal proceedings against Joined Cases C-267/91 and C-268/91 [1993] ECR I-6097  484 Konkurrensverket v TeliaSonera Sverige AB, Case C-52/09 [2011] ECR I-527, ECLI:​EU:​C:​2011:​83 80–81 Kraft Foods/Cadbury Case COMP/M.5644 [2010] OJ 29/4  251 LC Nungesser KG and Kurt Eisele v Commission Case C- 258/78 [1982] ECR 2015  361

xvi  Research handbook on methods and models of competition law L’Oréal and SA L’Oréal v PVBA ‘De Nieuwe AMCK,’ Case C-31/80 [1980] ECR 3775, ECLI:​EU:​C:​ 1980:​289  71–2 Lundbeck Case COMP/39.226 [2015] OJ C 80/13  197, 336–7 Lundbeck v Commission Case T-472/13, ECLI:​EU:​C:​2015:​117; [2013] OJ C 325/76 197, 321, 337,  341–3, 359, 362–3 MasterCard Inc and Others v European Commission, Case C-382/12 P [2014] ECLI: EU:​C:​2014:​2201  70–71 Maxima Latvija’ v Konkurences padome Case C-345/14 [2015] ECLI:​EU:​C:​2015:​784  192, 197 Metro SB-Großmärkte GmbH & Co KG vCommission of the European Communities, Case 26/76 [1977] ECR 1875 [ 1977] ECLI:​EU:​C:​1977:​167  71–2, 177–8 Métropole Télévision SA v Commission Case T-112/99 [2001] ECR II-2459  395 Microsoft v Commission Case T-201/04 [2007] ECR II-3601  361–2 Motorola (GPRS Standard Essential Patents) Case COMP/AT.39985 [2014] OJ C 344/6  43 Motosykletistiki Omospondia Ellados NPID v Greece Case C-49/07 [2008] ECR I-4863; [2009] 5 CMLR 11  394 Murphy v Media Protection Services Ltd Case 429/08, [2011] ECR I-9083  67, 172, 177–8, 197–8 Musique Diffusion Francaise v Commission, Joined Cases 100/80–103/80 [1983] ECR 1825; [1983] 3 CMLR 221  162, 400 Nederlandsche Banden Industrie Michelin v Commission of the European Communities, Case 322/81 [1983] ECR 3461; ECLI:​EU:​C:​1983:​313  74–5, 77–8, 86 Nintendo Cases IV/35.706 and IV/36.321 [2003] OJ L 255/33  175, 178 Nintendo v Commission Case T-13/03 [2009] ECR II-975  175 NV GB-Inno-BM SA v ATAB Case C-13/77 [1977] ECR 2115  479 Nynas/Shell/Harburg Refinery Case COMP/M.6360 [2013] OJ C 368/5  249 Olympic/Aegean Airlines Case COMP/M.5830 Commission Decision 2012/C 195/10 [2013] OJ C195/11  249 Oracle/PeopleSoft Case COMP/M.3216 Commission Decision 2005/621/EC [2005] OJ L218/6  250 Pavlov Case C-180/98 [2000] ECR I-6451  118–19 Pfleiderer AG v Bundeskartellamt, Case C-360/09 [2011] ECR I-5161; ECLI:​EU:​C: 2011:389  411 Philip Morris/Papastratos Case COMP/M.3191 [2003] OJ C 258/4  250 Pierre Fabre Dermo-Cosmétique SAS v Président de l’Autorité de la concurrence and Ministre de l’Économie, de l’Industrie et de l’Emploi, Case C-439/09, ECLI: EU:​C:​2011:​649, [2011] ECR-SC I-09419  67, 72, 173, 177–8, 198 Po/Yamaha COMP/37.975 (16 July 2003)  67 Polypropylene, In re, Decision IV/31.149 [1986] OJ L 230  158 Post Danmark A/S v Konkurrenceradet, Case C-23/14, ECLI:​EU:​C:​2015:​651  74–8, 80 Post Danmark A/S v Konkurrencerådet Case C-209/10 [2012] ECR I-172 26, 41, 75–7, 79–80, 209, 220 Poucet et Pistre v Assurances Generales de France Cases C-159–160/91 [1993] ECR I-637  394 Pre-Insulated Pipes Case T-21/99 [1999] ECR I-7163  405 Premier League Ltd v QC Leisure Case C-403/08 [2011] ECR I-9083  172, 178 Procureur du Roi v Dassonville Case C-8/74 [1974] ECR 837  484 Pronuptia de Paris v Schillgallis, Case 161/84 [1986] ECR 353  177–8 Protimonopolný úrad Slovenskej republiky v Slovenská sporiteña a.s. Case C-68/12, ECLI:​EU:​C:​2013:​ 71  201 Rewe-Zentral AG v Bundesmonopolverwaltung für Branntwein Case C-120/78 [1979] ECR 649 (Cassis de Dijon).  484 Rhone-Poulenc v Commissioner, Case T-1/89 [1991] ECR II-867  158 SA Binon & Cie v SA Agence et Messageries de la Presse Case 243/83 [1985] ECR 2015  178 Samsung ( Enforcement of UMTS Standard Essential Patents) Case COMP AT.39939 [2014] OJ C 350/8 43 Schneider Electric SA v Commission of the European Communities, Case T-77/02 [2003] ECR I-9027, ECLI:​EU:​T:​2002:​255  63 SELEX Sistemi Integrati SpA v Commission Case T-155/04 [2006] ECR II-4797; [2007] 4 CMLR 1096 394 Servier SAS v Commission Case T-691/14 [2018] ECLI:​EU:​T:​2018:​922  197, 321, 336, 359 Société Technique Minière (LTM) v Maschinenbau Ulm GmbH (MBU), Case 56/65 [1966] ECR 337; ECLI:​EU:​C:​1966:​38  65–6, 69, 172, 176–7

Table of cases  xvii Sodemare v Regione Lombardia Case C-70/95 [1997] ECR I-3395  394 Suiker Unie v Commissioner (Suiker) Joined Cases 40–48, 50, 54–56, 111, 113 & 114/73 [1975] ECR 1663  159–61 T-Mobile Netherlands BV, KPN Mobile NV, Orange Nederland NV and Vodafone Libertel NV v Raad van bestuur van de Nederlandse Mededingingsautoriteit, Case C-8/08 [2009] ECR I-4529; ECLI:​EU:​ C:​2009:​343  67–9, 119, 172, 174, 394, 396 Tele2Polska Case C-375/09 [2011] ECR I-3055  195 Telefónica, SA and Telefónica de España, SA v European Commission, Case T- 336/07, ECLI:​EU:​T:​ 2012:​172  82 Tetra Laval BV v Commission of the European Communities, Case T-5/02 [2002] ECR II-4381, ECLI:​ EU:​T:​2002:​264  63 Tetra Pak International SA v Commission of the European Communities Case C-333/94 [1996] ECR I-5951  216 Tomra Systems ASA et al v European Commission, Case C-549/10 P, ECLI:​EU:​C: 2012:221  74–5, 78 Toshiba Corp v Commission Case C-373/14 P, ECLI:​EU:​C:​2016:​26  197 Unilever/Sara Lee Body Care Case COMP/M.5658 [2010] OJ C108/06  251 United Brands Co v Commission Case 27/76 [1978] ECR 207  42, 216 Van Eycke v ASPA NV Case C-267/86 [1988] ECR 4769  479 Van Gend en Loos v Netherlands Inland Revenue Administration Case C-26/62 [1963] ECR 1  480 Visa International, Case No COMP/29.373 [2002] OJ L 318/17  65 Völk v Vervaecke Case 5/69 [1969] ECR 295  172 Volkswagen v Commission Case C-338/00 P, [2003] ECR I-9189  175 Volkswagen v Commission Case T-62/98 [2000] ECR II-2707  175 Volvo/cania Case COMP/M.1672,Commission Decision 2001/403/EC [2001] OJ L 143/74  250 Walter Rau Lebensmittelwerke v De Smedt PVBA Case C-261/81 [1982] ECR 3961  484 Wandoo Espana v Telefonica, C-COMP/38.784 (4 July 2007)  82 Windsurfing International, Inc v Commission Case C-193/83 [1986] ECR 611  360–61 Wouters v Algemene Raad Case C-309/99 [2002] ECR I-1577  118–19, 197, 394 Yamaha COMP/37.975 (16 July 2003)  175 Zinc Phosphate,COMP/E-1/37.027 [2003] OJ L 153/1  400

INDIA Alkali Manufacturers Association of India v American Soda Ash (1998) 3 Comp LJ 173 MRTPC  499 Novartis Ags v Union of India (2013) 6 SCC 1  325

JAPAN NTT East, Supreme Court of Japan, 17 December 2010, Heisei 21 (gyō-hi) No 348, 57(2) Shinketsush 215  269

MALAYSIA Megasteel, Case No MyCC/002/2012, 15 April 2016  My E.G. Services, Case No MyCC (ED) 700-1/1/2015, 24 June 2016 

269–70 269

xviii  Research handbook on methods and models of competition law

NEW ZEALAND Commerce Commission v New Zealand Bus Ltd [2006] HC WN CIV-2006-485-585 

242

SERBIA CPC v Amm Immovables d.o.o. (2015)  CPC v Eki Transfers and Tenfore (2010)  CPC v Frikom (2017)  CPC v Frikom (2012)  CPC v Imlek and Mlekara Subotica (2008)  CPC v N sport d.o.o. and retailers of sports shoes (2016)  CPC v Serbia Broadband (2008)  Primer C d.o.o./C market d.o.o. (2007)  Sunoko d.o.o./Hellenic Sugar Industry SA (2012) 

292 277 285 292 277 285 277 282–3 282–3

SOUTH AFRICA Competition Commission and Afrisam (South Africa) (Pty) Ltd [2011] ZACT 98 (16 November 2011) 425 Competition Commission and Aveng (Africa) Ltd [2013] ZACT 76 (23 July 2013)  429 Competition Commission and Aveng (Africa) Ltd t/a Duraset [2010] ZACT 56 (25 August 2010)  416 Competition Commission and Aveng (Africa) Ltd t/a Steeldale [2012] ZACT 32 (7 May 2012)  416 Competition Commission and Delatoy Investment (Pty) Ltd [2016] ZACT 37 (14 April 2016)  416 Competition Commission and DPI Plastics (Pty) Ltd [2012] ZACT 27 (4 July 2012)  416 Competition Commission and Fio-Tek Pipes and Irrigation (Pty) Ltd [2010] ZACT 72 (27 October 2010) 416 Competition Commission and Gralio Precast (Pty) Ltd [2011] ZACAC 7 (20 October 2011)  416 Competition Commission and Lafarge (Case No 23/CR/Mar12, 28 March 2012)  425 Competition Commission and Mailot Juane Trading (Pty) Ltd [2015] ZACT 44 (29 April 2015)  416 Competition Commission and Marley Pipes System (Pty) Ltd [2010] ZACT 24 (31 March 2010)  416 Competition Commission and Murry & Robert Ltd [2013] ZACT 75 (22 July 2013)  429 Competition Commission and Netstate (Pty) Ltd [2010] ZACT 29 (19 April 2010)  420–21 Competition Commission and RSC Ekusasa Mining (Pty) Ltd [2012] ZACT 82 (19 September 2012) 416 Competition Commission and Stanley’s Removals [2016] ZACT 89 (12 December 2016)  416 Competition Commission and Stefanutti Stocks Holdings Ltd [2013] ZACT 63 (22 July 2013)  429 Competition Commission and WBHO Construction (Pty) Ltd [2013] ZACT 74 (22 July 2013)  429 Competition Commission v Afrisam (South Africa) (Pty) Ltd [2011] ZACT 98 (16 November 2011) 424 Netstar (Pty) Ltd v Competition Commission (2011) 3 SA 171  420–21

UNITED KINGDOM Argos & Littlewoods v Office of Fair Trading (CAT Cases 1014 and 1015/1/1/03) [2006] EWCA Civ 1318  163 IB v The Queen [2009] EWCA Crim 2575  410 Imperial Tobacco Group plc v OFT [2011] CAT 41  173–4 JJB Sports/All Sports v Office of Fair Trading (CAT Cases 1021/1/1/03 and 1022/1/1/03) [2004] CAT 17  163, 187, 200–201 Tobacco Manufacturers v Office of Fair Trading (CAT Case CE/2596-03, 15 April 2010)  173–4

Table of cases  xix

UNITED STATES Advocate Health Care Network, In re FTC Docket No 9369  51, 53–4 AE Clevite, Inc 58 Fed Reg 35,459 (FTC 1 July 1993)  156 Albrecht v Herald Co, 390 US 145 (1968)  43–4, 195 American Tobacco Co v United States 328 US 781 (1946)  154 Aro Corp v Allied Witan Co, 531 F 2d 1368 (6th Cir 1976)  345 Aspen Skiing Co v Aspen Highlands Skiing Corp, 472 US 585 (1985); 105 S Ct 2847  44 Associated Press v United States, 326 US 1 (1945); 65 S Ct 1416  366 Bell Atlantic Corp v Twombly, 550 US 544 (2007)  155 Broadcast Music, Inc v CBS, 441 US 1 (1979); 127 S Ct 1955; 167 L Ed 2d 929  117 Brooke Group Ltd v Brown & Williamson Tobacco Corp, 509 US 209 (1993); 113 S Ct 2578 44, 209,  220–21, 520 Business Electronics Corp v Sharp Electronics Corp, 485 US 717 (1988); 108 S Ct 1515  44 Cabell Huntington Hosp, In re, FTC Docket No 9366  46–7, 51 California Dental Association v FTC, 526 US 756 (1999); 119 S Ct 1604  49, 53 California Retail Liquor Dealers Assn v Midcal Aluminium, Inc, 445 US 97 (1980); 100 S Ct 937  481 California v Sutter Health System, 84 F Supp 2d 1057 (ND Cal 2000)  46 Carbice Corp v American Patents Dev Corp, 283 US 27 (1931); 51 S Ct 334  348 Cardizem CD Antitrust Litigation, In re, 105 F Supp 2d 682 (ED Mich 2000)  352 Cardizem CD Antitrust Litigation, In re, 332 F 3d 896 (6th Cir 2003)  352 Ciprofloxacin Hydrochloride Antitrust Litigation, In re, 544 F 3d 1323 (Fed Cir 2008), cert denied, 129 S Ct 2828 (2009)  354, 357 City of Columbia v Omni Outdoor Advertising, Inc, 499 US 365 (1999); 111 S Ct 1344  453 Continental Television, Inc v GTE Sylvania Inc, 433 US 36 (1977); 97 S Ct 2549; 53 L Ed 2d 568 (1977)  44, 60, 87, 117, 195 Dollar Tree, Inc, In re, FTC File No 141-0207  53 Dr Miles Medical Co v John D Park & Sons Co, 220 US 373 (1911); 31 S Ct 376  44, 195 DraftKings, Inc v FanDuel Ltd, FTC Docket No 9375  51 Eastman Kodak Co v Image Tech Servs, Inc, 504 US 451 (1992); 112 S Ct 2072; 119 L Ed 2d 265  53 Ehrheart v Verizon Wireless, 609 F 3d 590 (3d Cir 2010)  346 Evanston Northwest Healthcare Corp, FTC Docket No 9315  46 F Hoffmann-La Roche v Empagran SA, 542 US 155 (2004); 124 S Ct 2359  498, 502–3 Fashion Originators’ Guild v FTC, 312 US 457 (1941); 61 S Ct 703  122 First National Bank v Cities Services Co 391 US 253 (1968); 88 S Ct 1575, 20 L Ed 2d 569 (1967) 154 Flat Glass II Antitrust Litigation, In re, FTC Docket No 11-658 (2012) WL 5383346  502 FTC v Actavis, 570 US 133 (2013)  321, 326 FTC v Actavis, Inc, 133 S Ct 2223 (2013)  49, 336–8, 341–3, 349–57, 356–7 FTC v Advocate Health Care Network, 841 F 3d 460 (7th Cir 2016)  40, 47, 49 FTC v Butterworth Health Corp, 946 F Supp 1285 (W D Mich 1996), aff’d mem, 121 F 3d 708 (6th Cir 1997)  51 FTC v Cement Inst, 333 US 683 (1948); 68 S Ct793  344 FTC v Cephalon, Inc, 551 F Supp 2d 21 (DDC 2008)  320–21 FTC v Freeman Hospital, 69 F 3d 260 (8th Cir 1995)  51 FTC v H J Heinz Co, 246 F 3d 708 (DC Cir 2001)  49–50, 247 FTC v Indiana Federation of Dentists, 476 US 447 (1986); 106 S Ct 2009  44, 49, 53 FTC v Lundbeck, Inc, 650 F 3d 1236 (8th Cir 2011)  52 FTC v Penn State Hershey Medical Center, 838 F 3d 327 (3d Cir 2016)  40, 47, 49 FTC v Phoebe Putney Health Sys, Inc, 568 US 216 (2013); 133 S Ct 1003; 185 L Ed 2d 43 (2013) 49, 233 FTC v Rambus, 522 F 3d 456 (DC Cir 2008), cert denied 129 S Ct 1318 (2009)  43 FTC v Staples, Inc, 190 F Supp 3d 100 (DDC 2016)  49–50 FTC v Superior Court Trial Lawyers Association, 493 US 411 (1990); 110 S Ct 768  49, 117 FTC v Tenet Healthcare Corp, 186 F 3d 1045 (8th Cir 1999)  51 FTC v Ticor Title Insurance Co, 504 US 621 (1992)  49 FTC v Watson Pharmaceuticals, Inc, 611 F Supp 2d 1081 (CD Cal 2009)  320–21

xx  Research handbook on methods and models of competition law FTC v Watson Pharmaceuticals, Inc, 677 F 3d 1298 (11th Cir 2012)  354–5 FTC v Whole Food Market, Inc and Wild Oats Markets, Inc, Case 09-1020 (23 January 2009)  251 Geneva Pharmaceuticals Technology Corp v Barr Laboratories, Inc, 386 F 3d 485 (2d Cir 2004)  47 Gibbons v Ogden, 22 US 1 (1824)  481 Gough v Rosmoor Corp, 585 F 2d 381 (9th Cir 1978)  117 Granholm v Heald, 544 US 460 (2005); 125 S Ct 1885; 161 L Ed 2d 796  481, 485–6 Graphic Products Distributors Inc v Itek Corp, 717 F 2d 1560 (11th Cir 1983)  87 Hartford Fire Ins Co v California, 509 US 764 (1993); 113 S Ct 2891  498, 502–3 Hecht v Pro-Football Inc, 570 F 2d 982 (DC Cir, 1977); cert denied 436 US 956 (1978)  366 Henry v Chloride, Inc, 809 F 2d 1334 (8th Cir 1987)  47 Hospital Board of Directors of Lee County, 38 F 3d 1184 (11th Cir 1994)  51 Hughes v Alexandria Scrap Corp, 426 US 794 (1976); 96 S Ct 2488  481 Independent Service Organizations Antitrust Litigation, In re, 203 F 3d 1322 (Fed Cir 2000)  348–9 Intel Corp v Advanced Micro Devices Inc, 542 US 241 (2004); 124 S Ct 2466; 159 L Ed 2d 355 (2004) 502 International Shoe Co v FTC, 280 US 291 (1930); 50 S Ct 89  249 Interstate Circuit, Inc v United States 306 US 208 (1939); 59 S Ct 467; 83 L Ed 610 (1939)  154 Jefferson Parish Hospital No 2 v Hyde, 466 US 2 (1984); 104 S Ct 1551  117 K-Dur Antitrust Litigation, In re, 686 F 3d 197 (3d Cir 2012); 2012 Trade Cas. (CCH) 77971; 103 U.S.P.Q.2d (BNA) 1497  355–7 Keifer-Stewart Co v Joseph E. Seagram & Sons Inc, 340 US 211 (1951)  195 Leegin Creative Leather Prods, Inc v PSKS, Inc, 551 US 877 (2007);127 S Ct 2705; 168 L Ed 2d 623  44, 60, 64, 67, 195–6, 200, 398 Lewis v BT Investment Managers, Inc, 447 US 27 (1980); 100 S Ct 2009  481 McWane Inc and Star Pipe Productions, Ltd, In re, FTC, Docket No 9351  242 Matsushita Electric Industrial Co v Zenith Radio Corp, 475 US 574 (1986); 106 S Ct 1348 44, 154–5, 502 MCI Communications Corporation v American Telephone and Telegraph Company, 708 F 2d 1081 (7th Cir, 1983)  366 Mercoid Corp v Mid-Continent Inv Co (Mercoid I), 320 US 661 (1943); 64 S Ct 268; 88 L Ed 376 348 Minnesota v Clover Leaf Creamery Co, 449 US 456 (1981); 64 S Ct 268  481 Monsanto Co v Spray-Rite Service Corporation 465 US 752 (1984)  154–5 Motorola Mobility LLC and Google Inc, In re, FTC Docket No 121-0120  43 Motorola Mobility LLC v AU Optronics Corp, No 14-8003 (7th Cir 2014)  502 Music, Inc v Columbia Broadcasting System, Inc, 441 US 1 (1979); 99 S Ct 1551; 60 L Ed 2d 1 (1979) 53 Musical Instruments and Equipment Antitrust Litigation (Guitar Center), In re, 798 F 3d 1186 (9th Cir 2015)  155 National Collegiate Athletic Association v Board of Regents of the University of Oklahoma, 466 US 85 (1984)  43–4, 53, 174 Nestlé Holdings, Inc; Dreyer’s Grand Ice Cream Holdings, Inc; and Dreyer’s Grand Ice Cream, Inc, In re, FTC, Docket No C-4082  251 New Energy Co of Indiana v Limbach, 486 US 269 (1988); 108 S Ct 1803  481 Nobelpharma AB v Implant Innovations, Inc, 141 F 3d 1059 (CA Fed 1998)  348–9 North Carolina State Board of Dental Examiners v FTC, 135 S Ct 1101 (2015)  49 Northern Pacific Railway Co v United States, 356 US 1 (1958); 78 S Ct 514; 2 L Ed 2d 545  117 NW Wholesalers, Inc v Pacific Stationery & Printing Co, 472 US 284 (1985); 105 S Ct 2613; 86 L Ed 2d 202 (1985)  117 Ohio v American Express Co, 585 US __ (2018)  223 Otter Tail Power Co v United States, 410 US 366 (1973) 366 Pacific Bell Telephone Co v linkLine Communications Inc, 555 US 438 (2009); 129 S Ct 1109; 172 L Ed 2d 836  221 Parker v Brown, 317 US 341 (1943); 63 S Ct 307  453, 481 Penn State Hershey Med Ctr, In re, FTC Docket No 9368  51 Pike v Bruce Church, Inc, 397 US 137 (1970); 90 S Ct 844  481 Precision Instrument Manufacturing Co v Automotive Maintenance Machinery Co, 324 US 806 (1945); 65 S Ct 993; 89 L Ed 1381  347 Professional Real Estate Investors, Inc v Columbia Pictures Industries, Inc, 508 US 49 (1993)  349

Table of cases  xxi Pzifer v Government of India, 434 US 308 (1978); 98 S Ct 584  502 Quality Trailer Products Corporation 115 FTC 944 (1992)  156 Reiter v Sonotone, 442 US 330 (1979)  41 Robert Bosch GmbH, In re FTC Docket No 121-0081  43 Sanford Health, In re, FTC Docket No 9376  51 Schering-Plough Corp, In re, 136 FTC 956 (2003)  320–21, 353–7 Schering-Plough Corporation, et al, Administrative complaint in re, Docket 9297 (2001)  350–52 Schlegal Mfg Co v USM Corp, 525 F 2d 775 (6th Cir 1975)  345 Société Nationale Industrielle Aérospatiale and Société de Construction d’Avions de Tourisme, Petitioners v United States District Court for the Southern District of Iowa, etc, 482 US 522 (1987); 107 S Ct 2542; 96 L Ed 2d 461  500 Southern Pacific Co v Arizona, 325 US 761 (1945); 65 S Ct 1515  481 Spectrum Stores Inc v Citgo Petroleum Corp, 632 F 3d 938 (2011)  502 Staples, Inc, In re, FTC Docket No 9367  51, 233–4, 251 State of California v Sutter Health System, 84 F Supp 2d 1057 (N D Cal 2000), aff’d, 217 F 3d 846 (9th Cir 2000), amended by, 130 F Supp 2d 1137 (N D Cal 2001)  51 State Oil Co v Khan, 390 US 145 (1968)  43–4, 195 Stolt Nielsen Transportation Group Ltd v United States, 352 F Supp 2d 553 (2006)  502 Tamoxifen Citrate Antitrust Litigation, In re, 277 F Supp 2d 121 (ED New York 2003)  352 Tamoxifen Citrate Antitrust Litigation, In re, 466 F 3d 187 (2nd Cir 2005); 429 F 3d 370  354 Theatre Enterprises, Inc v Paramount Film Distributing (Theatre Enterprises) 346 US 537 (1954); 74 S Ct 257  154 U-Haul International, Inc and AMERCO, In the Matter of, FTC File No 081-0157 (2010)  156 US v AB Electrolux, No 1:15-cv-1039, 139 F Supp 3d 390 (DDC 2015)  51 US v Aetna Inc, No 1:16-cv-1494 (DDC 21 July 2016)  49 US v Aluminum Co of America, 148 F 2d 416 (2d Cir 1945)  59, 496–7 US v Anthem Inc and Cigna Corporation, USCA 17-5024 (2017)  233–4 US v Anthem, Inc, No 1:16-cv-1493 (DDC 21 July 2016)  49 US v Apple Inc, No 12 Civ 2826 (SDNY 10 July 2013)  157–8 US v Arnold, Schwinn & Co, 388 US 365 (1967); 87 S Ct 1856  44, 59–60, 195 US v Baker Hughes Inc 908 F 2d 981 (DC Cir 1990)  199 US v Colgate and Co, 250 US 300 (1919); 39 S Ct 465  365 US v Container Corporation of America 393 US 333 (1969); 89 S Ct 510; 21 L Ed 2d 526 (1969)  157 US v General Dynamics Corp, 415 US 486 (1974)  44, 49 US v Glens Falls Newspapers, Inc, 160 F 3d 853 (2d Cir 1998)  346 US v Gosselin World Wide Moving NV, 333 F Supp 2d 497 (2004)  502 US v Grinnell Corp, 384 US 563 (1966); 86 S Ct 1689  207 US v Haliburton Co, No 1:16-cv-233,(D Del 6 April 2016)  51 US v Hsiung, No 12-10492 2015 US App LEXIS 1590 (9th Cir, 30 January 2015)  522 US v Long Island Jewish Medical Center, 983 F Supp 121 (EDNY 1997)  51 US v Marine Bancorp, Inc, 418 US 602 (1974); 94 S Ct 2856  49 US v Mercy Health Services, 902 F Supp 968 (N D Iowa 1995), vacated as moot, 107 F 3d 632 (8th Cir 1997)  51 US v Microsoft Corp 253 F 2d 34 (DC Cir 2001)  199 US v Nippon Paper Industries Co Ltd, 17 F Supp 2d 38 (1998)  502 US v Paramount Pictures, Inc 334 US 131 (1948); 68 S Ct 915  154 US v Philadelphia National Bank, 374 US 321 (1963); 83 S Ct 1715  40, 49–50 US v Terminal Railroad Association of St Louis, 224 US 383 (1912); 32 S Ct 507  366 US v Trenton Potteries, 273 US 392 (1927); 47 S Ct 377  117 US v VISA USA, Inc, 344 F 3d 229 (2d Cir 2003)  199 US v Von’s Grocery Co, 384 US 270 (1966); 86 S Ct 1478  43–5 Utah Pie Co v Continental Baking Co, 386 US 685 (1967); 87 S Ct 1326  59 Valassis Communications, Inc, In the Matter of, FTC Docket No C-4160 (2006)  156 Verizon Communications, Inc v Law Offices of Curtis V Trinko 540 US 398 (2004); 124 S Ct 872  42,  44, 164–5, 217, 367

xxii  Research handbook on methods and models of competition law Victrex plc, FTC Docket File No 141-0042 (27 May 2016)  53 Vitamin C Antitrust Litigation, In re [2011] US District Court EDNY 440 1:06-md-01738-BMC-JO 495, 500 Walker Process Equipment, Inc v Food Machinery & Chemical Corp, 382 US 172 (1965); 86 S Ct 347 349 Weyerhaeuser Company v Ross-Simmons Hardwood Lumber Company, 549 US 312 (2007); 127 S Ct 1069  520 White Motor Co v United States, 372 US 261 (1963); 83 S Ct 696  398 White v Massachusetts Council of Construction Employers, 460 US 204 (1983); 103 S Ct 1042  481 Whole Foods Market, Inc, and Wild Oats Markets, Inc, In re, FTC, Docket No 9324  251 YKK (USA) Inc, In the Matter of , 116 FTC 628 (1993)  156

Table of legislation

EUROPEAN LEGISLATION Directives Directive 96/9/EC on the legal protection of databases [1996] OJ L77/20  313 Directive 2004/48/E on the enforcement of intellectual property rights [2004] OJ L195  313 Directive 2006/111/EC on the transparency of financial relations between Member States and public undertakings [2006] OJ L 318/17 Art 2(b)  480 Directive 2011/77/EU amending Directive 2006/116 on the term of protection of copyright and certain related rights [2011] OJ L 265/1  313 Directive 2014/104/EU on rules governing antitrust damages [2014] OJ L 349/1 123–4,  391, 403

Regulations Regulation 17/62 (EC) implementing Articles 85 and 86 of the Treaty [1962] OJ 13/204  169, 400 Regulation 2790/1999 (EC) on the application of Article 81(3) of the Treaty to categories of vertical agreements and concerted practices [1999] OJ L336/21  170 Regulation 1/2003 (EC) on the implementation of competition rules [2003] OJ L 1/1 123, 168,  170, 392, 403, 409 Art 3  409 Art 3(3)  409 Art 7  214 Art 9  359, 391 Art 11(6)  409, 411 Art 12(3)  410–11 Art 23  119, 213–14 Art 23(2)  400 Art 23(2)(a)  398 Recital 8  409

Regulation 139/2004 (EC) on the control of concentration between undertakings (Merger Regulation) [2004] OJ L 24/1  63, 281 Art 2(2)  231–2 Art 2(3)  231–2 Art 10  441 Regulation 726/2004 (EC) on procedures for the authorisation and supervision of medicinal products for human and veterinary use [2004] OJ L 136/1 recital 13  340 Regulation 772/2004 (EC) on the application of Article 81(3) of the Treaty to categories of technology transfer agreements.[2004] OJ L 123/11  360 Regulation 330/2010 (EU) on the application of Article 101(3) TFEU to categories of vertical agreements and concerted practices [2010] OJ L102/1  169–70, 174–6, 395 Recital 15  176 Art 5  176 Art 6  176 Art 29(1)(2)  176 Regulation 1217/2010 (EU) on the application of Article 101(3) TFEU to categories of research and development agreements [2010] OJ L 335/36  395 Regulation 1218/2010 (EU) on the application of Article 101(3) TFEU to categories of specialisation agreements [2010] OJ L 335/43 OJ L335  395 Regulation 2018/302 on unjustified geo-blocking and other forms of discrimination based on customers’ nationality [2018] OJ L60I/1 192

Treaties European Economic Area Agreement 1994 Art 53  410 Art 54  208, 319 Treaty Establishing the European Community Art 2  24 Art 3  24 Art 3(1)(g)  73 Art 4  24

xxiii

xxiv  Research handbook on methods and models of competition law Art 81  392, 410 Art 81(1)  68 Art 82  64, 208, 319, 367 Art 86  208 Treaty on European Union Art 3  24 Art 4(3)  479–80 Protocol on the Internal Market  24 Treaty on the Functioning of the European Union  204 Art 9  168 Art 26  480 Art 26(2)  483–4 Art 34  483–4 Art 36  361, 484 Art 101 57, 61, 63–72, 78–9, 118, 158–9, 162, 167, 276, 318, 391–8, 409–410, 479–80 Art 101(1)  64–5, 70–71, 158, 167–203, 279,  336, 342, 359–61, 393–5 Art 101(3)  25, 65, 70–71, 119, 167, 174–7,  179, 192, 194–5, 198–201, 204–5,  254, 279–80, 361, 394–7, 408–9 Art 102  25, 43, 57, 61, 64, 72–83, 168,  208–210, 276, 281, 323, 367–8, 401, 479–80 Art 102(a)  42, 208 Art 106  454, 479–80 Art 106(1)  479–80 Art 106(3)  480 Art 107  454 Art 267  123 Art 345  361

INTERNATIONAL INSTRUMENTS Agreement on Trade-Related Aspects of Intellectual Property 1994 (TRIPS Agreement)  8, 313 Art 8  326 Art 31bis  314–15, 332 Art 31(k)  326 Art 40  326 ASEAN Regional Guidelines on Competition Policy 2020  16, 33 ASEAN Set of Agreed Principles 2007 Strategic Goal 5  16 Common Market for Eastern and Southern Africa (COMESA) Treaty 1993 Art 55  333–4 EU-CARIFORUM Economic Partnership Agreement 2008  329–30 Art 127  329

Art 128  329 Art 129  329 General Agreement on Tariffs and Trade 1947 (GATT)  510 Agreement on Implementation of Article VI of GATT 1994  517–20 Hague Convention on Taking Evidence Abroad in Civil or Commercial Matters 1970 500–501

NATIONAL LEGISLATION Australia Competition and Consumer Act 2010 (Cth) 83–4,  119, 234 s2  30 s5  133 s 44CA  373, 384 s 44CA(1)(b)  387 s 44CA(1)(d)  388 s 44CA(2)  387 s 44CA(3)  388 s 44F(1)  372–3 s 44ZZRG  133 s 45(1)(c)  152–3 s 45(2)  152–3 s 45A  84 s 45AD  120 s 45AF  120 s 45AG  120 s 45AJ  120 s 45AK  120 s 46  84–7, 209 s 47  84–5 s 75B(1)  124 s 76(1)(b)  120 s 77  124 s 77A  135 s 77B  135 s 79  133 s 86D  120 s 90(7)  30 s 155(1)  237 s 155(2)(b)(iii)  237 Competition and Consumer Amendment (Competition Policy Reform) Act 2017 (Cth)  134 Competition Policy Reform Act 1995 (Cth)  371 s 59  371–2 Constitution s 92  482–3 Federal Court Rules 2011 (Cth)  124 Federal Courts of Australia Act 1976 (Cth)  124

Table of legislation  xxv Legal Profession Uniform Law Application Act 2014 (Vic) s 183  124 Sch 1  124 Trade Practices Act 1974 (Cth)  83 s 44AA  371 s 44B  371–2 s 44CA  371 s 44F  371 s 44F(1)  372–3 s 44F(3)  373 s 44F(4)  373 s 44G(1)  372 s 44G(1A)  373 s 44G(2)  371, 373 s 44G(2e)  372 s 44G(7)  372–3 s 44H(1A)  373 s 44H(2)  373 s 44H(3)  372 s 44H(4)  371, 373 s 44H(4)(e)  372 s 44H(6C)  372–3 s 44H(8)  371 s 44K  371 s 44L  371 s 44S  372 s 44V  372 s 44V(2A)  372 s 44V(3)  372 s 44X  372 s 44ZP  372 s 44ZW  372 s 44ZY  372 s 44ZZD  372 s 50  232 s 86  85, 209 Trade Practices Amendment Act 1977 (Cth) s 27  232 Trade Practices Amendment (National Access Regime) Act 2006 (Cth) Sch 1  373

Brazil Constitution 1988 Art 170  307 Art 173(4)  307 Law No 4.137/62 on Antitrust  307 Law No 8.884/94 on Antitrust  307 Law No 12.529/2011 on the Protection of Competition  307–10

Brunei Darussalam Competition Order 2015 s xx 

28

Cambodia Competition Law 2018 (draft) Art 1 

32

Chile Competition Law No 211 of 1973 Art 48  Securities Market Law No 18,045 

297 298

China Administrative Licensing Law 2003 Presidential Order No 7, 27 August 2003

Art 32(4) 

440

Administrative Litigation Law 1990  465 Anti-Monopoly Law 2007  436–7 Art 1  28, 41 Art 2  91 Art 3  120 Art 5  106 Art 6  89–90 Art 7  210 Art 8  32, 90, 452, 458–9, 461, 477 Art 12  120 Art 13  89, 120–21, 465 Art 14  121–2 Art 15  121 Art 17  93, 105, 208 Art 17(1)  42 Art 19  209 Art 20  90, 438 Art 21  438 Art 22  95 Art 25  440 Art 26  440 Art 27  445–6 Art 28  95, 445–6 Art 29  443 Art 31  90 Art 32  459, 463–4, 477 Art 32–37  459, 463–4, 477–8 Art 33  459 Art 34  459 Art 36  461 Art 37  459, 461, 465 Art 46  121–2

xxvi  Research handbook on methods and models of competition law Art 47  136, 214 Art 48  136 Art 50  122, 125, 213 Art 51  459–60, 478 Art 53  211, 465 Art 53.1  465 Art 53.2  465 Constitution 1982 Art 6  28 Art 7  28 Interim Regulation on the Standards Applicable to Simple Cases of Concentrations between Business Operators [2014] MOFCOM Announcement No 12, 11 February 2014, Art 2  442–3 Law of Civil Procedure 1991 Art 52  126 Art 53  126 Art 54  126 Art 55  126 Measures on the Calculation of Sales Revenues for the Notification of Concentrations between Business Operators in the Financial Sector [2009] MOFCOM, CSRC, CIRC and PBOC Order No 10, 15 July 2009  438 Measures on the Notification of Concentrations between Business Operators [2009] MOFCOM Order No 11, 21 November 2009  439 Provisional Regulation on the Assessment of the Impact on Competition of Concentration between Business Operators [2011] MOFCOM Order No 55, 29 August 2011  445–6 Provisional Regulation on the Implementation of Divestiture of Assets or Businesses in Concentrations between Business Operators [2010] MOFCOM Order No 41, 5 July 2010  443 Provisions on Foreign Investors’ Merger with and Acquisition of Domestic Enterprises [2006] MOFCOM, SASAC, SAT, SAIC, CSRC and SAFE Order No 10, as amended by [2009] MOFCOM Order No 6 (22 June 2009)  437 Regulation on Anti-Price Monopoly (APM) 2010 Art 11  93–4 Regulation on Prohibiting the Abuse of a Market Dominant Position (AMD) 2010 93 Regulation on the Attachment of Restrictive Conditions to Concentrations between Business Operators (Trial) [2014] MOFCOM Order No 6, 4 December 2014  443 Art 7  443 Art 14  443

Rules on the Application of Law to Civil Disputes Involving Anticompetitive Conduct in 2012 (SPC Rules) Art 6  126 Art 8  125 State Council Regulation on the Notification Thresholds for Concentrations between Business Operators [2008] State Council Order No 529, 3 August 2008  438–9

Gambia Competition Act No 4 of 2007 Sch 1(5) 

328

Hong Kong Competition Ordinance 2012 

89, 102–3

Hungary Act on the Prohibition of Unfair and Restrictive Market Practices, Act LVII of 1996 Art 85  477–8

India Competition Act 2003  Preamble  Patents Act 1970 Art 3(d) 

42 32 325

Indonesia Law concerning the Ban on Monopolistic Practices and Unfair Business Trade, Law No. 5 of 1999 Art 2  31 Art 3  31

Japan Administrative Case Litigation Act 1962 Art 30  260–61 Anti-Monopoly Act 1947  255 Art 1  33, 253 Art 2(9)  254 Art 3  254 Art 7  263, 267 Art 19  254 Art 20  263 Art 24  262

Table of legislation  xxvii Art 25  Art 26  Art 26(2)  Art 31  Art 44(2)  Art 47  Art 77  Art 85  Art 85(2)  Art 89  Art 95  Art 101  Art 102  Art 103  Civil Code 1896  Art 703  Art 709  Art 719(1)  Art 724  Companies Act 2005  Transaction Law 1952 

262 262 263 257 258 265 260 260 262 261, 264 264 265 265 265 260 263 263 263 263 260 491

Mexico Federal Economic Competition Law 1993 Art 2  Art 9  Art 10  Art 56  Art 86 

Kenya

Myanmar

Competition Act 2010 s 28(1) 

Competition Law 2015 Ch II 

328

Korea

Namibia

Monopoly Regulation and Fair Trade Act 1980 Art 1  33 Art 3-2(1)  42

Competition Act 2003 s 30 

Latvia Competition Law 2001 s2 

32

Malawi Competition Act 1998 s 3(d) 

328

Malaysia Competition Act 2010  s2  s 4(1)  s 4(2)  s5  s 10(2)(d)  s 16 

28, 253 255 254 254 254 269 258

263 258 259 271 263 267 271 271 271 271 271

s 17(2)(b)  s 18(1)-(2)  s 39  s 40  s 40(1)  s 41  s 42  s 43(1)  s 43(2)  s 43(3)  s 43(4) 

28 42 304 42 305

32–3

328

Philippines Civil Code 1949 Art 28  Competition Act 2015  s2  s 12  s 14  s 15  s 26  s 27  s 28  s 29  s 30  s 31  s 35  s 37  s 39  s 41  s 44  s 45  s 51 

261 253 253 258 254 255 254 255, 269 271 264 264 258 267 271 261 264 260, 261 261 258, 260

xxviii  Research handbook on methods and models of competition law s 64(1)  260 Development Plan 2017-2022 31 Government Procurement Reform Act 2003  264 Philippines Competition Commission Rules of Procedure 2017 Rule II, Art I-II  266

Singapore

Russian Federation

Competition Act 1998  s 12A(3) 

Law on Protection of Competition, Law No 135-FZ of 2006  31, 454–5, 477 Art 7  455 Art 8  455 Art 15-16  477 Art 23(3)  478 Art 23(6)(a)  478

28

Competition Act 2004 (Ch 50B) 

South Africa 329, 414 29

Sweden Competition Act 2008 Art 1 

28

Thailand

Serbia

Trade Competition Act 2017 (BE 2560) 

Constitution 2006 Art 16(2)  287 Art 194  287 EU Stabilisation and Association Agreement 2008  274, 276, 287–9 Art 42  289 Art 73  287–8 Federal Antitrust Law 1996  275 Law on General Administrative Procedure 1997  284 Law on Protection of Competition 2005  274,  276–8, 285, 291–2 Law on Protection of Competition 2009 278, 285 Art 1  278 Art 3  278 Art 5(2)  281 Art 10  282 Art 10(1)  279 Art 10(2)  279 Art 11  279, 282 Art 12  279 Art 13  279 Art 15(1)  280 Art 16  281 Art 19(1)  281 Art 21  284 Art 23(2)  283 Art 24(1)  283 Art 61  282 Art 62(1)  282

Trade Practices Amendment (Infrastructure Access) Act 2010

Seychelles Fair Competition Act 2009 7(4)(c) 

328

Sch 2 

28

371–2

Ukraine Law on the Antimonopoly Committee 1993, Law No 3659-XII Art 25  478 Law on the Protection of Economic Competition 2011, Law No 2210-III Art 15–17  477 Art 46  478 Preamble  28

United Kingdom Competition Act 1998  Preamble 

162 28

United States Antitrust Penalties and Procedures Act, Pub L 93-528, 88 Stat 1708 s3  128 Celler-Kefauver Act 1950, Pub L 81-899, 64 Stat 1125  231 Clayton Antitrust Act 1914, Pub L 63-212, 38 Stat 730  231 s4  117–18 s5  231

Table of legislation  xxix s7  231 Constitution Art I, s 8, cl 1, 3  481 Drug Price Competition and Patent Term Restoration Act of 1984, Pub L No 98-417, 98 Stat 1585  339–40, 350 Federal Rules of Civil Procedure r 16  346 Federal Trade Commission Act 1914, Pub L 63-203, 38 Stat 717 s5  156, 208, 211, 344–5, 349–50 Hart-Scott-Rodino Antitrust Improvements Act 1976, Pub L 94-435, Stat 1383  232–3 Hatch Waxman Act 1984, Pub L No 98-417, 98 Stat 1585  339–40, 350 Robinson-Patman Act 1936, Pub L 74-692, 49 Stat 1526  44–5 Sherman Act 1890, 26 Stat 209, 15 USC s 1-7  13–14, 18–19, 43, 116, 128, 231, 345 s1  117, 127, 153–5, 195, 231 s2  83, 156, 207–9, 214, 219, 231, 366 Sonny Bono Copyright Term Extension Act 1998, Pub L No 105-298, 112 Stat 2827 313 United States Code 15 USC s1  127, 153–5, 231

s 45(a)(1)  344 s 45(c)(1)  353 s 52-53  231 21 USC s 355(j)(2)(A)(vii)(IV)  339–40 s 355(j)(8)(B)  339 35 USC s 271(e)(2)(A)  340 Voluntary Consensus Standards for Relief under Antitrust Laws Act 2004, Pub L 108-237, 118 Stat 668 s 215(a)  127 Webb-Pomerene Act 1918, 40 Stat 516,15 USC s 61-65  491

Vietnam Competition Law No 23/2018 

32

Zambia Competition and Consumer Protection Act No 24 of 2010 s 3(a)  328

PART I THE OBJECTS AND ECONOMICS OF COMPETITION LAW

1. Competition law in flux: established and emerging approaches to methodology Deborah Healey and Rhonda L. Smith

The dynamic environment of competition law has reverberated globally of late in unprecedented fashion. While the competition law environment has traditionally been one of change for reasons as diverse as differing legal systems, political economies and economic views, spectacular growth in the number of jurisdictions adopting competition laws, particularly since the new millennium, has emphasized just how many views there are on what constitutes an effective competition law, the methodology for its application and enforcement, and what other policies might support it. Coupled with the growth of the digital environment, which has thrown new challenges to the assessment of anticompetitive conduct, the competition law world is in a substantial state of re-examination and potential refinement. The extent to which major change is necessary is under consideration in many jurisdictions. There is general agreement about the types of conduct which should be subject to competition law, but global convergence will not happen in the near future, if ever, despite attempts to do so. In any event, identical provisions in all countries would still lead to dissimilar approaches and outcomes due to cultural values, legal systemic issues, political elements, agency differences and market distinctions.1 This edited collection was assembled with the view that while certain basic economic concepts and anticompetitive conduct should be at the core of competition law, there were many approaches to the way they would be adopted in each jurisdiction, particularly in developing countries. The book gives due consideration to the more traditional approaches to competition law and respects the experience of the jurisdictions which have applied the law for many years. However, governments are assumed to seek the best competition outcomes for markets in their jurisdictions, with the qualification that exceptions and goals not related to competition are likely to dilute the benefits of competition in their markets which might otherwise be achieved. This book thus provides an overview of the competition law world at a time of substantial change.

1 A number of international agencies work to raise the standard of competition law implementation and enforcement, encouraging the development and convergence of methodology. The International Competition Network’s mission is 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’, https://​ www​ .international​ competitionnetwork​ .org/​ . The United Nations Commission for Trade and Development focuses on ‘the creation of a safe, stable and attractive environment for enterprise and industry, through fair, sound and robust national competition’, https://​unctad​.org/​ en/​Pages/​DITC/​CompetitionLaw/​Competition​-Law​-and​-Policy​.aspx. The Organisation for Economic Co-operation and Development works to encourage ‘well-designed competition law, effective enforcement and competition-based economic reform’, https://​www​.oecd​.org/​daf/​competition/​. Despite detailed and strategic work by these organizations, the task of convergence is very difficult.

2

Competition law in flux: approaches to methodology  3 The chapters are analytical rather than descriptive, seeking to examine and assess the operation of competition laws, their methodologies and the issues faced, in a range of representational jurisdictions. The editors sought example jurisdictions from all continents in an attempt to globalize discussions around important ongoing issues. In a more practical sense, this book examines competition law under three main themes or perspectives, which the contributors were asked to consider in a global context. The first is the methodology of competition law: What does a competition law contain? What methods are used to apply the law within the jurisdiction? How are markets analysed? What are the accepted economic views on market competition and which economic approach is adopted? What other possible fields, such as empirical competition law or the sociology of competition law, does it consider in implementing and applying its competition law? What other factors are considered in its competition law determinations? Second, the book asks how competition law is applied procedurally within a jurisdiction. Is there an administrative or court-based system of enforcement or a mix of the two? Are courts specialized or general in nature? How is the competition regulator constituted and funded? What is its degree of independence? How are fines and other remedies imposed and enforced? Is there a private right of action and can compensation or damages be recovered by injured parties? Finally, the book has a comparative dimension. Some chapters provide internal comparison across jurisdictions, while others offer the opportunity for chapter-on-chapter comparison. Several chapters specifically compare aspects of the law and its enforcement in more than one jurisdiction, to give the flavour of the similarities and differences which exist even in jurisdictions which are closely linked politically or geographically. Each of the comparisons is both substantive, showing differences in the policy behind the individual competition laws, and procedural, showing differences in how the competition laws of jurisdictions apply and enforce the pillars of competition law: cartels and other horizontal and vertical arrangements; monopolies (or abuse of dominance); and mergers. The chapters seek to employ deep comparisons, although the capacity to do so necessarily varies with the length and depth of experience within a particular jurisdiction. Some chapters focus on areas where a jurisdiction takes a particularly unusual approach to a competition problem, or where a comparison in a particular area might be useful as an exemplar for other jurisdictions. Part I is entitled ‘The objects and economics of competition law’. It contains this chapter, followed by chapters which set the goals, economic background and methodology of the substantive chapters which follow. Chapter 2, ‘The ambit of competition law: comments on its goals’, is a detailed examination by Deborah Healey of current discussions surrounding the goals of competition law. It explores goals in the context of both ‘traditional’ models and templates, and the newer range of goals which have been adopted in developing jurisdictions and by more recent adopters of competition law. It examines these issues against the basic questions of why we have competition law and the methodology for its application. Individual goals and methodologies significantly impact the interpretation and application of competition laws and the discussion thus provides an important background to the book as a whole. The chapter charts these issues at a time when there are divergent views around what the goals should be, the effectiveness of existing models, and whether broader or multiple goals should be adopted or accepted as being effective. The analysis is pursued against the background of significant global developments,

4  Research handbook on methods and models of competition law such as the perceived rise of market power in large corporations globally, the rise of digital platforms, and other market features such as the impact on privacy of these developments. Substantial consideration is also given to the goals of developing nations, particularly in Asia, and their distinctions as against more traditional templates, to show that a one-size-fits-all approach is not being taken. The conclusion is that a number of divergent and distinct arguments are bundled into current debates. Most jurisdictions have the broad goal of fostering a strong and competitive environment, loosely for the benefit of consumers, whether this is stated expressly or is implied. Additional or alternative goals are deemed appropriate in many developing jurisdictions, but there is currently no template for them, nor particular means of determining priority between multiple goals in circumstances where they may conflict in their application or enforcement. The author is of the view that discussions about goals and application will certainly continue for some time and that younger jurisdictions may ultimately provide working examples of alternative goals to other jurisdictions, but not in the short term. In purely economics-based Chapter 3, ‘The cost of progress: hurdles facing antitrust’s economic advance’, Alan Devlin sets out a detailed analysis of the use of empiricism in competition law, focusing particularly on the US experience. The author argues that while the competitive implications of conduct are not always clear, purely doctrinal analysis should ebb as the body of useful empirical evidence grows. The decision by an enforcer in the US of whether to challenge a merger, for example, now invariably turns on data-driven analysis. As the author states, simulations based on willingness-to-pay scenarios can predict price effects; diversion ratios estimate substitutability; and gross upward pricing pressure indices measure a firm’s post-merger incentive to raise price in a market for differentiated products. These approaches can bypass the traditional assessment of market definition and may obviate inferences about a firm’s power to raise price from its share of a market. So, he concludes that the use of these econometric devices is generally to be applauded. However, the author identifies at least three difficulties with these processes: first, he states that there is disparity between how regulators analyse the competitive effects of a merger when they intervene, and how they approach proof of these effects in their court proceedings. Second, the relegation of traditional doctrinal considerations such as market definition may undermine the consistency and predictability of competition law decision-making. Third, the author states that economic progress in antitrust has yielded a more refined enquiry focused on price, output and dynamic efficiency effects. This arguably narrows the remit of antitrust law, fostering industrial concentration and entrenchment of economic power. The chapter uses cases to explore these important methodological issues. Chapter 4 by Geoff Edwards and Jennifer Fish is entitled ‘The relevance of economics in US, EU and Australian competition law’. It considers in detail the use of economic analysis in competition law and policy in these jurisdictions, comparing the individual approaches to form-based and economic effects-based analyses in each. Through in-depth consideration of important case law, the authors review trends and show that while the US is firmly focused on effects-based thinking, the EU has until relatively recently remained a form-based system. They identify the substantial evolution in the EU since 2000 towards a more effects-based approach but illustrate some notable backtracking on occasion. Australia, by way of contrast, has shifted to a more effects-based analysis, but the authors note that form-based thinking is still apparent in that jurisdiction. They conclude, however, that the trend towards economic uniformity in these jurisdictions is threatened by the re-emergence of populist antitrust theo-

Competition law in flux: approaches to methodology  5 ries based on ideas of bigness of large firms and concentration. These take a different approach to the current economic orthodoxy, particularly in the US. While the earlier chapters focused primarily on use of economics in the US and the EU, a novel Chapter 5 by Lin Ping and Yan Yu, ‘The use of economics in competition law enforcement in mainland China and Hong Kong’, carefully evaluates the economic approach in these two important jurisdictions. It focuses on China, which implemented its Anti-Monopoly Law in 2008, and compares its approach with that of its ‘one-country-two-systems’ neighbour Hong Kong, a more recent adopter of competition law. The approach of China is critically important since it is now the third major regulator in terms of numbers of merger reviews, following the US and the EU. Many of these mergers are global transactions either involving a Chinese company or a foreign company operating in China, or with impact in China. This merger conduct and these merger reviews have the potential to impact commerce worldwide. Other anticompetitive conduct by Chinese and foreign firms in China also has the capacity to impact global trade. For these reasons, China has the capacity to be a particularly influential jurisdiction, so this chapter is of significance. The contrast with the more recent application of the competition law in Hong Kong and the approach of its regulators is significant. The chapter compares the substantive provisions and the approach to enforcement and sanctions available under the two laws, before examining in depth the approach to economic methodology in the two jurisdictions. Guidelines issued by regulators are considered in assessing the substantive provisions and the methodologies for determining whether the laws have been breached. Important cases provide exemplars of economic methodology. Views on Hong Kong are substantially provided via the views of the regulator since there have been few cases to date. The authors conclude that China has shown strong willingness to use economic analysis in its competition law cases, mainly along international lines, and that the guidelines of the Hong Kong Competition Commission indicate that rigorous economic analysis in line with international standards will be employed in Hong Kong. Part II of the book, ‘The content of the law’, is focused on the content of individual laws and the way in which the pillars of competition law are drafted and assessed in multiple jurisdictions. Chapter 6, ‘Cartel prohibition and the search for deterrent penalties: the United States, the European Union, Australia and China compared’, by Mark Williams, surveys the very different cartel laws of these four jurisdictions in the context of analysing which model works best to successfully deter cartel conduct. The author concludes that despite improved detection techniques, increased fines and penalties, and the imposition of individual criminal liability, there appears to be little reduction in cartel activity. Detailed consideration is given to each law, addressing application, tests, enforcement mechanisms including leniency policies, and historical outcomes. Institutional and procedural issues are considered, as well as the impact of personal liability. Substantial comparisons are drawn and assessed. In conclusion, the author explains that the views of cartel cases within the four jurisdictions are basically similar, but differences reflect the ‘idiosyncrasies of the legal order’, differing drafting styles, and ‘other environmental factors’. The chapter advocates for individual civil sanctions in the EU and China and recommends that Australia should make more use of its existing armoury of individual sanctions to improve outcomes and deterrence. ‘Algorithm-driven collusive conduct’ is the subject of Chapter 7 by Rob Nicholls. The chapter examines the extent to which anticompetitive conduct which is cartel-like may be facilitated by data-driven technologies, particularly facilitation arising from the use of algo-

6  Research handbook on methods and models of competition law rithms. The author explains the implications in detail and in the context of important distinctions in the meaning of terms as between computer science, economics and competition law. He notes that even parallel conduct which might be facilitated by algorithms may result in collusion, particularly where the sector in question exhibits collusive oligopoly characteristics. Both horizontal and vertical price-fixing issues may be relevant, as well as hybrid collusion based on both vertical and horizontal conduct. These issues are important when one considers the increasing use of machine learning and artificial intelligence in businesses of many kinds. The chapter also addresses the issues which arise from the use of algorithm-based platform businesses. Finally, the chapter analyses the limits of the law in the US, the EU and Australia to deal with issues of this kind. Chapter 8 by Miguel de la Mano and Alison Jones is entitled ‘Vertical agreements under EU competition law: proposals for pushing Article 101 analysis, and the modernization process, to a logical conclusion’. The authors examine the analysis of vertical agreements in the EU in the context of modernization and the adoption of a more ‘effects-based’ approach based on a consumer welfare objective. They conclude that the current approach does not adequately reflect the economic logic of vertical restraints and relies heavily on presumptions of illegality. Distinctions in decision-making at a national level have not assisted the process and compound uncertainties. The chapter argues that change would be assisted by publication of carefully selected decisions and especially more complex effects cases, which would allow the EU courts to scrutinize the EC’s policies in order to clarify and develop the law. Chapter 9 by Rhonda Smith and Deborah Healey, ‘Unilateral conduct analysis: focus on harm in multiple guises’, reviews the law dealing with unilateral use of market power in multiple jurisdictions, focusing on the methodology employed to determine whether conduct is harmful to market competition. It addresses the US Chicago School approach to monopolization, with its more laissez-faire view of the issue and its consideration of whether conduct has a business rationale (which generally means that it is efficiency enhancing and not in breach of the law), and compares this with other approaches. In Europe and Asia, for example, there is more of a belief in fairness, which is reflected in the goals or policy objectives of the law, as noted in Chapter 2. This approach examines efficiency but is more sensitive to the fair distribution of wealth and recognizing incentives for innovation. Other jurisdictions, such as China, take a slightly different approach again. The focus on market power is at the heart of issues of competition law but market power is a matter of degree and different jurisdictions set different thresholds for addressing the issue, as explained. The two models for enforcement, administrative or court-based decision-making, are identified and their advantages and disadvantages are discussed. Other differences discussed include the role of private versus public enforcement, and the severity of penalties for anticompetitive unilateral conduct, which in part reflect differing philosophies as to the role of penalties. In the section dealing with methodology, the focus is on how differing jurisdictions assess market power – what indicators are used and which are rejected. While diminishing, there remains a striking difference between the economics-based approach in the US and the form-based approach of the EU, as discussed. Determining whether unilateral conduct is likely to be anticompetitive is prone to error, so tests have been developed to reduce this risk and the pros and cons of the various tests are discussed. The chapter concludes with a discussion of the potential problems that arise in addressing anticompetitive unilateral conduct in the digital environment. Applying established approaches

Competition law in flux: approaches to methodology  7 to market definition and the identification of market power will need to be modified if competition law enforcement is to be effective. Chapter 10, by Rhonda Smith, concerns mergers. Mergers may have anticompetitive consequences either by increasing the ability to collude, including tacitly, and/or by conferring or increasing the unilateral market power of the merged entity. These economic effects are far reaching and are not easily redressed post merger, which explains why competition authorities are so concerned about them. The treatment of mergers in three jurisdictions – the US, the EU and Australia – is the focus of the chapter. The merger control process in the three jurisdictions is compared and discussed, as are the processes used to analyse the competition effects. Some brief conclusions about the treatment of mergers under competition law are drawn. Globalization of markets and technology that significantly reduces search costs mean that many mergers that may have raised competition concerns in the past will no longer do so. However, firms proposing multi-country mergers face significant costs associated with the differing requirements and approaches in those countries. A key problem for merger enforcement has always been that the effects of a merger will not be known until some time in the future. Predicting the future performance of smaller firms that are the takeover targets of the digital giants such as Google is an increasing problem for competition authorities, especially when these mergers are assessed by a court. Chapter 11, ‘Competition law in Japan, Malaysia and the Philippines: an overview’, by Mel Marquis, provides a useful analysis of the laws of these three jurisdictions, one of which (Japan) is a long-standing competition law jurisdiction, while the others are younger jurisdictions. The author situates the three individual laws in their legal, institutional and cultural structures to identify issues which might impact the ability for competition law to effectively function. All three laws have as their aim the promotion and protection of competition, but the jurisdictions all see competition law as a means of encouraging economic development. Fairness is also a feature of the laws. The similarities and distinctions between the laws of the three jurisdictions are addressed in detail, highlighting the strengths and weaknesses of each law in contrast to the others. Each of the relevant pillars of the laws is discussed. Enforcement is surveyed from the angles of regulatory structure, procedure and effectiveness, as well as investigatory powers, enforcement history and levels of success. The guidelines of the individual regulators are noted. The chapter concludes that each jurisdiction has its own strengths and weaknesses. Chapter 12, by Dragan Penezic and Zoran Soljaga, ‘Building an efficient system of protection of competition in Serbia on its path to the EU’, traverses the development of competition law in Serbia. The authors emphasize the importance of competition policy for the development of an efficient and competitive market economy, along with other policies such as foreign trade and investment, privatization, public procurement and general liberalization. They outline challenges to the introduction of competition policy in Serbia, such as state presence in the market, the low level of awareness of competition law and policy among relevant stakeholders, a regulatory regime which was not procompetitive, and support for national champions and state-owned enterprises. The EU and the process of EU integration were important features of the adoption of the law. The chapter charts the introduction of the first competition law at the end of the 1990s, noting its strengths and weaknesses. The first ‘modern’ competition law was passed in 2005, but it was also found wanting. Finally, a new law was passed in 2009, and the authors explain this law in detail, noting its similarities with EU law and discussing its strengths and weak-

8  Research handbook on methods and models of competition law nesses. The chapter also explains the institutional organization of the regulator, its competencies and processes for review, and its enforcement history. It makes a number of suggestions for its ongoing development. Chapter 13 by Fernando Furlan, entitled ‘Merger review updates in Latin America’, provides a useful summary of recent changes to merger laws and merger review in Brazil, Chile and Mexico over the last few years. The author outlines the changes, noting that each jurisdiction has modified existing procedures based on experience and on international best practice. In each jurisdiction there has been strong political support which is essential to real success. Part III of the book, ‘Particular issues’, contains chapters which address several aspects of competition law that are relatively unique exemplars for specific competition law problems, or which address unusual provisions within particular laws. The first chapter in this part, Chapter 14, ‘The interface between intellectual property rights and competition law: implications for public health in sub-Saharan Africa’ by Mor Bakhoum, discusses the dual roles of competition law in relation to intellectual property laws. The author argues that competition law has roles both as an innovation instrument and as an access and dissemination tool in the context of the pharmaceutical sector. The author examines developments in intellectual property and competition law since the TRIPS Agreement, such as the increasing emphasis on intellectual property protection and commercialization since that time. Increased protection for intellectual property, as well as more aggressive application of competition law to intellectual property-related restraints to protect innovation and to enhance access and dissemination, are observable. The chapter then focuses on the interface of the two legal areas with public health, emphasizing intellectual property restrictions which affect both innovation and access, based on enforcement experience in the EU and the US and discussing reverse payment settlement cases. The author argues that competition law should, in a complementary fashion to the TRIPS Agreement, be a legal instrument fostering innovation and access in the sector. Sub-Saharan Africa is a particular focus of the discussion. Chapter 15 by Sven Gallasch is entitled ‘Pay for delay in perspective: the impact of adversarial and inquisitorial legalism on pharmaceutical antitrust enforcement’. It compares the treatment of pay-for-delay settlements in patent infringement cases and their treatment by competition authorities in the US and the EU. It asks the question: Is the EU system an imitation of the US system or is it is a more efficient version of the same process? The author examines in detail important cases on both sides of the Atlantic in the context of the contrasting legal systems. He concludes that both systems have advantages and disadvantages. Barrister Alice Muhlebach is the author of Chapter 16, ‘The Australian approach to third party infrastructure access under Part IIIA of the Competition and Consumer Act 2010’. The chapter outlines the legislative regime in place in the jurisdiction to deal with a firm that controls infrastructure which is essential to a competitor for participation in a related market. These provisions are in addition to actions for misuse of market power (abuse of dominance or monopolization), which may be available in the jurisdiction depending upon the circumstances. Recourse under the regime is not, however, dependent on a breach of competition law. The chapter briefly canvasses the treatment of this conduct in the US and the EU before focusing in depth on the Australian regime in the context of its history and operation. Using detailed case examples, the author identifies problem areas and demonstrates that the process is complex and lengthy. Despite this, a recent review of the Australian Competition and Consumer Act recommended its retention.

Competition law in flux: approaches to methodology  9 Part IV of the book deals with ‘Enforcement’. Chapter 17, ‘The EU method of antitrust enforcement’ by Andreas Stephan, sets out in detail what the author describes as ‘the world’s most common antitrust enforcement model’. The model is purely civil enforcement by the European Commission in a regime which is non-adversarial. Only when a party to an infringement decision appeals to the Court of Justice of the European Union is there truly independent adjudication of an issue. Penalties, however, are ‘unmistakably punitive in their scale and function’ and the European Court of Human Rights considers antitrust enforcement to be criminal in nature. The chapter reviews and assesses the methodology of EU antitrust enforcement in the context of hard-core cartels by focusing on the provisions of the law, the calculation of fines, and sanctions on a national level. The author concludes that there are significant shortcomings in the system, which should be addressed if the regime is to have a truly deterrent effect. Chapter 18 by Simon Roberts is entitled ‘Cartel enforcement: critical reflections from the South African experience’. The chapter details the introduction of competition law in the jurisdiction and the progress made with enforcement in cartels since that time by a relatively young regulator. The author credits several factors as contributors to enforcement success: the introduction of a corporate leniency policy (and its subsequent amendment); proactive investigation by the regulator in priority sectors; the use of settlements to identify additional offending conduct; and international cooperation, which has led to the identification of international cartels operating within the jurisdiction. The chapter discusses the distinction between hard-core cartels and other anticompetitive arrangements under the law and sets out in detail the outcomes of various cases, analysing the different reasons for this. The work also focuses on groups such as trade associations and industries, which have been more likely to undertake the conduct. Importantly, cooperation with neighbouring countries has assisted with enforcement of international cartels, as has cooperation with a wider range of jurisdictions globally. Difficulties of proof, such as showing that information exchange is a hard-core breach under the law, are also discussed. The author discusses in detail prioritization, screening and other methods used by the regulator, such as the role of settlements and leniency policy, and their impact on enforcement. These features are credited with much of the success of the regulator. Wang Xiaoye and Adrian Emch’s Chapter 19, ‘Procedure and substance in China’s merger control regime’, explains the way in which the merger provisions of the Anti-Monopoly Law of China are applied. This is particularly significant given the importance of mergers relating to China in a global context. The chapter notes the transition from the traditional regulator, the Ministry of Commerce, following the merger of several important agencies to create the State Administration for Market Regulation. The authors note, however, that procedural divisions remain the same and mergers are still handled centrally. Comparisons with EU methodology and procedure throughout underline the distinctions between the two systems in a detailed and helpful fashion. The authors discuss in depth regulations which relate to the imposition of remedies, using case examples to clarify particular provisions. The test for mergers is said to be basically a ‘substantial lessening of competition’ test, similar to that of the EU, although the Anti-Monopoly Law provisions expressly mention ‘public interest’ as a consideration, which is an important distinction. The authors also comment on theories of harm in a merger context. They conclude that Chinese regulators have an impressive track record over their short existence.

10  Research handbook on methods and models of competition law Part V, ‘Competition policy and other issues’, includes chapters on competition policy initiatives. The first two chapters deal with competition policy and the second two take a broader global view of competition law and policy. Chapter 20, ‘An effective way to keep power in an institutional cage: legislation and regulation of administrative monopoly’, is written by Xu Shiying. The chapter assesses the problem of administrative monopoly, which is of substantial significance in China. She defines administrative monopoly as a combination of public power and private interests in the use of administrative power to exclude or restrict competition in the Chinese economy, issues which are not generally in competition law provisions. She outlines steps taken to redress the issue in the fairly unique provisions of the Anti-Monopoly Law of China, noting their strengths and weaknesses. The author discusses various examples of administrative monopoly and the way they have been resolved, and also addresses the Fair Competition Review System, a system for ex ante review of laws and regulations for anticompetitive features, which has been implemented to regulate administrative monopoly overall. The second chapter addressing competition policy is Chapter 21 by Allan Fels and Wendy Ng, ‘Competition advocacy: a broader perspective’. It examines competition advocacy, which involves non-enforcement activities of competition regulators that ‘promote a competitive environment, help to build a competition culture, and dissuade governments from taking actions that restrict competition’. They describe it as an essential task and in this chapter focus on advocacy as a tool to overturn government restrictions on competition, which generally create more entrenched anticompetitive impact than private conduct. They focus particularly on two dimensions which they say are overlooked in debates about competition advocacy: the political challenge in dealing with government restraints and other ways of dealing with these issues, such as laws prohibiting anticompetitive laws and regulations, constitutional provisions, and legislative reform programmes as part of broader competition policy. The authors conclude that taking a broader approach to the issue of government restrictions on competition will help to improve the effectiveness of competition advocacy by regulators. The final two chapters deal with important global issues related to competition law and global competition generally. Chapter 22, ‘Export cartels in times of populist protectionism: challenges and options for young and small competition agencies’ by Pierre Horna and Leonila Papa, addresses the position of young and small competition regulators in the context of ‘populist protectionist rhetoric’ of larger jurisdictions and their export cartels. The authors define and discuss the operations of export cartels, their general support by governments, and their exemption from competition law. They predict that political reactions may hamper the effective enforcement of export cartels, which would damage smaller jurisdictions. The chapter outlines ways that regulators can pursue export cartels and reviews the core features of international cooperation in the area, as well as alternative solutions to address the issue. Hassan Qaqaya has written Chapter 23, ‘Complementarities and tensions between competition and trade law and policy’, which rounds off the volume. The author notes that the common thread between ‘free trade and competition policies is the elimination of market distortions and barriers to market entry’, with the goal of promoting economic efficiency and welfare. This chapter addresses the urgent need for a comprehensive set of rules dealing with unfair practices and anticompetitive conduct in international commerce in the context of trade and competition law rules and policies. The author underscores the difficulty with promoting market access and fair competition in the context of divergent rules for ‘dealing with private market power and “behind the border”

Competition law in flux: approaches to methodology  11 restraints to trade’, and state-sponsored unfair trade practices. He argues that reforms are needed to accommodate competition and anti-dumping laws through either harmonization or convergence or a short-term policy stance to maintain effectiveness. He recommends a number of incremental steps to address the challenges of the difficult situation which currently exists. In conclusion, the book addresses in detail aspects of competition law, enforcement and policy in a wide range of jurisdictions. It provides a window into developments globally and traverses the distinctions between approaches in various jurisdictions, and a range of broader issues and policy responses. While the goals of competition law vary with economic, cultural and historic circumstances, typically it has as its central tenet protecting and increasing competition. The US and, perhaps predominantly, the EU competition laws typically form the basis for the competition law of most countries, although new adopters of competition law are more likely to add innovative features. So, too, procedures adopted for competition law enforcement vary in terms of the body responsible for enforcement – the competition authority or the court – whether penalties are punitive, and the availability of exemptions from the law. Increasingly, the methodologies adopted for assessing potentially anticompetitive conduct are converging towards effect-based tests, with the exception of cartels, which are almost universally treated as per se offences. Despite access to a range of tests and increasingly sophisticated quantitative techniques, decision errors remain a risk, with most jurisdictions opting to reduce the risk of incorrectly condemning conduct (Type 1 error) at the cost of an increased risk of failing to identify anticompetitive conduct (Type 2 error). Going forward, the digital environment raises issues such as these: How can markets be defined when products have a zero price? Is a new approach to identifying market power needed? What implications do privacy issues have for competition? All jurisdictions will be challenged when applying their competition laws to these areas.

2. The ambit of competition law: comments on its goals Deborah Healey1

I INTRODUCTION This chapter explores the goals or objects of competition law. In short, why do we have competition laws? This seminal issue significantly impacts the interpretation and application of competition law and is an important backdrop to the totality of this book. The chapter surveys the goals of competition law at a time when there is heightened debate about the effectiveness of existing models, and whether they should be amended or more broadly interpreted than laws for efficiency and consumer welfare. Global developments, such as the rise of the digital economy,2 which bring new competitive dynamics and additional complex issues of assessment; a perceived growth in inequality arising from increased economic concentration and the rise of market power; and more basic concerns about the effectiveness of current competition law models, have prompted global reconsideration of goals and methodology in a crescendo of views that is unprecedented in recent competition law history. These developments are coupled with the parallel global growth and development of competition law in jurisdictions with differing political economies that adopt and enforce their own distinct laws and policies, with goals that may diverge from the more traditional. The chapter examines the goals of competition law in both the established competition law economies and in the newer adopters, and examines some of the current debates about goals and methodology.3 It does not seek to examine the issues in depth but rather to give form to current debates, as befits a volume that seeks to assess the methodology of competition law. Several of the debates are covered more fully in subsequent chapters. The chapter analyses novel approaches to express objects by recent adopters that challenge more established views. It examines the connection between goals and methodology, recognizing that competition law 1 The author is grateful to Professor Eleanor M Fox, Dr Rhonda Smith, Dr Maciej Bernatt and Professor Michael Jacobs for their very helpful comments and suggestions. Any errors may be attributed to the author alone. 2 It is beyond the scope of this chapter to deal with competition law issues arising from big data and other aspects of the digital economy. This has been addressed in a number of useful volumes, including Maurice E Stucke and Allen P Grunes, Big Data and Competition Policy (Oxford University Press 2016); Ariel Ezrachi and Maurice E Stucke, Virtual Competition: The Promise and Perils of the Algorithm-Driven Economy (Harvard University Press 2016); Mark R Patterson, Antitrust Law in the New Economy (Harvard University Press 2017); Ariel Ezrachi, Discussion Paper: The Goals of EU Competition Law and the Digital Economy (2018), https://​www​.beuc​.eu/​publications/​beuc​-x​-2018​-071​ _goals​_of​_eu​_competition​_law​_and​_digital​_economy​.pdf; Eleanor M Fox, ‘Platforms, Power and the Antitrust Challenge: A Market Proposal to Narrow the U.S.–Europe Divide’ (2019) 98(2) Nebraska Law Review 297. 3 It is impossible to discuss all the debates fully in one short chapter, as individual issues could be – and some are – the subject of separate volumes.

12

The ambit of competition law: comments on its goals  13 is particularly susceptible to changes in economic thinking and political ideology. Economic goals are particularly important in setting the trajectory of a competition law because they impact on the methodology for determining outcomes.4 It does not argue for a universal goal or set of goals, for this would assume the uniformity of national political circumstances, judicial approach and economic analysis at an illusory and non-existent level. Rather, the chapter surveys the more ‘standard approaches’ to competition law goals (to the extent that these exist), considering ongoing international experience. It then considers other goals that have been and might be adopted, focusing primarily on issues of fairness and inequality, which are examined in the context of the US, the EU and other newer jurisdictions. It concludes that fairness is a generalized term that does not have a uniform meaning in competition law but is highly valued in many jurisdictions, as indicated by its broad adoption as a goal in many places. An overarching goal of equality or equity is also mooted by some scholars and implemented in some laws, but a refined template for an approach to the issue appears to be some way off at this stage and the issue does not have traction outside developing countries.5 Finally, the chapter draws some conclusions about the implications of different jurisdictional choices in relation to goals and the prospects for change.

II

GOALS OF COMPETITION LAW: THE ORIGINATORS TO MORE RECENT TIMES

The goals, objects or objectives (together the ‘goals’, for the purposes of this chapter) of a law reflect the views of legislators about the purpose of the legislation at the time it was enacted. They assist with the interpretation of provisions in circumstances where more than one outcome is a possibility. The goals of competition or antitrust law were simple at the time of the enactment of the US Sherman Act of 1890, one of the earliest competition laws and an acknowledged leader in the field. It is well recognized that antitrust law at that time was seen as a tool to combat aggregations of market power, dealing solely with private market conduct. Judicial interpretation of the goals of US competition law is complicated because the Sherman Act itself was drafted in very simple terms. This means that the courts in a very practical sense are the ultimate arbiters of both the goals of the law and the methodology for its delivery, in contrast to the competition laws of many other jurisdictions where goals (and possibly methods) are more fully expressed. The long title of the Sherman Act is ‘An Act to protect trade and commerce against restraints and monopolies’ and this sets very broad parameters of interpretation for the courts. The Sherman Act itself does not elaborate on how this protection should be delivered or its specific measurement of success, apart from in the substantive prohibitions of the Act themselves. The prohibitions contained in the Sherman Act are also

4 David Gerber, ‘Comment on Parret’ in Daniel Zimmer (ed), The Goals of Competition Law (Edward Elgar Publishing 2012) 85, 93. 5 Eleanor M Fox, ‘South Africa, Competition Law and Equality: Restoring Equity by Antitrust in a Land Where Markets Were Brutally Skewed’, Competition Policy International (December 2019).

14  Research handbook on methods and models of competition law very broadly drawn, encompassing generalized tests without much direction as to priorities of application, leaving substantial discretion to regulators and courts.6 So, perceptions of the ‘problem’ to be addressed by antitrust law at the time of its original enactment initially defined and limited its scope, but in very broad terms. Antitrust law was conceived as a means to control the anticompetitive acts of private business firms and was focused on industrialists who were ‘building business empires at the expense of farmers, buyers and other small players and were undermining the vision of the social good’.7 The legislation was based on the idea that by controlling these industrialists, businesses of all sizes would compete and this would benefit society. The rest was left to the courts. The focus on private business by the originators of competition law ignored the fact that the state has a substantial impact on markets and left open the issue of competition law coverage of the state in the market. In the US, state and local ownership of business was relatively unusual, and thus the approach was logical and unsurprising. More recently and more widely, however, it is generally recognized that government restraints can have a significant and enduring negative impact on the market. While this is recognized in the US today, the impact of the state in US markets is still far less than in other jurisdictions that have more state-based commerce, and these are typically emerging markets and developing nations. Elsewhere, and often due to other issues which might affect competition, such as the more expansive involvement of the state in markets arising from history, or stage of economic development, other jurisdictions have taken a broader view of the goals of competition law. In a different approach in Europe during the establishment of what became the EU, the creation of a single market, requiring that European business would be competitive, was a key goal of competition law. The Messina Report, which was foundational to the adoption of the Treaties of Rome, the predecessors of the current competition laws of the EU, focused on ‘market power as a threat to the fundamental objectives pursued by the opening of domestic markets’, which were meant to contribute to a ‘more rational allocation of resources’ and a ‘general increase in welfare’ following the Second World War in Europe.8 In relation to competition, the primary concern was that anticompetitive agreements and mergers could reinstate the very trade barriers between Member States that the Treaties sought to dismantle. Ensuring that this did not occur required the prohibition of restraints on trade and competition by Member States and businesses in the internal market. This foundational purpose of EU law has significantly influenced the enforcement of its competition law. Competition was addressed in the rules setting up and governing the Union.9 Ultimately, these laws prohibited restraints on trade and competition by both Member States and businesses (including those owned by Member States and at other levels of government) in the single market. In the context of state acts, the EU law sought to distinguish between appropriate state acts designed to protect the public and those that were inappropriate and damaged competition, the latter being prohib6 See Eleanor M Fox, ‘Against Goals’ (2013) 81 Fordham Law Review 2157, 2157: ‘Congress wanted to rein in the power of the trusts, but it did not legislate how to do so. Rather, it deliberately left this task for the courts to develop on a case-by-case basis.’ 7 Eleanor M Fox and Deborah Healey, ‘When the State Harms Competition – The Role for Competition Law’ (2014) 79(3) Antitrust Law Review 769, 769. Parts of this chapter are drawn from research and ideas presented in that article and from ongoing research in that area. 8 See Eleanor M Fox and Damien Gerard, EU Competition Law (Edward Elgar Publishing 2017) xii. 9 The Treaty of Rome established the European Economic Community in 1957. The current law is set out in the Treaty on European Union and the Treaty on the Functioning of the European Union.

The ambit of competition law: comments on its goals  15 ited.10 The subsequent adoption of a more economic approach in the EU 11 supplemented its foundational goals and over time other goals have been identified.12 Economics and economic analysis are at the heart of competition law. Economic thinking, however, is evolving and not static. Challenges in the use of competition economics exist not only in their approach and application, but also in proof of assumptions and outcomes. Changes in the perception of the ways in which competition may be damaged and the sources of damage have prompted a broader approach to the goals of competition law in more recent times. Changing economic views and the adoption of competition laws in jurisdictions that have different political economies and differing roles for the state have led to the setting of different goals. Different approaches to development and different conceptions of what competition law can do in developing economies have also affected the immediate goals and scope of competition laws.13 In a different context, the fall of the Berlin Wall saw many previous command economies in Eastern Europe adopt market systems and competition laws. Many state-owned enterprises in those jurisdictions were privatized, with a large number maintaining their monopoly status and their existing advantages. Local governments in those jurisdictions were also powerful and important facilitators and players in trade. Competition laws adopted often sought to address existing trade blockages caused by local governments.14 The very significant political and economic changes which occurred in these jurisdictions produced different responses to the more traditional competition law template – not only was the state covered by the competition law, but additional provisions were often adopted to deal both with specific state-related power and competition issues, and with developmental issues. These laws often took a broader approach to the issue of goals. More recently, jurisdictions in Asia, Africa and South America began to implement competition laws. Many of these jurisdictions traditionally had significant state involvement in markets. Most were developing economies, and many had significant issues of poverty and market failure. This required a distinct approach to the law suitable to their own needs and produced different outcomes again in respect of the goals and application of the law.15 See Treaty on European Union, Arts 4(3), 106(6). This ‘modernization’ is often seen to have occurred following the enactment of Council Regulation 1/2003 [2003] OJ L1/1, but it is actually part of a ‘wider and deeper phenomenon’ in the 1990s that saw a move towards ‘a realignment of competition law in line with economic thinking on efficiency and welfare’. Alison Jones and Brenda Sufrin, EU Competition Law (Oxford University Press 2016) 37. The Commission ‘adopted an effects approach to applying the consumer welfare standard’. This policy has continued to date and is reinforced by documents such as the Guidelines on Article 101. See Jones and Sufrin, ibid, 38. 12 Neither the current chapters of the Treaty on the Functioning of the European Union nor past EU laws which dealt with competition ‘discuss specific objectives or goals’. Laura Parret, ‘The Multiple Personalities of Competition Law: Time for a Comprehensive Debate on Its Objectives’ in Daniel Zimmer (ed), The Goals of Competition Law (Edward Elgar Publishing 2012) 61, 63. However, as noted subsequently, there are some mentions in supporting documentation. 13 See, for example, Oliver Budzinski and Maryam HH Beigi, ‘Generating Instead of Protecting Competition’ in Michal S Gal, Mor Bakkhoum, Josef Drexl, Eleanor M Fox and David J Gerber (eds), The Economic Characteristics of Developing Jurisdictions: Their Implications for Competition Law (Edward Elgar Publishing 2015) 223. 14 See, for example, Roger Alan Boner, ‘Antitrust and State Action in Transition Economies’ (1998) 43 Antitrust Bulletin 71, cited in Fox and Healey (n 7). 15 See, for example, the Competition Act of South Africa, discussed further below. 10 11

16  Research handbook on methods and models of competition law The groundswell to enact competition laws which began in the new millennium and continues today also saw competition law adopted in nations employing methodologies that have been labelled ‘state capitalism’,16 where the state has a continuing important role in the economy, and well-developed industrial policy exists alongside competition law. Variants of this exist in different jurisdictions. This group, which includes a range of nations – among them, Middle Eastern countries, large trading nations such as China, and developed economies such as Singapore17 – have employed in different ways an approach that accepts significant government involvement in markets and fosters it to some extent while adopting many features of market economies, including a competition law. Finally, nations in the Association of Southeast Asian Nations (ASEAN) region agreed to adopt competition laws by 2015, and have set 2025 as the date for adoption of an ASEAN Set of Agreed Principles.18 ASEAN’s Regional Guidelines talk of competition policy, which is defined as competition law and ‘other public policies and general governmental directions aimed at introducing, increasing and/or maintaining competition’.19 The ASEAN Regional Guidelines are instructive of ASEAN’s views on the goals of competition law: The most commonly stated objective of Competition policy is the promotion and protection of the competitive process. Competition policy introduces a ‘level playing field’ for all market players that will help markets to be competitive. The introduction of a competition law will provide the market with a set of ‘rules of the game’ that protects the competition process itself, rather than competitors in the market. In this way, the pursuit of fair or effective competition can contribute to improvements in economic efficiency, economic growth and development and consumer welfare.20

ASEAN makes it clear, however, that each member state may decide which of the objectives it wishes to pursue, taking into account its own national competition policy needs.21 So, while there is some commonality of goals, the ASEAN Regional Guidelines recognize that the goals of jurisdictions may differ. Each ASEAN jurisdiction, in fact, has slightly differing goals, although strengthening and protecting competition are common themes throughout. Many of the newer jurisdictions adopting competition laws are classified as developing economies. As the World Bank concludes, competition policy ‘can make the difference between severe poverty and a dignified life’ in a developing economy.22 It can also ‘promote

16 See generally Ian Bremmer, The End of the Free Market: Who Wins the War between States and Corporations? (Portfolio 2010) 57. 17 Deborah Healey, ‘Application of Competition Laws to Government in Asia: The Singapore Story’ (2012) 2 KLRI Journal of Law and Legislation 57. Chapter 50B of the Competition Act states that it is an ‘Act to make provision about competition and abuse of a dominant position in the market and to set up the Commission of Singapore’. 18 Strategic Goal 5 in ASEAN’s Competition Action Plan is ‘Moving towards greater harmonization of competition policy and law’: ASEAN, An ASEAN Competition Action Plan (2016–2025) (2016) 8, https://​asean​.org/​storage/​2012/​05/​ACAP​-Website​-23​-December​-2016​.pdf. 19 ASEAN, Regional Guidelines on Competition Policy (2010), 2.1.2, 2.2.1ff, https://​www​.asean​ .org/ ​ w p ​ - content/ ​ u ploads/ ​ i mages/ ​ 2 012/ ​ p ublications/ ​ A SEAN ​ % 20Regional ​ % 20Guidelines ​ % 20on​ %20Competition​%20Policy​.pdf. 20 Ibid. 21 Ibid, 2.2.5. 22 World Bank, ‘Boosting Competition in African Markets Can Enhance Growth and Lift at Least Half a Million People out of Poverty’ (Media Release, 27 July 2016), https://​www​.worldbank​.org/​en/​ news/​press​-release/​2016/​07/​27/​boosting​-competition​-in​-african​-markets​-can​-enhance​-growth​-and​-lift​

The ambit of competition law: comments on its goals  17 opportunity, inclusiveness and indeed, efficiency’.23 The laws of developing economies reflect that status as well as the sometimes complex relationships that their economies have with the state. Many of the newer jurisdictions have goals that encompass more than the creation of a level playing field24 and they sometimes apply competition law to the state and to government businesses in a less comprehensive fashion.25 In summary, more than 130 jurisdictions worldwide have adopted competition laws.26 These range from the established jurisdictions such as the US, where major debates centre around the continuing suitability of the current ideology and methodology of competition law and where certainty is not as ingrained as is often suggested, to the EU, where a broader range of issues lie behind the EU and its objects, and where the economic approach is different from that of the US, although more similar than in previous times. There are jurisdictions which follow these approaches, with some adopting one line or another more strongly. There are newer adopters in Eastern Europe, Asia and South America, which have appropriated parts of the standard approach but have fine-tuned the competition law to their own needs. And, finally, there are other jurisdictions that adopt competition law in principle but adjust its application regularly to support industrial policy or other policy initiatives. So while at a basic level the competition laws of most jurisdictions have goals or objects that encourage or protect competition to assist the efficient use of resources, either expressly or impliedly assuming that this will benefit the economy and particularly consumers, a diversity of other goals that vary across jurisdictions is directed at other economic or social agendas. These other goals are judged by legislators to be suitable to that jurisdiction.27 Context is everything in competition law. More recent developments, such as the rise of digital platforms and the use of artificial intelligence to enable functions in many areas of life, raise issues that were unthought of in previous times. Issues of market power, innovation, barriers to entry, machine-based collusion, and impacts on privacy and consumer law are some of the areas now calling for in-depth examination in a digital context. Many ask whether current competition law goals and methodologies are effective to deal with all of this, and whether the

-at​ -least​ -half​ -a​ -million​ -people​ -out​ -of​ -poverty, cited in Eleanor M Fox, ‘Competition Policy at the Equity-Efficiency Intersection’ in Damien Gerard and Ioannis Lianos (eds), Reconciling Efficiency and Equity: A Global Challenge for Competition Policy (Cambridge University Press 2019) 441, 445. 23 Fox, ‘Competition Policy at the Equity-Efficiency Intersection’ (n 22), 445. 24 See below for further discussion. 25 See, for example, Healey, ‘Application of Competition Laws to Government in Asia’ (n 17). 26 Richard Whish and David Bailey, Competition Law (9th ed, Oxford University Press 2018) 1, noting reference material in relation to world competition laws at www​.oecd​.org/​competition and https://​ www​.international​competitionnetwork​.org/​. 27 The United Nations Conference on Trade and Development Model Law on Competition states that the goal of competition law is: To control or eliminate restrictive agreements or arrangements among enterprises, or mergers and acquisitions or abuse of dominant positions of market power, which limit access to markets or otherwise unduly restrain competition, adversely affecting domestic or international trade or economic development. (Chapter 1, Objectives or purpose of the law.) For further examples of objects or purposes of the law, see United Nations Conference on Trade and Development, Model Law on Competition: Substantive Possible Elements for a Competition Law, Commentaries and Alternative Approaches in Existing Legislation, Part II, Commentaries on Chapters of the Model Law and Alternative Approaches in Existing Legislation, 13ff, https://​unctad​.org/​en/​Docs/​tdrbpconf7d8​_en​.pdf.

18  Research handbook on methods and models of competition law introduction of multiple and diverse goals would aid in the application of competition law in these complex circumstances. On the one hand, some argue that existing rules and assumptions are flexible and sufficient to deal with new circumstances that have arisen. In defence of the single consumer welfare goal, they also argue that multiple goals may dilute the central economic message that has traditionally been at the heart of competition law.28 The adoption of a broader range of goals, such as the support of small business or workers, the creation of national champions, or the environment and sustainability, may create an issue of prioritization in enforcement and decision-making, where goals conflict in a particular case. If there is no guidance on their interpretation in these circumstances it may lead to uncertainty. Despite this, many nations confidently adopt multiple goals and some of those goals are outlined below. Failure to resolve issues of priority between more than one goal in application and enforcement may result in unclear or unpredictable application of the law and compromise the positive impact on competition that the law is intended to have within a jurisdiction. At the same time, departure from the single-goal economic approach has capacity to add to the time and cost of competition law enforcement in that more information and evidence need to be assessed. This chapter now examines the approach to goals in more depth. It begins with the established competition law jurisdictions of the US and the EU, and then expands to select jurisdictions, particularly those in Asia. The discussion encompasses views on methodology to the extent that they are relevant. The chapter concludes that, predictably, most competition laws have one or more goals that first and foremost preference the protection or encouragement of competition and a competitive environment. Other goals may follow. The ambit and enforcement of these vary according to the priorities of the individual jurisdiction.29

III

THE US AND THE ‘SINGLE GOAL’

As noted earlier, the original goals of the Sherman Act were focused on reining in the power of the industrial trusts, fighting high concentration and abuses of power, and keeping markets open.30 Goals were not much discussed. The question was how to protect competition. This approach morphed over time. By the middle of the twentieth century, the Harvard School of economists took the general view that antitrust law had been enacted to protect competition. They advocated an approach to goals and interpretation that focused on addressing 28 Others have theorized that there is no ‘satisfactory comprehensive definition’ of competition itself, which makes the question of goals even more difficult to determine. See Maurice E Stucke, ‘What Is Competition?’ in Daniel Zimmer (ed), The Goals of Competition Law (Edward Elgar Publishing 2012) 27. 29 Ibanez-Columo suggests that debates about goals of competition law are pointless and that, broadly, the real divide in views is between ‘discretionalists’ and ‘legalists’, although he accepts that the position is more complex than this. Discretionalists believe that courts and regulators ‘should be as unconstrained as possible to attain the result that is considered correct in any case’, whereas legalists downplay that outcome, but ‘place constraints on competition authorities’ in order ‘to ensure that intervention is consistent and predictable’. Subcategories exist under both these categories. Pablo Ibanez-Columo, ‘Discretionalists vs Legalists: The True Divide in the Competition Law Community?’, Chillin’Competition (28 May 2018), https://​chillingcompetition​.com/​2018/​05/​28/​discretionalists​-vs​ -legalists​-the​-true​-divide​-in​-the​-competition​-law​-community. 30 Fox, ‘Against Goals’ (n 6), 2158.

The ambit of competition law: comments on its goals  19 structure and high concentration in industries to preserve competitive markets. These features were said to lead to poor economic performance and reduced output, and to enable monopoly prices. They supported the protection of small business. In summary, the Harvard economists replicated to some extent the approach of the original drafters of the law but adopted a more sophisticated economic approach to the problems that had originally been identified. These Harvard School economists were focused on the structure–conduct–performance paradigm, which, put simply, held that ‘the structure of the market determines the firm’s conduct and that conduct determines market performance (for example profitability, efficiency, technical progress and growth)’.31 Under this view, structural remedies were said to be more effective than behavioural remedies in the control of mergers and acquisitions. Chicago School economists, who began to emerge in the 1960s, argued that the Harvard School approach was basically flawed and led to over-regulation. They believed that ‘consumer welfare narrowly conceived in explicitly economic terms should be the only objective’ of antitrust (competition) law, with allocative efficiency as the promoter of this goal. This approach, championed initially by economists such as George Stigler, is described as having replaced the Harvard School, with its fear of the corruption of government and society, with an approach that feared corruption of the marketplace.32 The focus was on the self-correcting nature of markets, light-handed application of the law, and a preference for under-enforcement rather than over-enforcement. In 1976 Richard Posner opined that economic efficiency should be the sole goal of antitrust policy,33 and in 1978, in his book The Antitrust Paradox, it is said that Robert Bork ‘rewrote history’ when he ‘asserted that the original purpose of the Sherman Antitrust Act was to maximise consumer welfare’, defined as total welfare.34 Chicago economists focused heavily on their view of economics, and antitrust law for the first time ‘became a branch of economics’.35 Some argue that the focus on the idea of consumer welfare was introduced as a tool to rein in activist antitrust at the time of the Reagan administration,36 but nonetheless, the views of the Chicagoans were universally adopted by the courts and regulators. The focus on efficiency and consumer protection, however, meant that Chicagoans and their views, when adopted by the courts, used their approach to economics for analysis and as the arbiter and standard for antitrust questions and disputes. Ultimately, while the Chicago paradigm has changed in some respects over the course of its use, the consumer welfare approach continues to be used by regulators, the judiciary and the public.37 Jones and Sufrin (n 11), 20. See Binyanmin Appelbaum, The Economists’ Hour: False Prophets, Free Markets and the Fracture of Society (Hachette Book Group 2019) 139. 33 Ibid, 145, quoting Richard Posner, Antitrust Law (University of Chicago Press 1976). 34 Appelbaum (n 32), 149. See also Douglas H Ginsburg, ‘Bork’s “Legislative Intent” and the Courts’ (2014) 79(3) Antitrust Law Journal 941. As to the certainty and simplicity of the Chicago School methodology in the context of the work of Bork, see Christopher R Leslie, ‘Antitrust Made Too Simple’ (2014) 79(3) Antitrust Law Journal 917. 35 Michael S Jacobs, ‘An Essay on the Normative Foundations of Antitrust Economics’ (1995) 74(1) North Carolina Law Review 219. 36 Fox, ‘Against Goals’ (n 6), where the author states at 2157 that ‘it is not the goal of antitrust unless the concept of “goal” reads ninety years out of antitrust history’. The same author argues that maximizing consumer welfare is a euphemism (at 2159). Also cited in Albert Allen Foer and Arthur Durst, ‘The Multiple Goals of Antitrust’ (2018) 63(4) Antitrust Bulletin 494, 496. 37 Foer and Durst (n 36), 495. 31 32

20  Research handbook on methods and models of competition law However, there are many critics of the Chicago School consumer welfare approach. Some believe, for example, that ‘[l]aw is not economics … The basic difference between Chicagoans and traditionalists is a difference of vision about what kind of society we are and should strive to be …’.38 Supporters of the Post-Chicago School, which followed the Chicagoans, take a less doctrinaire approach to the application of economics to competition law. They accept the same goals as the Chicago School economists, but their approach is more subtle, and they argue that firms do not react passively to market conditions but rather attempt to influence them by their strategic behaviour. They ‘recognise that economics may give indications of what questions to ask, but does not always yield definitive answers, and certainly not answers which are necessarily value-free’.39 They make use of game theory and examine the effect of strategic conduct in market situations to determine breach. They question some of the key assumptions of the Chicago economists, in many areas seeking a more nuanced inquiry about the actual impact of conduct on markets in particular cases, rather than adopting a blanket view of all conduct of that nature. For example, Chicago economists believe that vertical arrangements are rarely anticompetitive. Post-Chicago economists are far more suspicious of vertical arrangements, recognizing that there is potential harm in bundling, and argue that vertical mergers may result in anticompetitive market foreclosure of downstream acquirers from essential sources of supply.40 More recently, exponents of what has been described as ‘Populist Antitrust’, ‘Hipster Antitrust’ and ‘New Brandeis’ economics are also strong critics of the Chicago approach. They return to a focus on concentrated market structures and the idea of ‘bigness’, which were previously emphasized by the Harvard School, and have become more vocal in the last few years in their challenge to the current economic approach. They believe that market consolidation creates competition issues that are not always outweighed by price reductions or innovation, which were key features of Chicago thinking. They argue that the consumer welfare standard is too narrow and therefore does not protect anticompetitive harm to other market participants; they believe that on occasion some other benefits may trump consumer benefit. They also argue that the consumer welfare test as currently implemented does not effectively account for market developments such as multi-sided markets and digital platforms, or the general growth in inequality brought about by increasing market concentration. Their opponents say, however, that the Hipsters ‘propose … to divorce antitrust law from economic analysis, to abandon the well-established consumer welfare framework that introduced the rule of law to antitrust and to replace that standard with a vague and pliable socio-political approach’.41

38 Eleanor M Fox and Lawrence A Sullivan, ‘Antitrust Retrospective and Prospective: Where Are We Coming From? Where Are We Going?’ (1987) 62 New York Law Review 936. 39 Jones and Sufrin (n 11), 20. 40 Jacobs (n 35). 41 Elyse Dorsey, Jan M Rybnicek and Joshua D Wright, ‘Hipster Antitrust Meets Public Choice Economics: The Consumer Welfare Standard, Rule of Law and Rent-Seeking’ (2018) Antitrust Chronicle Special Edition, Year in Review 7, 7. The authors refer to the position prior to the adoption of the consumer welfare standard as an ‘incoherent hodgepodge of socio-political goals governing antitrust’.

The ambit of competition law: comments on its goals  21 There are thus many critics, but currently little by way of strong theoretical frameworks to replace the Chicago approach.42 A

Revision in the US?

Many critics would agree with the comments of Ariel Ezrachi, who has stated that ‘objectivity, clarity and analytical superiority’ are not always present in the current economic approach to competition analysis.43 Much ink has been spilled on the goals of competition law, with a number of key articles referred to here. In the context of the current debate, several of the issues are mentioned below. The immediate question when one critiques the current US approach to competition law economics is to consider which economics is appropriate. Chicago economists favour consumer welfare as the sole goal or objective of competition law, but others have questioned its utility in contrast to the possible alternate use of the ‘total welfare’ standard. 44 Others again favour the adoption of a broader welfare standard (not in the purely economic sense), for example, which would add to factors for analysis in relevant cases. New considerations such as the impact of conduct on the environment and employment could be added and could be given priority over the effect of the conduct on consumer prices if it were found that the other benefits outweighed consumer benefit in a particular case. This would inject other goals into the analysis. Another criticism is that in an era of ever-larger corporations, and possibly because of perceived under-enforcement of competition law under the Chicago approach, ‘a narrow focus on consumer welfare is causing other kinds of damage’.45 Of these current critics, Foer and Durst, for example, argue that a change in approach is warranted because market conditions have changed significantly since the adoption of the Chicago School economic efficiency/ consumer welfare single object approach. They list the ‘Too Big to Fail’ banking crisis; the rise of digital platforms; issues relating to privacy and big data; the growing number of international jurisdictions using multiple goals; Trump’s comments about big corporations; and the perception that, for many people, the capitalist system is not working well, as examples of changing market dynamics which raise questions about the current approach.46 Foer and Durst also argue that ‘a single-minded objective for antitrust should be replaced by recognition that there are multiple objects’ and agree with others such as Lina Khan that it should be about ‘the promotion and protection of systems that provide society’s best mixture of competition and co-operation, given its culture, history and political situation at a given point in time’.47 On

42 For comment on the New Brandesian views in their context of welfare economics, see Mark Glick, ‘The Unsound Theory behind the Consumer (and Total) Welfare Goal in Antitrust’ (2018) 63 Antitrust Bulletin 1; see also Mark Glick, ‘How Chicago Economics Distorts “Consumer Welfare” in Antitrust’ (2019) 64 Antitrust Bulletin 495. 43 Ariel Ezrachi, ‘Sponge’ (2017) 1 Journal of Antitrust Enforcement 5, 50. 44 See, for example, Louis Kaplow, ‘On the Choice of Welfare Standards in Competition Law’ in Daniel Zimmer (ed), The Goals of Competition Law (Edward Elgar Publishing 2012) 3ff. The author favours the use of a total welfare standard. 45 Appelbaum (n 32), 158. 46 See Foer and Durst (n 36), 495. 47 Ibid. See also Joseph E Stiglitz, People, Power and Profits: Progressive Capitalism for an Age of Discontent (Allen Lane 2019) 68ff; as to new technologies, see also 123–37.

22  Research handbook on methods and models of competition law methodology, they are particularly critical of the part of the Chicago School test which states that ‘if you cannot prove something you have to leave it alone. We are sceptical that this is too oversimplified a test, coming from an oversimplified model that misses too much of the richness of life.’48 They recognize, however, at a simple level, that the concept of a single goal is appealing as it would promote certainty and objectivity.49 Globally, others consider that competition law should aim to reduce inequality, which would ‘imply the inclusion of an explicit distributional perspective in the enforcement of competition law’.50 Among the suggestions considered by Ioannis Lianos is the adoption of economic and social equality as a goal of competition law, in addition to the usual goals of consumer welfare and efficiency ‘as part of a broader public interest standard’, under which these goals could be given a higher priority.51 He advocates for the injection of equality to balance efficiency in a theoretical framework that should be ‘devoid of populism, and entrenched in a fairness-driven competition law’.52 So he argues for the introduction of equality as a factor under a strong theoretical framework. In a similar vein to Foer and Durst, Lina Khan argues that ‘robust markets’ should be the goal of competition law. She continues that because the current framework in antitrust relies on ‘assumptions embedded in the Chicago School framework and the way competition is assessed’, it fails to register certain forms of anticompetitive harm and is therefore ‘unequipped to promote real competition – a shortcoming that is illuminated and amplified in the context of online platforms and data-driven markets’.53 Khan states that the debate about goals often obscures the most important factors: perspective and assumptions. She suggests that ‘analysis of competition and structure would offer better insight into the state of competition than do measures of welfare’54 and argues for a new test which would assess … whether a company’s structure creates anticompetitive conflicts of interest; whether it can cross-leverage market advantages across distinct lines of business; and whether the economics of online platform markets incentivizes predatory conduct and capital markets permit it. More specifically, restoring traditional antitrust principles to create a presumption of predation and to ban vertical integration by dominant platforms could help maintain competition in these markets. If, instead, we accept dominant online platforms as natural monopolies or oligopolies, then applying elements of a public utility regime or essential facilities obligations would maintain the benefits of scale while limiting the ability of dominant platforms to abuse the power that comes with it.55

Foer and Durst (n 36), 508. Ibid, 506. 50 Ioannis Lianos, ‘The Poverty of Competition Law’ in Damien Gerard and Ioannis Lianos (eds), Reconciling Efficiency and Equity: A Global Challenge for Competition Policy (Cambridge University Press 2019) 45, 47, citing, among others, Jonathan B Baker and Steven C Salop, ‘Antitrust, Competition Policy and Inequality’ (2015) 104 Georgetown Law Journal 1, 24. 51 See Lianos (n 50), 47. The volume itself contains a number of thoughtful and detailed considerations of this issue, which are beyond the scope of this chapter. 52 See Damien Gerard and Ioannis Lianos, ‘Introduction’ in Damien Gerard and Ioannis Lianos (eds), Reconciling Efficiency and Equity: A Global Challenge for Competition Policy (Cambridge University Press 2019) 1, 7 53 Lina M Khan, ‘Amazon’s Antitrust Paradox’ (2016) 126(3) Yale Law Journal 710, 737. 54 Ibid, 745. 55 Ibid, 803. 48 49

The ambit of competition law: comments on its goals  23 At a more foundational level, Eleanor Fox summarizes the discussion on goals thus: ‘The debate on goals is a stand-in for a different conversation … What do you, and what should we, as a country, want from markets and antitrust?’56 She posits that the answer is: We want our markets to be robust. We want our businesses to be efficient, effective, lithe, inventive and adaptable to change. We want an environment that will create incentives for businesses to strive the hardest they can in these directions. If businesses successfully do so, they will deliver to buyers and ultimate consumers what they need, want and are able to pay for, including the range of choice they desire; entrepreneurs are likely to have economic opportunity; American business is likely to be competitive in the world; the economy is likely to be strong and grow; and people in general are likely to be better off.57

It is clear from the above that there is no ‘grand unifying theory’ of consumer welfare when it is discussed in antitrust. Eleanor Fox also states on this score: [W]when people say ‘Antitrust is for consumer welfare’ they are obscuring the real debate. I am constantly bemused by the ardent defense of ‘The Consumer Welfare Standard’. There is no such thing. It is not a unified theory. It is a range of standards, and they all devolve from a principle goal of protecting and facilitating the competition process … The fact that ‘consumer welfare’ is an incomplete phrase should not obscure the critical importance of consumer welfare and consumers’ interests in antitrust analysis … Impact of conduct or transactions on people in their capacity as consumers is the first line of analysis, and often but not always the last line of analysis, and if a proposed enforcement should harm consumers, that should disqualify the enforcement.58

Finally, the members of the association of five major emerging national economies (Brazil, Russia, India, China and South Africa) known as BRICS have argued that there is a need for a new conceptual framework for competition law in the context of the digital economy, including: ‘first, call for a re-conceptualisation of the goals of competition law in the digital era, as competition law moves from the calm and predictable waters of “consumer welfare”, narrowly defined, to integrate considerations of income/wealth distribution, privacy and complex equality’.59 BRICS also advocates for the role of competition law in the digital era to be integrated into the broader debate about ‘new processes of value generation and capture in the era of digital capitalism and the complex economy which it has given rise to’.60 Thus, the question of the goals of competition law is currently the subject of fierce discussion in the US and elsewhere. In summary, a growing number of scholars in the US now raise doubts about the overall utility of the prevailing Chicago and post-Chicago Schools of economics to deliver the best outcomes for the country in circumstances where inequality is increasing and new forms of competition – for example, among digital platforms – require further consideration by competition authorities, although not all of the issues they raise can or should be dealt with by competition law. Globally, the issue of inequality/inequity is also on the agenda. Proponents Fox, ‘Against Goals’ (n 6), 2159. Ibid, 2159–60. 58 ‘Interview with Eleanor M Fox’, CPI Talks (7 November 2019). 59 BRICS Competition Law and Policy Centre, Digital Era Competition: A BRICS View (2019) 36, http://​bricscompetition​.org/​materials/​news/​digital​-era​-competition​-brics​-report/​. 60 Ibid. 56 57

24  Research handbook on methods and models of competition law of these alternative approaches have not, however, formulated practical guidelines that would enable their adoption in any clear or reasonably consistent fashion, which means that even if the views are attractive on some level, there is still much work to be done in developing criteria for the application of broader goals.61 The issues relating to multiple goals also apply. B

European Union Law

European competition law has taken a different approach to goals. As noted at the outset, early iterations of the EU law focused on the economic union and competition, with ‘fair competition’ prominently mentioned.62 In the Treaty Establishing the European Community, for example, the ultimate goal of the European Community was ‘the sustainable development of the economy: economic welfare within a single market’. In addressing this goal, a ‘degree of competitiveness’ was specifically mentioned.63 ‘Strengthening of the competitiveness of Community industry’ was identified in the list of activities of the Community in Article 3 ‘as an instrument with which to achieve’ the ultimate goals of Article 2. 64 Article 4 stated that ‘among other things, the activities of Member States … shall be based on the internal market, on the common objective that they shall act in accordance with the principle of an open market economy with free competition’.65 All of these references indicated that competition was very important and that competition provisions should be interpreted in accordance with these tenets. However, the enactment of the Lisbon Treaty, which entered into force in 2009, modified these presumptions. It amended the goals of the EU and deleted the previous reference to undistorted competition as an objective (goal), although the Protocol on the Internal Market still refers to ‘undistorted competition’ as ‘part of the internal market’.66 In practice, these changes may have little effect. Article 3 of the Treaty on European Union still mentions competition. The provisions on objectives, not surprisingly, ‘have been an important source for the interpretation on the Treaty provisions’.67 ‘Fairness’ has always been part of the objective, with the fair distribution of wealth from competitive markets an element of this.68 The approach is founded on the idea of a level playing field where firms can compete on the merits. But while fairness is a guiding principle, in each and every case the Commission looks into the

South Africa is currently preparing guidelines on goals of fairness. See below. This chapter does not revisit the competition laws of individual European countries commencing in the 1920s, particularly in Germany under the Ordo-Liberals. See David Gerber, ‘Comparative Competition Law’ in Mathias Reimann and Reinhard Zimmermann (eds), The Oxford Handbook of Comparative Law (2nd ed, Oxford University Press 2019) 1169, 1181. 63 Treaty establishing the European Community (Nice Consolidated Version), Art 2. 64 Parret (n 12), 62. 65 Article 4; see Parret (n 12). 66 Parret (n 12), 64. 67 Ibid. 68 ‘The aim of the competition rules lies in the assurance of free, fair, undistorted and at the same time effective competition on the common (internal) market.’ See Lubos Tichy, ‘Goals of Union Competition Law on Regulated Markets: Pharmaceutical Industry and Parallel Trade – Comment on Negrinotti’ in Daniel Zimmer (ed), The Goals of Competition Law (Edward Elgar Publishing 2012) 338, 339. 61 62

The ambit of competition law: comments on its goals  25 evidence, conducts rigorous economic analysis and considers the law and guidance provided by the European courts.69 This approach is carried through to the provisions of the competition law itself, with Article 101(3) of the Treaty on the Functioning of the European Union allowing for agreements that would otherwise breach the law if they improve production or distribution, or promote technical and economic progress, while allowing consumers a fair share of the resulting benefit.70 Conduct by an entity with the requisite degree of market power may constitute an abuse if it directly or indirectly imposes unfair purchase or selling prices, or unfair trading conditions.71 In interpreting the law, the EU has a consumer focus but it is demonstrated in a different way from the US, with more of a focus on ‘levelling the playing field’ and ‘protecting rivalry’.72 While the goals of competition law revolve around consumer welfare, they are not limited to it and the goals of fostering an effective competitive structure and efficiency are important.73 The views on the balance between law and economics may differ between regulators, and scholars differ on the ‘optimal balance between law and economics, as well as the latter’s ability to fully reflect the goals and values of European competition law’.74 Identified goals of the EU and its competition law include the previously noted market integration, economic freedom, economic efficiency, industrial policy, justice and non-discrimination.75 So, it has traditionally been a law with broadly focused goals. It is clear that the multitude of goals presents challenges: ‘They represent an amalgamation of values which often overlap but may also reveal friction. Indeed, their implementation calls for trade-offs between norms and may result in varying balancing points and ambiguity.’76 As to economic methodology, it has been said that the EU approach is ‘distinctly post-Chicago’.77 It has also been suggested that the approach taken to the interpretation of EU goals ‘is often different according to whether the subject is studied by lawyers or economists’.78 Lianos has summarized the position in the EU in the following way: after noting that there is ‘largely political consensus in Europe on the benefits of active competition law enforcement’, he states that ‘there is still disagreement over the role and extent of welfare analysis in EU competition law, as opposed to a rights and principles-based approach and the interplay between the different goals of competition law, assuming that one adheres to goals pluralism’.79 In an EU context, ‘fairness’ as a goal has long been criticized, particularly by the US, because it is a subjective term that does not add much to the debate on competition, although, 69 Johannes Laitenberger, former Director-General of the Directorate-General for Competition of the European Commission (20 June 2018), https://​ec​.europa​.eu/​competition/​speeches/​text/​sp2018​_10​_en​ .pdf. 70 Unless the agreement also eliminates competition in a substantial part of the market. 71 Treaty on the Functioning of the European Union, Art 102. 72 James Keyte, ‘Why the Atlantic Divide on Monopoly/Dominance Is So Difficult to Bridge’ (2018) 33(1) Antitrust 113, 114. See also Ezrachi, Discussion Paper (n 2), 3, 6. 73 Ezrachi, ‘Sponge’ (n 43), 10. 74 Ibid, 19. 75 Parret (n 12), 66–74. 76 Ezrachi, Discussion Paper (n 2), 6. 77 Keyte (n 72), 114. 78 Parret (n 12), 64. 79 Ioannis Lianos, ‘Some Reflections on the Question of the Goals of EU Competition Law’ (CLES Working Paper Series 3/2013) 2.

26  Research handbook on methods and models of competition law as has been said, competition law is ‘all about fairness’.80 However, US commentators believe that it injects a dangerous element of uncertainty into the application of competition law.81 Clearly, this depends upon the breadth given to the interpretation of the concept of fairness, and also to the prominence and regularity with which it is applied to decision-making. Fairness has been said to ‘echo a moral norm embodied in the [EU] competition rules’,82 and ‘economic reasoning is … fused into the norm of fairness’.83 It is not, however, to be confused in the EU with the protection of competitors as it is well accepted there that less efficient competitors may be forced from the market: ‘the value of fairness is not used to challenge such legitimate competition.’84 The focus on fairness in practice means that EU decision-makers ‘take into account not just the repercussions on the market, but also the impact on rivals and consumers of the actions under scrutiny’.85 Enforcement actions taken show that the European Commission looks beyond the immediate effects on consumers where appropriate, with the actions against Microsoft, Intel, Qualcomm and Google as examples where organizations have been fined and forced to change their behaviour.86 The EU, then, has a well-established regime recognizing multiple goals. Fairness is an accepted element, although the concept is variable. Inequality is not currently a goal or an element for consideration, although European scholars are heavily involved in inequality debates.

IV

THE GOALS OF COMPETITION LAW(S): A BROADER APPROACH

The fact that the newer competition law jurisdictions, particularly those that could be classified as developing economies, have often taken differing and broader approaches to their goals has been noted.87 While many jurisdictions have developed their own drafting style, provisions are often influenced by the templates of the more established jurisdictions. It is clear, however, that US law has not been such an influential template as has EU law. This is primarily because the EU template provides more detailed guidance to market participants, regulators and the judiciary on the types of conduct that will be caught by the prohibitions. The US law, by way of contrast and as noted earlier, tends to leave much of the work to the courts. 80 Alfonso Lamadrid de Pablo, ‘Competition Law as Fairness’ (2017) 8(3) Journal of European Competition Law and Practice 147: ‘not as a standalone goal but as an outcome of the current law’. See also Sandra Marco Colino, ‘The Antitrust F Word: Fairness Considerations in Competition Law’ (2019) 5, Journal of Business Law 329–45. 81 Nevertheless, it has been adopted in many established Asian jurisdictions and in many newer jurisdictions – as to which, see below. 82 Ezrachi, ‘Sponge’ (n 43), 12. 83 Ibid, 4. Privacy may also be a factor in fairness. 84 Ibid, 14, citing the Post Danmark case: C-209/10 Post Danmark A/S v Konkurrenceradet [2012] ECLI:​EU:​C:​2012:​172, para 20. 85 Marco Colino (n 80), 4. 86 Ibid, 11. 87 See, for example, Paulo Buccirossi and Lorennzo Ciari, ‘Western Balkans and the Design of Effective Competition Law: The Role of Economic, Institutional and Cultural Characteristics’ in Boris Begović and Dušan V Popović (eds), Competition Authorities in South Eastern Europe: Building Institutions in Emerging Markets (Springer Open 2018) 7, 17.

The ambit of competition law: comments on its goals  27 This is disadvantageous to newer jurisdictions in three ways: the competition statute does not signpost the prohibited conduct for those who are not familiar with competition law, and the emphasis on court decision-making is not practical in jurisdictions that lack strong rule of law or that are not equipped with courts that can easily interpret these economic rules. In addition, the US competition law template is not really appropriate for those jurisdictions that adopt an administrative system of enforcement. Even the term ‘competition law’, which is broader than antitrust, requires clarification from a US viewpoint. David Gerber’s view is that it has ‘no commonly accepted usage’ and may refer to laws with different objectives, such as ‘to protect competitors from unfair competition’.88 In considering the goals of these newer laws, one should not assume that the goals of the more established jurisdictions should have been replicated. Several considerations lead to this conclusion. First and foremost, the political economies and legal systems of newer jurisdictions are generally quite different. For example, the more established competition law jurisdictions have functioning markets in the main, but these may not exist in developing economies. But, second, cultural differences dictate that all aspects of antitrust may need to be amended, although ‘the differences are in degree rather than in kind’.89 Issues of political economy, including legal rules and cultures, also mandate that even were laws identical it is unlikely that identical outcomes – even in the very same set of facts and circumstances – would be likely to ensue in different jurisdictions. So, to quote Ezrachi, ‘fragmentation should not be understood as an improper application of competition law’.90 Josef Drexl also summarized the position well when he noted that while competition economics and industrial organization law ‘claim universal validity’, the question for developing nations is really about the role consumer welfare (in a broad sense) should be given in their national competition laws and policies, and that global antitrust should be ‘flexible enough to allow for national variations that are required in the national interest of individual countries’.91 As a general rule, in practice the newer jurisdictions use the more conventional economic goals as a prime basis for their laws, but, in addition, may apply other goals of social welfare and public interest that they judge to be appropriate. This means that there is often a tension in the prioritization and weightings of the various goals. The goals themselves may be simple or they may be broad and varied, but they can all be categorized as factoring in issues that a jurisdiction sees as being of importance to its competitive environment, growth and development, and to its citizens. These additional goals are a manifestation of what the lawmakers determine is best for their economy and their citizens.

Gerber, ‘Comparative Competition Law’ (n 62), 1170. Thomas K Cheng, ‘How Culture May Change Assumptions in Antitrust Policy’ in Ioannis Lianos and D Daniel Sokol (eds), The Global Limits of Competition Law (Stanford Law Books 2012) 205, 205. 90 Ezrachi, ‘Sponge’ (n 43), 53. 91 Josef Drexl, ‘Consumer Welfare and Consumer Harm: Adjusting Competition Law and Policies to the Needs of Developing Jurisdictions’ in Michal S Gal, Mor Bakhoum, Josef Drexl, Eleanor M Fox and David J Gerber (eds), The Economic Characteristics of Developing Jurisdictions: Their Implications for Competition Law (Edward Elgar Publishing 2015) 266, 267. 88 89

28  Research handbook on methods and models of competition law

V

NEWER AND MORE NOVEL MODELS

Thus, the US model of goals of economic efficiency and consumer welfare, and the wider goals of the EU, have given way to an even broader range of goals in other jurisdictions – although, in most, at the core is increased competition to advance consumer welfare at some level. A selection of the goals of newer jurisdictions is discussed below. Some have been chosen for their breadth, some as exhibiting novel goals, some as containing representative goals, and some at random. However, they do underscore the wide diversity of goals currently in operation as part of laws enacted in the name of competition. They are discussed in rough categories below. A

Focus on Purely Economic Goals

Of newer jurisdictions, some focus solely on economic goals although it is not certain how the laws will be interpreted in practice. The Competition Order 2015 of Brunei Darussalam focuses on economic goals – ‘to promote and protect competition in markets in Brunei Darussalam, to promote economic efficiency, economic development and consumer welfare’ and to provide for the functions of the competition regulator.92 Malaysia’s long title and preamble refer to promotion and protection of the process of competition, and the encouragement of efficiency, innovation and entrepreneurship, which promotes competitive prices, quality of products and services, and wider choices for consumers.93 Mexico talks of protection of the competitive process and free market access.94 The laws of Singapore, Sweden, Thailand and Ukraine and the more established law of the UK have goals or contain a preamble that focuses purely on the protection of competition or economic goals that do not specifically mention consumers or aspects of consumer welfare.95 So, solely economic goals still exist in some jurisdictions. B

Multiple Diverse Goals

Looking outside the narrow and more traditional, one of the broadest sets of goals within a competition law is contained in the Anti-Monopoly Law of China (AML). Article 1 of the AML states that the law is enacted for the purpose of preventing and restraining monopolistic conduct (which is defined to include cartels and other anticompetitive agreements, both horizontal and vertical, abuse of dominance and anticompetitive mergers), protecting fair market competition, enhancing economic efficiency, safeguarding the interests of consumers and the interests of society as a whole, and promoting the healthy development of the socialist market economy. This bundle of goals must be viewed against the background of the promotion of the ‘healthy development of the socialist market economy’, in a jurisdiction that regards state-owned enterprises (SOEs) as a foundational part of the economy under its constitution,96

Competition Order 2015 (Brunei Darussalam), s xx. Competition Act 2010 (Malaysia). The Act has no specific clause setting out goals. 94 Federal Economic Law (Mexico), Art 2. 95 Competition Act (Ch 50B) (Singapore), Long Title; Swedish Competition Act (2008: 579), Art 1; Thailand Trade Competition Act BE2560, Preamble; Law of Ukraine on the Protection of Economic Competition (2011), Preamble; Competition Act 1998 (UK), Preamble. 96 Constitution of the PRC, NPC, 1982, Arts 6, 7. 92 93

The ambit of competition law: comments on its goals  29 although the AML does purport to apply to SOEs with some qualifications. The statement of objects and other provisions of the AML also refer to ‘orderly markets’, which are not necessarily competitive markets (where competition is often vigorous and inefficient competitors may be damaged). The AML thus contains elements that give enforcers, rule-makers and jurists substantial discretion to consider issues not related to competition, such as industrial policy, in applying the law.97 The nature of decision-making and the way AML cases are reported in China mean that it is unlikely that goals are discussed in detail and it is difficult to unpick judgments to clarify the approach taken to the goals of the AML in individual cases. C

Promotion of SMEs and Broader Social Issues

Promotion of small and medium-sized enterprises (SMEs) is a common goal of competition law, which ties in with issues of fair competition often raised in Asian jurisdictions, discussed below. It is regularly linked to other goals. Among the selection of broad goals in the South African law, the promotion of SMEs sits with the more common goals of promotion of competition, efficiency, and adaptability and development of the economy. In addition, South Africa has the unique goals of the promotion of employment, the advancement of the social and economic welfare of South Africans, the promotion of a greater share of ownership, and the imperative to increase the ownership stakes of historically disadvantaged people. These broader goals, which introduce an array of public interest concerns and issues, were enshrined in the Competition Act of 1998, despite criticisms at the time from business and more established competition law regimes.98 The purpose was not, however, replicated in the prohibitions of the Act itself, which left ambiguity in the law,99 and over the first 20 years opportunities to fulfil the goal of economic opportunity were missed despite generally excellent enforcement.100 These broader goals have to date been applied mainly in merger review, and the merger factors in South Africa are supplemented by factors including ‘the ability of national industries to compete in international markets’.101 More recent amendments to the law have been described as pushing back even further against a consensus which focussed on consumer welfare to the detriment of the structure of the economy in general and a broader examination of market power … the minister considers that the consumer welfare standard is ill-equipped to deal with the architecture of market

97 See Deborah Healey and Eleanor M Fox, ‘State Restraints in China: A Different Case?’ in Adrian Emch and Wendy Ng (eds), Wang Xiaoye Liber Amicorum: The Pioneer of Competition Law in China (Concurrences 2019) 192. At a recent conference attended by one of the authors in China, a scholar suggested that the proposed amendments to the AML should include broadened goals. China also has an Anti-Unfair Competition Law, which contains some prohibitions in common with the AML, but also a broader range of provisions – including some that focus on consumer protection. It has far less emphasis on economics in its actual application. 98 Section 2. See Dennis M Davis, ‘Public Interest, Industrial Policy and Competition Law Remedies: The South African Experience’ in Paulo Burnier da Silviera and William E Kovacic (eds), Global Competition Enforcement: New Players, New Challenges (Wolters Kluwer 2019) 203, and particularly at 205. The Preamble of the 1998 law is more fulsome again on these issues. 99 Fox, ‘South Africa, Competition Law and Equality’ (n 5), 8. 100 Ibid, 3. 101 Competition Law (South Africa), s 12A(3).

30  Research handbook on methods and models of competition law power. In short, exclusive concerns about price and output are insufficient to deal with the skewed structure of a modern economy which perpetuates inequality and exclusion.102

It is not possible to address the amendments enacted in February 2019 fully here, but in relation to mergers the concept of transformation – using mergers to provide business opportunities for disadvantaged people – is an important element, particularly in the face of the critical problem of skewed ownership. Of the other amendments, ‘[m]uch of the pricing behaviour identified is unfair exploitation made possible by grossly unequal bargaining power. The authorities face significant challenges to identify “unfairness” in a manner that has standards and is justiciable.’ Commission Guidelines on the issue of ‘unfairness’ are forthcoming.103 Interestingly, in the context of public benefit and while Australia has a well-established competition law with more traditional specific economic goals (‘enhance the welfare of Australians through the promotion of competition and fair trading and provision for consumer protection’),104 the Competition and Consumer Act 2010 (Cth) contains provisions allowing the Australian Competition and Consumer Commission to grant administrative permission for a specified period (usually up to five years, with potential for review and continuation) for specific conduct (following individual application) if the conduct is likely to substantially lessen competition but this is outweighed by likely public benefits.105 This process was included in the law expressly in recognition of areas of discontinuity in markets where competition might not work effectively.106 Indonesia, a newer but well-established competition law jurisdiction, sets a diverse range of goals for its competition law. Some individual goals are economic – to ‘prevent monopolistic conduct and/or unfair conduct’, ‘create healthy business competition’ and ‘create effectiveness and efficiency in business transactions’. In the conduct of business activities in Indonesia,

Davis (n 98), 214. Fox, ‘South Africa, Competition Law and Equality’ (n 5), 6. 104 Competition and Consumer Act 2010 (CCA), s 2. The Australian mention of consumer protection must also be viewed in the context of the inclusion in the CCA of Australia’s provisions on consumer protection and the Australian Competition and Consumer Commission being the regulator for both competition and consumer protection laws. This provision has been criticized for failing to prioritize the individual objects: ‘If the objects are to be pursued, there must be some process for mutual qualification or reconciliation applied.’ See JD Heydon, ‘Is the Competition and Consumer Act 2010 (Cth) in Competition with Itself?’ (Paper presented at the 2013 Competition Law Conference, Sydney, 4 May 2013) 2, referred to in SG Corones, Competition Law in Australia (6th edn, Lawbook Co 2014) 33. New Zealand, Peru and Poland strike a similar balance to that of Australia without mentioning consumer protection, and without the authorization process mentioned below. See Commerce Act 1986 (New Zealand), Art 1A; Act of 16 February 2007 on Competition and Consumer Protection (Poland), Art 1. 105 CCA, s 90(7). In the context of mergers in a little-used provision under the authorization provisions, the public benefit in a significant increase in the real value of exports, the substitution of domestic goods for imported goods, and ‘all other relevant matters that relate to the international competitiveness of any Australian industry’ (s 90(9A)) are specifically designated as public benefits, in addition to the public benefits relevant to other provisions. Public benefits are unspecified in the CCA in relation to the other provisions of the CCA, but the Australian Competition and Consumer Commission has accepted a wide range of public benefits in its determinations. 106 Independent Committee of Inquiry into Competition Policy in Australia, National Competition Policy: Report by the Independent Committee of Inquiry into Competition Policy in Australia (1993) (Hilmer Review), http://​ncp​.ncc​.gov​.au/​docs/​National​%20Competition​%20Policy​%20Review​ %20report​,​%20The​%20Hilmer​%20Report​,​%20August​%201993​.pdf. 102 103

The ambit of competition law: comments on its goals  31 the goals call for ‘economic democracy by considering the balances between entrepreneurs’ interest and public interest’. The goals also refer to ‘securing equal business opportunities for large, middle and small class entrepreneurs’, which goes back to protection for SMEs but includes other businesses as well.107 It is unclear whether this express protection takes these goals further than levelling the playing field, which would protect other businesses of all sizes against conduct that breaches the law. The Philippines Competition Law has a very fulsome expression of goals in its section 2, entitled ‘Declaration of Policy’. In summary, the first part of the provision refers to ‘market competition as a mechanism for allocating goods and services’ and recognizes that ‘liberalisation of key sectors needs reinforcement by measures safeguarding competitive competitions’, serving ‘the interests of consumers’, so the provision contains key economic underpinnings. The provision then refers to equal opportunities for the promotion of entrepreneurial spirit, the encouragement of private investment, the facilitation of technological development, and resource productivity. Relying on what it refers to as ‘key constitutional goals’, which are set out, it provides that the state shall enhance economic efficiency and promote fair competition, establish a national competition policy,108 prevent economic concentration with anticompetitive impact, and penalize other forms of anticompetitive agreements (in summary) ‘with the objective of protecting consumer welfare and advancing domestic and international trade and economic development’.109 D

Creation of a Single Economic Market

The creation of a single economic market is often listed as a goal in jurisdictions where the state previously had, or continues to have, a large role in markets at central, provincial or local levels. Russia focuses on the protection of competition by prohibiting the usual range of anticompetitive conduct that is nominated in Article 1, entitled ‘Subject and Goals of this Federal Law’.110 Article 1 also talks of the ‘restriction or removal of competition’ by a wide range of government bodies, including the federal executive, states, local government, ‘state off-budget funds and the Central Bank of the Russian Federation’. Finally, the law talks of ensuring a common free market zone, free movement of commodities, protection of competition and creation of conditions for the efficient functioning of commodity markets. This means that this competition law is aimed not only at the private sector and SOEs, but also at a variety of other government bodies, and it includes provisions focusing on the creation of a single market within Russia.

107 Law of the Republic of Indonesia No 5 of 1999 Concerning the Ban on Monopolistic Practices and Unfair Business Trade, Arts 2, 3. 108 Which it has done: see Deborah Healey, ‘Competitive Neutrality and the Role of Competition Authorities: A Glance at Experiences in Europe and Asia-Pacific’ in Paulo Burnier da Silveira and William Evan Kovacic (eds), Global Competition Enforcement: New Players, New Challenges (Wolters Kluwer 2019) 187, 193. 109 The Philippines Development Plan (2017–2022), a broad economic initiative, supports the law and other policies aimed at bolstering competition. See Healey, ‘Competitive Neutrality and the Role of Competition Authorities’ (n 108). 110 See Federal Law No 135-Fz of 26 July 2006 on Protection of Competition (Russia).

32  Research handbook on methods and models of competition law This single market imperative is well established in the EU, as previously discussed, but is also contained in the ‘Administrative Monopoly’ provisions of the AML of China,111 although these AML provisions are not referred to in the clause setting out the broad AML goals, which is noted above. E

Economic Development

In several developing jurisdictions, such as India, the goals are focused on the protection of competition and consumers but are prefaced by qualifications relating to the economic development of the country.112 Some of the jurisdictions previously mentioned also include goals around economic development. F Fairness As noted previously, fairness or fair competition as a goal is rejected in the US. Other jurisdictions, such as the EU and particularly those in Asia, have goals of fairness or ‘fair competition’. These include the more established competition jurisdictions of Taiwan, Japan and Korea.113 But it is not completely clear what ‘fair competition’ involves – it appears to have more than one meaning. It may mean fairness in the sense of rejecting conduct that is anticompetitive under the law, giving an unfair advantage over other market players; hence, the elimination of the prohibited conduct levels the playing field for ‘fair’ competition. Alternatively, it might mean fair competition in the sense of some broader conception of fairness, which would operate in parallel or in addition to other goals, be they economic or otherwise. For example, Vietnam talks of the law ‘providing for competition-restricting acts, unfair competition acts’, which seems to suggest a broader meaning for the term ‘unfair’.114 Cambodia’s draft Law on Competition 2018 talks in Article 1 (Purpose) of ‘encouraging fair business relations’,115 along with other more usual economic goals such as promoting economic efficiency, but also promoting the establishment of new business. Laos also mentions ‘lawful and fair competition’ and the prevention of ‘unfair business competition’. These three laws seem to inject a broader element of unfairness, although again they may be referring to fairness in the sense that a corporation breaching the competition law has an unfair business advantage. Some goals in the law of Laos are more economic in nature, but the law contains more unusual elements, such as the ‘protection of the interests of the State’ and the ‘sustainability of the national socioeconomic development’, elements which are reminiscent of the AML of China. Myanmar focuses on acts that injure the public interest through breaches of its competition provisions, but also refers to ‘endangering free competition in economic activities for the purposes of national economic development’, which suggests that it goes further than just being about non-compliance with the law. In this context, the Myanmar law also mentions the ‘control of unfair market competition’ in wording that seems merely to use that term in the context of breach of the spe-

113 114 115 111 112

AML (China), Art 8, Ch 5. See Competition Act 2003 (India), Preamble. Section 2 of the Competition Law of Latvia 2001 also talks of ‘free, fair and equal competition’. Competition Law No 23/2018 (Vietnam), effective 1 July 2019. Article 1.

The ambit of competition law: comments on its goals  33 cific provisions of the law, so without a broader meaning being given to ‘unfair’.116 In respect of these last jurisdictions, which are all ASEAN member countries, it is useful to note that the ASEAN policies appear to see the elimination of anticompetitive conduct itself as producing ‘fair competition’.117 Korea and Japan, which have longer-established and well-enforced competition laws, both refer to the promotion of ‘free and fair competition’, as well as the ‘development of the national economy’ and other economic goals.118 Japanese law has some other similarities with the AML of China, which no doubt was influenced by Japanese law. In addition to free and fair competition, the Japanese law states that its purpose is to ‘stimulate the creative initiative of enterprise, encourage business activity, heighten the level of employment and actual national income, and thereby promote the democratic and wholesome development of the national economy as well as secure the interests of general consumers’. The detailed objects clause also talks of ‘unreasonable’ restraints and ‘unjust’ restrictions. It thus emphasizes the development of the national economy in the context of its democratic underpinnings, in addition to the usual economic goals. Commentary on the law of Japan is instructive. Iwakazu Takahashi has described reference to the development of the national economy as the ‘ultimate object’ of the Act, with ‘fair and free competition’ being the ‘direct object of the Act’. Takahashi states: ‘Fair competition’ means competition not using unfair methods such as dumping, boycotting, discriminatory pricing, etc, and ‘free competition’ means the freedom to compete in a market, to decide conditions of transaction, to enter into a market, etc. The ‘democratic and wholesome development of the national economy’ includes the achievement of ‘economic efficiency’.119

The author concludes: the relationship between the direct object and the ultimate object in Article 1 is that the achievement of ‘democratic and wholesome development of the national economy’, including the achievement of economic efficiency and consumer interest, is the economic result of fair and free competition. By understanding the relationship of both words in this way, there is no collision between ‘freedom’ and ‘economic efficiency’ or ‘consumer benefit’ in the Japanese Antimonopoly Act.120

This detailed explanation of the wording of the goals of the Japanese competition law clarifies that ultimately the detailed wording is not too far removed from the purely economic approach to goals envisaged in the US and itself appears to emphasize that failure to comply with the law creates circumstances of unfairness in relation to other market participants.

Competition Law 2015 (Myanmar) Ch II. See ASEAN, Regional Guidelines (n 19) and relevant commentary. 118 Monopoly Regulation and Fair-Trade Act (Korea), Art 1; Act on Prohibition of Private Monopolization and Maintenance of Fair Trade (Act No 54 of 14 April 1947) (Japan), Art 1. 119 Iwakazu Takahashi, ‘On the Difference of Methodology in Jurisprudence and Economics – Comment on Kunzler’ in Daniel Zimmer (ed), The Goals of Competition Law (Edward Elgar Publishing 2012) 217. 120 Ibid. 116 117

34  Research handbook on methods and models of competition law

VI

IMPLICATIONS OF BROADER GOALS

It is difficult to cleanly group the goals of competition laws. The material below looks at the idea of multiple goals from three separate perspectives under which different questions arise. A

The US Position

To recap, the criticism in the US of the choice of a single goal and its current economic methodology is strident but, for a number of reasons, there are equal numbers of advocates for its maintenance. Advocates believe that it just makes economic sense to continue with a system which they say provides clear economic answers in most situations. They also argue that the current approach is effective to deal with new problems that arise in markets. In contrast, and as has been discussed above, many scholars advocate for a broader approach to goals, which would raise the issues discussed below. B

Multiple Goals

It is difficult to draw definitive conclusions but at this point there seem to be more laws with multiple goals than there are with a single goal. Several issues arise from this. Multiple objectives without direction on priority risk conflicts and inconsistent application of the law, particularly in newer jurisdictions: ‘The interests of different stakeholders may severely constrain the independence of competition policy authorities, lead to political intervention and compromise and adversely affect one of the major benefits of the competitive process, namely economic efficiency.’121 While conflicts between economic efficiency and other goals may be balanced by thoughtful decision-making, the impact on competition overall will depend upon the frequency with which the other goals are applied. If other goals are preferenced over competition in a small number of situations, this is not likely to detract significantly from market competition, but if other goals are regularly prioritized, the position will be different and the potential benefits arising from an effective competition law are likely to be diluted. Objects other than the encouragement of competition in order to promote the efficient use of resources are categorized as ‘supplementary objectives’ by the Organisation for Economic Co-operation and Development (OECD), which has also stated: ‘It has been noted that these “supplementary objectives” tend to vary across jurisdictions and over time. The latter reflects the changing nature and adaptability of competition policy to address current concerns of society while remaining steadfast to the basic objectives.’122

121 OECD Global Forum on Competition, ‘The Objectives of Competition Law and Policy, Note by Secretariat’ (29 January 2003, CCNM/GF/COMP (2003)3) 2–3, http://​www​.oecd​.org/​daf/​competition/​ 2486329​.pdf. 122 Ibid, 2.

The ambit of competition law: comments on its goals  35 C

Legal Certainty

Legal certainty is a key element of any successful law but certainty within a particular jurisdiction is the measurement in most other areas of legal regulation.123 While it can be forcefully argued that certainty and uniformity within a jurisdiction are crucial, it is far less likely that decisions made across jurisdictions in competition law will concur – although no doubt it would be more convenient for multinational corporations were this to be the case. The reliance on economic analysis in competition law simplistically suggests that outcomes in similar circumstances should be the same or similar. Increased globalization and cross-border transactions highlight differences in the roles and application of competition law in different jurisdictions, which may not be as obvious in other legal areas. But it is important to emphasize that different legal priorities afforded competition law, as well as widely diverging political economies and legal systems, stage of development, importance of rule of law, and other national economic priorities, all suggest that outcomes are likely to be divergent. To suggest that there is a well-defined and superior template provided by a purely economic model neglects important features of both law and economics. D Fairness The debate about the role of fairness in competition law has intensified over the last few years. Attainment of procedural fairness in the sense of due process has become more important worldwide, driven largely by the US.124 The discussion on fairness in the current context can be cast along four main trajectories. The first is the traditional view in the US, where Chicago economists have no interest in, or see no need for, the introduction of a fairness goal. The second, again in the US, favours the adoption of fairness as an overarching mantra by those advocating for broader change to antitrust goals and standards within the jurisdiction. The third trajectory is that of the EU, where it is already recognized as one of a number of goals of competition law and is applied in appropriate circumstances. Finally, the fourth path sees fairness in the form of ‘fair competition’ forming part of the goals of newer regimes. Each of these four trajectories takes a different view of fairness as a goal, with differing implications. But, more generally, the issue of what constitutes a fair outcome in a case and whether this should even be an issue for competition law has been discussed at length, with the issue largely unresolved. However, one approach has been expressed by Professor Eleanor Fox: ‘Equity in antitrust means tilting the scales away from powerful business and for powerless people, away from protecting the (freedom of) the incumbent and toward protecting the access and opportunity of young challengers.’125 This is particularly true in the context of developing economies. The question of what would constitute a reliable and predictable template for a test of fairness is also problematic. The concept of what ‘fair competition’ means has been raised in the

Ezrachi, ‘Sponge’ (n 43), 69. See, for example, discussion of current developments in D Daniel Sokol, ‘The New Procedural Fairness in Competition Law: Global Developments’ (2019) 10(4) Journal of European Competition Law & Practice 197–8,  https://​doi​.org/​10​.1093/​jeclap/​lpz016. 125 See Fox, ‘Competition Policy at the Equity-Efficiency Intersection’ (n 22), 443. 123 124

36  Research handbook on methods and models of competition law discussion of newer competition laws, but the answer to what this actually involves appears to vary between jurisdictions and circumstances. Whether the use of fairness means merely that issues other than economic efficiency should be relevant to decision-making, whether the use of the term ‘fair’ merely refers to markets where all market participants are observing competition law, or whether it is an additional factor that has a different meaning is unresolved. Decisions of the newer jurisdictions are unlikely to enunciate views on the concept of fairness that they apply in written decisions, which means that it will take some time for the impact of decision-making in these areas to enlighten the issue. E Inequality The adoption of a goal of addressing inequality that would apply equality as an overarching goal or as one of several goals is intriguing as it may divert focus from the traditional economic foundations of competition law. Proponents argue that current economic approach(es) do not give the appropriate weighting to efficiency in markets allowing the growth of ever-larger corporations that hold very substantial market power, and that the benefits of efficiencies in this context do not flow through to consumers, thus entrenching inequality. Addressing inequality as a goal of competition law requires further detailed examination and significant research, and the development of an appropriate and effective framework, before it can be fully considered.

VII CONCLUSIONS Most laws reflect the norms of the society in which they operate. Competition law introduces economics as a foundational tool of interpretation and application. The stage of development and views on the value of competition itself affect the prioritization of the law in individual jurisdictions. Markets are dynamic. Conduct and the views of its impact on competition change over time. The foregoing material shows that there have never been truly uniform goals for competition law globally, nor general agreement as to how laws should be interpreted or to whom they should apply. Far from being uniform, the issues arising in relation to goals differ as between jurisdictions, although there is some overlap of ideas and discussion. In the US, questions relate to the goal of consumer welfare and whether it should be broadened with resultant amendment to economic methodology, both generally and particularly considering ongoing market developments such as digitalization and fears about growing inequality. In the EU, where multiple goals are more accepted, current questions appear to focus more on issues of inequality and whether fostering equality should be incorporated in some way into competition law analysis. In other jurisdictions, particularly those of later adopters, the issues are about the priority of goals, their interpretation and how often non-economic goals will be applied. There is little indication of approach from decided cases at this point. However, the goals of most jurisdictions focus primarily on the economic – expressed in a variety of ways – and in many jurisdictions the other goals pan out from that initial focus. The goals of a number of jurisdictions specifically mention consumers, and a number mention the development of the domestic economy or a competitive economy more generally. The often unwritten and unifying feature, then, is that all believe that effective competition will bring benefits to their economy, which will flow through to consumers.

The ambit of competition law: comments on its goals  37 It is expected that the current arguments about approaches to goals and methodology in their various forms will continue for quite some time. How goals are implemented and tested in newer jurisdictions will continue to develop and may yet shine light for the more traditional jurisdictions to follow.

3. The cost of progress: hurdles facing antitrust’s economic advance Alan Devlin

I INTRODUCTION Empirical methodologies have honed antitrust into a precision instrument.1 Those tools help agencies to distinguish procompetitive and anticompetitive acquisitions, estimate competitive effects, predict consumer injury, and fashion remedies that allow otherwise efficient mergers to proceed.2 Intervention decisions by US enforcement agencies are increasingly data focused and strictly policed by the judiciary, which has refined antitrust doctrine into a nuanced body of law. In the United States – if not elsewhere – competition law has become ‘a branch of applied economics’.3 Anyone doubting that proposition need merely peruse this book, which explores the nuances of contemporary antitrust policy. Economic progress has transformed competition law, but the march of progress has not always been smooth. The modern trend towards empiricism creates a complex interplay between economic theory, legal doctrine, and hypothesis testing. The difficulty lies in a knowledge problem. Retrospective studies can reliably estimate the price effects of some consummated mergers.4 They can also allow economists to test and refine their predictive models.5 Similarly, with appropriate data, industrial-organization experts can build merger simulations that predict a deal’s market impact.6 But antitrust’s evidentiary content is far from complete. Statistical issues limit the inferences that one can reliably draw even from robust data. Unsurprisingly, there are myriad questions to which empirical answers remain elusive.

1 Although this observation holds true for the modern application of competition law across jurisdictions, this chapter focuses on the US experience. The views expressed in this chapter belong to the author alone, and should not be attributed to Latham & Watkins, its clients, lawyers or staff. 2 See generally ABA Section of Antitrust Law, Econometrics: Legal, Practical, and Technical Issues (2nd edn, American Bar Association 2014); Jonathan B Baker and Daniel L Rubinfeld, ‘Empirical Methods Used in Antitrust Litigation: A Review and Critique’ (1999) 1 American Law and Economics Review 386. 3 Richard A Posner, The Problematics of Moral and Legal Theory (Belknap Press 1999) 229. 4 See, for example, In re Evanston Northwestern Healthcare Corp, Docket No 9315, Opinion of the Commission (6 August 2007), https://​www​.ftc​.gov/​sites/​default/​files/​documents/​cases/​2007/​08/​ 070806opinion​.pdf. 5 See, for example, Daniel Greenfield, Nicholas Kreisle and Mark Williams, Simulating a Homogeneous Product Merger: A Case Study on Model Fit and Performance (Working Paper No 327, October 2015) https://​www​.ftc​.gov/​system/​files/​documents/​reports/​simulating​-homogenous​-product​ -merger​-case​-study​-model​-fit​-performance/​rwp​_327​.pdf. 6 See, for example, United States v H & R Block, Inc, 833 F Supp 2d 36, 64, 71, 86–8 (DDC 2011). But cf United States v Oracle Corp, 331 F Supp 2d 1098, 1169–70 (ND Cal 2004) (rejecting government economists’ proffered merger simulation as unreliable).

38

The cost of progress: hurdles facing antitrust’s economic advance  39 Dominant-firm conduct remains particularly challenging where enforcers routinely weigh the static price benefits of intervention against the dynamic costs that may inure from diminished incentives. Those long-term effects are theorized, unobservable, and thus impossible to reduce to a utilitarian calculus to weigh against positive short-term effects. Yet, dynamic efficiencies are critically important to social welfare and hence warrant emphasis. Even in the more data-focused area of merger review, important knowledge gaps remain. For instance, there is no consensus about how marginal changes in structural concentration affect competitive outcomes across industries.7 Yet, contemporary merger-review principles assume that a delta in market concentration sheds light on competitive effects.8 There is a firm evidentiary basis for that view in certain markets, like those for generic drugs. Nevertheless, the empirical foundation for a broader structure–conduct–performance causal chain is limited.9 This is all to say that antitrust’s evidentiary core is growing, but not yet fully mature. Where empirical techniques cannot answer the antitrust question, theory rises to the fore. Modern industrial organization proffers myriad price-theoretic and game-theoretic models founded on various assumptions. In each case, agencies try to employ the most suitable model, which is the one most likely to produce accurate predictions. But models bearing distinct assumptions often yield divergent outcomes such that predictive modelling is a tricky and potentially unreliable process. As empirical testing becomes more sophisticated and carries a broader reach, economists can refine their deductive theories. Simultaneously, models become defunct when they predict effects that econometricians can measure or estimate using available data. In short, theories ebb when evidence of actual effects flows. Of course, their borders are porous – if not imprecise – because statistical challenges routinely limit the inferences that one can reliably draw from econometric work. Complicating the relationship between theory and evidence is the law itself, which evolves over time to reflect the latest learning. But the lag between economic thinking and doctrinal law can be pronounced. Although private-firm litigation has played a uniquely important role in the evolution of US antitrust law, it has traditionally fallen to the Department of Justice (DOJ) and the Federal Trade Commission (FTC) to develop the law in a more economically robust direction. There lies a tension born of advances in modern antitrust economics. To a degree, econometric techniques, empirical learning, and economic theories have outgrown the substantive law that the agencies enforce. Among other difficulties, the substantive analysis employed within an agency to justify intervention often differs from the proof offered up in litigation to establish a violation. Specifically, structuralist inferences embedded in still-good law provide 7 See Dennis W Carlton and Jeffrey M Perloff, Modern Industrial Organization (4th edn, Pearson 2004). 8 See, for example, US Department of Justice and Federal Trade Commission, Horizontal Merger Guidelines, 5.3 (2010); European Commission, Guidelines on the Assessment of Horizontal Mergers under the Council Regulation on the Control of Concentrations between Undertakings 2004 OJ C 31/5, [14]–[21]. 9 See, for example, Samuel Peltzman, ‘The Gains and Losses from Industrial Concentration’ (1977) 20(2) Journal of Law and Economics 229; Harold Demsetz, ‘Industry Structure, Market Rivalry, and Public Policy’ (1973) 16 Journal of Law and Economics 1. For recent discussion of this point by contemporary policymakers, see Maureen K Ohlhausen, ‘Does the U.S. Economy Lack Competition?’ (2016) 1 Criterion Journal on Innovation 47; Joshua Wright, ‘Concentration in an Industry Does Not Mean It Lacks Competition’ New York Times (14 November 2016). But cf, for example, Richard Schmalensee, ‘Inter-Industry Studies of Structure and Performance’ in Richard Schmalensee and Robert D Willig (eds), 2 Handbook of Industrial Organization (North-Holland 1989).

40  Research handbook on methods and models of competition law the DOJ and the FTC with a significant advantage in court.10 Understandably, the agencies are not enthused to abandon the ‘Philadelphia National Bank presumption’, not least when they face courts all too eager to put them to their proof and rule against them when necessary.11 And the PNB presumption might benefit the public by allowing the DOJ and the FTC more readily to block deals that they deem problematic based on a more refined economic analysis. But that possible benefit bears costs, including to the integrity of an economic approach, that, if justified, should be universalized. Further, disparate analysis within an agency and in court may dilute judicial scrutiny properly brought to bear on enforcers that would – over time – yield superior outcomes. Empirical advances thus put some doctrinal considerations in tension. More importantly, they threaten to obscure basic understandings that inform contemporary policy. Among other issues, for example, should an unwavering commitment to an evidentiary approach require substantiating any inference drawn from structural considerations? And is today’s approach too narrow, overlooking pertinent economic effects? Recent years have seen widespread calls for enforcers to adopt a harder line against mergers that increase concentration and to clamp down on firms that enjoy entrenched dominance. Such calls, which may reflect populist considerations, put ongoing economic advances into question. This chapter explores how antitrust methodologies have produced an increasingly empirical competition policy. Above all, however, it explains that those empirical advances come at a price.

II

TOWARDS METHODOLOGICAL RIGOUR: HOW ECONOMICS TRANSFORMED ANTITRUST

Modern competition policy is more rigorous and evidence based than ever before due to advances in industrial-organization economics. Nevertheless, it is important to recognize the contextual element. To measure the utility of a methodology, we need to know the end to which an agency wields a methodology. Unfortunately, enforcers approach antitrust issues differently. The scope of this book is, of course, global. But the problems that emerge from methodological advances are local, due to a lack of antitrust conformity at the universal level. Hence, here in Section II, this chapter first observes the differences that fracture symmetric enforcement across jurisdictions. Choosing US law as its test subject, Section II then traces the development in antitrust economics to the modern day. It concludes by illustrating the real-world benefits of those economic advances by reference to recent cases. Section III, however, reveals the important conceptual and practical difficulties that have emerged as the price of success. Hence, even within a single jurisdiction that has zeroed in on accepted principles and guiding norms, new empirical methodologies raise their own challenges.

See United States v Philadelphia National Bank, 374 US 321, 364 (1963). Two district courts in the past year, for instance, ruled against the FTC in economically sophisticated challenges to healthcare mergers. The appellate courts reversed in both instances based on economic errors: Federal Trade Commission v Advocate Health Care Network, 841 F 3d 460 (7th Cir 2016); Federal Trade Commission v Penn State Hershey Medical Center, 838 F 3d 327 (3d Cir 2016). 10 11

The cost of progress: hurdles facing antitrust’s economic advance  41 A

Applying Competition Policy to What End? Why We Need Clarity of Goals Before Evaluating Antitrust Methodologies

Despite convergence achieved through the International Competition Network, the Organisation for Economic Co-operation and Development, and other inter-agency forums, as well as the American Bar Association  Section of Antitrust Law, global competition law remains fractured. Antitrust enforcers pursue distinct goals, employ dissimilar legal standards, and use divergent methodologies across countries. So, too, certain institutional designs raise problems of due process, transparency, and accountability.12 None of this is surprising. The rapid proliferation of antitrust laws to encompass more than 140 countries today has produced an abundance of relatively inexperienced agencies responsible for developing modern competition policies in short order. Competition law is analytically complex. But many states complicate matters further by infusing their antitrust objectives with the pursuit of industrial policies. China expressly looks to its Anti-Monopoly Law to develop its ‘socialist market economy’.13 South Africa wields its antitrust laws with an eye to promoting employment.14 And, perhaps most importantly, the European Union uses its competition laws to integrate the single market.15 Few jurisdictions employ their antitrust rules exclusively to protect the competitive process as an end in itself. The result is a set of normative principles that, even for an individual state, resist cohesion. Muddled competition objectives, replete with conflicting goals that enforcers must weigh on an ad hoc basis, inevitably arise. There is agreement, at a high level of abstraction, that consumer welfare drives antitrust policy.16 That reality feeds oft-repeated claims of consensus and convergence. But proclaimed adherence to that principle is itself misleading, not least because consumer welfare is a malleable concept. It is subject to plausible readings that can support irreconcilable enforcement decisions. An agency could justify its intervening in a market to require a dominant firm’s break-up, or to impose draconian sharing obligations, by arguing that a less concentrated industry structure will benefit consumers through lower prices. But an enforcer inclined to stay its hand in the same circumstances would appeal to the same metric. It might claim that not intervening will protect returns on investment that drive firms to compete in the first place, thus benefitting consumers over time. ‘Consumer welfare’ depends on contestable priors not only about static and dynamic effects, but also about which market features qualify in the cal12 Contemporary due-process concerns are perhaps most often raised with respect to certain emerging antitrust jurisdictions in Asia. See, for example, US–China Business Council, Competition Policy and Enforcement in China (2014). Still, the framework of EU competition enforcement has drawn criticism for combining investigatory, prosecutorial and adjudicative functions in a single body. See, for example, Ian S Forrester, ‘Due Process in EC Competition Cases: A Distinguished Institution with Flawed Procedures’ (2009) 34 European Law Review 817. The US Federal Trade Commission’s administrative-litigation process, known as Part 3, has faced similar criticism, though the latest evidence suggests that critics may have exaggerated its due-process limitations. See Maureen K Ohlhausen, ‘Administrative Litigation at the FTC: Effective Tool for Developing the Law or Rubber Stamp?’ (2016) 12(4) Journal of Competition Law & Economics 623. 13 Anti-Monopoly Law of the People’s Republic of China, Art 1. 14 The Competition Act, No 89 of 1998, Ch 1, [2(c)]. 15 See, for example, Joined Cases 56 and 58/64, Consten and Grundig – Etablissements Consten SA and Grundig-Verkaufs-GmbH v EEC Commission [1966] ECR 299, [516]–[517]. 16 See, for example, Reiter v Sonotone, 442 US 330, 343 (1979); Case C-209/10, Post Danmark A/S v Konkurrencerådet [2012] ECR I-172, [20], [22], [24], [42], [44].

42  Research handbook on methods and models of competition law culus of creditable anticompetitive effects. Antitrust’s lodestar can thus become the end point of the analysis – that is, a conclusion – rather than the guiding metric. Worse, most countries do not even agree on the meaning of competition policy. In the author’s view – consistent with the modern US approach – antitrust has an exclusive purview. It is to protect the supply-side and demand-side market constraints that spurs firms both to respond to consumer demand and cabin their pricing and output decisions.17 In that respect, competition policy is purely market focused. It supplements market processes to ensure their freedom from artificial restraints of trade, exclusionary practices, and power-enhancing mergers. But it does not replace them. If antitrust jumps beyond firm conduct to define violations by the desirability of market outcomes, it ceases to be a competition mechanism and becomes regulatory. That is why US law never finds an antitrust violation based on a high price – no matter how egregious – but focuses only on conduct that lifts a material market constraint.18 Similarly, it imposes liability on a monopolist for refusing to deal in vanishingly few circumstances.19 In each case, the question is whether the accused firm erased a limit that its rivals imposed on its market power. There is, however, no universal consensus on this core point. In fact, most countries interpret their competition policies in contravention of that principle. For many, the prevailing standard for judging the legality of dominant-firm practices is whether an ‘abuse’ occurred – an enquiry that goes beyond anticompetitive effects to capture exploitation. EU competition law thus finds violations based on ‘unfair purchase or selling prices’.20 Although the European Commission has exercised commendable restraint in wielding that provision, it remains a basic principle of EU law. Indeed, Margrethe Vestager reaffirmed in December 2016 that the Commission prohibits certain market-pricing decisions.21 She highlighted excessive pricing in gas supply, the pharmaceutical industry, and standard-essential technologies as good candidates for direct intervention. This point matters for two reasons. First, as a firmly established element of EU law, it reveals an inconsistency at the heart of antitrust. The world’s two most important antitrust jurisdictions answer differently the most basic question – namely, what defines a violation of competition law. Second, EU law today is probably the world’s most influential competition regime. Emerging antitrust jurisdictions look disproportionately to Europe for guidance. Most adopt similar limits on dominant-firm pricing.22 As they have done so, however, few have been as restrained as the European Commission in avoiding regulatory intervention under the rubric of ‘competition enforcement’. We have seen the Anti-Monopoly Law agencies in China 17 Accord Maureen K Ohlhausen, ‘What Are We Talking about When We Talk about Antitrust?’ (Remarks at the Concurrences Review Dinner, 22 September 2016). 18 See Verizon Communications Inc v Law Offices of Curtis V Trinko, 540 US 398, 407 (2004). 19 Ibid 409. 20 Treaty on the Functioning of the European Union (TFEU), Art 102(a). 21 Margrethe Vestager, ‘Protecting Consumers from Exploitation’ (Chillin’Competition Conference, 21 November 2016), http://​ec​.europa​.eu/​commission/​2014​-2019/​vestager/​announcements/​protecting​ -consumers​-exploitation​_en. 22 Cf TFEU, Art 102(a) and Case 27/76, United Brands Co v Commission of the European Communities [1978] ECR 207, [250] with Anti-Monopoly Law of the People’s Republic of China, Ch III, Art 17(1); Indian Competition Act, Ch II, [4(a)]; Korean Monopoly Regulation and Fair Trade Act, Art 3-2(1); and Japanese Act on Prohibition of Private Monopolization and Maintenance of Fair Trade, Ch I, Art II (9)(vi)(b). Cf Mexican Federal Economic Competition Law, Book One, Title I, Art 9; Book Two, Ch 3, Art 56.

The cost of progress: hurdles facing antitrust’s economic advance  43 and the Korea Fair Trade Commission issue antitrust-IP guidelines (some in draft form) that require owners of valuable patents to license them.23 The Competition Commission of India has looked askance at a patentee’s negotiating royalties based on end-device prices.24 And competition agencies the world over deploy antitrust law to condemn potential breaches of contract in the standard-setting space, without first identifying substitute technologies displanted by deception or other anticompetitive conduct.25 US antitrust law under the Sherman Act is largely, if not completely, at odds with those enforcement actions.26 This is all to say that even the most refined empirical tools provide only answers. They do not identify the right questions. For that reason, it is perhaps not feasible – and certainly less productive – to attempt to chart the course of methodological improvement in antitrust enforcement on a global scale. Hence, this chapter looks to the world’s most experienced and economics-focused antitrust jurisdiction, the United States. Even there, it is most unclear that the tale of economic advance is inevitable. For reasons explained below, ‘ever forward’ may not be on the cards. B

The Evolution of US Antitrust Economics

For those versed in antitrust economics, there is much to celebrate in competition policy’s evolution since the 1970s. The Harvard School’s approach rooted itself in the structure–conduct– performance (SCP) paradigm through the 1960s.27 Unfortunately, it produced a body of law

23 State Council Anti-Monopoly Commission, Anti-Monopoly Guidelines on Intellectual Property Abuse, III (ii) 2 (2015); State Administration for Industry and Commerce, Regulation on the Prohibition of Conduct Eliminating or Restricting Competition by Abusing Intellectual Property Rights, Art 7 (2005); SAIC, Guidelines for Anti-Monopoly Enforcement against Abuse of Intellectual Property Rights, Art 24 (7th edn, 2016); Korea Fair Trade Commission, Review Guidelines on Unfair Exercise of Intellectual Property Rights (2016). 24 See, for example, J Gregory Sidak, ‘FRAND in India: The Delhi High Court’s Emerging Jurisprudence on Royalties for Standard-Essential Patents’ (2015) 10 Journal of Intellectual Property Law & Practice 609. 25 See, for example, Case C-170/13, Huawei Technologies Co Ltd v ZTE Corp [2015] ECR 477, [53], [71], passim; Case AT.39985 – Motorola – Enforcement of GPRS Standard Essential Patents, Commission Decision of 29.04.2014; AT.39939 – Samsung – Enforcement of UMTS Standard Essential Patents, Commission Decision of 29.4.2014. 26 See, for example, US Department of Justice and Federal Trade Commission, Antitrust Guidelines for the Licensing of Intellectual Property (2016) (imposing no limits on a patentee’s pricing, observing no general duty to license, and expressing no restriction on firms’ ability to negotiate royalties at the component or device levels). In the standard-setting space, the influential DC Circuit has held that deceiving a standard-setting organization (SSO) to avoid a RAND-licensing commitment cannot violate section 2 unless the deception led the SSO to abandon a substitute technology: FTC v Rambus, 522 F 3d 456 (DC Cir 2008), cert denied 129 S Ct 1318 (2009). This requirement has no apparent analogue in Europe or in Asia. Notably, although Samsung’s efforts to enjoin Apple led the European Commission to impose obligations to stave off an Article 102 TFEU violation, the DOJ closed its investigation of Samsung. Presumably due to the limits on liability under section 2, the FTC has controversially relied on its standalone section 5 authority to challenge alleged RAND-licensing violations. See In re Motorola Mobility LLC and Google Inc, File No 121-0120, Comp (3 January 2013); In re Robert Bosch GmbH, File No 121-0081, Compl (26 November 2012). 27 See, for example, Herbert Hovenkamp, Federal Antitrust Policy: The Law of Competition and Its Practice (4th edn, West Publishing 2015) [1.7]; Leonard W Weiss, ‘The Structure-Conduct-Performance Paradigm and Antitrust’ (1979) 127 University of Pennsylvania Law Review 1104.

44  Research handbook on methods and models of competition law that lacked a rigorous foundation for its ‘inhospitality tradition’. Mergers to less than 8 per cent market share, maximum resale price limits, and the ‘Nine No-Nos’ – vertical restraints in patent-licensing contracts – faced condemnation.28 Monopoly leverage was an axiom in construing dominant-firm conduct.29 Even the Robinson-Patman Act, an anti-price-discrimination statute enacted to protect inefficient firms at the consumers’ expense, found urgent enforcement at the hands of the FTC.30 The Chicago School ushered in the modern economic approach, employing price theory to predict competitive effects.31 The ensuing revolution saw the US courts discard unsubstantiated rules against vertical restraints.32 The courts also adopted a sceptical perspective on predatory-pricing claims, limited antitrust-imposed duties to deal with rivals, and began to emphasize the primacy of empirical competitive effects over doctrinal dogmatism.33 So, too, the FTC and the DOJ enacted Horizontal Merger Guidelines of increasing economic nuance,

28 United States v Von’s Grocery Co, 384 US 270 (1966); Albrecht v Herald Co, 390 US 145 (1968) overruled by State Oil Co v Khan, 390 US 145 (1997); Bruce B Wilson, ‘Patent and Know-How Licence Agreements: Field of Use, Territorial, Price and Quantity Restrictions’ (Fourth New England Antitrust Conference, 6 November 1970) 9. 29 For a classic Chicago School attack on the then-prevalent concept of monopoly leverage, see Aaron Director and Edward H Levi, ‘Law and the Future: Trade Regulation’ (1956) 51 Northwestern University Law Review 281. 30 Mary L Azcuenga, The Robinson-Patman Act: A Perspective from the FTC, 13 May 1993, 1–2, https://​www​.ftc​.gov/​system/​files/​documents/​public​_statements/​509581/​ma51393​.pdf. 31 See Richard A Posner, ‘The Chicago School of Antitrust Analysis’ (1979) 127 University of Pennsylvania Law Review 925, 925–8; Joshua D Wright, ‘Abandoning Antitrust’s Chicago Obsession: The Case for Evidence-Based Antitrust’ (2011) 78 Antitrust Law Journal 301, 303–9. 32 See Leegin Creative Leather Prods, Inc v PSKS, Inc, 551 US 877 (2007) (holding that the rule of reason governs vertically imposed minimum prices and overruling Dr Miles Medical Co v John D Park & Sons Co, 220 US 373 (1911)); State Oil Co v Khan, 522 US 3 (1997) (finding that vertical maximum price fixing is subject to the rule of reason and overruling Albrecht v Herald Co, 390 US 145 (1968)); Continental Television, Inc v GTE Sylvania Inc, 433 US 36 (1977) (holding that courts must analyse vertical geographic restraints under the rule of reason and overruling United States v Arnold, Schwinn & Co, 388 US 365 (1967)); see also Business Electronics Corp v Sharp Electronics Corp, 485 US 717, 726 (1988) (presuming that the rule of reason applies and requiring a ‘demonstrable economic effect’ to invoke the per se rule); NCAA v Board of Regents of University of Oklahoma, 468 US 85, 100–1 (1984) (refusing to subject horizontal price and output restraints to per se rule where some such restrictions were necessary for the product to be available at all). 33 On predatory pricing, see Brooke Group Ltd v Brown & Williamson Tobacco Corp, 509 US 209 (1993); Matsushita Electric Industrial Co v Zenith Radio Corp, 475 US 574 (1986). Although the Supreme Court let stand a monopolization claim founded on a dominant firm’s abruptly terminating a consumer welfare-enhancing joint product with its rival (see Aspen Skiing Co v Aspen Highlands Skiing Corp, 472 US 585 (1985)), it subsequently limited the scope of section 2 liability to that decision: Verizon Communications Inc v Law Offices of Curtis V Trinko, 540 US 398, 409 (2004). As to the US judiciary emphasizing direct evidence of competitive effects over doctrinal considerations, see Federal Trade Commission v Indiana Federation of Dentists, 476 US 447, 460–61 (1986); United States v Gen Dynamics Corp, 415 US 486 (1974) (although structural showing may have demonstrated undue concentration, post-merger evidence showed no section 7 violation).

The cost of progress: hurdles facing antitrust’s economic advance  45 effectively abandoning certain merger principles still embedded in Supreme Court law.34 And the FTC has effectively stopped enforcing the anti-consumer Robinson-Patman Act.35 While the noninterventionist teachings of the Chicago School seldom found a warm reception outside the United States, they placed deductive economic models at the centre of robust competition policy. So, too, Chicago’s devastating critique of the SCP literature laid the groundwork for the more reliable empirical studies in which industrial-organization economists engage today.36 To be sure, in reacting to the law’s then-uncritical proscriptions, some early Chicago scholars were too dismissive of harmful unilateral exclusion and vertical foreclosure.37 But in using microeconomics to evaluate firm conduct, the Chicago School created the beginning of modern antitrust policy. The hallmark of subsequent events has been refinement. The post-Chicago literature produced game-theoretic models showing that, under strict assumptions, a dominant firm could rationally exclude competition through vertical restraints such as exclusive-supply contracts, predatory pricing, product tying, rebates and other conduct.38 Alas, some commentators misconstrue those models as general cases, when they are in fact possibility theorems due to their strong assumptions about scale economies, entry barriers, and cost justifications.39 Indeed, recent economic work emphasizes that vertical restraints and mergers generally produce lower 34 Cf, for example, United States v Von’s Grocery Co, 384 US 270 (1966) (merger of two firms with joint market share of 7.5 per cent violated section 7, despite the existence of ongoing competition in the market due to a trend of rising concentration) with US Department of Justice, 1968 Merger Guidelines, para I.5, at 6 (agency unlikely to move against a merger with less than 8 per cent market share). The DOJ issued subsequent merger guidelines in 1982 and 1984. Together with the FTC, it issued joint horizontal merger guidelines in 1992 (as revised in 1997). The latest iteration of the agencies’ horizontal merger guidelines was issued in 2010: Federal Trade Commission and US Department of Justice, Horizontal Merger Guidelines (2010). 35 See, for example, D Daniel Sokol, ‘Analyzing Robinson-Patman’ (2015) 83 George Washington Law Review 2064, 2071–6; David A Hyman and William E Kovacic, ‘Can’t Anyone Here Play This Game? Judging the FTC’s Critics’ (2015) 83 George Washington Law Review 1948, 1958–9, 1965–78. 36 See, for example, Carlton and Perloff (n 7). 37 For instance, in his classic work The Antitrust Paradox, Robert Bork effectively demolished the Warren Court’s antitrust jurisprudence, but in doing so made excessive claims, such as that tying arrangements, exclusive dealing, and vertical mergers can almost never harm competition: Robert Bork, The Antitrust Paradox (Simon & Schuster 1993). For a collection of essays critical of the Chicago School’s sceptical approach to antitrust questions, see Robert Pitofsky (ed), How the Chicago School Overshot the Mark: The Effect of Conservative Economic Analysis on US Antitrust (Oxford University Press 2008). In the author’s view, however, certain aspects of those essays mischaracterize the Chicago School’s true contributions. Accord Daniel A Crane, ‘Chicago, Post-Chicago, and Neo-Chicago’ (2009) 76 University of Chicago Law Review 1911, 1914, passim. 38 Famous post-Chicago economic models showing rational exclusion via vertical arrangements include Michael D Whinston, ‘Tying Foreclosure, and Exclusion’ (1991) 80 American Economic Review 837; Eric B Rasmusen, J Mark Ramseyer and John Shepard Wiley Jr, ‘Naked Exclusion’ (1991) 81 American Economic Review 1137; Janusz A Ordover, Garth Saloner and Steven C Salop, ‘Equilibrium Vertical Foreclosure’ (1990) 80 American Economic Review 127; Oliver Hart and Jean Tirole, ‘Vertical Integration and Market Foreclosure’ (Brookings Papers: Microeconomics, 1990) 205; Paul Milgrom and John Roberts, ‘Predation, Reputation and Entry Deterrence’ (1982) 27 Journal of Economic Theory 280; see also Michael H Riordan, Competitive Effects of Vertical Integration (2006), http://​www​.columbia​ .edu/​~mhr21/​papers/​Competitive​_Vert​_Int​.pdf. 39 See, for example, Daniel P O’Brien, ‘The Antitrust Treatment of Vertical Restraints: Beyond the Possibility Theorems’ in Swedish Competition Authority (ed), The Pros and Cons of Vertical Restraints (Konkurrensverket 2008) 40–1.

46  Research handbook on methods and models of competition law prices and higher output.40 Hence, enforcement actions targeting such conduct properly remain limited in the modern era. Still, post-Chicago models that show rational vertical exclusion under certain assumptions are important, not least for showing that competitive effects are almost invariably fact dependent. Indeed, today myriad theoretical models predict that similar conduct will produce competitively benign or harmful outcomes, depending on the assumptions and market conditions at issue.41 The ultimate gauge of a theory, however, is validation.42 Here, empiricism reigns supreme, both in testing a model’s predictions and in rendering theory obsolete when direct measurement of competitive effects is possible. Industrial-organization economists have made great advances in measuring and predicting effects. The key to their progress has been to discard inter-industry, cross-sectional studies of the SCP tradition, which produced unreliable measurements due to endogeneity, simultaneity bias, the use of industry-classification codes instead of economically defined markets, and the use of accounting returns as proxies for economic rents.43 Those problems may be insurmountable. That is why the ‘New Industrial Organization’ literature focuses on discrete antitrust markets, estimating model parameters and price effects.44 Empirical work has contributed enormously to contemporary antitrust policy, often revealing market realities to be at odds with the then-prevailing model of choice. Take two illustrative examples, both from the healthcare industry. The first involves anticompetitive healthcare mergers. After a string of losses, the FTC conducted a section 6(b) retrospective study of four consummated mergers.45 That study revealed relevant geographic markets to be more localized than previously thought. Previous enforcement and judicial decisions had employed the Elzinga-Hogarty test, focusing on patient inflows and outflows to identify the contours of the geographic market.46 Retrospective analysis revealed that even the presence of hospital systems within the same vicinity did not prevent negative price effects. FTC economists refined their two-step bargaining model and empirical methodologies to more accurately capture the competitive realities of healthcare systems.47 Since then, the FTC has prevailed in all of its hospital-merger challenges, save for one in 2016 when West Virginia passed a law 40 See, for example, Francine Lafontaine and Margaret Slade, ‘Vertical Integration and Firm Boundaries: The Evidence’ (2007) 45 Journal of Economic Literature 629; James C Cooper, Luke M Froeb, Dan O’Brien and Michael G Vita, ‘Vertical Antitrust Policy as a Problem of Inference’ (2005) 23 International Journal of Industrial Organization 639. 41 See, for example, Joshua D Wright, ‘Abandoning Antitrust’s Chicago Obsession: The Case for Evidence-Based Antitrust’ (2011) 78 Antitrust Law Journal 301–2, 313–16. 42 The Chicago School was a leading proponent of empirical validation of any antitrust theory. See, for example, Frank H Easterbrook, ‘Vertical Arrangements and the Rule of Reason’ (1984) 53 Antitrust Law Journal 135, 151. 43 See Carlton and Perloff (n 7). 44 See ibid; see also Timothy F Bresnahan and Richard Schmalensee, ‘The Empirical Renaissance in Industrial Economics: An Overview’ (1987) 35 Journal of Industrial Economics 371. 45 Timothy J Muris, ‘Everything Old Is New Again: Health Care and Competition in the 21st Century’ (7th Annual Competition in Health Care Forum, 7 November 2002) 19–20. 46 See, for example, California v Sutter Health System, 84 F Supp 2d 1057, 1069 (ND Cal 2000). For a description of the Elzinga-Hogarty test, see Kenneth G Elzinga and Thomas F Hogarty, ‘The Problem of Geographic Market Delineation Revisited: The Case of Coal’ (1978) 23 Antitrust Bulletin 1. 47 Gregory Vistnes, ‘Hospitals, Mergers, and Two-Stage Competition’ (2000) 67 Antitrust Law Journal 671; Evanston Northwest Healthcare Corp, FTC Docket No 9315, Opinion of the Commission (6 August 2007).

The cost of progress: hurdles facing antitrust’s economic advance  47 exempting the challenged transaction from federal antitrust law.48 Importantly, the FTC’s empirical work has found favour with the judiciary. In 2016 alone, two federal Courts of Appeals strongly embraced the teachings of the agency’s econometric work.49 The second example lies in understanding the competitive dynamics of generic-drug competition. As branded and generic drugs are, by definition, functionally interchangeable, one might reasonably expect them to belong in the same market.50 And, given product fungibility, one might also expect Bertrand competition, such that even limited generic entry would lead market prices to approach the competitive level. Econometric work has revealed that not to be the case.51 That result is an important example of how empiricism properly dominates theory in advancing competition policy. With those advances, US antitrust law has become a precision tool. Impressionistic evaluation of deals and practices is off the table. Save for product tying, to which a qualified per se rule still attaches, the law on vertical restraints now aligns with modern economic theory. The agencies challenge relatively few vertical and conglomerate mergers, doing so only if the acquiring firm would have the incentive and ability to harm competition.52 And intervention decisions are evidence based. Sophisticated econometric techniques inform agency challenges. For differentiated products, for instance, DOJ and FTC economists often use market data to compute diversion ratios to estimate substitutability, using consumers’ revealed preferences instead of subjective assessments of functional substitutability. In turn, estimated diversion ratios can inform willingness-to-pay calculations in healthcare mergers and, with other data, upward-pricing-pressure models. Merger simulations can compute price effects. Even without advanced econometric techniques, agency staff look closely at customer evidence that the merging parties are each other’s closest rivals or that customers have secured price decreases by threatening to shift supply from one of the merging parties to another. And the agencies likewise adhere to an evidentiary approach in fashioning remedies to otherwise efficient or innocuous mergers, which threaten to harm competition in a subset of the markets in which the parties operate. In those settings, the DOJ and the FTC often use tailored divestitures to maintain horizontal competition at the pre-merger level. Importantly, the agencies’ remedial

48 In re Cabell Huntington Hospital Inc, Docket No 9366, Statement of the Federal Trade Commission (6 July 2016). 49 Federal Trade Commission v Advocate Health Care Network, 841 F 3d 460 (7th Cir 2016); Federal Trade Commission v Penn State Hershey Medical Center, 838 F 3d 327 (3d Cir 2016). 50 But cf, for example, Geneva Pharmaceuticals Technology Corp v Barr Laboratories, Inc, 386 F 3d 485, 497 (2d Cir 2004); Henry v Chloride, Inc, 809 F 2d 1334, 1342–3 (8th Cir 1987) (finding relevant antitrust markets limited to generic versions of a branded drug, which lies outside the market due to distinct consumer demand). 51 See David Reiffen and Michael R Ward, ‘Generic Drug Industry Dynamics’ (2005) 87 Review of Economics and Statistics 37; see also Luke M Olson and Brett W Wendling, The Effect of Generic Drug Competition on Generic Drug Prices During the Hatch-Waxman 180-Day Exclusivity Period (FTC Working Paper No 317, 2013) https://​www​.ftc​.gov/​reports/​estimating​-effect​-entry​-generic​-drug​-prices​ -using​-hatch​-waxman​-exclusivity. 52 See, for example, US Department of Justice, Antitrust Division Policy Guide to Merger Remedies (2011) 56, 12–13, 16–17; James A Keyte and Kenneth B Schwartz, ‘Getting Vertical Mergers through the Agencies: “Let’s Make a Deal”’ (2015) 29 Antitrust 10; Steven C Salop and Daniel P Culley, Potential Competitive Effects of Vertical Mergers: A How-To Guide for Practitioners (Georgetown Law Faculty Publications and Other Works 2014) 1392.

48  Research handbook on methods and models of competition law approach faces empirical scrutiny. In 2017, for example, the FTC released a retrospective study of 90 matters in which it imposed remedies and identified learning points from its findings.53 The result is a competition policy unrecognizable compared to what held sway during the Warren Court era. But the neoclassical and empirical revolution of US antitrust law is not free of cost. Section III, which follows, explains the challenges that emerge from advances in industrial organization, empiricism, and modern agency practice.

III

CHALLENGES BORN OF ADVANCES IN ANTITRUST METHODOLOGIES

There is much to celebrate in today’s competition law. We have recounted the economic revolution that took antitrust into the modern age, explaining its increasing empirical focus and observing how the proliferation of complex theoretical models has enhanced depth of understanding. Unfortunately, coincident with the many gains to date, difficulties have arisen. This section focuses on four problems. First, a tension exists between the agencies’ responsibility to drive the economic content of antitrust law forward and the desire to win the cases that they bring. Second, as methodologies become more sophisticated, they displace doctrinal considerations imperfectly. The extent to which newer models and econometric work may depart from the older, conventional, doctrinal approach is unclear. The result is a tension between achieving accuracy, limiting gamesmanship, and enhancing predictability. Third, although the economics of industrial organization has bestowed a world of knowledge on the antitrust bar, its breadth and nuance raise their own challenges. Among other issues, enforcers face a model-selection problem. When models point in different directions and their assumptions tie imperfectly to the case at hand, parties can choose models that support their desired outcomes. This phenomenon exacerbates the well-known problem of duelling experts. Finally, criticism in 2016 and 2017 has targeted advances in antitrust law directly. Widespread claims of growing monopoly power blame recent competition policy for being myopic. For instance, The Economist argues that modern antitrust’s scalpel-like approach misses some important economic effects, including entrenchment on the part of major technology firms and rising industrial concentration driven by conglomerate mergers and ones bearing little horizontal overlap.54 A

Developing the Law versus Winning in Court

The US agencies’ guiding metric in enforcing antitrust law is the public interest. At first blush, that mandate is simple. Responsible for protecting the competitive process, the FTC and the DOJ oppose mergers, practices and conduct that harm consumers by weakening or removing market constraints on firm behaviour. In executing that responsibility, the agencies must not only identify meritorious cases, but also win them. Otherwise, their ability to negotiate effective remedial consents and to convince firms to abandon problematic business practices would suffer. Parties, even the government, bargain in the shadow of law. Federal Trade Commission, Remedy Study, https://​www​.ftc​.gov/​policy/​studies/​remedy​-study. ‘Too Much of a Good Thing’, The Economist (26 March 2016); ‘The Problem with Profits’, The Economist (26 March 2016). 53 54

The cost of progress: hurdles facing antitrust’s economic advance  49 Importantly, however, the FTC and the DOJ have an unusual ability to influence the path of the law itself. Hence, both their case selection and the manner in which they litigate their cases matter. This consideration affects the enforcers’ public-interest calculus. Specifically, the goal of winning litigated cases – and thus achieving credibility – may undermine the distinct objective of guiding the evolution of substantive antitrust doctrine. Pushing the boundaries of the law, especially in grey areas involving novel business conduct or industry conditions, involves an inescapable measure of risk. For that reason, competition agencies losing some of their enforcement actions is consistent with a healthy enforcement agenda. The agencies have done much to guide the judiciary in the right direction. The FTC, for example, has litigated seven antitrust cases before the Supreme Court in the last 30 years and has won six of them.55 And although it would benefit from some procedural reforms, the FTC’s Part 3 administrative-litigation process – unique within the US antitrust enforcement system – is a particularly well-suited vehicle for the development of the law.56 The best example of the agencies’ success in guiding antitrust law, however, is also a story of unfinished business. The US Supreme Court last heard a substantive antitrust-merger case in 1974.57 As noted in Section II above, the Court’s jurisprudence from that era sorely lacks economic content. The DOJ and the FTC took up the mantle from there, crafting Horizontal Merger Guidelines that reflect the latest learning from the economics literature. Although the agencies have no mandate to overrule Supreme Court law, their prosecutorial discretion allows them to focus on meritorious cases and explain their criteria in public materials. Their merger guidelines have been tremendously influential, finding favour across the full spectrum of the lower courts. One could wax poetic about the DOJ’s and the FTC’s contributions to merger law. But the more interesting observation is the agencies’ hesitation to develop substantive antitrust-merger law in a direction that would challenge their litigation efforts, but be truer to the underlying economics. The answer lies in a 54-year-old Supreme Court decision, Philadelphia National Bank, under which a merger giving the acquiring firm a market share exceeding 30 per cent is presumptively unlawful.58 That rule has become a staple of antitrust-merger litigation. Invariably, when the DOJ or the FTC challenges an unconsummated acquisition, it alleges a relevant market in which the acquiring entity would enjoy a market share exceeding the 30 per cent threshold. Often, their complaints invoke the PNB rule.59 As they must, the courts accept the presumption if the government establishes its prima facie case.60 The merging 55 North Carolina State Board of Dental Examiners v Federal Trade Commission, 135S Ct 1101 (2015); Federal Trade Commission v Actavis, Inc, 133 S Ct 2223 (2013); Federal Trade Commission v Phoebe Putney Health System, Inc, 133 S Ct 1003 (2013); California Dental Association v Federal Trade Commission, 526 US 756 (1999); Federal Trade Commission v Ticor Title Insurance Co, 504 US 621 (1992); Federal Trade Commission v Superior Court Trial Lawyers Association, 493 US 411 (1990); Federal Trade Commission v Indiana Federation of Dentists, 476 US 447 (1986). 56 See Ohlhausen (n 12). 57 United States v Marine Bancorp, Inc, 418 US 602 (1974); United States v General Dynamics Corp, 415 US 486 (1974). The Court’s 2013 decision in Phoebe Putney resulted from an antitrust-merger challenge, but the decision turned on the limits of state-action immunity: Phoebe Putney Health System, Inc, 133 S Ct 1003 (2013). 58 United States v Philadelphia National Bank, 374 US 321 (1963). 59 See, for example, United States v Aetna Inc, No 1:16-cv-1494, [31] Compl (DDC 21 July 2016); United States v Anthem, Inc, No 1:16-cv-1493, Compl, [31] (DDC 21 July 2016). 60 From 2016 alone, see, for example, Federal Trade Commission v Penn State Hershey Medical Center, 838 F 3d 327, 347 (3d Cir 2016); Federal Trade Commission v Staples, Inc, 190 F Supp 3d 100,

50  Research handbook on methods and models of competition law parties then face the unenviable task of rebutting the presumption through evidence that ‘the market-share statistics give an inaccurate account of the merger’s probable effects on competition in the relevant market’.61 The PNB rule would be laudatory if it rested on a sound foundation. Specifically, if the industrial-organization literature taught that a 30 per cent market share generally yields unilateral or coordinated effects, then the rule would be uncontroversial. To be sure, the concentration associated with a firm’s growing beyond a 30 per cent share is associated with negative price effects in certain markets. But there is, as yet, a limited economic basis for applying that principle across markets and industries more generally. Yes, basic models of oligopolistic competition, such as output decisions (Cournot) and price decisions with differentiated goods (Bertrand), have long formalized the intuition that a drop in the number of competitors in an already-concentrated market will produce worse outcomes.62 But it is clear that some highly concentrated markets are fiercely competitive, even with as few as two or three competitors. It would seem that there is much more to competitive outcomes than an increase in concentration alone. The 2010 Horizontal Merger Guidelines recognize as much. They observe that the ‘measurement of market shares and market concentration is not an end in itself, but is useful to the extent it illuminates the merger’s likely competitive effects’.63 Although market shares and Herfindahl–Hirschman Index (HHI) numbers feature prominently in the agencies’ guidance, the ‘purpose of these thresholds is not to provide a rigid screen’.64 Rather, ‘[m]arket concentration is often one useful indicator of likely competitive effects of a merger’.65 Despite those qualifications, however, the Guidelines do employ a structural approach. Importantly, they eschew any reference to a trigger of a 30 per cent market share. Still, they draw an equivalent presumption when an HHI increase of over 200 occurs in a highly concentrated market.66 As under the PNB rule, a merging firm may rebut that presumption with ‘persuasive evidence showing that the merger is unlikely to enhance market power’.67 Section II traced US antitrust’s evolution towards an empirically driven policy. It would be a natural development to jettison a universal structural presumption like the rule in PNB. And, indeed, it is to their credit that the agencies do not use a 30 per cent threshold to decide whether a scrutinized merger may substantially lessen competition. They look instead at empirical issues, including customer testimony, evidence of price competition between the merging parties, sophisticated econometric modelling using available data, and, for consummated deals, any post-merger price effects. To be sure, the DOJ and the FTC compute HHI figures and ask whether the merger surpasses the relevant thresholds identified in the Horizontal

128 (DDC 2016); see also Federal Trade Commission v Advocate Health Care Network, 841 F 3d 460, 475 n 5 (7th Cir 2016). 61 Staples, 190 F Supp 3d at 115 (quoting Federal Trade Commission v H J Heinz Co, 246 F 3d 708, 715 (DC Cir 2001)). 62 Carlton and Perloff (n 7). 63 US Department of Justice and Federal Trade Commission, Horizontal Merger Guidelines (2010) [4]. 64 Ibid, [5.3]. 65 Ibid. 66 Ibid. 67 Ibid, [2.1.3], [5.3].

The cost of progress: hurdles facing antitrust’s economic advance  51 Merger Guidelines. But, in practice, the deals that they challenge typically involve market shares far surpassing those limits.68 The reality today is that the DOJ and the FTC prove their cases in court differently from how they decide whether a merger harms competition. From a litigation perspective, the agencies obviously like the PNB rule. Antitrust-merger cases are complex, and staff – facing dogged opposition from leading counsel and hired economists – can struggle to convince lay judges to enjoin a deal. Although they enjoy an enviable track record, the agencies have faced periods of sustained losses in the past.69 The PNB rule helps the government’s cause, effectively granting the agencies a measure of deference in implicit recognition of their expertise. To the extent that the rule allows the government to enjoin anticompetitive mergers that would otherwise pass muster, it benefits social welfare. There are other benefits, too. A presumptive rule, like the 30 per cent share trigger, improves the administrability of antitrust law by avoiding all-encompassing utilitarian calculi for which lay judges may be ill-prepared. It also introduces a measure of predictability for companies and the lawyers who advise them. Nevertheless, there is a cost to staying the course. An asymmetry between in-house and in-court analysis for identifying anticompetitive mergers should trouble those who celebrate modern antitrust law’s economic advance. The agencies could guide the courts to develop a more nuanced jurisprudence in this area. Although they operate within the law, the FTC and the DOJ have considerable leeway in how they present their cases. Further, as the means of proof begin to mirror the agencies’ internal economic analysis, judicial review should improve the underlying methodologies. That proposition holds true even if the courts occasionally – and inevitably – get some decisions wrong. For the agencies are prone to err, too, and being put to one’s proof promotes superior advocacy and further refinement of the appropriate analysis. In short, structural inferences untethered to price or other competitive effects observed in a market stifle antitrust’s arc of progress. That point does not imply that the agencies should discard market-share and concentration figures. Rather, it suggests that competition-law 68 For some recent examples, see In re Sanford Health, FTC Docket No 9376, [30], 21 June 2017 (deltas in highly concentrated relevant markets ranged from 2793 to 4800); DraftKings, Inc v FanDuel Ltd, FTC Docket No 9375, Compl [47], 19 June 2017 (‘the Merger would result in a post-Merger HHI of at least 8,100 and an increase in concentration much greater than 200 points’); United States v Haliburton Co, No 1:16-cv-233, Compl, [18]–[62] (D Del 6 April 2016) (post-merger HHI in markets running from 7700 to 2600 and deltas ranging from 3700 to 350); In re Advocate Health Care Network, FTC Docket No 9369, Compl, [30], 17 December 2015 (post-merger HHI of 3517 and a delta of 1423); In re Staples, Inc, FTC Docket No 9367, Compl, [43], 7 December 2015 (post-merger HHI of over 4900 and a delta well over 200); In re The Penn State Hershey Med Ctr, FTC Docket No 9368, Compl, [30], 7 December 2015 (post-merger HHI of 4500 and a delta of over 2000); In re Cabell Huntington Hosp, FTC Docket No 9366, Compl, [39], 5 November 2015 (post-merger HHI of over 5800 and a delta of over 2800); United States v AB Electrolux, No 1:15-cv-1039, Compl, [30] (DDC 1 July 2015) (post-merger HHIs ranging from 5400 to 2700 and deltas from 2000 to 450). 69 In the healthcare-merger space, for example, the government lost seven cases in a row between 1994 and 2000. See Hospital Board of Directors of Lee County, 38 F 3d 1184 (11th Cir 1994); Federal Trade Commission v Freeman Hospital, 69 F 3d 260 (8th Cir 1995); Federal Trade Commission v Butterworth Health Corp, 946 F Supp 1285 (W D Mich 1996), aff’d mem, 121 F 3d 708 (6th Cir 1997); United States v Mercy Health Services, 902 F Supp 968 (N D Iowa 1995), vacated as moot, 107 F 3d 632 (8th Cir 1997); United States v Long Island Jewish Medical Center, 983 F Supp 121 (EDNY 1997); Federal Trade Commission v Tenet Healthcare Corp, 186 F 3d 1045 (8th Cir 1999); State of California v Sutter Health System, 84 F Supp 2d 1057 (N D Cal 2000), aff’d, 217 F 3d 846 (9th Cir 2000), amended by, 130 F Supp 2d 1137 (N D Cal 2001).

52  Research handbook on methods and models of competition law enforcers should continue to develop the principles embraced by the Horizontal Merger Guidelines – namely, ‘to examine whether other competitive factors confirm, reinforce, or counteract the potentially harmful effects of increased concentration’.70 And it observes that continuing reliance on PNB contrasts with how the DOJ and the FTC have otherwise pushed the frontier of antitrust towards the latest learning of industrial-organization economics. B

What Role for Doctrine?

Antitrust’s rich theoretical base and refined methodologies bring other challenges. Although substantive antitrust law evolves in light of new learning, it lags that learning. Empirical advances supplant doctrinal analysis, but only partially. The result is relatively accurate intervention and – if the fact finder absorbs the latest methodological advance – improved merits decisions, but a disjointed relationship between doctrine and liability outcomes. This is another instance of how welcome improvements in antitrust’s economic thinking may bring in their wake important costs. Here, improvement comes at the cost of predictability. First, to the extent that modern antitrust is indeed a field of applied economics, then liability will often turn on which applied economic theories ultimately convince the agency, the court or the jury. To anticipate the outcome of a challenged transaction, a firm and its lawyers must review the literature and draw their own conclusions. But the industrial-organization field has grown so vast that it raises a model-selection problem.71 Generating so many models, each making different assumptions and yielding distinct predictions, it is increasingly difficult to identify which of the myriad theories best captures the dynamic of interest. And yet the choice matters because candidate models may point in opposite directions based on subtle changes in the underlying assumptions.72 That phenomenon exacerbates the issue of duelling with equally credentialled economists serving as expert witnesses. When economists disagree on the appropriate model with which to analyse a restraint, reject one another’s proposed markets, and estimate inconsistent market effects, they raise unintended questions not just about predictability, but also about the administrability of non-rule-based antitrust policy.73 Firms wishing to avoid antitrust liability under today’s regime face other difficulties. Market-share cut-offs still limit liability, but they may be hard to calculate ex ante. Intuitive analysis of functional substitutability may not reliably predict the relevant market because, for instance, empirical data may reveal distinct consumer preferences or low cross-price elasticity of demand.74 A firm may have a strong case as to its share of the market, but the Horizontal Merger Guidelines (n 63), [5.3]. See, for example, Joshua D Wright, ‘Evidence-Based Antitrust Enforcement in the Technology Sector’ (Competition Law Center, 23 February 2013) 3. 72 Consider, for example, the long-debated question whether monopoly or competition best drives dynamic efficiency (innovation). That classic Schumpeter-Arrow debate has produced a vast literature with few generally applicable principles, but a range of models that variously identify one of monopoly or competition as the superior determinant of R&D investment depending on specific, nuanced assumptions. See, for example, Richard J Gilbert, ‘Looking for Mr. Schumpeter: Where Are We in the Competition–Innovation Debate?’ in Adam B Jaffe, Josh Lerner and Scott Stern (eds), Innovation Policy and the Economy (Vol 6, MIT Press 2016) 159, 164. 73 Cf, for example, Rebecca Haw, ‘Adversarial Economics in Antitrust Litigation: Losing Academic Consensus in the Battle of the Experts’ (2012) 106 Northwestern University Law Review 1261. 74 See, for example, Federal Trade Commission v Lundbeck, Inc, 650 F 3d 1236 (8th Cir 2011) (finding that two drugs used to treat the same condition belonged in separate markets because neonatol70 71

The cost of progress: hurdles facing antitrust’s economic advance  53 government may bypass a detailed market assessment altogether should it identify direct evidence of anticompetitive effects.75 Mergers of competitors that sell differentiated goods may pass muster – or not – based on how economic statistics like gross-upward pricing-pressure indices (GUPPIs) illuminate likely effects.76 The results of such analysis may not be obvious to merging firms ex ante. Firms cannot use economic theory summarily to defeat antitrust claims when the evidence may not support the model’s predictions.77 And that is to say nothing of the larger uncertainty of outcomes under the rule of reason. Still, one can exaggerate the value of predictability. Legal thinkers distinguish between rules and standards, the latter of which forgo specificity for the benefit of more informed decision-making in future cases as unforeseeable facts become known. And the price of hard law is often rules bearing higher Type I error costs, which displease the business community and may harm competitive investment at the margin. Indeed, those seeking business certainty would do well to look to the world of US antitrust law before the economics revolution of the 1970s. Per se laws against vertical restraints and hostility to mergers in industries with rising concentration presented rules around which firms could readily conform. The rule of reason’s proliferation to encompass the near entirety of business conduct – even horizontal boycotts often avoid per se condemnation78 – allows companies to push the envelope and experiment with new business arrangements without fear of summary condemnation. Firms surely would rather be free to operate in legal grey zones than face categorical prohibitions from doing so. Diminished predictability nevertheless remains a cost, even if it is preferable to a rule-focused approach. The business community strives for legal certainty and seldom celebrates matters in which a fact-specific and economic analysis conducted under a broad legal standard suggests liability.79 An underappreciated reality of modern US competition policy is that sophisticated, evidence-driven methodologies yield answers that may not always be discernible ex ante. There is thus a need for ongoing research as to how the prevailing degree of uncertainty in the outcome of antitrust scrutiny affects firm behaviour. Related to this consideration are ongoing questions about how far the law will bend to accommodate empirical analysis. An interesting example from 2016 arose in Advocate-North Shore, where the FTC unsuccessfully challenged a healthcare merger before a district court in Illinois. The FTC’s appeal focused on geographic market definition. The agency explained that, under the Horizontal Merger Guidelines’ hypothetical-monopolist test, how an economist identifies a candidate market is irrelevant as long as that market satisfies the test. The Seventh Circuit agreed. The district court had rejected the market proposed by the FTC’s economist because the candidate market allegedly lay on an unreliable foundation. The appellate court found that consideration to be irrelevant because, if a hypothetical monopolist could profitably

ogists chose their preferred drug without regard to price and holding that ‘functionally similar products may be in separate markets’). 75 See Federal Trade Commission v Indiana Federation of Dentists, 476 US 447 (1986). 76 For a recent consent that relied on GUPPIs, see In re Dollar Tree, Inc, FTC File No 141-0207, Statement of the Federal Trade Commission (13 July 2015). 77 See Eastman Kodak Co v Image Tech Servs, Inc, 504 US 451 (1992). 78 See, for example, California Dental Association v Federal Trade Commission, 526 US 756 (1999); see also NCAA v Board of Regents of the University of Oklahoma, 468 US 85 (1984); Broadcast Music, Inc v Columbia Broadcasting System, Inc, 441 US 1 (1979). 79 See, for example, US Chamber of Commerce, Letter to Hon Edith Ramirez, Chairwoman of the Federal Trade Commission, regarding Victrex plc, FTC File No 141-0042 (27 May 2016).

54  Research handbook on methods and models of competition law increase price within the proposed area, then it was a relevant geographic market. From an economic perspective, it is difficult to argue with the Seventh Circuit’s analysis. But it does raise intriguing questions concerning whether an arbitrarily or randomly picked candidate market would suffice if it ultimately met the hypothetical-monopolist test. And what if economic analysis revealed a relevant market that, econometric techniques aside, appeared implausible or even gerrymandered? The merging hospitals in Advocate-North Shore raised arguments related to these questions, even though they did not carry the day. One wonders, however, whether facts in future cases might create more difficult questions when contextual observation, intuition and empirical methodologies point in different directions. Ultimately, these issues suggest that, as the law progresses towards an ever-more-evidentiary approach, the role of doctrinal analysis may subside further, with attendant costs and benefits. C

Have Empirical Advances Left Antitrust Too Narrow?

As antitrust has become more targeted, its prohibitory scope has narrowed. Warren Court-era principles of halting upward trends in industrial concentration and limiting the accumulation of economic power achieved through size or scale find scant favour in today’s antitrust enforcement. For those predisposed to neoclassical economic theory and focused on price effects, that reality is a good thing. Yet while many commentators applaud the expulsion of industrial-policy issues from competition policy, not everyone is convinced. An enforcement ideology that only prohibits mergers likely to raise market prices will clear most conglomerate and vertical mergers – many of which eliminate double-marginalization effects, align incentives, and achieve scope economies. Even deals that involve problematic horizontal overlaps in a subset of the markets in which the parties operate will often pass muster if targeted divestitures will cure the likely harm. The result is that most mergers and acquisitions proceed unchecked under today’s antitrust standards. From a price-theoretic viewpoint, that result is a feature, not a bug, of antitrust law. Competition policy rightly focuses on harm to the competitive process. The year 2016, however, saw a new-found hostility to industrial consolidation, as well as to the growth and perceived entrenchment of powerful firms – including in the technology sector. Populist winds joined with the traditional wariness of those on the left towards free-market policies and economic concentration.80 Combined with high mergers and acquisitions activity in the aftermath of the Great Recession, calls for reinvigorated antitrust intervention have ensued. The Council of Economic Advisers observed indicia of rising concentration in US industries and rising profits.81 It inferred that diminished competition might be to blame. So, too, The Economist

80 The basis for the political rallying cry to effect a more interventionist antitrust policy is no mystery. The modern economy is changing. Rapid technological advance, disruptive innovation, automation, globalization and other economic phenomena have produced market conditions that not all warmly receive. Income inequality has been a major political issue as a gulf emerges between the educated workers who are primed for the new economy and those who are less well placed to adapt. Meanwhile, M&A activity has been at near-record highs in recent years. Coupled with the fallout from the Great Recession of 2008–09, such conditions have led to widespread dissatisfaction with industrial and macroeconomic policies. Populist winds increasingly blow on both sides of the Atlantic. Competition policy has been caught up in the maelstrom, as various prominent commentators accuse the agencies of anaemic enforcement. 81 Council of Economic Advisers Issue Brief, Benefits of Competition and Indicators of Market Power (April 2016) 4.

The cost of progress: hurdles facing antitrust’s economic advance  55 argued that the US economy has a monopoly problem driven in part by anaemic enforcement of the antitrust laws.82 The newspaper worried that modern antitrust’s sophistication has blinded enforcement agencies to market effects that economists’ models do not capture.83 Those claims raise difficult empirical questions, which exceed this chapter’s scope. A panoply of challenges frustrate attempts empirically to link industry (rather than relevant-market) concentration to firm rents and pricing behaviour – especially insofar as one attempts to extrapolate causal sources and directions, and all the more so when using cross-sectional studies.84 Suffice to say that there is a crucial debate to be had. This section, however, asks a more basic question: Has antitrust’s economic transformation carried unintended consequences by virtue of its narrower field of enquiry? Does focusing on price effects close enforcers’ eyes to larger economic issues? And, if so, is there a gap in industrial economic policy for which no tool other than antitrust is presently available? Despite his scepticism that any such problem exists – let alone merits a change in direction for competition policy – the author poses these questions for three reasons. First, they reflect contemporary debate, which itself flows from evolving antitrust economics. Regardless of whether such criticisms have merit, they suggest that further refinement in neoclassical antitrust economics, methodologies, and remedies is by no means guaranteed. Progress is contingent. Second, and on that point, today’s antitrust enforcers and policymakers must explain to stakeholders why the Chicago, post-Chicago, and neo-Chicago schools teach the same core lesson. They focus on price and dynamic-efficiency effects, use economic models founded on robust organizing principles, and embrace empiricism. The consumer-welfare model driving the modern US antitrust approach warrants celebration, not least because it forgoes subjective, politically charged notions of which business types and industry structures are optimal. Rather, it relies on consumers’ revealed preferences in market transactions to determine values. Third, the author wishes to pour cold water on the suggestion that today’s antitrust methodologies have gone seriously awry and need revision so as to limit industry concentration or firm size as an end in itself. Returning to the questions posed two paragraphs above, consider a premise common to them all. Each one presupposes that social ills accompany industrial consolidation that does not visit negative price, output, or innovation effects on consumers. To the extent that current merger enforcement fails to stop mergers having such effects, of course, then that would indeed be a shortcoming. Yet the failure would be of execution and methodology, not of substantive policy. The agencies continue to absorb new learning, test their practices, and review substantive critiques by others.85 But changing course to arrest mergers The Economist (n 54). Ibid. 84 See, for example, Ohlhausen (n 9). 85 A timely example is the claim, most prominently advanced by Professor John Kwoka, that the divestitures routinely used by the DOJ and the FTC fail to protect price competition: John Kwoka, Mergers, Merger Control, and Remedies: A Retrospective Analysis of US Policy (MIT Press 2015). Leading members of the FTC’s Bureau of Economics scrutinized those claims and found them empirically unsound: Michael Vita & F. David Osinski, John Kwoka’s Mergers, Merger Control, and Remedies: A Critical Review, 82 Antitrust L.J. 361 (2018). But cf. John E. Kwoka, Jr., Mergers, Merger Control, and Remedies: A Response to the FTC Critique (Apr. 6, 2017). Notably, the FTC scrutinized its use of divestitures and other remedies through a retrospective study, reviewing all 90 remedial orders entered by the Commission between 2006 and 2012. The Commission committed to revise its remedial practices to reflect material learnings. The FTC’s Merger Remedies 2006–2012, A Report of the Bureaus of Competition and Economics (January 2017). 82 83

56  Research handbook on methods and models of competition law and practices bearing no price or quality effects, but threatening only to contribute to growing consolidation, would be to abandon the core economic principles that have transformed US antitrust law into a rigorous and relatively politically value-free enterprise.86 Contrary to The Economist’s implicit call, we should not discard the scalpel and embrace the hammer in tailoring our competition policies to the economics of twenty-first-century industries.

IV CONCLUSION This book is, in many ways, a celebration of modern competition law’s economic sophistication. Antitrust enforcers around the world have at their fingertips an array of empirical tools that allow them to predict future effects and measure existing ones with greater accuracy and specificity than ever before. Although the Chicago School’s laissez-faire market ideology never found universal favour outside the United States, its focus on price effects, empiricism, and consumer welfare has been tremendously influential. Although jurisdictions differ as to substantive and procedural competition law, dialogue and refined methodologies have achieved a measure of consensus – at least on a few core points. This chapter embraces an evidentiary approach to competition law, focused on price and innovation effects. Due to the myriad differences between countries’ competition laws, it focuses on the world’s oldest antitrust jurisdiction and the one where economics finds the greatest welcome: the United States. Outlining the course of progress over the last several decades, the chapter observes that advances in antitrust economics have come at a cost. First, the economic revolution is by no means complete, even in the United States. The enforcement agencies have made great strides in promoting industrial-organization thinking into their agendas and, in turn, into changes in doctrinal law. Nevertheless, antiquated doctrine from an earlier age still governs merger review, creating a quandary for the DOJ and the FTC between (1) abandoning law that helps them win in court, and thus might serve the short-term public interest, and (2) eschewing legal presumptions untethered to modern industrial organization, working more closely to align antitrust law and economics, and voluntarily embracing a higher litigation burden in challenging problematic deals. Second, as methodological advances have cemented an empirical approach, rule-based doctrine has ceded ground. The result may be a net plus for society, but it bears a cost. Among other issues, firms may struggle to predict the outcome of an antitrust review. The ebb and flow of empiricism and doctrine also leave unanswered fundamental questions about the limits of methodological techniques in answering legal questions. Finally, developments in the past year alone suggest that a price-theoretic view of antitrust policy – though firmly established in US law – and ongoing refinement of the neoclassical economic approach are no longer guaranteed. That reality should warn those who embrace today’s antitrust methodologies not to take modern competition policy for granted.

86 For a more detailed argument on a related point by the author, see Alan Devlin, ‘Antitrust in an Era of Market Failure’ (2010) 33 Harvard Journal of Law and Public Policy 557.

4. The relevance of economics in US, EU and Australian competition law Geoff Edwards and Jennifer Fish

I INTRODUCTION Economics has played a central role in antitrust policy in the US since the early 1970s, with rule of reason analyses – focused on the effects of mergers and conduct on consumer welfare – increasingly preferred to form-based presumptions of illegality. By contrast, in Europe and Australia economics has enjoyed some time in the sun, but form-based legalism has maintained a stronger grip. Indeed, until quite recently it seemed that European competition policy had reached the ‘end of economics’ outside the realm of merger assessment: the reform agenda commenced by the European Commission around the turn of the century – with its aim to move European competition policy from its ordo-liberal form-based roots to more economic effects-based analysis – appeared to have stalled, and legal formalism appeared entrenched with respect to both Article 101 and Article 102 of the Treaty on the Functioning of the European Union (TFEU). Meanwhile, in Australia, legal formalism took such a grip early in this century as to distort the interpretation of the prohibition on unilateral abuse of market power beyond any economic coherence. Recent decisions in Europe (in Cartes bancaire,1 Coty2 and Intel,3 in particular) and a significant change in the Australian unilateral abuse of market power legislation have raised the prospect that we are on the cusp of an economic renaissance in the enforcement of competition law in these countries. However, the re-emergence in recent years of populist antitrust, with an antipathy towards the ‘bigness’ of large firms and concentrated market structures, represents strong headwinds and even threatens the established economic orthodoxy in the US. This chapter proceeds as follows. In Section II, we begin with a brief overview of the development of competition policy in the US and explore the foundations of its evolution from form-based presumptions to economic-based rules of reason, which today represent a global benchmark for an economic effects-based approach. In Section III, we turn to European competition policy. We begin with an examination of its ordo-liberal origins, which to a large extent explain the grip that form-based presumptions have maintained over European competition policy during a period when the US moved quickly to effects-based analysis. We then observe the modernization process of the European Commission, which commenced around 1 Judgment of the Court (Third Chamber), 11 September 2014, Groupement des cartes bancaires (CB) v European Commission, Case C‑67/13 P, ECLI:​EU:​C:​2014:​2204 (hereafter Case C‑67/13 P, Cartes bancaires, 2014:2204). 2 Judgment of the Court (First Chamber), 6 December 2017, Coty Germany GmbH v Parfümerie Akzente GmbH, Case C-230/16, ECLI:​EU:​C:​2017:​941(hereafter Case C-230/16, Coty, 2017:941 or just Coty). 3 Judgment of the Court (Grand Chamber), 6 September 2017, Intel Corp. v European Commission, Case C-413/14 P, ECLI:​EU:​C:​2017:​632 (hereafter Case C-413/14 P, Intel, 2017:632).

57

58  Research handbook on methods and models of competition law the turn of the century – with the aim of moving Europe towards more economic approaches – and review the battles that have subsequently played out between form and effects under Articles 101 and 102 TFEU. Finally, in Section IV, we examine how tensions between legal form and economic effects have played out in Australia, with a particular focus on Australia’s unique prohibition on unilateral anticompetitive conduct.

II

FROM FORM TO EFFECTS IN US ANTITRUST POLICY

The US, with its consumer-welfare-focused rule of reason approach, is widely considered to offer the global benchmark of an economic effects-based approach to the enforcement of competition law. However, it was not always so. Throughout the 1960s, there was what Nobel laureate Oliver Williamson described as an ‘inhospitality tradition’ in US antitrust policy,4 in which monopolization motives were attributed to any form of non-standard or unfamiliar business practice (including vertical integration, customer and territorial restrictions, tying and bundling). Another Nobel laureate, Ronald Coase, referred to this tradition in the following terms: One important result of this preoccupation with the monopoly problem is that if an economist finds something – a business practice of one sort or another – that he does not understand, he looks for a monopoly explanation. And as in this field we are very ignorant, the number of ununderstandable practices tends to be very large, and the reliance on a monopoly explanation frequent.5

According to Williamson, the possibility that these practices had economizing purposes and effects was ignored and, worse, efficiencies tended to be regarded as unfair competitive advantages. This tradition of inhospitality developed in the period from the mid-1930s until the 1960s. It did so as a rejection of largely permissive approaches that had been applied from around 1915. As the US emerged from the great depression, the concentration of American industry that had been promoted by the government to ease the economic strain that US industry had been under came to be seen as a barrier to economic restoration. In political terms, the mid-1930s saw US antitrust enter a populist era.6 Conveniently, economic scholarship from the 1930s to 1960s focused on the structure-conduct-performance (SCP) paradigm often associated with Bain, which ascribed monopoly explanations to all non-standard or unfamiliar market practices and emphasized market structure, market shares and concentration measures, as well as barriers to entry for competitors. The SCP paradigm confirmed and fuelled the formalistic approach of the US 4 See Oliver E Williamson, The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting (The Free Press 1975) 19; Oliver E Williamson, ‘Antitrust Lenses and the Uses of Transaction Cost Economics Reasoning’ in Thomas Jorde and David Teece (eds), Antitrust, Innovation, and Competitiveness (Oxford University Press 1992) 143. 5 Ronald H Coase, ‘Industrial Organization: A Proposal for Research’ in Victor R Fuchs (ed), Economic Research: Retrospect and Prospect, Vol 3, Policy Issues and Research Opportunities in Industrial Organization (National Bureau of Economic Research 1972) 67. 6 There are parallels of the populism of the 1930s in today’s ‘hipster’ antitrust movement and growing calls, on both sides of the Atlantic, for greater intervention by competition authorities to resolve a wide range of concerns with the behaviours of large businesses, particularly in the digital sector.

The relevance of economics in US, EU and Australian competition law  59 courts during this period, which resulted in the adoption of a large number of per se rules regarding behaviour, from tying to non-price vertical restraints, capacity expansions to meet new demand (United States v Aluminum Co of America),7 and localized price cuts by a national baker (Utah Pie Co v Continental Baking Co).8 In Williamson’s terms, monopolization (not competition or economization) was viewed as the ‘main case’. In 1967 the US reached peak interventionism in the Schwinn case.9 This case concerned whether a bicycle manufacturer with 13 per cent of the market could lawfully impose territorial restrictions on franchisee retailers who were required to provide minimum services to customers. The US Department of Justice (DOJ) argued that these restraints were anticompetitive in nature, and that in industries where brands were highly differentiated, intrabrand competition was of great importance and retail competition should not be restricted. The US Supreme Court agreed with the DOJ and made the use of exclusive territories a per se violation of the US antitrust laws. That Schwinn likely imposed the restrictions for efficiency rather than monopolizing purposes passed by both the DOJ and the Supreme Court like a bicycle in the night. The correction from this nadir of antitrust enforcement was rapid. The late 1960s and the 1970s saw the development in the US of a greater appreciation of efficiency explanations for mergers and unilateral conduct, and in a rather short space of time US antitrust policy completed a 180-degree turn from per se prohibitions and deep suspicions of concentrations to considerable permissiveness and deference to efficiency explanations. The causes of the about face are complex. Many attribute it to the activism of Chicago School scholars and attorneys such as Bork, Posner and Easterbrook, who applied stylized economic theory to question the per se prohibitions that the courts had created over the previous 30 years and argued forcefully for the recognition of efficiency explanations. The Chicago School no doubt assisted the change in approach of the Supreme Court by providing simple economic principles that judges could apply to conduct, although post-Chicago scholarship has demonstrated these principles to be over-simplistic at times.10 Changes in leadership within the DOJ may also have contributed. Williamson attributes the change to the brief tenure of Donald F Turner as Assistant Attorney General of the Antitrust Division at the DOJ from 1965 to 1968. According to Williamson, Turner initiated a move within the DOJ to the more economic approach that has characterized US antitrust policy since the 1970s, by encouraging greater contributions by economists within the division (including Williamson himself) and re-setting the DOJ’s enforcement course towards practices that harmed consumers rather than cases that the lawyers thought could be won. The shift in focus of the DOJ was also perhaps an expediency that reflected the turning of the long-term economic and political cycle in the US. In 1969, Richard Nixon took the White House and made a number of significant appointments of Supreme Court judges with sympathy for efficiency explanations, including Chief Justice Warren E Burger. At the same time, US industry was facing a crisis of lack of competitiveness globally, and there was a shift in the

United States v Aluminum Co of America, 148 F 2d 416 (2d Cir 1945). Utah Pie Co v Continental Baking Co, 386 US 685 (1967). United States v Arnold Schwinn & Co, 388 US 365 (1967). 10 For a Harvard School perspective, which discounts the Chicago School as the cause of the change, see Louis Kaplow, ‘Antitrust, Law & Economics, and the Courts’ (1987) 50(4) Law and Contemporary Problems 181. 9 7 8

60  Research handbook on methods and models of competition law national mood from populist reactions against large businesses, to support for their attempts to compete internationally. On this perspective, Chicago School scholars and other agents, such as Turner within the DOJ and Williamson’s own school of transaction cost economics (which presented robust efficiency explanations for vertical integration and many forms of contracting), were merely the supporting cast in the right place at the right time. Whatever the cause, the revolution in relation to vertical arrangements was completed within ten years: the Supreme Court’s 1977 decision in GTE Sylvania11 acknowledged efficiency explanations and replaced Schwinn’s per se prohibition of territorial restrictions with a rule of reason. In particular, the Court considered that a manufacturer might wish to maintain an exclusive franchise system with vertical market divisions in order to permit its franchisees to capture returns from efforts to develop their markets without appropriation of those efforts by free-riders. The triumph of economic effects-based assessments over per se presumptions was completed three decades later, in 2007, with the Supreme Court’s Leegin decision,12 which overruled a precedent that retail price maintenance was per se illegal, which had stood since 1911. Along the way, US merger policy has also become permissive, with efficiency explanations coming to be seen as the ‘main case’ explanation, and interventions considered only when anticompetitive effects, assessed case by case on the basis of careful economic analysis, are seen to outweigh efficiencies. Concerns have been raised in recent years that the pendulum may have swung too far and that the continued application of Chicago and post-Chicago economic scholarship has contributed to the creation of concentrated market structures and large dominant firms, particularly in the digital sector. Those that belong to what has been coined the ‘hipster’ antitrust movement have called for the abandonment of the consumer welfare standard and insistence on effects-based economic analysis in favour of broader public policy goals and a return of analytical focus on the size of firms and market structure. The coming years will reveal whether this populist movement will swing the US towards a new era of inhospitality and whether this will be as rapid as was the shift to the consumer welfare standard.

III

FORM VERSUS EFFECTS IN EUROPEAN COMPETITION POLICY

A

Ordo-liberal Origins of European Competition Policy

European competition policy, born in the aftermath of the Second World War, has been influenced to a large extent (both at its point of creation and subsequently in judicial interpretation) by the German ordo-liberal school of thought. Understanding ordo-liberalism is therefore important for understanding European competition policy. Ordo-liberalism is grounded in a particular political and economic ideology rather than in rigorous economic theory.13 Indeed, challenges presented by economic theory have resulted in

Continental Television v GTE Sylvania, 433 US 36 (1977). Leegin Creative Leather Products, Inc v PSKS, Inc, 551 US 877 (2007). 13 See Pierre Larouche and Maarten Pieter Schinkel, ‘Continental Drift in the Treatment of Dominant Firms: Article 102 TFEU in Contrast to 2 Sherman Act’ in Roger D Blair and D Daniel Sokol (eds), The Oxford Handbook of International Antitrust Economics, Vol 2 (Oxford University Press 2015) 160. 11 12

The relevance of economics in US, EU and Australian competition law  61 considerable evolution of ordo-liberalism over time.14 Its core, however, has remained, which is a view of free markets (meaning freedom of action for market participants) as vital for the success of a liberal democracy, but at the same time as inherently fragile and vulnerable to aggregations of economic power (whether public or private).15 The protection of the freedom to compete for all firms is consequently central to ordo-liberal prescriptions for competition policy.16 This is sometimes referred to by ordo-liberals as the protection of the process of competition, by which is meant the protection of a market structure composed of many firms and the rivalry between them that produces choices for consumers. This is not necessarily the same as the protection of a process that delivers the best outcomes for consumers. Originally, in the writings of the Freiburg School and Eucken in particular, ordo-liberal prescriptions extended beyond the prohibition of cartels to direct intervention to reduce economic power (by breaking up dominant firms) and, where this was not possible, forcing firms with economic power to behave ‘as if’ they were subject to competition.17 However, by the time that the TFEU was negotiated in the 1950s, Eucken’s highly interventionist form of ordo-liberalism had been replaced by a much softer version.18 According to Peter Behrens, ‘second generation’ ordo-liberals at that time advocated the creation of rules of the game focused on prohibiting exclusionary practices in order to establish a ‘system of undistorted competition’, rather than regulatory interventions to constrain exploitative conduct or break firms up.19 While there is much debate regarding both what ordo-liberals prescribe and the contribution of ordo-liberalism to European competition policy,20 the formalistic approach of the European courts to the application of Articles 101 and 102 likely owes much to ordo-liberalism. The ordo-liberal emphasis on the ‘freedom to compete’ and the process of competition lends itself to the adoption of per se prohibitions on conduct considered likely to harm competitors and the market structure, without a need to examine the ultimate impact of the behaviour on consumers. Ordo-liberals assume that the preservation of unconcentrated market structures and rivalry, in itself, will deliver greater choice for consumers and, as such, they see no need to go further to posit and test coherent theories of how consumers would ultimately be harmed. Even Behrens – who is generally at pains to defend ordo-liberalism from what he sees as a large number of misunderstandings regarding its relationship with and influence on European competition policy – accepts it to be ‘true that ordoliberal reasoning has led to per se prohibitions of certain types of abuses’.21 Behrens views the formalistic nature of Article 102 positively, as reflecting both the ordo-liberal goal of preserving competition and consumer choice, and administrative pragmatism: he argues that whereas the structure of competition and consumer choice are measurable (presumably by simply measuring the number of competitors), the more economic consumer welfare standard lacks measurability.

14 Peter Behrens, ‘The Ordoliberal Concept of “Abuse” of a Dominant Position and Its Impact on Article 102 TFEU’ in Paul Nihoul and Iwakazu Takahashi (eds), Abuse Regulation in Competition Law (Proceedings of the 10th ASCOLA Conference Tokyo 2015) 8–9, https://​ssrn​.com/​abstract​=2​ 658045. 15 Larouche and Schinkel (n 13), 162. 16 Ibid. 17 Behrens (n 14), 3–4; Larouche and Schinkel (n 13), 162. 18 Behrens (n 14), 4–5. 19 Ibid, 8–11, 29. 20 For a review of the debate and a particular perspective on it, see ibid. 21 Ibid, 28.

62  Research handbook on methods and models of competition law Behrens challenges the frequent allegation that the ordo-liberal approach protects competitors rather than competition.22 In doing so, he quotes Schweitzer approvingly. Schweitzer posits that the challenge is to ‘distinguish those acts with exclusionary effects that result from legitimate competition on the merits from exclusionary acts which cannot be justified as normal acts of competition but which, on the contrary, exploit the special power that a dominant firm possesses so as to entrench the firm’s position in the marketplace’(emphasis added).23 Behrens goes on to conclude that, rather than protect competitors at all times, Article 102 TFEU ‘assumes an individual right of each competitor not to be excluded by illegal acts at the expense of consumers’ choice, irrespective of whether the exclusion results in verifiable inefficiencies or in a verifiable decrease of consumer welfare’ (emphasis added).24 The difficulty that advocates of a more economic, consumer welfare-focused approach have with this viewpoint is that, without a clear theory of harm that allows for an examination of the effects of conduct on consumers, illegal acts are undefined and largely in the eye of the beholder: authorities and courts will too easily and often regard behaviour that is unorthodox (as well as behaviour that is commonplace, but not well understood) as ‘abnormal’ or ‘illegal’ on the basis that it may harm a competitor. In very simple terms, proponents of a more economic approach fear false positives (findings of anticompetitive conduct when there is no harm to consumers) much more than ordo-liberal proponents of the form-based approach. The ordo-liberal approach would not, however, have seemed out of step internationally at the time that the Treaty of Rome was drafted in the 1950s. As discussed, the formalism of the SCP paradigm and the inhospitality tradition were dominant in antitrust policy on the other side of the Atlantic from the mid-1930s to the late 1960s. A formalistic approach with a focus on the protection of competitors against what was seen as the problematic might of large businesses would have seemed very much concordant with US antitrust policy at the time. B

The European Commission’s ‘Modernization’ Process

The striking difference between European and US competition policies over the last 50 years or so is that, whereas the US experienced an upheaval in the 1970s and beyond, from form-based rules to effects-based rules of reason, Europe did not – or at least did not until much later, and then only partially. Whereas in the US the promotion of consumer welfare became recognized as the goal of antitrust, and authorities and courts came to assign considerable weight to efficiency explanations and procompetitive rationales for behaviour previously seen as objectionable by nature, until at least 2000 the emphasis in Europe remained on the form of the conduct (within both the courts and authorities), with many innocuous agreements, vertical restraints and mergers above certain thresholds effectively presumed to be illegal. Many factors may have contributed to Europe’s contrasting path, including the ongoing strength of ordo-liberal perspectives in Europe; the lack of proximity to the activism by Chicago, post-Chicago and transaction cost proponents of a more economic approach; a less economic tradition and fewer trained economists in European competition policy; and

Ibid, 26. Ibid, 28, quoting H Schweitzer, ‘The History, Interpretation and Underlying Principles of Section 2 Sherman Act and Article 82 EC’ in CD Ehlermann and M Marquis (eds), European Competition Law Annual 2007 – A Reformed Approach to Article 82 EC (Hart Publishing 2008) 134, 142. 24 Behrens (n 14), 28. 22 23

The relevance of economics in US, EU and Australian competition law  63 that Europe did not experience the same political shift towards capitalism and faith in the self-healing nature of free markets as the US experienced in the 1970s and 1980s. Whatever the cause, by 2000 there was widespread realization in Europe of a considerable divergence from the US and the inadequacy of the European Commission’s approach. In the following years, a number of events provided the catalyst for change. First, in 2001, the European Commission blocked the GE/Honeywell conglomerate merger, citing concerns that amounted essentially to efficiency offences (that is, that the merged firm would become so much more efficient than its competitors as to compromise their competitiveness).25 The merger was cleared by the US authorities (and 11 others) and considerable tensions arose between the US and Europe over what US authorities and politicians viewed as overly interventionist European merger policy. Shortly afterwards, in 2002, the European Court of First Instance (now the General Court) annulled three Commission merger decisions (Airtours,26 Schneider27 and Tetra Laval28) with stinging criticism of what was viewed as inadequate economic analysis. This acted as a shock treatment, jolting the Commission into the realization that systemic change was needed in order for the Commission to develop stronger economic foundations for its decisions. The rhetoric of the modernization process was to move from ‘form-based’ presumptions to a requirement for clear theories of harm (to consumers) and detailed assessments of actual or likely effects. The key elements of the modernization within the Commission have been the following: ●● the appointment of the Competition Directorate’s first Chief Economist and the creation of a dedicated economics team, which now includes more than 30 economists (2003); ●● the creation of the Economic Advisory Group on Competition Policy (EAGCP) – a group of leading competition economists that advises the Competition Directorate and the Chief Economist in relation to internal debates within the Commission (2003); ●● changes in the merger standard (from ‘dominance’ to ‘significant impediment to effective competition’) to emphasize effects over form (2004);29 and ●● Commission guidelines that emphasize economics, effects and consumer welfare (rather than the form of the conduct and the protection of competitors): ––

Guidelines on Vertical Restraints under Article 101 place an emphasis on consumer welfare and include a block exemption for non-‘hardcore’ vertical restrictions below

25 Case COMP/M.220, General Electric/Honeywell v Commission, 3 June 2001. Other decisions that diverged from the US were Case No IV/M.877, Boeing/McDonnell Douglas, 8 December 1997, OJ L 336, 8.12.1997, 16–47; Case No IV/M.938, Guinness/Grand Metropolitan, 15 October 1997, OJ L 288, 27.10.1998, 24–54. 26 Judgment of the Court of First Instance (Fifth Chamber, extended composition) of 6 June 2002, Airtours plc v Commission of the European Communities, Case T-342/99, ECLI:​EU:​T:​2002:​146. 27 Judgment of the Court of First Instance (First Chamber) of 22 October 2002, Schneider Electric SA v Commission of the European Communities, Case T-77/02, ECLI:​EU:​T:​2002:​255. 28 Judgment of the Court of First Instance (First Chamber) of 25 October 2002, Tetra Laval BV v Commission of the European Communities, Case T-5/02, ECLI:​EU:​T:​2002:​264. The ECJ later dismissed an appeal by the Commission: see Judgment of the Court (Grand Chamber) of 15 February 2005, Commission of the European Communities v Tetra Laval BV, Case C-12/03 P, ECLI:​EU:​C:​2005:​87. 29 Regulation (EC) No 139/2004 of 20 January 2004 on the Control of Concentration between Undertakings (The EC Merger Regulation) [2004] OJ L 24/1.

64  Research handbook on methods and models of competition law

–– –– ––

certain market share thresholds (2000, updated 2010);30 Guidelines on Horizontal Agreements under Article 101 (2001, updated 2011);31 Horizontal and Non-Horizontal Merger Guidelines (2004 and 2008, respectively);32 and Guidance on Enforcement Priorities under Article 102, which affords a central role to the ‘as efficient competitor’ (AEC)/‘equally efficient operator’ (EEO) test in the Commission’s assessment of which price-based exclusionary conduct matters to pursue under Article 102 (2009).33

The clearest impact of the modernization process has been in the context of merger policy and enforcement. Merger investigations by the Commission today are typically firmly grounded in economics, with a strong focus on the effects on consumers, and sophisticated economic and econometric modelling is now routinely conducted. By contrast, in the context of Articles 101 and 102 TFEU, the modernization process has struggled to overcome the grip of legal formalism. A clear example of the persistence of form over effect is that resale price maintenance (RPM) continues to be regarded as a ‘hardcore’ restriction under Article 101, presumed to be illegal regardless of market shares and regardless of the many potential efficiency explanations for vertical restraints including RPM. The contrast with the US following the Leegin decision is stark here. Another example is the treatment of territorial and customer restrictions on passive (that is, unsolicited) sales and parallel trade, although EU single market goals play a role here. As the following subsections explain, tendencies have remained for the Commission (particularly the legal division responsible for defending decisions on appeal before the courts) and for the courts to prefer form-based presumptions over a whole-hearted embracement of effects-based analysis. Indeed, not too long ago there was a sense among the economic and legal community that Europe had reached the ‘end of economics’, with a number of decisions of the General Court (notably in Cartes bancaire with respect to alleged anticompetitive agreements under Article 101 and in Intel with respect to loyalty rebates under Article 102) leaving no room for economic analysis of the effects of the conduct on consumers. However, a number of recent judgments from Europe’s highest court, the European Court of Justice (ECJ), have breathed new life into effects-based analysis under Articles 101 and 102. C

Form versus Effects under Article 101

Article 101(1) TFEU prohibits ‘all agreements between undertakings, decisions by associations of undertakings and concerted practices which may affect trade between member states

Guidelines on Vertical Restraints [2010] OJ C 130/1. Guidelines on the Applicability of Article 101 of the Treaty on the Functioning of the European Union to Horizontal Co-Operation Agreements [2011] OJ C 11/1. 32 Guidelines on the Assessment of Horizontal Mergers under the Council Regulation on the Control of Concentrations between Undertakings [2004] OJ C 31/5; Guidelines on the Assessment of Non-Horizontal Mergers under the Council Regulation on the Control of Concentrations between Undertakings [2008] OJ C 265/6. 33 Guidelines on the Commission’s Enforcement Priorities in Applying Article 82 of the EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings [2009] OJ C 45/7. Article 82 of the EC Treaty is now Article 102 TFEU. 30 31

The relevance of economics in US, EU and Australian competition law  65 and which have as their object or effect the prevention, restriction or distortion of competition within the internal market’ (emphasis added). The standard of proof is significantly lower when a case is brought as a ‘by object’ case, where a less fulsome assessment of effects is required. As a result, the object limb of Article 101(1) has long been seen by competition authorities and parties seeking to challenge agreements as an attractive shortcut to infringement findings.34 The scope for competition authorities and courts to classify agreements as anticompetitive by object, without a full examination of the effects of the conduct, has consequently become a key battleground between European lawyers and economists (or, more strictly, between those willing to follow a form-based approach to competition policy and those keen to analyse the overall economic effects of behaviour on competition and, ultimately, consumers). There are, of course, sound rationales for a system whereby types of conduct that are considered rarely to be of a nature that benefits consumers may be classified as an infringement without extensive investigation of the conduct’s effects. These include to provide legal certainty and deterrence in relation to that conduct and to achieve administrative expediency. However, such a system will only work well if the conduct falling within the classification system is capable of being clearly defined. The very concept of an infringement by object – where ‘object’ is supposed to be an objective matter to be revealed by the nature of the agreement in its context, rather than what was in the minds of the allegedly infringing firms – reflects ordo-liberal ambitions that conduct can easily be classified as harmful to the structure of competition or otherwise by its very form, without further investigation. While this ambition may work adequately for extreme forms of conduct such as naked horizontal cartels between competitors – where there is no legitimate reason for the conduct other than to extract consumer surplus for the firms in question – the reality is that if one takes the time to understand behaviour in depth, as economists are inclined to do, very little other conduct can be so conveniently classified: the uncomfortable truth is that grey areas abound. Even horizontal price fixing cannot always be presumed to have an anticompetitive object, as the Commission recognized when pursuing Visa International in relation to the setting of multilateral interchange fees (MIFs).35 The system will also only maintain a proportionality to its rationales if it is not subject over time to creeping capture of an ever-widening range of conduct. Particularly during the early part of this century, a tendency for the courts and authorities to expand the range of conduct classified as ‘object’ restrictions led to a significant sidelining of economic analysis of Article 101 cases and regular triumphs of form over effects. The following paragraphs elaborate on this history. A test for delineating ‘object’ infringements was first laid down by the ECJ in Société Technique Minière (STM) in 1966.36 In that case, the ECJ established the test for whether an 34 In principle, even when a restriction is found to be an ‘object’ restriction under Article 101(1), there is scope for defendants to claim an exemption from infringement under a limited ‘efficiencies’ defence in Article 101(3). However, the scope for application of Article 101(3) is narrow and the Commission rarely undertakes a thorough evaluation of efficiencies under Article 101(3). 35 Case No COMP/29.373, Visa International, OJ L 318, 22.11.2002, [64]–[69]. The Commission concluded that although the MIFs restricted the freedom of banks to decide their own prices, this amounted to a restriction of competition by effect, not a restriction of competition by object. 36 Judgment of the Court of 30 June 1966, Société Technique Minière (LTM) v Maschinenbau Ulm GmbH (MBU), Case 56/65, ECLI:​EU:​C:​1966:​38 (hereafter STM).

66  Research handbook on methods and models of competition law agreement can be classified as an object infringement as whether an analysis of the agreement, having regard to its purpose and economic context, reveals ‘the effect on competition to be sufficiently deleterious’ (emphasis added).37 The term ‘sufficiently deleterious’ suffers from an analytical imprecision that permeates much of ordo-liberal thinking: it allows significant scope for subjective judgments based on inductive intuition rather than the application of careful analytical reasoning.38 Moreover, how harmful the conduct must be to satisfy the sufficiency condition is not clear on its face. For both reasons, the test is consequently ripe for arbitrary in-the-eye-of-the-beholder judgments and expansions or narrowings of scope. Shortly afterwards, in the joined Consten and Grundig cases, the ECJ emphasized that some clauses in an agreement may be by their nature so restrictive of competition to be classified as object restrictions, and found that export bans (that is, parallel import restrictions) fell into that category.39 A variant of this test (‘by their very nature’) was applied by the ECJ in its 2008 Beef Industry decision, which concerned an agreement to restrict output.40 This ‘nature’ test implies that to be an object restriction, the agreement must have only an anticompetitive aspect without any redeeming procompetitive rationale. However, its practical application can lead to greater capture of agreements than the ‘sufficiently deleterious’ test because it allows a decision-maker to reach a conclusion that an agreement has a restrictive nature on the basis of a partial frame of reference, without considering the effects of the agreement in its overall context.41 Over time, the analytical looseness of these tests and a dominance of formalism over economic analysis has resulted in European courts and authorities classifying as ‘by object’ a range of agreements (including various vertical arrangements) going well beyond unambiguously harmful hardcore horizontal cartels, including agreements where economic theory identifies potential positive effects for competition and consumers that deserve a broader assessment. This has been the result of narrow assessments of the nature or the effects of these agreements, without due regard to the potential for the agreements to have overarching efficiency or procompetitive consumer-welfare-enhancing rationales.42 Ibid, [249]. As Matthew O’Regan has eloquently explained, identifying an object restriction under Article 101 is much like ‘duck’ and ‘elephant’ hunting. Inductive reasoning to identify ‘object’ restrictions under Article 101 often takes the form or either ‘if it looks like a duck, swims like a duck and quacks like a duck, then it probably is a duck’ (the duck test) or ‘it is difficult to describe, but you know it when you see it’ (the elephant test). Matthew O’Regan, ‘Restrictions by Object: Duck and Elephant Hunting with the Court of Justice’, Kluwer Competition Law Blog (20 October 2014), www​.kluwercompetitionlawblog​ .com. 39 Judgment of the Court of 13 July 1966, Établissements Consten SàRL and Grundig-Verkaufs-GmbH v Commission of the European Economic Community, Joined Cases 56/64 and 58/64, ECLI:​EU:​C:​1966:​ 41 (hereafter Joined Cases 56/64 and 58/64, Consten & Grundig, 1966:41), [301]. 40 Judgment of the Court (Third Chamber) of 20 November 2008, Competition Authority v Beef Industry Development Society Ltd and Barry Brothers (Carrigmore) Meats Ltd, Case C‑209/07, ECLI:​ EU:​C:​2008:​643 (hereafter Beef Industry), [17]. 41 The suggestion that some conduct is an object restriction by its very nature suggests that not even a rudimentary assessment of the effects of the conduct in its context is required. 42 Much of this went undebated for many years because the EU developed a largely formalistic system for the enforcement of Article 101 governed by block exemption regulations that classified certain types of clauses as prohibited and others as authorized. With the focus on whether a particular agreement fell within the prohibited or authorized classifications, there was little consideration of the nature of the restriction in its entire context and its effects. 37 38

The relevance of economics in US, EU and Australian competition law  67 One example is export bans, which are common around the world – in particular in the pharmaceutical sector, which is characterized by large fixed costs of research and development and small marginal costs. Export bans typically have a procompetitive rationale in achieving an efficient recovery of fixed costs across consumers with different willingness to pay. In essence, export bans serve to preserve price discrimination schemes, which are well understood in economics to have a priori ambiguous consumer welfare effects, and the potential to considerably enhance overall consumer welfare by allowing customer groups to be served that would not be served without the price discrimination. However, the violence that export bans do to the unique European concept of the free internal market marked them for classification as ‘by object’ restrictions, despite their broader effects.43 Territorial restrictions on passive sales have much the same nature – they may enhance efficiency and consumer welfare but also offend the concept of the free internal market – and have met the same fate.44 Another example is selective distribution agreements, which have also been found to infringe Article 101 by object,45 notwithstanding their potential – well recognized by economists – to represent efficient output-expanding arrangements (for example, by precluding free-riding by some distributors on the promotional efforts of others and thereby preserving incentives for those efforts). Yet another is RPM,46 which in the US, following Leegin,47 is now subject to rule of reason analysis in recognition of its potential to have legitimate efficiency and consumer-welfare-enhancing explanations. In the mid to late 2000s, as enforcement of Article 101 was devolved to the Member States, there was growing discomfort – among both competition economists and lawyers – about the ease with which the competition authorities and courts were defining agreements as restrictions by object. Something of a crisis point was reached following an ECJ judgment in T-Mobile in 2009, which appeared to significantly lower the threshold for object classifications. In T-Mobile, the ECJ held that an agreement can be regarded as having an anticompetitive object if it has merely the potential or is capable of having a restrictive effect on competition.48 43 Joined Cases 56/64 and 58/64, Consten & Grundig, 1966:41; Judgment of the Court (Third Chamber) of 6 October 2009, Joined Cases GlaxoSmithKline Services Unlimited v Commission of the European Communities (C-501/06 P) and Commission of the European Communities v GlaxoSmithKline Services Unlimited (C-513/06 P) and European Association of Euro Pharmaceutical Companies (EAEPC) v Commission of the European Communities (C-515/06 P) and Asociación de exportadores españoles de productos farmacéuticos (Aseprofar) v Commission of the European Communities (C-519/06 P), ECLI:​EU:​C:​2009:​610 (hereafter Joined Cases C-501/06 P, C-513/06 P, C-515/06 P and C-519/06 P, GSK, 2009:610 or just GSK). 44 See Judgment of the Court (Grand Chamber) of 4 October 2011, Joined Cases Football Association Premier League Ltd and Others v QC Leisure and Others (C-403/08) and Karen Murphy v Media Protection Services Ltd (C-429/08), ECLI:​EU:​C:​2011:​631. 45 Judgment of the Court (Third Chamber) of 13 October 2011, Pierre Fabre Dermo-Cosmétique SAS v Président de l’Autorité de la concurrence and Ministre de l’Économie, de l’Industrie et de l’Emploi, Case C-439/09, ECLI:​EU:​C:​2011:​649 (hereafter Case C-439/09, Pierre Fabre, 2011:649 or just Pierre Fabre). 46 See Commission Decision of 16 July 2003 in COMP/37.975, Po/Yamaha, and, more recently, Commission Decision of 24 July 2018 in Case AT.40465, Asus (and similar cases decided on the same date concerning Denon & Marantz (AT.40469), Philips (AT.40481) and Pioneer (AT.40482)). 47 See note 12 above. 48 Judgment of the Court (Third Chamber) of 4 June 2009, T-Mobile Netherlands BV, KPN Mobile NV, Orange Nederland NV and Vodafone Libertel NV v Raad van bestuur van de Nederlandse Mededingingsautoriteit, Case C-8/08, ECLI:​EU:​C:​2009:​343 (hereafter T-Mobile), [31] and [43].

68  Research handbook on methods and models of competition law This was widely interpreted as a much lower threshold than the ‘sufficiently deleterious’ and ‘by nature’ standards, and gave extra impetus to competition authorities across Europe to bring cases under the object limb as a shortcut to infringement findings in relation to an ever-widening range of conduct, without having to engage in full effects assessments. This included ‘by object’ classifications of many vertical agreements in the context of the rising digital economy (including online hotel booking most-favoured-nation clauses in a number of Member States). The T-Mobile decision also confirmed the continuing dominance of the ordo-liberal aim of preserving market structures, without confirming harmful effects on consumers.49 The ECJ stated:50 like the other competition rules of the Treaty, [Article 101] is designed to protect not only the immediate interests of individual competitors or consumers but also to protect the structure of the market and thus competition as such. Therefore, contrary to what the referring court would appear to believe, in order to find that a concerted practice has an anti‑competitive object, there does not need to be a direct link between that practice and consumer prices.

The situation was compounded by the General Court’s 2012 decision in Cartes bancaire. In addition to repeating the lower threshold set by the ECJ in T-Mobile – that it is sufficient for a finding of a restriction by object if the agreement has the potential or is capable of restricting competition51 – the General Court was interpreted as stating that the concept of restrictions by object should not be applied narrowly.52 This was in the context of a case that concerned a restriction on one side of a two-sided platform where, seen in its broader context, the restriction benefitted customers on the other side and had at least the potential of being necessary for

Ibid, [36]–[39]. Ibid, [38]–[39]. Shortly afterwards, in its 2009 GSK decision, the ECJ further emphasized this point (Joined Cases C-501/06 P, C-513/06 P, C-515/06 P and C-519/06 P, GSK, 2009:610, [62]–[64]): With respect to the Court of First Instance’s statement that, while it is accepted that an agreement intended to limit parallel trade must in principle be considered to have as its object the restriction of competition, that applies in so far as it may be presumed to deprive final consumers of the advantages of effective competition in terms of supply or price, the Court notes that neither the wording of Article 81(1) EC nor the case-law lend support to such a position. First of all, there is nothing in that provision to indicate that only those agreements which deprive consumers of certain advantages may have an anti-competitive object. Secondly, it must be borne in mind that the Court has held that, like other competition rules laid down in the Treaty, Article 81 EC aims to protect not only the interests of competitors or of consumers, but also the structure of the market and, in so doing, competition as such. Consequently, for a finding that an agreement has an anti-competitive object, it is not necessary that final consumers be deprived of the advantages of effective competition in terms of supply or price (see, by analogy, T-Mobile Netherlands and Others, cited above, paragraphs 38 and 39). It follows that, by requiring proof that the agreement entails disadvantages for final consumers as a prerequisite for a finding of anti-competitive object and by not finding that that agreement had such an object, the Court of First Instance committed an error of law. 51 Judgment of the General Court (Seventh Chamber) of 29 November 2012, Groupement des cartes bancaires ‘CB’ v European Commission, Case T-491/07, ECLI:​EU:​T:​2012:​633 (hereafter Case T-491/07, Cartes bancaires, 2012:633), [125]. 52 Ibid, [124]. 49 50

The relevance of economics in US, EU and Australian competition law  69 the continued existence of the platform.53 At this point, in the view of many in the competition law and economics community, Europe had reached the ‘end of economics’ in Article 101 cases, with more and more conduct with ambiguous effects being summarily caught by the ever-creeping expansion of a formal net. Against this background, the ECJ’s 2014 Cartes bancaire decision, on appeal from the General Court, was hailed as a much-needed correction, and revived hope that serious economic analysis will have a role to play in future Article 101 cases. This was the first time that the ECJ had set aside a Commission finding of a restriction by object. Notably, the ECJ criticized the General Court for relying on the ECJ’s own looser potential or capable standard (from T-Mobile), and instead re-emphasized the ‘sufficiently deleterious’ and ‘by nature’ tests from STM and Beef Industry, among other cases.54 The ECJ observed a fundamental error in the General Court’s finding of a restriction by object given that the General Court had also found on the facts that the clauses in dispute pursued a legitimate objective of combating free-riding in the context of indirect network effects and essential interactions between actors on each side of a two-sided market.55 In essence, in the view of the ECJ, these facts – in particular, the two-sided nature of the market – revealed sufficient grounds to require an effects assessment. The ECJ also admonished the General Court for suggesting that the concept of an object restriction should not be conceived narrowly.56 The essence of the ECJ decision can be found in the following statement:57 in the light of that case-law, the General Court erred in finding … that the concept of restriction of competition by ‘object’ must not be interpreted ‘restrictively’. The concept of restriction of competition by object can be applied only to certain types of coordination between undertakings which reveal a sufficient degree of harm to competition that it may be found that there is no need to examine their effects otherwise the Commission would be exempted from the obligation to prove the actual effects on the market of agreements which are in no way established to be, by their very nature, harmful to the proper functioning of normal competition.

The ECJ’s opinion may have been influenced by the following passage from the advisory opinion of Advocate General Wahl:58

53 The Commission had found restrictions by object and effect. The General Court upheld the Commission’s object finding and did not go on to consider its effect finding. 54 Case C‑67/13 P, Cartes bancaires, 2014:2204, [56]–[57] and [49]–[52]. In doing so, the ECJ did not acknowledge any misstep of its own. Instead, it held that the General Court had erred by ‘in part’ failing to have regard to the case law, pointing to the higher standards in its STM and Beef Industry decisions. 55 Case C‑67/13 P, Cartes bancaires, 2014:2204, [69]–[87]. 56 Ibid, [58]. See also AG Wahl’s opinion on the case, which called for a ‘cautious attitude’ to and ‘strict interpretation’ of the notion of restriction by object (Opinion of AG Wahl in Case C-67/13 P, Cartes bancaires, 2014:1958, [59], [74]). In fairness to the General Court, when suggesting that the concept of an object restriction should not be conceived narrowly, it may simply have meant that object restrictions should not necessarily be limited to a closed list of hardcore restrictions – that is, that an open mind should be had as to whether a particular restriction may be a new kind of hardcore restriction. Understood in this way, the ECJ’s (and AG Wahl’s) criticism of the General Court in this respect may have been somewhat unjustified. 57 Case C‑67/13 P, Cartes bancaires, 2014:2204, [58]. 58 Opinion of AG Wahl in Case C-67/13 P, Cartes Bancaires, 2014:1958, [53]–[58]. In the particular context of the Cartes bancaire case, AG Wahl considered (as did the ECJ) that the General Court failed

70  Research handbook on methods and models of competition law Considering an agreement or a practice to restrict competition on account of its very object has significant consequences, at least two of which should be highlighted. First of all, the method of identifying an ‘anticompetitive object’ is based on a formalist approach which is not without danger from the point of view of the protection of the general interests pursued by the rules on competition in the Treaty. Where it is established that an agreement has an object that is restrictive of competition, the ensuing prohibition has a very broad scope, that it is to say it can be imposed as a precautionary measure and thus jeopardise future contracts, irrespective of the evaluation of the effects actually produced. This formalist approach is thus conceivable only in the case of (i) conduct entailing an inherent risk of a particularly serious harmful effect or (ii) conduct in respect of which it can be concluded that the unfavourable effects on competition outweigh the pro-competitive effects. To hold otherwise would effectively deny that some actions of economic operators may produce beneficial externalities from the point of view of competition. In my view, it is only when experience based on economic analysis shows that a restriction is constantly prohibited that it seems reasonable to penalise it directly for the sake of procedural economy. Only conduct whose harmful nature is proven and easily identifiable, in the light of experience and economics, should therefore be regarded as a restriction of competition by object, and not agreements which, having regard to their context, have ambivalent effects on the market or which produce ancillary restrictive effects necessary for the pursuit of a main objective which does not restrict competition. Second, such classification relieves the enforcement authority of the responsibility for proving the anticompetitive effects of the agreement or the practice in question. An uncontrolled extension of conduct covered by restrictions by object is dangerous having regard to the principles which must govern evidence and the burden of proof in relation to anticompetitive conduct. Because of these consequences, classification as an agreement which is restrictive by object must necessarily be circumscribed and ultimately apply only to an agreement which inherently presents a degree of harm. This concept should relate only to agreements which inherently, that is to say without the need to evaluate their actual or potential effects, have a degree of seriousness or harm such that their negative impact on competition seems highly likely. Notwithstanding the open nature of the list of conduct which can be regarded as restrictive by virtue of its object, I propose that a relatively cautious attitude should be maintained in determining a restriction of competition by object. Only conduct whose harmful nature is proven and easily identifiable, in the light of experience and economics, should therefore be regarded as a restriction of competition by object, and not agreements which, having regard to their context, have ambivalent effects on the market or which produce ancillary restrictive effects necessary for the pursuit of a main objective which does not restrict competition.

On the very same day as its Cartes bancaire decision was released, the ECJ released a separate decision in MasterCard that limited further the scope of an already barely used Article 101(3) efficiencies exemption. Article 101(3) provides that Article 101(1) does not apply to an agreement that ‘contributes to improving the production or distribution of goods or to promoting technical or economic progress, while allowing consumers a fair share of the resulting benefit’. In MasterCard, which was another case concerning a two-sided payment card platform, the ECJ defined separate markets for acquiring and issuing (the activities on each side of the platform) and held that if there is a restriction in one market, then beneficial effects flowing from the restrictive measure on a separate but connected second market cannot compensate for the restrictive effects in the first.59 From this, it seems that Article 101(3) can only be used to to have proper regard to the two-sided nature of the market and to take into account the contribution of the system on both sides of the market: [146]–[150]. 59 Judgment of the Court (Third Chamber), 11 September 2014, MasterCard Inc and Others v European Commission, Case C-382/12 P, ECLI:​EU:​C:​2014:​2201.

The relevance of economics in US, EU and Australian competition law  71 exempt an agreement from Article 101(1) if the benefit claimed is in the same market as the restriction. The rationale for delimiting an efficiency defence in such a way is difficult for an economist to comprehend and sits rather uneasily with Article 101(3)’s apparent promise of exempting agreements that contribute ‘to improving the production or distribution of goods or to promoting technical or economic progress, while allowing consumers a fair share of the resulting benefit’. However, when the Cartes bancaire and MasterCard decisions are read together, it is possible to conclude that the narrow interpretation of Article 101(3) in the context of two-sided markets may not matter much. This is because both decisions point European competition authorities and courts towards taking the entire economic effects of agreements in two-sided markets into account within Article 101(1). Indeed, this appears to be how, in the UK, the High Court has approached such agreements in the context of follow-on damages claims brought against MasterCard.60 If anything, this should result in sounder economic analysis of the overall effects of agreements in two-sided markets than a two-stage process of first identifying a restriction on one side under Article 101(1) and then assessing whether there are exculpatory efficiencies on the other or in the round under Article 101(3). It will be interesting to see whether, in time, the principle of holistic assessment of all the effects under Article 101(1) will be extended beyond two-sided market contexts to other agreements that have traditionally been considered by object, but are understood in economics to have, a priori, ambiguous effects, including vertical arrangements such as selective distribution agreements, RPM and restrictions on parallel trade.61 The ECJ’s most recent Article 101(1) decision in the Coty case62 is not promising in this respect. While many have hailed the Cartes bancaire decision as demonstrating an ECJ pivot to economic analysis over form-based presumptions, the Coty decision demonstrates that there is much nuance and precedent for the ECJ to navigate, and any pivot with respect to vertical restraints will be gradual at best. The Coty case concerned a selective distribution agreement in the context of premium goods. The assessment of such agreements under Article 101(1) has long been within a framework of a set of (formalistic) safe harbour conditions laid down by the ECJ in its Metro and L’Oréal decisions (in 1977 and 1980 respectively). These decisions established that the use of a selective distribution system for luxury goods is not a restriction by object and complies with Article 101(1) if resellers are chosen on the basis of objective criteria of a qualitative nature that are laid down uniformly for all potential resellers and applied in a non-discriminatory fashion; the characteristics of the product in question necessitate such a distribution network in order to preserve its quality and ensure its proper use; and the criteria laid down do not go

60 The High Court found in favour of MasterCard on the basis of an argument under Article 101(1) that MasterCard’s system would collapse without the allegedly restrictive agreements. In other words, a full holistic effects assessment was applied under Article 101(1) – assessed against a counterfactual without the agreements – to find that the agreements were not in fact restrictive of competition in an overall sense. 61 For example, it will be interesting to see whether future challenges to restrictions on parallel trade will fail at the Article 101(1) hurdle on the basis of a more holistic assessment that, in the absence of those restrictions, the manufacturer’s ability to fund future innovation would be severely impacted with adverse consequences for future consumers. 62 Case C-230/16, Coty, 2017:941.

72  Research handbook on methods and models of competition law beyond ‘what is necessary’.63 In a somewhat controversial decision, the ECJ in Pierre Fabre held that selective distribution agreements falling outside of the Metro and L’Oréal safe harbour (for example, those that discriminate between different types of resellers) will be classified as restrictions by object.64 In Coty, a decision handed down in late 2017, the ECJ passed up an opportunity to clarify that selective distribution agreements – including complete bans on online sales – have a priori ambiguous overall effects and should be assessed as ‘effect’ rather than object restrictions. Instead, the ECJ merely distinguished the online sales restrictions in Coty from those in Pierre Fabre, noting that in Coty the restrictions were only in relation to discernible third-party online platforms such as Amazon and eBay, whereas in Pierre Fabre the restrictions represented a complete ban on online sales of cosmetic products. This distinction appears to have saved the restrictions in Coty from classifications as restrictions by object. However, complete bans on online sales appear to remain susceptible to findings of restrictions by object, notwithstanding their vertical rather than horizontal nature and their potential for beneficial overall effects. D

Form versus Effects in Rebate Cases under Article 102

The structural focus of competition law in Europe on the number of competitors, rather than consumer welfare, is evident also in the context of the Article 102 prohibition on abuse of a dominant position. For example, in the seminal Hoffman-La Roche decision in 1979, the ECJ stated its view of abuse of dominance under Article 102 as follows:65

63 See Judgment of the Court of 25 October 1977, Metro SB-Großmärkte GmbH & Co KG v Commission of the European Communities, Case 26/76, ECLI:​EU:​C:​1977:​167 (Metro), [20], and Judgment of the Court of 11 December 1980, NV L’Oréal and SA L’Oréal v PVBA ‘De Nieuwe AMCK’, Case 31/80, ECLI:​EU:​C:​1980:​289 (L’Oréal), [15] and [16]. 64 See Case C-439/09, Pierre Fabre, 2011:649. Many competition law practitioners interpreted paragraph 46 of this decision as a blanket classification of all selective distribution systems that seek to maintain a luxury image as ‘by object’ restrictions. However, read in context together with paragraph 47, and in light of the ECJ’s clarifications in its subsequent Coty decision, it appears that the statement in paragraph 46 should be read as limited to the particular goods and the particular nature of the restriction in Pierre Fabre (in particular, the complete rather than partial ban on online sales) going beyond what was considered necessary. For reference, paragraphs 46 and 47 of the Pierre Fabre decision are reproduced below: The aim of maintaining a prestigious image is not a legitimate aim for restricting competition and cannot therefore justify a finding that a contractual clause pursuing such an aim does not fall within Article 101(1) TFEU. In the light of the foregoing considerations, the answer to the first part of the question referred for a preliminary ruling is that Article 101(1) TFEU must be interpreted as meaning that, in the context of a selective distribution system, a contractual clause requiring sales of cosmetics and personal care products to be made in a physical space where a qualified pharmacist must be present, resulting in a ban on the use of the internet for those sales, amounts to a restriction by object within the meaning of that provision where, following an individual and specific examination of the content and objective of that contractual clause and the legal and economic context of which it forms a part, it is apparent that, having regard to the properties of the products at issue, that clause is not objectively justified. 65 Judgment of the Court of 13 February 1979, Hoffmann-La Roche & Co AG v Commission of the European Communities, Case 85/76, ECLI:​EU:​C:​1979:​36 (hereafter Case 85/76, Hoffman La Roche, 1979:36 or just Hoffmann-La Roche), [91].

The relevance of economics in US, EU and Australian competition law  73 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.

This statement has been consistently cited by the ECJ in subsequent Article 102 decisions and continues to condition the European approach to abuse of dominance assessments. In its 2007 British Airways decision, the ECJ also restated its position from its earlier Continental Can judgment that there is no need for an examination of effects on consumers under Article 102 – it is sufficient to examine only effects on (the structure of) competition:66 Article [102] EC is aimed not only at practices which may cause prejudice to consumers directly, but also at those which are detrimental to them through their impact on an effective competition structure, such as is mentioned in Article 3(1)(g) EC. The Court of First Instance was therefore entitled, without committing any error of law, not to examine whether BA’s conduct had caused prejudice to consumers, … but to examine … whether the bonus schemes at issue had a restrictive effect on competition ….

Rebates under Article 102 are an area of European competition law where ordo-liberal presumptions of harm based on effects on competitors have been particularly persistent. Indeed, until quite recently it was understood that there was an irrefutable per se presumption that loyalty rebates offered by a dominant firm were anticompetitive under Article 102 (‘by their nature’). However, as explained in this section, the ECJ has recently clarified that this presumption is rebuttable. In doing so, the ECJ has paved the way for economic effects-based assessments of all rebate schemes. From an economic perspective, allegations of exclusionary rebate schemes require a similar assessment to any other allegation of price-based exclusionary conduct, including allegations of predatory pricing (that is, allegations that a firm has priced below cost with a view to excluding competitors from the market and subsequently raising prices). In each case, customers are offered lower prices, at least in the short term, and there is the potential for this to intensify competition and result in long-term benefits for consumers. For this reason, from an economic perspective there should be no anticompetitive presumption. However, in each case there may also be the potential for competitors to be excluded if they cannot profitably match the low prices. The key challenge when assessing such allegations is therefore to distinguish healthy competition delivering long-term benefits for consumers from harmful exclusionary behaviour that will result in long-term harm to consumers. In other words, from an economic perspective, rebate cases are of a nature that their effects are in principle ambiguous a priori, and a full rule of reason assessment is needed to understand whether the behaviour is procompetitive or anticompetitive. Various economic approaches have been developed to delineate good pricing behaviour from bad, including the profit sacrifice test and the equally efficient competitor test. Applied to a rebate case, the equally efficient competitor test asks whether a competitor that is as effi66 Judgment of the Court (Third Chamber) of 15 March 2007, British Airways plc v Commission of the European Communities, Case C-95/04 P, ECLI:​EU:​C:​2007:​166 (hereafter Case C-95/04 P, British Airways, 2007:166), [106]–[107].

74  Research handbook on methods and models of competition law cient as the dominant firm could profitably match the effective prices that the dominant firm is charging over the contestable part of the customer’s demand.67 Another important economic consideration is the share of the market affected by the conduct. If the affected share is small, then, even if competitors cannot match the dominant firm’s effective prices, it is unlikely that the conduct will foreclose competitors because there will be sufficient share of the market for competitors to realize minimum efficient scale. In Europe, however, the approach to rebate cases has evolved along a largely formalistic path. In a line of cases from Hoffman-La Roche in 197968 to Post Danmark II in 201569 (passing through Michelin in 1983,70 British Airways in 200771 and Tomra in 201272), the ECJ has established and maintained the idea of a three-way classification of rebate schemes. First, rebates based purely on volumes purchased are presumed to be legal under Article 102 (see Michelin, Tomra and Post Danmark II). This reflects that economies of scale can result in cost savings (that is, lower average unit costs) when supplying larger volumes to a customer. Second, loyalty rebates (also sometimes called fidelity rebates or exclusivity rebates) that incentivize customers to purchase all or a large share of their requirements from the firm offering the rebates have been regarded as by their nature abusive under Article 102 (see Hoffman-La Roche, Michelin, Tomra and Post Danmark II). This is notwithstanding that a loyalty rebate scheme may not preclude rival firms from profitably competing for most or all of the contestable demand (that is, matching the effective prices set by the dominant firm), or the scheme may affect only a small part of the market and leave sufficient customers for competitors to achieve efficient scale and remain as effective constraints in the market. Third, the ECJ has, on occasion, encountered rebate schemes that it did not see as falling clearly within either of the first two categories. The ECJ has held that for rebates in this third category, there should be an assessment of all the relevant circumstances, including the nature

67 For example, suppose that the customer wishes to purchase 50 units of product 1 and 50 units of product 2. Suppose also that firm A is the only firm that can supply product 1. The contestable demand for this customer is then just the 50 units of product 2. Next, suppose that firm A offers the customer prices of $1 per unit for the first 50 units that the customer purchases and then, if the customer purchases 51 or more units, a price of $0.50 per unit. This means that if the customer purchases just one unit of product 2 from Firm A, the customer will save $25 on its purchases of product 1. For Firm B to make the customer indifferent between purchasing all 50 units of product 2 from it versus all 50 units of product 2 from Firm A, Firm B will need to offer the customer a price for all 50 units of product 2 of ($0.50 × 50) – $25 = $0. The effective price for the customer’s contestable demand in this case is therefore $0. Note, however, that in this example the customer will get the $25 rebate on the purchases of product 1 if the customer takes only one unit of product 2 from Firm A. Therefore, the effective price for the last 49 units of product 2 is $0.50 × 49 = $24.50. As long as Firm B can produce 49 units of product 2 for less than $24.50, Firm B will not be foreclosed from 98 per cent of the contestable demand. 68 Case 85/76, Hoffman La Roche, 1979:36. 69 Judgment of the Court (Second Chamber) of 6 October 2015, Post Danmark A/S v Konkurrenceradet, Case C-23/14, ECLI:​EU:​C:​2015:​651 (hereafter Case C-23/14, Post Danmark II, 2015:651 or just Post Danmark II). 70 Judgment of the Court of 9 November 1983, NV Nederlandsche Banden Industrie Michelin v Commission of the European Communities, Case 322/81, ECLI:​EU:​C:​1983:​313 (hereafter Case 322/81, Michelin I, 1983:313 or just Michelin I). 71 Case C-95/04 P, British Airways, 2007:166. 72 Judgment of the Court (Third Chamber), 19 April 2012, Tomra Systems ASA et al v European Commission, Case C-549/10 P, ECLI:​EU:​C:​2012:​221 (hereafter Case C-549/10 P, Tomra, 2012:221 or just Tomra).

The relevance of economics in US, EU and Australian competition law  75 of the rebate scheme, the thresholds triggering the rebates, whether those thresholds were standardized or individualized to each customer, the extent of dominance of the firm in question, the conditions of competition, and the share of the market affected (see British Airways, Tomra and Post Danmark II).73 Retroactive rebate schemes offer discounts to customers with respect to volumes already purchased if the customer purchases additional volumes. The ECJ has considered retroactive rebates to fall within the third category and they are consequently assessed on the basis of ‘all the circumstances’. The ECJ’s stricter per se approach to second-category loyalty rebates is in tension with the economic fact that a loyalty rebate scheme may be seen as a retroactive rebate scheme and vice versa. The difference between them is largely, if not entirely, in the form in which they are expressed by the firm in question. It follows that they also have similar economic effects. Both can be benign (for example, if competitors can profitably match effective prices over contestable demand or if the share of the market that is affected is small) and both can have the potential to anticompetitively foreclose competitors if: (i) a part of demand is contestable and another part is not; (ii) competitors cannot profitably compete for (that is, match the effective prices for) the contestable volumes; and (iii) a large share of the market is affected. In light of the subsequent decision in Intel (more on this below), the Post Danmark II decision – which concerned retroactive rebates – appears to represent the high watermark of the ECJ’s form-based approach to rebates. The ECJ confirmed the three-way classification from previous decisions and focused on whether the rebate scheme harmed competitors, without a broader consciousness or inclination to investigate the effects of the scheme on customers and consumers. The ECJ also made a number of statements in relation to the assessment of ‘third category’ rebates that are troubling from an economic perspective. First, the ECJ observed Post Danmark’s high share and its statutory monopoly in relation to much of the market and held that the equally efficient competitor test (that is, a comparison of effective prices and Post Danmark’s own costs) was of no relevance in that context when assessing a rebate scheme because ‘the structure of the market makes the emergence of an as-efficient competitor practically impossible’.74 The first thing that is troubling here is that, even if a competitor may only be able to achieve a small share of the market, an economic effects-based approach would not summarily conclude that it would be ‘practically impossible’ for the competitor to be as efficient as (or more efficient than) Post Danmark. Indeed, statutory monopolies are often characterized by inefficiencies. Second, when read together with its Post Danmark I decision75 – which endorsed the use of a comparison of prices and costs when assessing a predatory pricing allegation in a market context very similar to the context in Post Danmark II – the ECJ appears to be suggesting that where a dominant firm has a statutory monopoly over a large share of the market, price-cost testing can be relevant for assessing a predatory pricing

73 Note that while these matters require assessment in such cases, the ECJ held in Tomra in 2012 that a rebate system can be abusive even if it does not fail the equally efficient operator test, because it can be sufficient to demonstrate the existence of a loyalty mechanism: see Case C-549/10 P, Tomra, 2012:221, [73]–[74], [79]. See also Case 322/81, Michelin I, 1983:313, [81]–[86]. 74 Case C-23/14, Post Danmark II, 2015:651, [59]. 75 Judgment of the Court (Grand Chamber), 27 March 2012, Post Danmark A/S v Konkurrenceradet, Case C-209/10, ECLI:​EU:​C:​2012:​172 (hereafter Case C-209/10, Post Danmark I, 2012:172).

76  Research handbook on methods and models of competition law allegation, but it is irrelevant when assessing a rebate scheme. It is difficult to understand the rationale for a stricter threshold for the relevance of price-cost tests in rebate cases. Second, again in the context of Post Danmark’s high share and statutory monopoly over much of the market, the ECJ suggests that the equally efficient competitor test is of no relevance because ‘the presence of a less efficient competitor might contribute to intensifying … competitive pressure [in the market] and, therefore, to exerting a constraint on the conduct of the dominant undertaking’.76 The suggestion here is that a dominant firm with significant market share and some statutory protection should set its prices so as to hold an umbrella over the head of somewhat less efficient competitors so that they may remain viable into the future. It is true that less efficient competitors may constrain a dominant firm and deliver beneficial outcomes to consumers compared to the dominant firm setting unconstrained prices. However, a requirement that a dominant firm’s discounting should be constrained and that it should price above its own costs to sponsor the continuing viability of inefficient competitors is deeply problematic. First, it would deny customers the benefits of lower-priced offers from the dominant firm and risk inefficient outcomes: even in regulatory contexts, the protection of inefficient competitors in the form of higher prices to be paid by consumers is only advisable if those competitors are anticipated to become efficient over time. Second, a requirement to protect inefficient competitors exposes dominant firms to significant legal uncertainty: just how high must they price to avoid an infringement decision? A virtue of the equally efficient competitor standard is that dominant firms can assess their pricing against a bright line and measurable standard (that is, against their own costs). The Post Danmark II requirement to price somewhere above cost is also inconsistent with the ECJ’s prior decisions in AKZO and Post Danmark I, where it is suggested that, in the absence of anticompetitive intent, pricing down to the firm’s own average variable (or incremental) costs will not be considered abusive.77 Third, the ECJ held that ‘fixing an appreciability (de minimis) threshold for the purposes of determining whether there is an abuse of a dominant position is not justified’ and there is ‘no need to show that’ the anticompetitive effect of a rebate scheme operated by a dominant undertaking ‘is of a serious or appreciable nature’.78 This statement suggests that a rebate scheme that affects only a very small proportion of the market and does not risk foreclosing any competitor may nonetheless be considered abusive. This sits rather uneasily with the ECJ’s confirmations earlier in the decision that all of the circumstances must be considered, including ‘the number of customers concerned’,79 and ‘[t]he fact that a rebate scheme, such as that at issue in the main proceedings, covers the majority of customers on the market may constitute a useful indication as to the extent of that practice and its impact on the market, which may bear out the likelihood of an anticompetitive exclusionary effect’.80 At the same time as the ECJ was maintaining a commitment to its form-based approach to rebates, the European Commission was endeavouring to adopt a more economic approach in terms of its enforcement practice. The Commission’s 2009 Guidance on Enforcement Priorities under Article 102 was a major, much debated and ultimately well-received com-

78 79 80 76 77

Case C-23/14, Post Danmark II, 2015:651, [60]. See, in particular, Case C-209/10, Post Danmark I, 2012:172, [37]–[38]. Case C-23/14, Post Danmark II, 2015:651, [73]–[74]. Ibid, [66]. Ibid, [46], [50].

The relevance of economics in US, EU and Australian competition law  77 ponent of the Commission’s modernization process. In that Guidance, the Commission committed to assess the economic effects of alleged exclusionary pricing conduct, including rebate schemes, and pursue only cases for which it considered there to be such effects, rather than rely on the form-based prescriptions given by the ECJ. A key plank of the Commission’s economic approach within its Guidance is the use of the equally efficient competitor test. In the context of rebate schemes, this requires comparing effective prices over contestable sales to the dominant firm’s own costs.81 The Intel case provided the Commission with an opportunity to demonstrate its Guidance and the equally efficient competitor test in action in the context of a loyalty rebate case: the Commission devoted around 150 pages of its decision to that test and found that an equally efficient competitor would have been unable to compete with Intel for the contestable part of demand. Intel’s appeal of the Commission’s decision to the General Court was therefore seen as a test case of the Commission’s more economic approach to rebates: would the courts engage with and endorse the more economic approach? However, in arguments before the General Court, the Commission led with the ECJ’s form-based prohibition of loyalty rebates, rather than its assessment of the economic effects.82 Following the ECJ’s three-way rebate classification system, the General Court in 2014 found the rebates in question to be loyalty-inducing rebates as in Hoffman-La Roche,83 and held that such rebates violate Article 102 without the need to assess all the circumstances because loyalty rebates are by their very nature capable of restricting competition.84 Referring to the ECJ decisions in Hoffman-La Roche, Michelin I and British Airways, the General Court found that85 [t]he question whether an exclusivity rebate can be categorised as abusive does not depend on an analysis of the circumstances of the case aimed at establishing a potential foreclosure effect. … it is only in the case of rebates falling within the third category that it is necessary to assess all the circumstances, and not in the case of exclusivity rebates falling within the second category.

In its submissions to the General Court, Intel observed the inconsistency of a per se presumption against loyalty rebates while other price-based exclusionary conduct allegations such as predatory pricing and margin squeeze are subject to ‘all the circumstances’ assessments (as confirmed by the ECJ in Post Danmark I and Deutsche Telekom). The General Court, however, distinguished loyalty rebates on the basis that they include an exclusivity requirement for their granting, unlike the other cases to which Intel referred.86 The General Court also concluded that because exclusivity-inducing rebates by their nature make it (even just slightly) more difficult for rivals to compete in the market, they are abusive under Article 102, regardless of the size of the rebates and whether an equally efficient com-

For an explanation of effective prices over contestable demand, see n 67 above. Indeed, the Commission submitted to the General Court that its equally efficient operator test did not form part of the legal analysis of its decision. See Judgment of the General Court (Seventh Chamber, Extended Composition), 12 June 2014, Intel Corp v European Commission, Case T‑286/09, ECLI:​EU:​T:​ 2014:​547 (hereafter Case T‑286/09, Intel, 2014:547), [140]. 83 Ibid, [75]–[79]. 84 Ibid, [85]. 85 Ibid, [80]–[84]. 86 Ibid, [98]–[99], [152]. 81 82

78  Research handbook on methods and models of competition law petitor could profitably compete against them,87 the duration of the rebates,88 the share of the market affected,89 and the share of customer demand affected.90 Similar to the ECJ in Post Danmark II a year later, the General Court asserted no de minimis exemption:91 the possible smallness of the parts of the market which are concerned by the practices at issue is not a relevant argument. Where the course of conduct under consideration is that of an undertaking occupying a dominant position on a market where for this reason the structure of competition has already been weakened, any further weakening of the structure of competition may constitute an abuse of a dominant position [cite to ECJ in Hoffmann-La Roche, [123]]. The Court of Justice has therefore rejected the application of an ‘appreciable effect’ criterion or a de minimis threshold for the purposes of applying Article 82 EC (Opinion of Advocate General Mazák in Case C‑549/10 P Tomra, paragraph 73 above, point 17). Furthermore, the customers on the foreclosed part of the market should have the opportunity to benefit from whatever degree of competition is possible on the market and competitors should be able to compete on the merits for the entire market and not just for a part of it [cite to ECJ in Tomra, paragraphs 42 and 46]. A dominant undertaking may not therefore justify the grant of exclusivity rebates to certain customers by the fact that competitors remain free to supply other customers.

In the General Court’s view, therefore, the presumption that loyalty rebates are restrictive of competition was complete and irrefutable.92 On this basis, the General Court concluded that effects analysis in general, including equally efficient competitor testing in particular, is unnecessary in such cases.93 The treatment of loyalty rebates as per se abusive, when other forms of price-based exclusionary conduct (including third category rebate schemes, predatory pricing and margin squeeze) were subject to an ‘all the circumstances’ test (including the use of price-cost tests for predatory pricing in particular) had for some time been seen by many in the European competition community as an unsatisfactory anomaly in European competition law.94 Following the General Court decision, there was a strong sense that economics would forever remain largely

Ibid. At [108], the General Court stated: [I]t is not the level of the rebates which is at issue in the contested decision but the exclusivity for which they were given. Thus, the rebate must only be capable of inducing the customer to purchase exclusively, irrespective whether the competing supplier could have compensated the customer for the loss of the rebate if that customer switched supplier. 88 Ibid, [111]. 89 Ibid, [116]–[117]. 90 Ibid, [132]–[134]. 91 Ibid, [116]–[117]. 92 Although the General Court acknowledged the principle from Hoffman-La Roche that an objective justification may excuse conduct that is considered restrictive of competition under Article 102, the concept of objective justification is limited in European law and does not extend to arguments around whether the conduct was or was not restrictive of competition in its effect. 93 Regarding equally efficient operator testing, see Case T‑286/09, Intel, 2014:547, [143]. The General Court went further to state, referring to the ECJ in Michelin I and Tomra, that equally efficient operator testing is not necessary even in third category rebate cases where assessments of ‘all the circumstances’ are required: see [144]–[146]. 94 This was expressed well in the opinion of Advocate General Wahl in the lead-up to the ECJ’s hearing of Intel’s appeal from the General Court decision. AG Wahl argued that there is no objective reason for loyalty rebates to be treated more strictly under Article 102 than third category rebates; the economic literature explains that loyalty rebates are not always harmful and their effects are context dependent; and loyalty rebates deserve similar treatment as other forms of price-based exclusionary 87

The relevance of economics in US, EU and Australian competition law  79 marginalized in Article 102 cases, much as it appeared also, at that time, destined to remain marginalized under Article 101. For some, the ‘end of economics’ was complete, with doubts whether economics could add value in the future to European competition law outside of the assessment of mergers.95 Reports of the death of economics were, however, premature. The ECJ decision, on further appeal in Intel, breathed new life into Article 102. With reference to Hoffman-La Roche, the ECJ observed that it has previously held that loyalty rebates engaged in by a dominant firm are per se abusive under Article 102.96 However, the ECJ then stated that the case law must be clarified in relation to a situation ‘where the undertaking concerned submits, during the administrative procedure, on the basis of supporting evidence, that its conduct was not capable of restricting competition and, in particular, of producing the alleged foreclosure effects’.97 In particular, the ECJ clarifies that, if the defendant makes such a submission, the Commission is required to analyse the capability of foreclosure with reference to the extent of dominance, the share of the market covered by the challenged rebates, the conditions and arrangements for the granting of the rebates, their duration and amount, and the possible existence of a strategy to exclude equally efficient competitors.98 The ECJ further clarifies that if the Commission carries out such an analysis (which it must do if the defendant makes such a submission), the General Court must examine all of the defendant’s arguments challenging the validity of the Commission’s findings concerning the foreclosure capability of the rebates in question. The ECJ therefore appears to have found an avenue for effects assessments of loyalty rebates without explicitly overturning its own prior case law: loyalty rebates remain prima facie abusive, but if the defendant puts forward a substantiated economic defence of no capability of anticompetitive foreclosure, the Commission must examine this. This may include an assessment of effective prices compared to costs (that is, an EEO test). As a practical matter, since defendants who consider that their rebates are not anticompetitive will almost always put forward a substantiated defence, the Intel decision appears to have shifted the case law on loyalty rebates from a form-based per se prohibition to a provision with an effective burden on the Commission to establish a contravention on the basis of an ‘all the circumstances’ effects-based assessment. In doing so, the ECJ appears to have largely aligned the required assessment of loyalty rebates with that required for other rebate schemes and other forms of price-based exclusionary conduct, such as predatory pricing and margin squeeze. All of these are now, at least as a practical matter, subject to effects-based assessments. The only difference seems to be that for loyalty rebates there is a burden on defendants to provide substantial supporting evidence that their rebate scheme is not capable of foreclosure, before the burden shifts to the Commission to evaluate all the circumstances. Two other aspects of the ECJ’s Intel decision are welcome from the perspective of those favouring effects-based assessments. First, the ECJ states firmly (with reference to its decision in Post Danmark I, a predatory pricing case) that Article 102 does not seek to ensure

conduct. Opinion of Advocate General Wahl on 20 October 2016 in Case C-413/14 P, Intel, ECLI:​EU:​ C:​2016:​788, [89]–[105] and in particular [90], [93]–[95], [99]–[103]. 95 The apparent irrelevance of the Commission’s economic analysis in Intel even threatened to marginalize the Chief Economist Team within the Commission in relation to non-merger matters. 96 Case C-413/14 P, Intel, 2017:632, [137]. 97 Ibid, [138]. 98 Ibid, [139].

80  Research handbook on methods and models of competition law that competitors less efficient than the dominant firm should remain on the market because ‘[c]ompetition on the merits may, by definition, lead to the departure from the market or the marginalisation of competitors that are less efficient and so less attractive to consumers from the point of view of, among other things, price, choice, quality or innovation’.99 However, as noted above (in relation to Post Danmark II) and below (in relation to TeliaSonera), other ECJ decisions appear less clear on this in the context of rebates and margin squeeze cases, respectively. Second, the ECJ may be taken in the Intel decision to have clarified that a de minimis defence does exist for rebates cases under Article 102. As mentioned above, in Post Danmark II the ECJ appeared to deny any such defence. Without referring to Post Danmark II, the ECJ in Intel clarifies that where the defendant makes a submission supported by evidence that the rebate scheme was not capable of foreclosure effects, the Commission must evaluate this, taking into account, among other things, ‘the share of the market covered by the challenged rebates’ as well as their duration and amount. E

Regulatory Creep under Article 102

On a number of occasions, the ECJ has applied Article 102 as a tool to soften competition from dominant firms in order to facilitate inefficient entry. This use of competition law to preserve scope for inefficient competitors is more in the nature of the mandate of an economic regulator and somewhat alien to the understanding of the role of competition authorities in other parts of the world. One example, discussed already, is the ECJ’s Post Danmark II decision. In rejecting the application of the equally efficient competitor test in the context of an assessment of a retroactive rebate scheme, the ECJ demonstrated a concern to preserve scope for less efficient competitors to enter and survive to act as an ongoing constraint on an incumbent player with a large share and a statutory monopoly. In doing so, the ECJ was effectively requiring firms in a position similar to Post Danmark to set effective prices within rebate schemes above their own costs, and thereby engaging in a form of industry engineering reserved in other parts of the world for economic regulators rather than competition authorities. Another example can be found in the TeliaSonera case – one of a number of Article 102 cases brought before the ECJ concerning allegations of abusive margin squeezes in the telecoms sector. In an earlier decision concerning Deutsche Telekom, the ECJ endorsed the Commission’s and the General Court’s application of the EEO standard and, in doing so, stated that for assessing pricing practices under Article 102 ‘it is necessary to adopt a test based on the costs … of the dominant undertaking itself’.100 One of the attractions of the EEO standard in Article 102 is that it offers legal certainty to the dominant firm since it will know its own costs and can therefore test its prices, and the margins they allow, with some confidence. The EEO standard therefore provides the dominant firm with a standard that it can apply ex ante in order to avoid ex post sanction. By contrast, a dominant firm will typically not have information on its competitors’ costs, let alone clear guidance as to how inefficient a ‘reasonably efficient’ competitor is. It also seems unreasona Ibid, [133]–[134]. Judgment of the Court (Second Chamber) of 14 October 2010, Deutsche Telekom AG v European Commission, Case C-280/08 P, ECLI:​EU:​C:​2010:​603 (hereafter Deutsche Telekom), [198]. 99

100

The relevance of economics in US, EU and Australian competition law  81 ble in an ex post investigation to expect a dominant firm to have set prices in the past so as to allow some ‘headroom’ for downstream competitors that are less efficient (in the absence of ex ante guidance from a regulator to this effect). The legal certainty benefit of the EEO standard was recognized by the ECJ in Deutsche Telekom.101 From an economic perspective, another attraction of the EEO standard is that it promotes efficient outcomes (at least in a static sense). Under the EEO standard, only downstream competitors that are at least as efficient as the dominant firm will be protected. Moreover, consumers will not be at risk of paying high prices in order to shelter inefficient competitors and upstream investment will not be at risk of under-recovery. In TeliaSonera, however, the ECJ appears to have set a lower bar for infringement in order to provide scope for inefficient entry. The ECJ first confirmed that, in general, the EEO test is the correct approach from a competition law perspective:102 ‘In order to assess the lawfulness of the pricing policy applied by a dominant undertaking, reference should be made, as a general rule, to pricing criteria based on the costs incurred by the dominant undertaking itself and on its strategy.’ However, the ECJ then conceded that the costs of competitors (which may be higher) may be relevant in certain circumstances:103 It cannot be ruled out that the costs and prices of competitors may be relevant to the examination of the pricing practice at issue in the main proceedings. That might in particular be the case where the cost structure of the dominant undertaking is not precisely identifiable for objective reasons, or where the service supplied to competitors consists in the mere use of an infrastructure the production cost of which has already been written off, so that access to such an infrastructure no longer represents a cost for the dominant undertaking which is economically comparable to the cost which its competitors have to incur to have access to it, or again where the particular market conditions of competition dictate it, by reason, for example, of the fact that the level of the dominant undertaking’s costs is specifically attributable to the competitively advantageous situation in which its dominant position places it.

The ECJ also stated that if the margin for an equally efficient competitor is positive104 ‘it must then be demonstrated that the application of [the] pricing practice was, by reason, for example, of reduced profitability, likely to have the consequence that it would be at least more difficult for the operators concerned to trade on the market concerned’. It is unclear where this leads the legal assessment of margin squeeze allegations. Overall, the ECJ’s approach in TeliaSonera suggests that even in the context of cases where effects analysis is employed (complete with price-cost assessments), the ordo-liberal origins of European competition law and emphasis on market structure (in preference to analysis of overall long-term effects on consumers) remain alive and kicking. The use of Article 102 for regulatory goals is, moreover, a well-established feature of the European Commission’s Article 102 enforcement practice, at least in the telecoms sector. In relation to margin squeeze allegations, the Commission’s Guidance on its Enforcement Priorities under Article 102 specifies that the Commission will apply long run average incremental cost (LRAIC) as its cost standard in the context of margin squeeze cases, whereas Ibid, [202]. Judgment of the Court (First Chamber) of 17 February 2011, Konkurrensverket v TeliaSonera Sverige AB, Case C-52/09, ECLI:​EU:​C:​2011:​83 (hereafter TeliaSonera), [41]. 103 Ibid, [45]. 104 Ibid, [74]. 101 102

82  Research handbook on methods and models of competition law for predatory pricing the Commission will apply average avoidable cost (AAC). LRAIC will typically be higher than AAC due to sunk costs being counted in the former but not the latter. The Guidance therefore applies a stricter standard to pricing abuses in margin squeeze contexts. The authors consider that this is likely explained by the fact that margin squeeze cases are common in the (highly regulated) telecoms sector, where the Commission also plays a sector-regulatory role that includes promoting the entry of competitors that often must invest in significant sunk costs. Whether conscious or not, the use of a different standard for pricing cases that predominantly arise in a particular sector appears to be a case of ‘regulatory creep’. Also notable is that in the Telefónica case the Commission examined whether Telefónica’s retail prices could be replicated on the basis of each of its wholesale products taken one by one, as opposed to an optimal mix of three available products, some of which could only be accessed by competitors that had already made a substantial investment in their own infrastructure. The Commission considered this to be appropriate, given the observed method of entry and expansion in broadband markets. In its view, new entrants should not be prohibited from climbing the investment ladder as a result of a margin squeeze on those wholesale products that they are most likely to access for any given level of infrastructure investment. As noted by the Commission in its decision:105 Due to the risks involved in investments that entail high sunk costs, alternative operators are likely to follow a step-by-step approach to continuously expanding their infrastructure investments. When climbing the ‘investment ladder’ … alternative operators seek to obtain a minimum ‘critical mass’ in order to be able to make further investments. It is therefore necessary that there should not be any margin squeeze in relation to any ‘step’ of the ladder, i.e. in relation to any wholesale product. If there was such a margin squeeze, new entrants that are climbing the ladder of investment, would be foreclosed.

On appeal, Telefónica argued that the Commission ought to have applied its margin squeeze test by reference to the EEO principle, whereby an equally efficient competitor should be defined as one with the economies of scale and the network costs of Telefónica. It argued that such alternative operators would (and did) use an optimal combination of wholesale products, some of which were not subject to a margin squeeze.106 The General Court rejected Telefónica’s argument, agreeing with the Commission that all levels of Telefónica’s wholesale products should be accessible so as to allow smaller competitors seeking access to Telefónica’s wholesale products to climb the ‘investment ladder’:107 the applicants maintain that the ‘ladder of investments’ theory does not require that all levels be accessible. That argument must, however, be rejected. As the Commission correctly observes, the process that enables alternative operators to invest gradually in their own infrastructure can constitute a viable strategy only where there is no margin squeeze practice at the different levels of the ladder. In fact, the margin squeeze imposed by Telefónica probably delayed the entry and growth of its competitors and their ability to achieve a sufficient level of economies of scale to justify investments in their own infrastructure and the use of local loop unbundling …

C-COMP/38.784, Wandoo Espana v Telefonica, 4 July 2007, [392]–[393], 110. Judgment of the General Court (Eighth Chamber), 29 March 2012, Telefónica, SA and Telefónica de España, SA v European Commission, Case T‑336/07, ECLI:​EU:​T:​2012:​172 (hereafter Case T‑336/07, Telefonica, 2012:172), [200]. 107 Ibid, [189]. 105 106

The relevance of economics in US, EU and Australian competition law  83 In a particular context, it could well be the case that forcing dominant firms to price high in the short term to allow inefficient entry will lead to long-term benefits for consumers that more than compensate in the form of greater rivalry and, ultimately, efficiency and lower prices. This is a concept familiar from regulatory policy debates. However, given the legal uncertainty for the dominant firm that is introduced when such a rule is implemented in an ex post competition law context, and the difficulty for a competition authority of assessing and balancing both the short-term and the long-term effects, it is not surprising that in other jurisdictions the promotion of inefficient entry and ‘ladders of investment’ are matters reserved for sector regulators.

IV

FORM VERSUS EFFECTS IN AUSTRALIAN COMPETITION POLICY

Australia’s competition law is considerably more codified than the competition laws of the US and Europe. An early version of the current Competition and Consumer Act 2010 (Cth) (CCA) – the Trade Practices Act 1974 – was already considerably more codified than the corresponding US and European legislation. Rafts of subsequent amendments have resulted in even greater codification, in preference to reliance on judicial precedents. For example, whereas the US courts developed a per se rule against RPM under (the broadly drafted) section 2 of the Sherman Act, the drafters of the Trade Practices Act saw fit to codify an equivalent per se prohibition on RPM. Per se prohibitions were also enacted in relation to horizontal price fixing and market sharing and a form of exclusive dealing referred to as third-line forcing. Other horizontal and vertical conduct, however, was left to be judged by the courts under rules of reason. A counterbalance to this formalism in the legislation is that Australian competition law has for some time allowed for the authorization of conduct that would otherwise be prohibited per se. This has facilitated rigorous economic assessments of the likely effects of certain conduct on a case-by-case basis.108 The authorization provisions of the CCA recognize that, in certain circumstances, conduct that might otherwise be viewed as a restriction of competition may have the ability to enhance efficiency and consumer welfare, and therefore be in the public interest. Over time, a wider scope of conduct has been capable of being authorized on the basis of its potential to enhance efficiency. For example, the decision to extend the authorization regime to cover RPM was made in 1995 in light of economic theory that demonstrated that, in certain circumstances, such arrangements can enhance economic efficiency.109

108 As an alternative to authorization, a notification may be given to the Australian Competition and Consumer Commission (ACCC) in respect of exclusive dealing conduct (prohibited under section 47), collective bargaining conduct (prohibited under the cartel provisions and section 45) and price signalling (prohibited under Division 1A). 109 The option for authorization of RPM was introduced in 1995, following a recommendation of the Hilmer Committee, which noted that ‘economic theory associated with RPM [presents] a convincing argument that RPM can, in certain circumstances, enhance economic efficiency’: Independent Committee of Inquiry, Parliament of Australia, National Competition Policy (1993) 58. See also Geoff Edwards, ‘When Should Resale Price Maintenance Be Authorised? Guidelines for Use in Authorisation Determinations’ (1996) 4(4) Trade Practices Law Journal 161.

84  Research handbook on methods and models of competition law Much like the European Commission, the Australian Competition and Consumer Commission (ACCC) has exhibited a tendency to bring cases that pursue infringements in form, without a convincing theory of harmful effects. One example is the Flight Centre case. This case concerned attempts by Flight Centre, a travel agent, to realize most-favoured-nation terms in its relationships with certain airlines.110 This case concerned vertical arrangements in a supply chain, notwithstanding that Flight Centre and the airlines were also in competition downstream in the sale of airline tickets. Despite the well-known potential for vertical restraints to deliver efficiencies, the ACCC pursued this case under the per se provisions of section 45A of the CCA (which, at the relevant time, deemed price fixing between competitors to be prohibited), rather than under section 45 (which would have required the ACCC to establish a purpose or effect of substantially lessening competition). The ACCC was ultimately successful before the High Court, which found that Flight Centre had breached the per se prohibition in section 45A. That the ACCC preferred the much easier section 45A route for a case that concerned vertical arrangements (where effects are typically complex and ambiguous a priori) rather than reserving this provision for hardcore horizontal cartels (where effects are typically straightforward) is understandable from the perspective of efficient and effective legal strategy. However, it is disappointing insofar as it reflects a competition authority pursuing form over effect in a case where the effects were ambiguous a priori. Another example is the Pfizer case, in which the ACCC alleged that Pfizer had acted with the ‘purpose’ of deterring or preventing other suppliers from engaging in competitive conduct (under section 46 of the CCA, which at the time prohibited the taking advantage of substantial market power for a number of proscribed purposes) and the ‘purpose’ of substantially lessening competition (under section 47 of the CCA, which prohibits anticompetitive exclusive dealing). As Flick J of the Federal Court observed, ‘purpose’ in sections 46 and 47 of the CCA is to be ascertained subjectively, not objectively.111 Notably (as his Honour also observed),112 the ACCC did not allege that the conduct had the ‘effect’ of substantially lessening competition, which was an option that was open to it under section 47 of the CCA. His Honour held that Pfizer did not act with the alleged (subjective) purposes. Interestingly, there are suggestions in the judgment that the lack of any prospect of harmful effects (because Pfizer would inevitably face intense competition from generic manufacturers in any event) had a bearing on his Honour’s finding of no anticompetitive purpose. For example, Flick J stated that the ACCC’s failure to establish a purpose of substantially lessening competition ‘was contributed to by the failure of the ACCC to direct any real attention to the likely state

110 Specifically, Flight Centre attempted to induce a number of international airlines to cease discounting airline tickets through their direct-to-consumer internet channels, in a context in which the airlines were not simultaneously and correspondingly lowering the wholesale prices paid by Flight Centre for those tickets and so the margins that Flight Centre could earn (if it were to match the airlines’ direct-to-consumer prices) were being eroded. Interestingly, had Flight Centre instead attempted to induce the airlines to maintain a constant relationship between the airlines’ direct-to-consumer prices and the wholesale prices it paid to the airlines (without seeking to influence directly the airline’s direct-to-consumer prices), the economic effect may have been exactly the same, yet it seems far less likely that the ACCC would have brought a section 45A case. 111 Australian Competition and Consumer Commission v Pfizer Australia Pty Ltd [2015] FCA 113, [46]–[51] (Flick J). 112 Ibid, [325] (Flick J).

The relevance of economics in US, EU and Australian competition law  85 of future competition in the market ‘with and without’ Pfizer’s impugned conduct’.113 On appeal, the Full Federal Court upheld Flick J’s decision on purpose, noting that it ‘was clear to Pfizer that, no matter what it did and no matter what strategy it adopted, immediately after 18 May 2012, it was going to face intense competition in the atorvastatin market from the major generics manufacturers and their aligned wholesalers’.114 One wonders, therefore, whether the Court is signalling to the ACCC the need for any case under section 46 or 47 to be grounded in a theory of harmful effects. The most contentious and debated provision in Australian competition law is the prohibition on unilateral anticompetitive conduct contained in section 46 of the CCA. This provision, until very recently, prohibited the taking advantage of substantial market power with the purpose of eliminating or substantially damaging a competitor. Although Australia’s competition law was no doubt informed in its gestational phases in the 1960s and 1970s by competition laws and thinking in the US and Europe, it is difficult to discern in any of the contemporaneous documents a clear school of thought regarding the objective that the legislators sought to achieve when enacting section 46. It is possible that Australia’s geographic remoteness obscured the major drivers and shifts of competition policy overseas (in Europe, the ordo-liberal ideology, and in the US, the shift occurring from the SCP paradigm and hostility towards unusual or not-well-understood conduct towards a less interventionist consumer welfare standard). Some elements of ordo-liberalism are nonetheless apparent in the drafting of section 46. In particular, the concern for the welfare of competitors, market structure and the process of competition (rather than consumer welfare) can be seen in the range of proscribed purposes included within the versions of section 46 that applied from 1974 until late 2017. Those versions provided that a corporation that has a substantial degree of power in a market shall not take advantage of that power for a number of purposes, all of which suggest that the purpose of the section is to protect individual competitors.115 The ‘take advantage’ terminology included within section 46 of the Trade Practices Act 1974 (Cth) also likely had its genesis in early interpretations of Article 86. Until the early 1980s, many in Europe understood Article 86 to prohibit only behaviour that a firm would not engage in without dominance (similar to the concept that a firm only contravenes the provision if it ‘takes advantage’ of its dominance). This appeared to be confirmed in the 1979 decision of

113 Ibid, [451] (Flick J). See also [413]–[414] in relation to the purpose element of the section 46 allegation: Pfizer unquestionably knew that it would face intense competition as from either February or May 2012. Its revenue would unquestionably be impacted … … contrary to the facts in Baxter, in the present proceeding it was inevitable that there would be intense competition for the supply of generic atorvastatin to community pharmacies after May 2012 (if not earlier). 114 Australian Competition and Consumer Commission v Pfizer Australia Pty Ltd [2018] FCAFC 78, [560]. 115 The version of section 46 applicable from 1988 until 2017 stated these purposes as follows (earlier versions were similar): (a) eliminating or substantially damaging a competitor of the corporation or of a body corporate that is related to the corporation in that or any other market; (b) preventing the entry of a person into that or any other market; or (c) deterring or preventing a person from engaging in competitive conduct in that or any other market.

86  Research handbook on methods and models of competition law the ECJ in Hoffman-La Roche, which suggested that only conduct ‘different from … normal competition’ is abusive.116 However, in 1983, in the Michelin I case, the ECJ introduced the concept of a special responsibility on dominant firms ‘not to allow [their] conduct to impair undistorted competition’.117 This special responsibility has subsequently been interpreted as meaning that a dominant firm may infringe Europe’s prohibition on abusive unilateral conduct even if engaging in conduct in which non-dominant firms would engage. Following this decision, European and Australian approaches to unilateral abuse of market power diverged fundamentally. In its seminal decision in Queensland Wire, the High Court of Australia (HCA) rejected the notion that firms with a substantial degree of market power have a special responsibility not to engage in certain forms of conduct or to ensure that competitors are not harmed. Rather, the HCA held that such firms are only prohibited from engaging in conduct that would not be commercially rational had they been operating in a competitive market.118 That is, a firm will not be considered to have ‘taken advantage’ of substantial market power unless some causal connection can be established between the market power and the conduct. In Queensland Wire, the HCA also expressed its view that it is the welfare of consumers that is the ultimate object of section 46 and the Act more generally. This provided the HCA and lower courts with a platform to examine economic effects free from form-based straightjackets and aligned Australian competition law more with the US than with Europe. In particular, Mason CJ and Wilson J stated that ‘the object of s. 46 is to protect the interests of consumers, the operation of the section being predicated on the assumption that competition is a means to that end’.119 In the subsequent Boral case, McHugh J stated that [w]hile conduct must be examined by its effect on the competitive process, it is the flow-on result that is the key – the effect on consumers. Competition policy suggests that it is only when consumers will suffer as a result of the practices of a business firm that s 46 is likely to require courts to intervene and deal with the conduct of that firm.120

The decisions of the HCA in Queensland Wire and Melway opened the door to consideration under section 46 of the efficiencies of unilateral conduct, and introduced a broad presumption that conduct that a firm without market power would engage in is likely to be economically Case 85/76, Hoffman La Roche, 1979:36, [91]. Case 322/81, Michelin I, 1983:313, [57]. 118 Queensland Wire Industries Pty Ltd v Broken Hill Pty Co Ltd (1989) 167 CLR 177. See the judgment of Mason CJ and Wilson J at 192: It is only by virtue of its control of the market and the absence of other suppliers that BHP can afford, in a commercial sense, to withhold Y-bar from the appellant. If BHP lacked that market power – in other words, if it were operating in a competitive market – it is highly unlikely that it would stand by, without any effort to compete, and allow the appellant to secure its supply of Y-bar from a competitor. See also the reasons of Dawson J at 202: That being so, there can be no real doubt that B.H.P. took advantage of its market power in this case. It used that power in a manner made possible only by the absence of competitive conditions. Inferences in this regard can be drawn from the fact that BHP could not have refused to supply Y-bar to QWI if it had been subject to competition in the supply of that product. 119 Ibid at 191. See the judgments of Deane J at 194 and Mason CJ and Wilson J at 191. 120 Boral Besser Masonry Ltd v Australian Competition and Consumer Commission (2003) 215 CLR 374, [261]. 116 117

The relevance of economics in US, EU and Australian competition law  87 efficient. In Melway, the HCA expressly recognized that Melway’s refusal to supply, which was a consequence of its adoption of an exclusive distribution system, may have restricted intrabrand competition, but also had the potential to promote interbrand competition. Notably, the HCA referred to US rather than European precedents on vertical restraints (including the US Supreme Court’s decision in GTE Sylvania, discussed earlier) and acknowledged that the overall effect of such restraints on competition and consumers can be positive.121 Central to the HCA’s decision to exonerate Melway was that Melway had engaged in the same distribution system even at a time when it did not have market power.122 Although the HCA was not explicit, its preservation of the approach to ‘take advantage’ established in Queensland Wire – to excuse conduct that would be rational behaviour in the absence of market power – was a preservation of an important avenue within section 46 for arguments on efficiencies and effects on consumers (rather than effects on competitors per se) to ultimately determine decisions under Australia’s abuse of market power provision. By incorporating considerations of efficiencies and effects on consumers into the interpretation of the ‘take advantage’ element of section 46, the Australian provision against unilateral conduct developed into one that had a much greater focus on economics and effects than its European equivalent. However, the HCA appeared to lose its way in 2003 with its judgment in Rural Press.123 A different kind of legal formalism triumphed in this case, with the HCA indicating a preparedness to exonerate any conduct that a firm without substantial market power ‘could’ engage in, rather than ‘profitably could’ or ‘would’. Since any firm ‘could’ engage in almost any conduct (even though it may not be profitable to do so), the ‘take advantage’ test became overly permissive. In short, economics became sidelined by an overly permissive legal formalism, in contrast to the overly restrictive form that prevailed in the US until the 1970s and that has dominated in Europe for much longer. The reaction to this and other formalistic judgments that have followed has been a number of legislative amendments to the prohibition on anticompetitive unilateral conduct, including, in November 2017, the replacement of ‘take advantage’ and the proscribed purposes with a substantial lessening of competition test. It remains to be seen how the new provision will be interpreted by the ACCC, private litigants and the courts and whether it will result in greater focus on economic analysis of effects.

121 Melway Publishing Pty Ltd v Robert Hicks Pty Ltd (2001) 205 CLR 1. See the majority judgment of Gleeson CJ, Gummow, Hayne and Callinan JJ at [19], [20]. It is clear that the majority considered that effects on competition should be considered by reference to effects on consumer welfare. For example, in support of the observation that vertical restraints can have positive effects on competition, the majority included the following quote from the United States Eleventh Circuit Court of Appeals in Graphic Products Distributors Inc v Itek Corp, which is focused on effects on consumers: We note first that a vertical restraint on trade, almost by definition, involves some reduction in intrabrand competition. When a manufacturer restricts a dealer to selling only within a certain territory, or only to certain customers, or only from certain locations, it is necessarily restraining intrabrand competition. However, this may or may not have a negative effect on the welfare of the consumer … The effects of a restraint of intrabrand competition on consumer welfare cannot be viewed in isolation from the interbrand market structure. A restriction of intrabrand competition may – depending on the interbrand market structure – either enhance or diminish overall competition, and hence consumer welfare (emphasis added). 122 Melway Publishing Pty Ltd v Robert Hicks Pty Ltd (2001) 205 CLR 1. See the majority judgment of Gleeson CJ, Gummow, Hayne and Callinan JJ at [68]. 123 Rural Press Ltd v Australian Competition and Consumer Commission (2003) 216 CLR 53.

88  Research handbook on methods and models of competition law

V CONCLUSIONS This chapter has reviewed the evolving relevance of economic analysis in competition law and policy in the US, the EU and Australia. This evolution has reflected battles in each country between legal form-based approaches focused on preserving particular market structures and economic effects-based analyses focused on consumer welfare. During the 1970s, the US experienced a rapid shift from the populist structural paradigm that had prevailed since the 1930s to the consumer welfare standard. However, a new wave of populism is rising and may yet bring with it a reversal in years to come. Competition policy in the EU remained largely form based and unwelcoming of economics for much longer. However, since the turn of the century there has been some evolution towards a more economic approach, albeit in fits and starts. In Australia, effects-based analysis and the consumer welfare standard enjoyed some favour, particularly in the latter part of the last century, but reliance on form and structure in preference to assessments of effects on consumer welfare remains common.

5. The use of economics in competition law enforcement in mainland China and Hong Kong Lin Ping and Yan Yu

I INTRODUCTION China’s Anti-Monopoly Law (AML) of 2007 came into effect in 2008, while Hong Kong’s general competition law came into effect in 2015. As a special administrative region of China, Hong Kong established its Competition Ordinance (CO) independently under its common law system, whereas the AML is part of China’s continental legal system. Some interesting comparisons can be made between the two. This chapter reviews the two competition law regimes of Hong Kong and China in the context of examining the use of economic analysis in implementing the respective laws.

II

CHINA’S ANTI-MONOPOLY LAW

China’s AML was enacted on 30 August 2007 and came into effect on 1 August 2008. Considered as the ‘economic constitution’ of China, the AML sets out its objectives and general principles in very broad terms. In particular, chapter 1 states that ‘[t]he Law is enacted for the purpose of preventing and restraining monopolistic conduct, protecting fair competition in the market, enhancing economic efficiency, safeguarding the interests of consumers and social public interest, and promoting the healthy development of the socialist market economy’. China has made great progress in enforcing the AML in all areas covered by the law, including monopoly agreements, abuse of dominant position, mergers, and – a unique feature of China’s competition law – administrative monopoly.1 Article 13 of the AML provides that competing undertakings are prohibited from making agreements on price-fixing, output restriction, market sharing, restrictions on products or technology development, boycotts, and any other horizontal monopoly agreements as determined by the enforcement agencies. Article 13 does not, however, make it clear whether any or all of the agreements are per se illegal. The AML also prohibits two types of vertical agreements (namely fixing and setting minimum resale prices) and other vertical arrangements as determined by the enforcement agencies.2 Article 6 prohibits firms with a dominant position from abusing that position to eliminate or restrict competition.3 Six types of abuse are specifically prohibited. These include unfair pricing (selling or buying goods at unfairly high or low prices), below-cost sales, refusal to

1 So far, there has been very little practical use of economic analysis on administrative monopoly cases in China. Therefore, in this chapter we do not refer to those AML cases. 2 AML, Art 14. 3 Ibid, Art 6.

89

90  Research handbook on methods and models of competition law deal, exclusive or designated dealing, tying or imposing other unreasonable transactional terms, and discriminatory dealing. Other abusive behaviour by dominant undertakings determined by the enforcement authorities is also prohibited. The AML applies to mergers and acquisitions. Under the AML, the concept of ‘concentrations of undertakings’ refers to a variety of circumstances, including (1) mergers; (2) acquisitions of control of other undertakings through purchasing shares or assets; and (3) acquisitions of control of other undertakings, or of the ability to exercise decisive influence over other undertakings, through contract or other means.4 Article 31 of the AML refers to a possible national security review for concentrations involving foreign parties. Article 8 of the AML generally prohibits the abuse of administrative powers by public authorities (including administrative agencies and organizations empowered by laws or regulations for public affairs administration) to eliminate or restrict competition, a type of behaviour widely referred to as ‘administrative monopoly’.5 Following an earlier three-pronged enforcement strategy, in March 2018 the new State Administration for Market Regulation (SAMR) took over the roles of the three existing enforcement agencies in order to achieve greater efficiency. There is a right of private action for private parties directly or indirectly harmed by anticompetitive acts on both a stand-alone and a follow-on basis.6 According to the official data released by SAMR,7 by the end of October 2018, 2,437 merger cases were reviewed by China’s anti-monopoly agency, of which 37 were approved with conditions and two were prohibited. There were approximately 165 monopoly agreement cases and 55 abuse of market dominance cases. The number of private antitrust cases in China has also gradually increased, although private antitrust enforcement is still in the early stages. By the end of 2017, 700 first-instance civil monopoly cases had been accepted, 630 of which had been closed. The cases involved various industries, including transportation, insurance, medicine, food, household appliances, power supply, and information networking. More than 90 per cent of the 700 cases concerned abuse of market dominance.8

Ibid, Art 20. Ibid, Art 8. 6 See the People’s Supreme Court of China, Judicial Interpretation on the Application of Laws to Anti-Monopoly Private Actions (Judicial Interpretation) issued by the People’s Supreme Court of China (8 May 2012), https://​www​.mayerbrown​.com/​en/​perspectives​-events/​publications/​2012/​05/​china​ -antimonopoly​-law​-finalised​-judicial​-interpretr (in English); http://​www​.court​.gov​.cn/​fabu​-xiangqing​ -3989​.html. 7 See ‘Achievements in Anti-Monopoly Law Enforcement during Its First Yen Years’, http://​ finance​.people​.com​.cn/​n1/​2018/​1116/​c1004​-30404971​.html. 8 Figures on private enforcement are from International Law Office, ‘A New Era – 2018 Anti-Monopoly Law Enforcement Highlights’ (February 2019), https://​ www​ .internationallawoffice​ .com/​Newsletters/​Competition​-Antitrust/​China/​AnJie​-Law​-Firm/​A​-new​-era​-2018​-Anti​-Monopoly​-Law​ -enforcement​-highlights. 4 5

Economics in competition law enforcement in mainland China and Hong Kong  91

III

THE ROLE OF ECONOMICS IN THE ENFORCEMENT OF CHINA’S AML

This section reviews the use of economic analysis in the enforcement of China’s AML in the areas of (i) market definition; (ii) monopoly agreements; (iii) abuse of dominance; and (iv) merger control. A

Definition of ‘Relevant Market’: China

On 24 May 2009 China’s Anti-Monopoly Commission announced its Guidelines Concerning the Definition of Relevant Markets.9 These recognize that behaviour that would, or is likely to, eliminate or restrict competition occurs within a market.10 The Guidelines acknowledge that defining the relevant market ‘plays an important role in key issues’, such as:11 ●● identifying existing and potential competitors; ●● determining market share and position of business operators and the degree of market concentration; and ●● analysing the impact of the business operator behaviour on competition and judging whether it is illegal. As a result, defining the relevant market is usually the starting point for competition analysis.12 In line with international standards, the Guidelines define the relevant product and geographical markets based on demand- and supply-side substitution. They also consider temporal aspects. Importantly, they accept use of the hypothetical monopolist test, applying a 5–10 per cent SSNIP (small but significant non-transitory increase in price) within one year. Standard economic frameworks and analyses (for example, SSNIP test, critical loss analysis, etc.) have been used in cases under the AML. B

Use of Economics in Prohibiting Monopoly Agreements

The previous regulator, the Price Supervision and Anti-Monopoly Bureau of the National Development and Reform Commission (NDRC), imposed fines on a number of monopoly agreement violations of the AML, including the auto-part cartel and vertical resale price maintenance (RPM) cases involving infant milk formula and the Chinese liquor market. The economics behind overt cartel agreements is relatively simple: these horizontal agreements always reduce consumer welfare and hence should be per se illegal without the need for any competitive assessments. Nonetheless, some agreements, even among competitors, may be procompetitive (for example, agreements that induce collaborative research and development, set industry standards, or enhance product quality or environmental protection). The AML may allow these types of agreements if participants can provide evidence justifying their efficiency. Economic 9 See Anti-Monopoly Commission, Guidelines Concerning the Definition of Relevant Markets (24 May 2009), http://​www​.gov​.cn/​zwhd/​2009​-07/​07/​content​_1355288​.htm. 10 Ibid, Art 2. 11 Ibid. 12 Ibid.

92  Research handbook on methods and models of competition law analysis helps verify procompetitive effects and assess their trade-off vis-à-vis anticompetitive effects. This has yet to occur. However, there have been some controversies with regard to the competitive effect of vertical monopoly agreements under the AML. As stated above, two types of vertical agreements are prohibited under the AML – namely, fixing and setting a minimum level of the resale prices. Economic theory tells us that RPM can lead to both procompetitive efficiency benefits and anticompetitive harms. In practice, the debate among economists often lies in whether an RPM provision would have an overall adverse effect on competition and consumers. However, many jurisdictions, including the EU and the UK (but not the US), still consider RPM to be a ‘hard-core’ price-fixing conduct. Agreements fixing and setting minimum resale prices are prohibited under the AML but there was originally confusion on the standard of proof required to show that an RPM agreement constituted a vertical monopoly agreement under the AML. The question was whether China would treat RPM as per se illegal (with the possibility to appeal) or would allow for economic analysis. A mixed message emerged from the Shanghai High People’s Court in its ruling on the landmark Johnson & Johnson case, the first RPM private litigation case in China. There, the Court viewed it as necessary to demonstrate that an RPM agreement would have a significant adverse effect on competition in the relevant market in order for it to constitute a vertical monopoly agreement. However, it was a prima facie concern that the decision reduced the ability of Johnson & Johnson’s distributors to set resale prices flexibly.13 The economic arguments put forward by the Shanghai High People’s Court, together with the decisions by NDRC on certain RPM agreements, raised the question of whether the courts and competition authorities in China would put substantial emphasis on economic analysis in these types of cases going forward. Fortunately, in the Johnson & Johnson judgment, the Shanghai High People’s Court did set out a framework which allowed room for debate on economic arguments. This suggested that there is scope for putting forward credible economic arguments supported by factual evidence. More recently, the Supreme People’s Court of China has affirmed a decision of the Hainan Provincial Price Bureau that a per se approach should be applied in RPM cases while the market is still underdeveloped in China, while recognizing that RPM agreements may have a procompetitive or anticompetitive effect. Reference was made to the potential costs of taking the alternative viewpoint. A business operator may still rebut the allegation by proving that the particular agreement does not restrict competition or by invoking an exemption under Article 15. These views are in line with the Interim Provisions on Prohibition of Monopoly Agreements issued by SAMR on 1 July 2019. In relation to the decision in Rainbow v Johnson & Johnson, the Supreme People’s Court stated that it is not improper for the court to determine whether to uphold the plaintiffs’ claims based on whether the monopoly agreement eliminates or restricts competition.14

13 See, for example, Yan Yu, ‘The Use of Economic Analysis in RPM Cases in China: Is There Gold at the End of the Rainbow?’ CPI Antitrust Chronicle (28 October 2013). 14 Hainan Provincial Price Bureau v Hainan Yutai Scientific Feed Company (December 2018; published 24 June 2019). See the case study published by China Market Regulation News on 16 July 2019, http://​www​.cicn​.com​.cn/​zggsb/​2019​-07/​16/​cms119187article​.shtml.

Economics in competition law enforcement in mainland China and Hong Kong  93 C

Use of Economics in Abuse of Market Dominance Cases

As previously mentioned, three agencies were previously responsible for enforcement, with two of these responsible for enforcement of abuse of dominance, and a number of cases were decided by them. NDRC was responsible for enforcing the provisions on price-related abusive conduct, and the State Administration for Industry and Commerce (SAIC) for all other types of abusive conduct. The general provisions of Article 17 of the AML on abuse of dominance are supplemented by the Regulation on Anti-Price Monopoly (APM) and the Regulation on Prohibiting the Abuse of a Market Dominant Position (AMD), issued by NDRC and SAIC respectively in December 2010, although SAMR is now responsible for all enforcement.15 These regulations provide specific guidance on the economic reasoning to be adopted in determining market dominance, as well as in assessing various types of abuse of market dominance conduct. We discuss below how two types of abusive conduct are assessed in the review of two high-profile cases, Qualcomm and Tetra Pak. Article 11 of NDRC’s APM Regulations lists factors that NDRC will consider in determining whether price is excessively high or low. These include whether: ●● the price of the product is significantly higher/lower than that charged or faced by other operators; ●● the price change exceeds the normal range when production costs are relatively stable; and ●● the price change is significantly larger than the change in costs when production costs are relatively unstable. There is no definition of production costs mentioned in the factors above. Moreover, neither the AML nor NDRC’s APM regulation has any specific requirement to demonstrate recoupment, consumer harm, or the rival’s exit. Whether such requirements will be considered in future court and enforcement decisions remains uncertain. 1 The Qualcomm case (2015) On 10 February 2015, NDRC fined Qualcomm 6.08 billion Chinese yuan (approximately US$975 million) for abusive patent licensing practices, the largest competition fine ever given in China.16 Qualcomm is a leading chipmaker in the US and holds a significant number of standard essential patents (SEPs). NDRC found that Qualcomm was dominant in several markets – the licensing of SEPs for the CDMA, WCDMA and LTE wireless communication standards, and the supply of baseband chips. First, Qualcomm refused to provide customers with a list of all patents included in its comprehensive licensing package, resulting in customers being charged for patents that had already expired. Second, Qualcomm imposed unfair cross-licensing conditions: it forced customers to grant Qualcomm free licences for

15 See NDRC, Regulation on Anti-Price Monopoly (December 2010), http://​jjs​.ndrc​.gov​.cn/​zcfg/​ t20110104​_389399​.htm; SAIC, Regulation on Prohibiting the Abuse of a Market Dominant Position (December 2010), http://​www​.saic​.gov​.cn/​fldyfbzdjz/​zcfg/​zcfg/​201101/​t20110107​_103379​.html. 16 See the Chinese decision on NDRC (2015): Administrative Penalty Decision, http://​www​.ndrc​.gov​ .cn/​xzcf/​201503/​t20150302​_754177​.html. See also Susan Ning, Kate Peng and Lingbo Wei, ‘Qualcomm Investigation Finally Closed: Some Changes in Business Model in Addition to an RMB 6.088 Billion Fine’ (2015), https://​www​.chinalawinsight​.com/​2015/​02/​articles/​corporate/​antitrust​-competition/​.

94  Research handbook on methods and models of competition law the customers’ patents while refusing to lower its royalties in response. Third, the royalty rate was set at a high level and applied to the net wholesale price of the mobile devices concerned. NDRC found that this conduct resulted in Qualcomm imposing excessive royalties on Chinese mobile device manufacturers in violation of the AML. NDRC also found that Qualcomm forced customers to accept the licensing of its non-essential patents (in which Qualcomm holds no dominant position) in order to obtain a licence for its SEPs. Finally, NDRC found that Qualcomm imposed unreasonable conditions on the sale of baseband chips because Chinese customers were forced to accept a non-challenge clause prohibiting them from challenging the validity of Qualcomm’s patents. Not much information was disclosed in NDRC’s decision on how much and what economic analysis was used in the case. The considerations in its decision differed from the factors listed in Article 11 of its APM Regulations determining unfairly high price, which are listed above. This approach differs from that of the judge in Huawei v IDC, another recent AML case on the licensing of SEPs,17 which emphasized the relatively low royalties that IDC charged other mobile device manufacturers (for example, Apple and Samsung) in concluding that the royalties Huawei was charged were unfairly high. It therefore seems that NDRC adopted a qualitative, as opposed to quantitative, approach, accepting charging for already expired licences and forcing free grant-backs as sufficient to indicate unfair pricing. 2 The Tetra Pak case (2016) On 16 November 2016, following a thorough investigation lasting more than four years, SAIC imposed a fine of US$97 million on Tetra Pak after concluding that the food packaging company had abused its dominant market position in the liquid food aseptic carton packaging industry from 2009 to 2013.18 A key concern was Tetra Pak’s loyalty discount scheme in the packing material market. SAIC considered two main elements of this scheme: retroactive loyalty discounts and customized volume-targeted discounts. The former were offered retroactively when a customer’s cumulative total purchases reached a certain threshold within a year. Additional discounts were available for customers purchasing two or more types of packaging material. Tetra Pak also applied volume discounts, usually when a customer’s purchases during a given period reached the fixed volume target set individually based on customer background and profile. SAIC’s economic reasoning was based on the leveraging theory adopted by the EU in its investigation of Intel. SAIC considered that Tetra Pak’s retroactive discount scheme could induce customers to purchase as many products as possible within a given period in order to reach the thresholds and enjoy higher discounts. Likewise, volume-targeted discounts ‘locked in’ a customer’s demand. Both had the effect of foreclosing competitors from the market. SAIC acknowledged that loyalty discounts could lower prices in the short run, but in the long run would reduce profits of other packaging material manufacturers and lead to underuti-

17 See Guangdong People’s Court Decision 305, 粤高法民三终字第305号 (2013), http://​www​ .mlex​.com/​China/​Attachments/​2014​-04​-18​_AXRC879FW8P38IO7/​guangdonghpc​_IDChuawei​_SEP​ _18042014​.pdf. For a discussion of the case, see Michael Han and Kexin Li, ‘Huawei v. InterDigital: China at the Crossroads of Antitrust and Intellectual Property, Competition and Innovation’, Competition Policy International (3 December 2013). 18 See the Tetra Pack Administrative Penalty Decision, http://​www​.competitionlaw​.cn/​info/​1025/​ 23864​.htm.

Economics in competition law enforcement in mainland China and Hong Kong  95 lization. This was likely to harm competitors and ultimately adversely influence competition and customers in the market. This case indicates that China is using state-of-the-art economic theories in its enforcement activities.19 Similar approaches will be adopted in dealing with other abuse of market dominance cases, which by nature require a rule of reason analysis and hence sound economic reasoning and analysis. D

Use of Economics in Merger Reviews

As the agency traditionally overseeing merger control in China, MOFCOM (the previous merger regulator) handled hundreds of mergers and acquisitions (and joint ventures) annually. The majority were cleared without any conditions, mostly following the simplified notification process.20 Each year, a few cases required in-depth review involving detailed economic analysis. 1 Economic principles underpinning merger reviews in China Article 28 of the AML states: If a concentration of undertakings has or may have an effect of eliminating or restricting competition, the Anti-Monopoly Law Enforcement Authority under the State Council shall decide to prohibit such concentration. If the undertakings are able to prove that the positive impact of such concentration on competition apparently outweigh the negative impact, or such concentration conform to the requirements of social or public interests, the Anti-Monopoly Law Enforcement Authority under the State Council may decide not to prohibit the concentration of undertakings.21

As elsewhere, in China, mergers are generally assessed according to whether they create or strengthen dominance, and thus have an adverse effect on competition. Although not explicitly stated, it is widely understood that the restriction on competition from a merger needs to be substantial to justify a prohibition under the AML, which again is consistent with the general practice in other jurisdictions. MOFCOM characterized mergers as horizontal or non-horizontal (vertical). The former involve undertakings that offer substitutable products or services in the same market – that is, a direct overlap in business – while the latter involve undertakings that do not compete directly in one market and instead offer products or services that are complementary and independent, though related to one another. Non-horizontal mergers can then be further segmented into vertical mergers and conglomerate mergers. In vertical mergers, the parties operate at different levels of a supply chain – for example, GE China/China Shenhua Coal JV (2011).22 In conglomerate mergers, the parties 19 For a detailed description and discussion of the economic reasoning used in the case, see Xiao Fu and Guofu Tan, ‘Abuse of Market Dominance under China’s Anti-Monopoly Law: The Case of Tetra Pack’ (2019) 54(2) Review of Industrial Organization 409. 20 Anti-Monopoly Bureau of MOFCOM, 关于经营者集中简易案件申报的指导意见（试行）(18 April 2014), http://​fldj​.mofcom​.gov​.cn/​article/​c/​201404/​20140400555353​.shtml. 21 See AML, Art 22 (3 August 2008), http://​ english​ .mofcom​ .gov​ .cn/​ article/​ policyrelease/​ businessregulations/​201303/​20130300045909​.shtml. 22 Notice of the Ministry of Commerce No 74 of 2011 on the Conditional Approval of the Decision of the Antitrust Review of the Joint Venture between General Electric (China) Co Ltd and China

96  Research handbook on methods and models of competition law operate in two separate but related markets – for example, Coca Cola/HuiYuan (2009). In the latter example, MOFCOM alleged that Coca Cola had the ability to leverage its market power in the carbonated drinks segment into the fruit juice segment, thus foreclosing competition. The deal was subsequently blocked by MOFCOM.23 A transaction may involve both horizontal and non-horizontal relationships. This categorization is not just a formality. It sets an important benchmark in helping to identify competition concerns and relevant theories of harm, which affects the economic frameworks and the appropriate theoretical and empirical approach to the case. Despite there being only a few published decisions by MOFCOM since 2009, the economic principles underpinning AML merger control are generally in line with economic thinking in other jurisdictions. China has given considerable weight to standard economic frameworks and analytical approaches relating to theories of harm from mergers in other jurisdictions. These have been widely adopted in merger reviews in China throughout recent years. 2 An economic approach to horizontal merger assessments in China Compared with non-horizontal mergers, horizontal mergers are often subject to tighter scrutiny as such mergers eliminate a direct competitor from the market and are more likely to restrict competition. In reviewing these mergers, MOFCOM generally focused on ‘unilateral effects’ – that is, the increased ability and incentive of the merged entity to profitably raise prices.24 Most of the mergers cleared with remedies by MOFCOM raised unilateral effects concerns. Notable examples include AB InBev/SAB Miller (2016)25 and United Technologies/Goodrich (2012).26 Market shares may provide a useful proxy for the competitive constraints imposed by rivals post merger, and indeed form an important consideration in MOFCOM’s competitive assessment. There are, however, no guidelines in China that indicate a market share threshold below which no competition concerns are raised. Some past MOFCOM decisions suggested that unilateral effects could arise even if the parties’ combined market share is less than 25 per cent. Theoretically, the use of market shares as the initial filter stems from the following: the higher the merged entity’s market share, the greater the benefit it will enjoy from a price increase or an output restriction on the installed customer base. Many other techniques, in addition to market shares calculation, have been implemented in merger cases reviewed by MOFCOM – for example, the upward pricing pressure tests. In particular, in the Thermo Shenhua Coal Liquefaction Co Ltd (10 November 2011), http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201111/​ 20111107855595​.shtml. 23 See Anti-Monopoly Bureau of MOFCOM, 商务部关于禁止可口可乐公司收购中国汇源公司 审查决定的公告 (18 March 2009), http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​200903/​20090306108494​ .shtml. 24 See Simon Bishop and Mike Walker, The Economics of EC Competition Law: Concepts, Application and Measurement (Sweet & Maxwell 2010) 367. 25 Notice of the Ministry of Commerce of the People’s Republic of China No 38 of 2016 on the Additional Restrictive Conditions Authorized the Investic AB Group to Acquire Ownership of the Milwaukee Brewery Company in South Africa (29 July 2016), http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​ 201607/​20160701369044​.shtml. 26 Notice of the Ministry of Commerce of the People’s Republic of China No 35 of 2012 on the Approval of the Joint Technology Acquisition Gundricky Operators to Concentrate Antitrust Review Decision on Additional Restrictive Conditions (15 June 2012), http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​ 201206/​20120608181083​.shtml.

Economics in competition law enforcement in mainland China and Hong Kong  97 Fisher/Life Technology merger, MOFCOM included the results of the illustrated price rise tests and margin-concentration analysis undertaken by their external economists.27 MOFCOM recognizes that the scope for unilateral effects from a merger depends not just on market shares, but also on other factors, particularly dynamic responses from smaller competitors expanding production or new suppliers entering in response to higher prices, increasing supply, and curtailing price increases. Such a threat, if sufficiently credible, could undermine the merged entity’s ability to raise price unilaterally. MOFCOM has shown a willingness to consider different dynamic aspects of competition in most cases, and its analytical framework for unilateral effects assessment is not fundamentally different from the international standard. However, published decisions are relatively brief and it remains to be seen how much weight will be given to these dynamic factors in practice. In the Wilmar/Kemira JV case, for example, MOFCOM unconditionally cleared the proposed joint venture between two international players competing directly in China and operating complementary businesses. The parties overlapped horizontally in the production and supply of AKD wax in China (with limited overlap in the export of AKD wax into overseas markets). Hence, a classic unilateral effects concern was the reduction of direct competition between the parties in supplying AKD wax in China. To address this concern, the parties undertook empirical economic analyses similar to those commonly used in other major antitrust jurisdictions. For example, a switching analysis showed that customer diversion between Wilmar and Kemira was not disproportionately higher than the diversion between Wilmar (or Kemira) and other third parties. Further, the results of a critical loss analysis illustrated that rivals of the parties would have sufficient spare capacity to absorb any lost sales by the parties if (hypothetically) the parties were to raise prices post merger. These empirical analyses were submitted to MOFCOM, together with other quantitative and qualitative evidence, to support the conclusion that other rivals are likely to exert credible competitive constraints on the parties post transaction. 3 Coordinated effects A merger may lead to the increased likelihood of explicit or implicit collusion (that may or may not be lawful in and of itself).28 A price increase (or output restriction) resulting from coordinated effects depends not only on the merged entity’s own action, but also on whether other market participants consider competing more or less vigorously will yield higher profit. Until now, coordinated effects concerns have rarely been mentioned in Chinese cases decided by MOFCOM. For example, in Baxter/Gambro (2013), MOFCOM concluded that the parties competed horizontally in the market for the supply of continuous renal replacement therapy (CRRT) and related products. Both the global and Chinese CRRT markets were already highly concentrated. In the Chinese CRRT market, NIPRO, Gambro and Baxter each had a market share of 26 per cent, 19 per cent, and 3 per cent respectively. Prior to the merger, NIPRO was the contract manufacturer for Baxter. The merged entity post merger, together

27 Notice of the Ministry of Commerce of the People’s Republic of China No 3 of 2014 on the Approval of the Limited Anti-Monopoly Examination Decision Relating to the Acquired Restricted Technology Companies by Thermo Fisher Scientific, http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201401/​ 20140100461603​.shtml. 28 See, for example, Einer Elhauge and Damien Geradin, Global Competition Law and Economics (Hart 2007).

98  Research handbook on methods and models of competition law with NIPRO, would be the two main competitors in the market with a combined market share of up to 48 per cent. The manufacturing contract between NIPRO and Baxter contained information about production costs and quantities, which implied that NIPRO and the merged entity might have the ability and incentive to coordinate with each other to reduce competition post merger. Moreover, barriers to entry to the CRRT market were high due to investment costs, the lengthy process in establishing a sales network, and the need for patent licences.29 Further, in Primearth EV Energy (PEVE)/Toyota China/Keliyuan (2014), MOFCOM also concluded that Panasonic had a share of 19.5 per cent in PEVE in the global Vehicle NiCd (nickel-cadmium) battery market. The proposed joint venture therefore created mutual benefits for Panasonic, PEVE, and Keliyuan, implying that all three companies may have had the incentive to reduce competition in the market.30 A credible theory of harm based on coordinated effects requires that firms in the coordinating group (assuming identified) are: ●● able to reach an understanding as to price, output or capacity; ●● able to sustain such an understanding, despite an inherent incentive to cheat – thus, proper monitoring and punishment mechanisms are required; and ●● likely to change the market structure to the extent that it may either trigger the players to change their previously competitive strategies and tip the market towards a collusive outcome, or enhance collusion that was already taking place absent the merger. These conditions form a widely accepted analytical framework for assessing coordinated effects in jurisdictions such as the EU and the US. The economics literature supports the view that some industries are more prone to tacit collusion than others. Traditionally, factors conducive to tacit collusion were identified as including fewer firms, symmetry among competitors, inelastic demand, and lack of maverick firms. However, the development of economic understanding suggests that some factors might be more important than others, and some may be procompetitive depending on the nature of competition in the market. MOFCOM’s reluctance to pursue coordinated effects affirms the general lack of robust tools to identify tacit collusion. Although some economic reasoning was presented in Baxer/ Gambro, as discussed above, MOFCOM’s analytical framework is unclear. It remains to be seen whether more weight will be placed on these theories of harm going forward, and, more importantly, how empirical questions will be approached. So far, SAMR has not blocked any mergers; however, several have been cleared with conditions.

29 Notice of the Ministry of Commerce of the People’s Republic of China No 58 of 2013 on the Additional Restrictions Approving the Announcement of the Decision of the United States Baxter Int’l to Acquire the Concentration Antitrust Review by Swedish Campbell (Anti-Monopoly Bureau of the Ministry of Commerce (13 August 2013), http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201308/​ 20130800244176​.shtml. 30 Notice of the Ministry of Commerce of the People’s Republic of China on the Decision of the Ministry of Commerce on Limiting the Concentration of Anti-Monopoly Examination (Anti-Monopoly Bureau of the Ministry of Commerce, 2 July 2014), http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201407/​ 20140700648291​.shtml.

Economics in competition law enforcement in mainland China and Hong Kong  99 4 Economic framework for assessing non-horizontal mergers in China Non-horizontal mergers are less likely to be anticompetitive and may increase efficiency. MOFCOM has raised foreclosure concerns in some cases, particularly when one or both parties have a relatively strong market power. Generally speaking, two types of foreclosure concerns may arise from a vertical merger: ●● foreclosure of downstream rivals to its upstream products by the now vertically integrated entity – this could raise rivals’ costs and reduce competition in the downstream market; and ●● customer foreclosure, which arises when acquisition of an important downstream customer locks away customers from upstream rivals, raising downstream rivals’ costs and allowing the merged firm to increase downstream prices. Conglomerate mergers are mergers that are neither horizontal nor vertical. In practice, the focus is on mergers between companies supplying complementary products or weak substitutes.31 The main theory of harm underlying conglomerate effects is that the merging parties will leverage their (strong) position in one of the markets in question in order to foreclose rivals in the other market. Competition authorities typically consider the tying or bundling of products or, in some circumstances, the ‘portfolio effect’. Based on available information, foreclosure concerns have been raised by MOFCOM in decisions requiring remedies to secure a clearance. Cases of note include NXP/Freescale Semiconductors (2015)32 and Essilor/Luxottica (2018),33 which is discussed in Appendix B. Although not articulated in published decisions, it is understood that the analytical framework adopted by MOFCOM for assessing non-horizontal mergers was in line with the economic framework widely used in other jurisdictions – namely, the ‘ability–incentive–effects’ framework. This was applied in numerous merger cases in China.34 E

Practical Issues in Use of Economics in China

The importance of economics has been widely acknowledged by all competition agencies in China and by private practitioners. All three agencies have obtained expert economic opinions and advice in the process of their assessment. Although MOFCOM had some internal economic capacity (as several of its case handlers have economic backgrounds), it engaged external economists, as well as local academic economists, to provide advice on non-trivial cases. Over time, MOFCOM showed a real willingness to rely on external expertise in this area via hiring external economists to assist with competition assessment, especially high-profile global mergers, to identify plausible theories of harm and to assist MOFCOM in gathering evidence to assess these theories based on economic principles and market facts.

31 See European Commission, Guidelines on the assessment of non-horizontal mergers under the Council Regulation on the control of concentrations between undertakings (2008/C 265/07) 5–18 [91]. 32 Notice of the Ministry of Commerce Notice No 64 of 2015 on the Decision of Concentrating Antitrust Review by NXP Acquiring 37 per cent of Freescale’s Acquisition of Restricted Conditions (27 November 2015), http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201511/​20151101196182​.shtml. 33 For a summary of the decision, see https://​mp​.weixin​.qq​.com/​s/​GHzG1YjOWHdHagUaRY3Q. 34 See European Commission, Guidelines on the assessment of non-horizontal mergers under the Council Regulation on the control of concentrations between undertakings (2008/C 265/07) 5–18 [91].

100  Research handbook on methods and models of competition law In the Thermo Fisher/Life Technology merger, MOFCOM published the results of the illustrated price rise tests and margin-concentration analysis, illustrating MOFCOM’s willingness to explore various quantitative economic techniques beyond simple market share or Herfindahl-Hirschman Index (HHI) assessments.35 However, it is debatable whether such static tests would provide reliable assessment of competition concerns due to the dynamic nature of the industry in question. Other cases where external economists were used include Thermo Fisher/Life Technology Merck/AZ Electronic36 and Western Digital/Hitachi,37 which were cleared with conditions, and Lam/KLA-Tencor and Applied Materials/TEL, which were ultimately abandoned by the parties. MOFCOM’s willingness to engage external experts shows that it recognized the importance of applying economic thinking and conducting economic analysis in merger control. It has emphasized the need for facts-based economic analysis in merger investigations.38 External economists regularly interact with MOFCOM’s case team during the investigation process. This provides an effective way for MOFCOM to gain practical economic expertise. All these developments encouraged merging parties to put forward economic evidence and analysis, particularly in high-profile and complex mergers. NDRC and SAIC also favoured capacity building in economics. In particular, the Tetra Pak decision by SAIC showed that economics plays an important role in the whole investigation process, and in fact a large part of the final decision was economic reasoning and factual evidence. F

Acceptance of Economic Evidence in Courts

Private litigation is a key area of development that has significantly influenced antitrust debates in China. Compared to the decisions made by administrative authorities, court judgments often set out the parties’ arguments, the evidence, and the reasoning behind the court’s ruling in much greater detail. Parties subject to AML litigation can choose to submit expert economic evidence to the court in the form of written reports and can call expert witnesses to provide oral testimony at the hearing. There is currently no restriction on the use of foreign economic experts testifying before Chinese courts, indicating a willingness to consider such evidence. Two landmark decisions, Rainbow v Johnson & Johnson and Qihoo 360 v Tencent, discussed below, show attempts by the courts to adopt economic arguments but suggest that there are still flaws in the economic reasoning presented in the judgments. The case of Rainbow v Johnson & Johnson, decided in August 2013, was the first AML private litigation involving vertical agreements. The Shanghai Higher People’s Court considered that the reduction of 35 Notice of the Ministry of Commerce of the People’s Republic of China on Announcement on the Decision of Restricting the Antitrust Resolution of the Operator of Thermo Fisher (15 January 2014), http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201401/​20140100461603​.shtml. 36 Notice of the Ministry of Commerce of the People’s Republic of China on Announcement of the Decision of Concentrating Anti-Monopoly Examination by Merck Agile on the Acquisition of Asahi Electronic Material Co (30 April 2014), http://​fldj​.mofcom​.gov​.cn/​article/​201404/​20140400569060​ .shtml. 37 Western Digital [2012] MOFCOM Public Announcement No 9, 2 March 2012. 38 See, for example, 新浪财经, 商务部反垄断局审查处官员谈经济分析在经营者集中审查里的 应用 (6 March 2017), http://​finance​.sina​.com​.cn/​meeting/​2016​-11​-10/​doc​-ifxxsmic5900228​.shtml.

Economics in competition law enforcement in mainland China and Hong Kong  101 intra-brand competition would be sufficient for the effects-based test for RPM, despite there being no strong evidence suggesting that inter-brand competition was affected. This is a major flaw in the economic reasoning of the court’s conclusion. Qihoo 360 v Tencent was the first AML judgment by the Supreme People’s Court (SPC) of China, as well as the first case in which international economists testified. The SPC judgment provides significant insights into its largely effects-based approach to an abuse of dominance assessment. It demonstrates the ability of the SPC judges to engage in complex economic issues and to handle the evidence presented by the economic experts of both parties. Moreover, the SPC rectified several fundamental errors in the economic reasoning of the lower court judgment. Both parties submitted economic reports to address the fundamental questions, such as how to apply economic principles of relevant product and geographic market definition in the context of free goods distributed online, presenting data analysis of various measures of market shares, and examining the factors in assessing whether Tencent was dominant and the extent to which the conduct in question could harm consumers and competition. The scope and format of the economic reports in this case were very similar to reports used in an equivalent context in the EU. The SPC placed high prominence on economic reports and evidence throughout the process. The presence of third-party economists has been increasing in AML litigation cases, in particular in matters involving SEPs, such as the InterDigital/Huawei case and the Hitachi Metals dominance case in 2015. There are promising signs that economic reasoning and evidence are being considered by courts in China and that they could shape the debate and may even influence the outcome of a case. More importantly, the economic principles that have been considered by the Chinese courts are not fundamentally different from those in other jurisdictions.39 G

Uncertainties Do Not Undermine the Importance of Economics

Although the economic principles acknowledged by the Chinese authorities and courts are not fundamentally different from those in other jurisdictions, it is necessary to be aware of the objective constraints in engaging with economic evidence in cases under the AML due to unique features in China. First, there is still limited transparency in antitrust investigations in China and we need to observe a greater quantum of cases to assess whether the agencies and the courts will adopt a consistent and coherent approach on the economic principles. Second, where the criteria for decisions go wider than classic competition factors in China, outcomes may well differ from those that might come out of the EU or the US, for example in multi-jurisdictional mergers. Agencies may focus on different empirical questions in merger cases. This may be driven by the fact that competition dynamics in the industry in China differ from those in other countries, as the market definition may differ from Chinese customers’ perspectives and the competitive constraints in place may differ from those in other countries,

39 See the SPC ruling in Qihoo 360 v Tencent, https://​cgc​.law​.stanford​.edu/​zh​-hans/​judgments/​ spc​-2013​-min​-san​-zhongzi​-4​-civil​-judgment/​; the ruling by Guangdong High People’s Court in Huawei/InterDigital Communications, Inc, https://​www​.mlex​.com/​China/​Attachments/​2014​-04​ -18​_AXRC879FW8P38IO7/​guangdonghpc​_IDChuawei​_SEP​_18042014​.pdf; and the ruling by the Shanghai High People’s Court in Rainbow v Johnson & Johnson, https://​www​.pharmamedtechbi​.com/​~/​ media/​Supporting​%20Documents/​Pharmasia​%20News/​2013/​August/​JNJ​.pdf.

102  Research handbook on methods and models of competition law which may lead to China-specific competition concerns. Moreover, industry policies and other non-competition factors may also be relevant. Inevitably, the broader set of policy criteria we find in Chinese laws will make the decisions less predictable than would be the case with a pure competition test. Indeed, facts-based economic analysis that explains market reality should be universal and independent of differences between legal systems. Despite such uncertainties, the use of economics will help shift the debate towards one that is more competition oriented and challenge the validity of and motivation for complaints that appeal to the criteria that extend beyond the normal competition criteria.

IV

THE USE OF ECONOMICS IN ENFORCEMENT OF THE COMPETITION ORDINANCE IN HONG KONG

This section provides a brief overview of Hong Kong’s competition law regime, as well as discussing the use of economics as reflected in the relevant guidelines and some market studies. A

Hong Kong’s Competition Ordinance (2012)

Hong Kong’s CO was enacted in June 2012 but only became effective on 14 December 2015. The publication of guidelines and the making of subsidiary legislation occurred during this transition period. The CO prohibits three types of anticompetitive conduct described under the ‘competition rules’: the First Conduct Rule (FCR), the Second Conduct Rule (SCR) and the Merger Rule. While the FCR prohibits agreements and concerted practices that restrict competition, the SCR prohibits an undertaking with a substantial market power from abusing that power by engaging in conduct that restricts competition. The Merger Rule that prohibits anticompetitive mergers and acquisitions applies only to the telecommunications sector.40 The CO provides for the establishment of the Hong Kong Competition Commission (HKCC), with its key function to enforce the CO.41 Unlike the administrative model adopted in Europe, the competition regime in Hong Kong is prosecutorial: the HKCC is empowered to bring cases to the Competition Tribunal (the Tribunal), which determines whether the alleged conduct contravenes the CO and, if so, the level of penalty or any other appropriate remedies. A business that infringes a conduct rule can be fined up to 10 per cent of its turnover in Hong Kong. The Tribunal also has other broad powers to disqualify directors and impose penalties on individuals, award damages to aggrieved parties, make injunction orders, and terminate or vary an agreement. Follow-on private actions are allowed under the CO. The FCR does not apply to an agreement between undertakings (or a concerted practice engaged in by undertakings) in any calendar year if the combined turnover of the undertakings

40 The absence of a merger control regime in the CO was a result of substantial debates in Hong Kong before the enactment of the CO on whether merger control is needed for a small economy like that of Hong Kong. 41 The Communication Authority shares concurrent jurisdiction with the HKCC in respect of anticompetitive conduct of undertakings in the telecommunication sector.

Economics in competition law enforcement in mainland China and Hong Kong  103 does not exceed HK$200 million (approximately US$25.8 million).42 The SCR does not apply to conduct engaged in by an undertaking with an annual turnover of less than HK$40 million (approximately US$5.2 million). There is also a general exclusion for agreements that enhance overall economic efficiency. B

Enforcement Activity

The HKCC had received over 2,000 complaints and enquiries as of February 2017. Half related to the FCR, mainly involving alleged cartel conduct. On 23 March 2017 the HKCC announced that it had commenced its first proceedings in the Tribunal against five IT companies for alleged bid-rigging in a tender to supply and install a new server.43 The HKCC brought its second case in August 2017, alleging that ten renovation companies had engaged in a price-fixing and market-sharing agreement related to the provision of renovation services at a public housing site in Hong Kong.44 In September 2018 the HKCC brought a third case alleging infringement of the FCR against three construction companies and two individuals for price-fixing in relation to home renovation services at a housing estate, the first action against individuals.45 C

The Use of Economics

There has been very limited use of economics in enforcement cases in Hong Kong, mainly due to the small number and nature of the cases taken to date. However, economics is used in the Commission’s guidelines and in market studies conducted. For example, the Guideline on the FCR contains standard theories of harm of various horizontal and vertical agreements. The Guideline on the SCR details the methodology for defining markets using the concepts of demand- and supply-side substitution, as well as the standard hypothetical monopoly test (the SSNIP test). The Guideline on the SCR contains a standard economic framework on market power assessment and on conducts that may constitute an abuse of dominance. Two examples illustrate the use of economics in Hong Kong competition analysis.

42 These exclusionary provisions do not apply to agreements considered as ‘serious anticompetitive conduct’. The list of serious anticompetitive conduct includes price-fixing, market-sharing, production/ sales quota, bid-rigging, and RPM in some cases. 43 See HKCC, ‘Competition Commission Takes Big-Rigging Case to Competition Tribunal’ (Press Release, 23 March 2017), https://​www​.compcomm​.hk/​en/​media/​press/​files/​20170323​_Competition​ _Commission​_takes​_bid​_rigging​_case​_to​_Competition​_Tribunal​_e​.pdf. See also Competition Commission v Nutanix Hong Kong Ltd (Competition Tribunal, Lam J, 17 May 2019). 44 See HKCC (2017) 5 Competition Matters, https://​www​.compcomm​.hk/​en/​media/​newsletter/​files/​ Competition​_Matters​_Dec2017​.pdf. See also Competition Commission v W Hing Construction Company (Competition Tribunal, Lam J, 17 May 2019). 45 See HKCC, ‘Competition Commission Takes Renovation Cartel Case to Competition Tribunal’ (Press Release, 6 September 2018), https://​www​.compcomm​.hk/​en/​media/​press/​files/​Competition​ _Commission​_takes​_renovation​_cartel​_case​_to​_Competition​_Tribunal​_EnglishPR​.pdf.

104  Research handbook on methods and models of competition law 1 Resale price maintenance Resale price maintenance has historically been a major competition issue in Hong Kong, especially in industries such as supermarkets.46 Applying modern economic theory, the FCR’s Guideline states that RPM can restrict competition by:47 (i) facilitating coordination through enhanced price transparency in the market; (ii) undermining suppliers’ incentives to lower prices to distributors and distributors’ incentives to negotiate lower wholesale prices; (iii) limiting ‘intra-brand’ price competition, a particular concern when there are strong or well-organized distributors operating in a market – RPM facilitates coordination between distributors on the downstream market affected by the RPM; (iv) preventing the emergence of new market participants at the distributor level and generally hindering the expansion of distribution models with an aggressive pricing strategy (for example, the emergence of discounter distributors); and (v) excluding smaller suppliers from the market when implemented by a supplier with market power. Distributors have the incentive to promote only the product affected by the RPM, causing harm to consumers. Regarding theory (iii), the Guideline on the FCR clarifies that the HKCC interprets the FCR as prohibiting not only restrictions on inter-brand competition but also restrictions on intra-brand competition.48 Lin argues that there should perhaps be greater emphasis on protection of intra-brand competition in a small economy like Hong Kong (where retailing markets tend to be concentrated due to geographic constraints) compared to large economies where protection of inter-brand competition is perhaps more important.49 The Guideline on the FCR recognizes that RPM may lead to efficiencies, especially for ‘experience’ goods or complex products. In particular, the Guideline states that RPM may help address the so-called ‘free riding’ problems at the distribution level, where the extra margin guaranteed by the RPM structure encourages parties to provide certain sales services that benefit consumers. While recognizing the possible procompetitive effect of RPM, the HKCC emphasizes that ‘the Commission would expect to see compelling evidence of an actual free rider problem’. 2 Exclusive dealing The SCR’s Guideline on exclusive dealing states that the HKCC will have particular concerns when the following criteria are met: ●● an undertaking with a substantial degree of market power has imposed exclusive purchasing obligations on many customers; ●● no consumer benefits can be derived from such obligation; and

46 For more information, see, for example, Ping Lin, ‘Treatment of Resale Price Maintenance in Hong Kong’, CPI Antitrust Chronicle (30 September 2015). 47 See HKCC, Guideline on the First Conduct Rule (2015) [6.72], https://​www​.compcomm​.hk/​en/​ legislation​_guidance/​guidance/​first​_conduct​_rule/​files/​Guideline​_The​_First​_Conduct​_Rule​_Eng​.pdf. 48 See Ping Lin, ‘Treatment of Resale Price Maintenance in Hong Kong’, CPI Antitrust Chronicle (30 September 2015). 49 Ibid.

Economics in competition law enforcement in mainland China and Hong Kong  105 ●● the obligations cause foreclosure because of the locking-up of a significant portion of the relevant market. The Guideline specifically identifies conditional rebates (loyalty discounts or retroactive rebates) offered by a firm with substantial market power as a potential example of abusive conduct. They may foreclose the market since buyers switching portions of their demand to an alternative supplier would lose the rebate in respect of all products purchased, not only the incremental amount sourced from alternative suppliers.50 This same reasoning was adopted by China’s SAIC in Tetra Pak. We can therefore conclude that competition authorities in both Hong Kong and mainland China are up to date with the economic reasoning regarding retrospective discount. D Summary Early indications are that standard economic reasoning is likely to be used in future enforcement cases.

V

COMPARISON OF COMPETITION LAWS AND ENFORCEMENT IN MAINLAND CHINA AND HONG KONG

A

Fairness versus Efficiency

While belonging to one country, mainland China and the Hong Kong Special Administrative Region differ substantially in terms of economic, legal, cultural, and social aspects. One fundamental difference between their competition laws is that the AML contains a deeply rooted appreciation of fairness, whereas Hong Kong’s CO aims solely at market efficiency. The value placed on fairness is apparent in Article 17 of the AML; the provision prohibiting ‘unfairly high selling prices’ as an abusive conduct, can be understood as coming from the traditional Chinese value of protection of fairness in the marketplace. In Hong Kong, with its free market economy worldwide,51 entrepreneurship and efficiency are regarded as the driving forces for economic and social development. Therefore, there has been no strong voice calling for fairness in the area of competition law. Even at the public consultation stage prior to the passage of the CO, small and medium-sized enterprises (SMEs) were most concerned about the perceived danger of abusive conduct from large businesses using the competition law, as well as increased compliance costs.

50 See HKCC, Guideline on the Second Conduct Rule (2015) [5.31], https://​www​.compcomm​.hk/​en/​ legislation​_guidance/​guidance/​second​_conduct​_rule/​files/​Guideline​_The​_Second​_Conduct​_Rule​_Eng​ .pdf. 51 See Hong Kong Government, ‘Hong Kong Is Ranked as the World’s Freest Economy Again’ (Press Release, 16 February 2017), http://​www​.info​.gov​.hk/​gia/​general/​201702/​16/​P2017021600208​ .htm.

106  Research handbook on methods and models of competition law B

Protecting SOEs versus Protecting SMEs

The second difference between China’s AML and Hong Kong’s CO lies in the protection of enterprises of different sizes. The AML has provisions aimed at protecting large-scale domestic enterprises, especially state-owned enterprises (SOEs),52 whereas the CO contains provisions to protect SMEs. In fact, one reason that it took China 13 years to pass the AML was the concern that a competition law would harm SOEs and slow the rate of industrial concentration.53 The AML contains Articles that encourage concentration. Article 5 provides that ‘undertakings may via fair competition and voluntary combination combine with one another in accordance to the law, so as to expand scale and increase their competitiveness in the market’. Such a provision is rarely seen in other jurisdictions but has been interpreted by many as the intention of the Chinese lawmakers to protect large enterprises in China. In Hong Kong, however, SMEs account for around 90 per cent of employment, resulting in the exclusion of undertakings with an annual turnover of less than HK$200 million (approximately US$25.8 million) from the cartel provisions, and undertakings with an annual turnover not exceeding HK$40 million (approximately US$5.2 million) from the abuse of market dominance provisions. Political economy considerations are the key to understanding the above differences between the competition laws in China and Hong Kong. C

Lack of General Merger Control in Hong Kong

Another obvious difference between the CO of Hong Kong and the AML of China (and almost all other competition law regimes in the world) is that the CO does not have a merger control section except for the telecommunications sector. In particular, there appears to be a misunderstanding that merger-specific anticompetitive effects that arise post merger can be controlled by abuse of dominance provisions of the competition law, and it is well accepted that the unilateral effects of a horizontal merger led to increased prices, reduced production quality and services, and/or slower rates of innovation. China’s merger control regime has been very active in controlling such anticompetitive effects of mergers. That said, given the presence of the extraterritorial provisions under the AML’s regulation, it is unlikely that undertakings operating in mainland China will have the incentive to incorporate and merge in Hong Kong. D

Perfectionism versus Gradualism

Lastly, it may be argued that mainland China and Hong Kong have adopted two different approaches to enforcing their respective competition laws. While Hong Kong chose to establish a set of guidelines before its law came into force, the mainland’s enforcement of the AML follows the so-called ‘learning by doing’ procedure, where the three competition agencies

52 See, for example, Allan Fels, ‘China’s Anti-Monopoly Law 2008: An Overview’ (2012) 41(1/2) Review of Industrial Organization 7. 53 See, for example, Ping Lin and Yue Qiao, ‘Understanding the Economic Factors That Have Affected China’s Anti-Monopoly Law’ in Michal S Gal, Mor Bakhoum, Josef Drexl, Eleanor M Fox and David J Gerber (eds), The Economic Characteristics of Developing Jurisdictions (Edward Elgar Publishing 2015).

Economics in competition law enforcement in mainland China and Hong Kong  107 began to handle cases prior to issuing specific guidelines. This approach is consistent with China’s gradualism approach to economic reform during the past four decades. By the time the CO came into effect in late 2015, substantial information had been disseminated within the business community regarding the objectives and coverage of the CO, how HKCC would interpret and enforce the various conduct rules of the CO, and the legal risks of infringement. This approach substantially increased the degree of understanding of the public and business of the CO and its enforcement before the law went into effect. E

Enforcement Structure and Use of Economics

The structure of competition law enforcement in Hong Kong differs significantly from that in China. The HKCC has the legal power to investigate and bring cases to the Tribunal, which decides whether the alleged conduct constitutes a contravention of the CO. In mainland China, the enforcement agency has the power to both investigate and decide on cases (although the undertakings have the right to appeal for administrative reassessment under the AML). Hong Kong’s two-tier enforcement structure means that the reassessment of any cases brought to the Tribunal should lead to more careful and rigorous economic analysis of the HKCC before it brings the case to the Tribunal.

VI

CONCLUDING REMARKS

In this chapter, we reviewed the AML and the CO in the context of the use of economics in their enforcement. Since the AML came into effect in 2008, both the enforcement agencies and the courts in China have gradually applied economic analysis in all areas of its enforcement – especially in merger control and the abuse of market dominance. This is a remarkable achievement by the Chinese agencies, given their very limited staff capacity. While Hong Kong has had very few formal antitrust cases since its CO came into effect in 2015, the well-drafted guidelines of the HKCC during the preparatory phase incorporated modern economic reasoning. There is good reason to believe that Hong Kong will employ economic analysis in line with international standards in future enforcement.

market as a highly concentrated market.

upward pricing pressure.

Regular cargo transportation services and

refrigerated cargo transportation services (by line), agreements ASPA 1, 2 and 3 (collectively covers

Maersk Line/Hamburg

Südamerikanische

chloride.

have a combined share of approximately 70–75%

in the world. The ASPA 1, 2 and 3 vessel agreement parties

Coast route) to consolidate market power and harm The ASPA vessel-sharing agreement parties have

of shipping container (global) and marine cargo

transportation agency services (China).

(2017)c

In the refrigerated cargo transportation services on the Far East to South American West Coast route, Maersk has a share of 25–30% while Hamburg Süd has a share of 15–20%.

their operation, particularly having detrimental effect on the level of competition within the agreement. Post consolidation, the merged entity would have

Far East to South American West Coast route.

refrigerated cargo transportation services on the

harm competition in the market for the supply of

the market power to unilaterally raise prices and

75–80% in certain segments).

the vessel-sharing agreements, which may affect

competition. It also strengthens their power within a combined share in excess of 55% (and as high as

route.

and Asia 2 (the Far East to South American East

oil tanker transportation services (global), supply

Dampfschifffahrts-Gesellschaft

the Far East to South American West Coast route) in the Far East to South American West Coast

The merging parties could use vessel-sharing

control 50% of the supply of potassium chloride

by the parties and Mosaic, the merged entity would

markets outside of North America, jointly owned

further reduce Chinese buyers’ negotiating power. When including Capontex, a potash supplier to

give them the power to limit competition and

and the parties’ high combined market share would market share in the global supply of potassium

chloride.

The global market for potash, especially potassium High concentration in the relevant product market Potash Corporation of Saskatchewan has a 30%

MOFCOM characterized the A4 laser printer

would remove a close competitor and lead to an

printers in China.

Saskatchewan (2017)b

Agrium/Potash Corporation of

while Samsung’s was 5–10%.

similar product ranges; therefore, the merger

HP’s share in A4 printers in 2016 was 45–50%,

The parties’ market shares

printing material (eg cartridges) used by A4 laser

The market for A4 laser printers and the market for Market investigation showed that the parties had

MOFCOM’s concerns

Affected product/geographic market

Case notified (year)

Examples of more recent merger cases (with conditional clearance) that were subject to unilateral effects concern

HP/Samsung (2017)a

Table 5A.1

APPENDIX A

108  Research handbook on methods and models of competition law

Bard’s market share in the relevant market is almost 50%. Given this and a delta HHI of 270 resulting from the merger, Becton’s share is approximately 2.7%.

merged entity could allow it to harm competition

The market for core needle biopsy (CNB) devices Consolidation of market power would lead to unilateral effects that harm competition and consumers. Innovation could be adversely affected by the

examined the Chinese market separately).

in China.

Industries (2017)d

Becton/Bard (2017)e

high.

to obtain permit indicate that barriers to entry are

in design and manufacturing, and the time needed

period for new products, high technical difficulties

Lack of consumer switching due to long adaption

continue with this project.

the merger could reduce the parties’ incentive to

project that would directly affect Bard sales. Thus,

merger, as Becton is conducting a major R&D

new entrants to the parties.

technological requirements lead to limited threat of

High barriers to entry due to complexity and

and discriminatory pricing.

and consumers by conducts such as price increases

consumers. The increase in market power of the

assembling and testing services (MOFCOM also

Engineering/Siliconware Precision

25–30%, both globally and in China.

The merger would lead to a reduction in choice for The merged entity will have a combined share of

The global market for outsourced semiconductor

The parties’ market shares

MOFCOM’s concerns

Affected product/geographic market

Case notified (year)

Advanced Semiconductor

Economics in competition law enforcement in mainland China and Hong Kong  109

the Chinese non-selective herbicide market with

of eliminating or restricting competition in the

seeds, sterilizing coating agent for corn seeds,

Concerning vegetable seeds markets, Bayer and Monsanto each had a market share of 45–50% and 10–15% in the market of long-day onion seed in

well as the global markets of traits (ie corn, soybean cotton and oilseed rape traits) and digital agricultural.

and digital agriculture.

seeds undercutting process for sales in China respectively. The merged entity will have a market share of more than 60% in these markets post merger. Concerning traits markets, Bayer and Monsanto each had a market share of 0–5% and 55–60% in

markets. Bayer may conduct unilateral acts of eliminating and restricting competition such as price hikes in the markets of vegetable seeds in China. It may negatively impact the technical progress and weaken buyers’ bargaining power in the relevant traits market.

15–20% and 55–60% in the global cotton traits

China, as well as the global markets for traits.

market, indicating a combined share of 90–95%.

and 50–55% in the global oilseed rape traits

Finally, they each had a market share of 35–40%

It will further enhance Bayer’s dominant position. market, indicating a combined share of 70–80%.

Bayer and Monsanto each had a market share of

of non-selective herbicide and vegetable seeds in

effective competitors in the short term for markets indicating a combined share of 55–65%. Further,

Due to high barriers to entry, there will be no new the global corn traits and soybean traits markets,

of 45–50% and 15–20% in the market of carrot

the control of Bayer in the aforementioned

Specifically, the concentration will further enhance China respectively. They each had a market share

other competitors in the market.

seeds (eg long-day onion seeds) in China, as

corn seeds in China. The global markets of traits

a combined market share far higher than that of

The merging parties were ranked first and third in

The concentration has (or may have) the effects

The markets of non-selective herbicide, vegetable

insecticide coating agent for corn seeds and hybrid markets of non-selective herbicide and vegetable

The parties’ market shares

MOFCOM’s concerns

Affected product/geographic market

Case notified (year)

Bayer/Monsanto (2018)f

110  Research handbook on methods and models of competition law

The parties’ market shares The merged entity will be the biggest lenses

MOFCOM’s concerns The concentration may reinforce the merged

40–45% market share in mid-to-high-end

sunglasses market.

The merged entity will be the biggest mid-to-high-end sunglasses wholesaler with a 35–40% share.

future. Specifically, both Luxottica and Essilor have spent a significant amount of money on R&D to reinforce their market position. Despite

the counterfactual).

behaviours in the future absence of the merger (ie

merged parties to engage in procompetitive

constraint and (ii) reduce the incentives for the

point of view, the merger will (i) eliminate this

other in the future. Therefore, from a dynamic

them a significant competitive constraint for each

and frames respectively would have likely made

their leading market position in optical lenses

overlap in most of the relevant wholesale markets,

the merged parties having limited horizontal

low-end wholesale lenses market.

potential competition between the parties in the

More importantly, the concentration may eliminate wholesale lenses market and 30–35% share in

significantly higher than its rivals. It will have

markets and the mid-to-high-end wholesale

a

Notes: Notice of the Ministry of Commerce of the People’s Republic of China No 58 of 2017 on the Additional Restrictive Conditions to Approve the Announcement of Hewlett-Packard Co Ltd on the Decision of Concentrating Anti-Monopoly Examination of Some Business Cases of Samsung Electronics Co Ltd (Anti-Monopoly Bureau, 5 October 2017), http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201710/​20171002654063​.shtml; b Notice of the Ministry of Commerce No 75 of 2017 on the Approval of the Centralized Antitrust Resolution by the Operator of the Combination of Jiayang Company and Saskatchewan Potash Corporation Limited with Additional Restrictive Conditions (Anti-Monopoly Bureau, 6 November 2017), http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201711/​20171102666641​.shtml; c Notice of the Ministry of Commerce Notice No 77 of 2017 on the Additional Restrictive Conditions to Approve the Decision of Maersk Line to Take Over the Concentration Antitrust Review by the Owners of Hamburg South American Shipping Group (Anti-Monopoly Bureau, 8 November 2017), http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201711/​20171102667566​.shtml; d Notice of the Ministry of Commerce Notice No 81 of 2017 on the Additional Restrictive Conditions to Approve the Announcement on the Concentration of Anti-Monopoly Examination by the Operator of ASIMCO Semiconductor Manufacturing Co Ltd on the Acquisition of the Equity Ownership of Silicon Products Precision Industry Co Ltd (Anti-Monopoly Bureau, 24 November 2017), http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201711/​20171102675701​.shtml; e Notice of the Ministry of Commerce Notice No 92 of 2017 on the Additional Restrictions Approval of the Centralized Anti-Monopoly Review of the Becton-Dickinson Company and the United States Bud Company Merger (Anti-Monopoly Bureau, 27 December 2017), http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201712/​20171202691390​.shtml; f MOFCOM Announcement No 31 of 2018 on Anti-monopoly Review Decision Concerning the Conditional Approval of Concentration of Undertakings in the Case of Acquisition of Equity Interests of Monsanto Company by Bayer Aktiengesellschaft Kwa Investment Co, http://​english​.mofcom​.gov​.cn/​article/​policyrelease/​announcement/​201803/​20180302719967​.shtml.

and of retail optical products in China.

wholesale lenses, frames and sunglasses in China, entity’s market power in the two wholesale lenses wholesaler in China, with market shares

Affected product/geographic market

Markets of mid-to-high-end and low-end

Case notified (year)

Essilor/Luxottica (2018)

Economics in competition law enforcement in mainland China and Hong Kong  111

SAN Switch Integrated Circuits, while Brocade has a share of 70–80% globally and 40–50% in China in FC SAN Switch. Broadcom has a share of 40–50% globally (and in

Switch suppliers, obtained by Brocade through its operations of producing FC SAN Switch Integrated Circuits to other FC SAN Switch

The market for A4 laser printers and the market for Bundling printer with printing material is

HP/Samsung (2017)b

MOFCOM characterized the A4 laser printer market as a highly concentrated market.

such a strategy and thus harm competition and consumer interest in the printing material market.

while Samsung’s was 5–10%.

would have the ability and incentive to carry out

printers in China.

HP’s share in A4 printers in 2016 was 45–50%,

a possibility post merger as the merged entity

printing material (eg cartridges) used by A4 laser

constrain the merged entity.

is unlikely to be the threat of new entrants to

create high barriers to entry, and thus there

High capital requirement and technical difficulties

practices.

inappropriately, and using bundling and tying

system components, using confidential information

limit competition by reducing compatibility of

In the FC HBA market, the merged entity could

the FC SAN Switch market.

suppliers, to benefit itself and harm competition in China) in FC HBA.

Broadcom has a share of 30–40% globally in FC

confidential information of other FC SAN

and FC HBA.

Systems (2017)a

The parties’ market shares

The global market for Fibre Channel SAN Switch The merged entity could inappropriately use

MOFCOM’s concerns

Affected product/geographic market

Case notified (year)

Examples of more recent merger cases (with conditional clearance) that were subject to foreclosure concerns

Broadcom/Brocade Communications

Table 5B.1

APPENDIX B

112  Research handbook on methods and models of competition law

Essilor/Luxottica (2018)d

the Chinese non-selective herbicide market with

and incentive to conduct bundling sales of seeds,

seeds, sterilizing coating agent for corn seeds,

and of retail optical products in China.

wholesaler in China with market shares

The merged entity will be the biggest lenses

market, indicating a combined share of 90–95%.

and 50–55% in the global oilseed rape traits

Finally, they each had a market share of 35–40%

market, indicating a combined share of 70–80%.

15–20% and 55–60% in the global cotton traits

Bayer and Monsanto each had a market share of

indicating a combined share of 55–65%. Further,

the global corn traits and soybean traits markets,

each had a market share of 0–5% and 55–60% in

Concerning traits markets, Bayer and Monsanto

merger.

share of more than 60% in these markets post

respectively. The merged entity will have a market

seeds undercutting process for sales in China

of 45–50% and 15–20% in the market of carrot

China respectively. They each had a market share

wholesale lenses market and 30–35% share in the prices and gross margins of the parties) indicated

mid-to-high-end sunglasses wholesaler with a 35–40% share.

conducts. Doing so would be likely to foreclose rivals and lessen/eliminate competition from the relevant markets.

The merged entity will be the biggest a profit-maximizing intention to engage in such

that it would be profitable for a merged entity with low-end wholesale lenses market.

have 40–45% market share in the mid-to-high-end

switching ratio (evaluated based on the wholesale

frames into a complete pair of glasses. The critical significantly higher than their rivals. They will

wholesale lenses, frames and sunglasses in China, optical lenses together with Luxottica’s optical

The merged entity might bundle/tie Essilor’s

Monsanto each had a market share of 45–50% and

agriculture for bundling sales.

Markets of mid-to-high-end and low-end

Concerning vegetable seeds markets, Bayer and

risks that Bayer will utilize the platform of digital

and digital agriculture.

10–15% in the market of long-day onion seed in

competitors in the market.

non-selective herbicides. It may also increase the

corn seeds in China. The global markets of traits

a combined market share far higher than other

The merging parties were ranked first and third in

The concentration provides Bayer with the ability

The markets of non-selective herbicide, vegetable

insecticide coating agent for corn seeds and hybrid traits, and agrochemical products concerning

The parties’ market shares

MOFCOM’s concerns

Affected product/geographic market

Case notified (year)

Bayer/Monsanto (2018)c

Economics in competition law enforcement in mainland China and Hong Kong  113

Affected product/geographic market

their relative small sizes.

bundling/tying in the wholesale markets, given

to withstand any foreclosing effect from potential

(from retailers) is also considered to be difficult

lower than the merged entity. Further, buyer power

market shares in individual markets are much

in the optical lenses market only) and (ii) their

the merged entity (eg Carl Zeiss is mainly active

(i) the scope of their market presence is less than

to a potential bundling/tying conduct given that

The parties’ rivals would face difficulty in reacting

other.

the fact that they are not complements to each

between these two products are plausible, despite

conducts indicated that the bundling and/or tying

sunglasses and (ii) the economic analysis of the

merged entity’s market position in lenses and

sunglasses. The results from evaluation of (i) the

Essilor’s optical lenses together with Luxottica’s

Similarly, the merged entity might bundle/tie

MOFCOM’s concerns

The parties’ market shares

a

Notes: Notice of the Ministry of Commerce Notice No 46 of 2017 on the Authorization of Broadcom Co Ltd to take over the Equity Ownership of Brocade Communications Systems Limited for the Centralized Antitrust Review Decision (Anti-Monopoly Bureau, 22 August 2017), http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​ 201708/​20170802632065​.shtml; b Notice of the Ministry of Commerce Notice No 58 of 2017 on the Additional Restrictive Conditions to Approve the Announcement of Hewlett-Packard Co Ltd on the Decision of Concentrating Anti-Monopoly Examination on Some Business Cases of Samsung Electronics Co Ltd (Anti-Monopoly Bureau, 5 October 2017), http://​fldj​.mofcom​.gov​.cn/​article/​ztxx/​201710/​20171002654063​.shtml; c MOFCOM Announcement No 31 of 2018 on Anti-monopoly Review Decision Concerning the Conditional Approval of Concentration of Undertakings in the Case of Acquisition of Equity Interests of Monsanto Company by Bayer Aktiengesellschaft Kwa Investment Co, http://​english​.mofcom​.gov​.cn/​article/​policyrelease/​announcement/​201803/​20180302719967​.shtml; d For a summary of this decision, see https://​mp​ .weixin​.qq​.com/​s/​GHzG1YjOWjWHdHagUaRY3Q.

Case notified (year)

114  Research handbook on methods and models of competition law

PART II THE CONTENT OF THE LAW

6. Cartel prohibition and the search for deterrent penalties: the United States, the European Union, Australia and China compared Mark Williams1

I INTRODUCTION The search for the most effective way to prohibit and deter cartel activities has proved elusive. Despite the 125-year history of the Sherman Act prohibition of cartels2 with significant corporate fines and individual criminal and civil penalties, cartel conduct is still evident in the US.3 Research estimates that only a small fraction of cartels in the US and globally are detected4 and so the substantive prohibitions and remedies sufficient to effectively deter cartel conduct remain elusive. Most other countries globally now prohibit cartel activity but have adopted a wide variety of legal prohibitions with idiosyncratic exemptions and exclusions and have also adopted a variety of remedies. In the last two decades, international pressure on cartels has increased, especially following the endorsement by Member States of the recommendation of the Organisation for Economic Co-operation and Development (OECD) to enhance effective action against hard-core cartels in 1999.5 However, despite the adoption of improved detection techniques, increased corporate civil fines, and the imposition of individual criminal penalties by many countries, there appears to be little reduction in estimated cartel activity.6 1 The author gratefully acknowledges the invaluable research assistance of James Matheson and William Georgiou, which significantly contributed to the completion of this chapter. 2 15 US Code 26 Stat 209. 3 In the US, the Cartels section of GCR Antitrust Review of the Americas 2018 noted that recently closed cartel cases in 2016 included a very wide range of cartel conduct in markets related to auto parts, interest rates, foreign exchange rates, real estate auctions, pharmaceuticals, packaged seafood, electrical capacitors, ocean shipping, online wall décor, environmental services, construction, public school bus services and their location services. So, despite increased enforcement efforts related to personal criminal liability and much higher corporate fines, cartels continue to flourish in the US. See http://​ globalcompetitionreview​.com/​insight/​the​-antitrust​-review​-of​-the​-americas​-2018/​1147274/​united​-states​ -cartels. 4 Jeffrey E Zimmerman and John M Connor, ‘Determinants of Cartel Duration: A Cross-Sectional Study of Modern Private International Cartels’ (2005), https://​ssrn​.com/​abstract​=1​158577; John M Connor, ‘Recidivism Revealed: Private International Cartels 1990–2009’ (2010) 6(2) CPI Journal 101–27. 5 OECD, ‘Recommendation of the Council Concerning Effective Action against Hard Core Cartels’ (C(98)35/FINAL, 13 May 1998), http://​www​.oecd​.org/​daf/​competition/​2350130​.pdf. 6 Given that cartels are secret in nature, accurate estimates of active cartels are somewhat speculative. However, Connor notes the precipitate rise of global cartel enforcement activity. This evidence gives credence to the assertion that cartel activity operating in various parts of the world and across borders remains high. See John M Conner, ‘The Rise of ROW Anti-Cartel Enforcement’ (2015) 9 CPI Antitrust Chronicle, Competition Policy International, https://​www​.competitionpo​licyinternational​.com/​ the​-rise​-of​-row​-anti​-cartel​-enforcement/​.

116

Cartel prohibition and the search for deterrent penalties  117 This chapter seeks to consider and contrast the substantive cartel prohibitions of the US, the EU, Australia and China, comparing the institutional and procedural similarities and differences in those jurisdictions, and their respective remedy regimes. Finally, it will be argued that, despite the OECD recommendation for greater emphasis on individual criminal sanctions, other than in the US, due to its particular justice system, an emphasis on individual civil liability of corporate directors and senior managers is likely to be the most effective way to deter the formation and operation of hard-core cartels.

II

SUBSTANTIVE CARTEL PROHIBITIONS

In the US, section 1 of the Sherman Act, provides: Every contract, combination in the form of trust or otherwise, or conspiracy, in restraint of trade or commerce among the several States, or with foreign nations, is declared to be illegal.

This admirably simple provision clearly includes a prohibition of cartel activity. Decisions of the courts have confirmed that cartel activity is a per se offence and that mere proof of conspiracy is sufficient to found liability without consideration of whether anticompetitive harm is caused.7 The type of conduct categorized as a per se offence is not detailed in legislation but rather derives from individual decisions of the courts. Over the long period of enforcement of the Sherman Act, the courts have concluded that horizontal price-fixing,8 horizontal market division,9 and concerted refusal to deal10 are all per se offences. Consequently, the rule of reason is not engaged when dealing with cartel conspiracies.11 Section 1 of the Sherman Act then proceeds to provide the criminal penalties for breach: Every person who shall make any contract or engage in any combination or conspiracy hereby declared to be illegal shall be deemed guilty of a felony, and, on conviction thereof, shall be punished by fine not exceeding $10,000,000 if a corporation, or, if any other person, $350,000, or by imprisonment not exceeding three years, or by both said punishments, in the discretion of the court.

Thus, both corporations and individuals are criminally liable for fines and imprisonment. Moreover, a cartel conspirator also incurs civil liability to injured parties who may rely on a criminal conviction to ground a civil claim, or where no criminal prosecution has been undertaken by the Department of Justice (DOJ). By virtue of section 4 of the Clayton Act (1914), US federal law provides a cause of action to ‘any person injured in his business or property by reason of anything forbidden in the anti-trust laws’ and a successful plaintiff is entitled to 7 See Northern Pacific Railway Co v United States, 356 US 1, 5 (1958) – conduct has a pernicious effect of competition with no redeeming features; Broadcast Music, Inc v CBS, 441 US 1, 19–20 (1979) – such conduct would always, or almost always, restrict competition and decrease output. 8 United States v Trenton Potteries, 273 US 392, 397–8 (1927). 9 Continental T.V., Inc. v GTE Sylvania Inc, 97 S Ct 2549 (1977). 10 FTC v Superior Court Trial Lawyers Association, 493 US 411 (1990) (collusive effects); NW Wholesalers, Inc v Pacific Stationery & Printing Co, 472 US 284 (1985) (exclusionary effects). 11 See Jefferson Parish Hospital No 2 v Hyde, 104 S Ct 1551, 1556ff (1984); Gough v Rosmoor Corp, 585 F 2d 381, 386–9 (9th Cir 1978) – per se rule forecloses analysis of the purpose or effect of a restraint of trade.

118  Research handbook on methods and models of competition law recover ‘threefold the damages by him sustained, and the cost of suit, including a reasonable attorney’s fee’. This section clearly provides a powerful incentive to private litigants to enforce antitrust law and, indeed, most antitrust litigation in the US is a result of private lawsuits. Moreover, most individual states within the US also have antitrust statutes that largely mirror federal law, providing yet another incentive to enforce anti-cartel cases vigorously.12 Thus, the substantive cartel prohibitions in the US assist effective enforcement by providing: ●● a relatively low legal threshold test; ●● both criminal and private civil remedies; and ●● substantial procedural incentives for private claimants. However, the US system does not provide for federal civil antitrust penalties in cartel cases, although individual states do have such provisions.13 Turning to the substantive prohibitions of cartels in the EU, Article 101 of the Treaty on the Functioning of the European Union (TFEU) provides: 1. The following shall be prohibited as incompatible with the internal market: all agreements between undertakings, decisions by associations of undertakings and concerted practices which may affect trade between Member States and which have as their object or effect the prevention, restriction or distortion of competition within the internal market, and in particular those which: (a) directly or indirectly fix purchase or selling prices or any other trading conditions; (b) limit or control production, markets, technical development, or investment; (c) share markets or sources of supply; (d) apply dissimilar conditions to equivalent transactions with other trading parties, thereby placing them at a competitive disadvantage; (e) make the conclusion of contracts subject to acceptance by the other parties of supplementary obligations which, by their nature or according to commercial usage, have no connection with the subject of such contracts. 2. Any agreements or decisions prohibited pursuant to this Article shall be automatically void. 3. [Provides an exemption regime largely irrelevant to hard core cartels.]

Thus, with regard to cartels, the prohibition is principles based and declares unlawful ‘agreements between undertakings’ that ‘may affect trade between Member States’ and that ‘have as their object or effect the prevention, restriction or distortion of competition’ and, in particular, those which ‘(a) directly or indirectly fix purchase or selling prices or any other trading conditions’; ‘(b) limit or control production, markets, technical development, or investment’; or ‘(c) share markets or sources of supply’. These three specific iterations provide for the illegality of hard-core cartel agreements. Several points should be noted about the EU anti-cartel provisions. First, they only apply to ‘undertakings’ and not individuals; thus, the EU provisions are narrower in application than those of the Sherman Act. The concept of ‘undertaking’ has been broadly defined by the

12 For a brief discussion of US state antitrust laws, see Gibbs Law Group LLP, ‘State Antitrust Laws’, https://​www​.classlawgroup​.com/​antitrust/​state​-laws/​. 13 For a full discussion of antitrust civil penalties in the US federal and state systems, see Harry First, ‘The Case for Antitrust Civil Penalties’ (2009–10) 76 Antitrust Law Journal 127. 

Cartel prohibition and the search for deterrent penalties  119 European Court of Justice (ECJ) to include a very wide range of bodies14 involved in economic activity, regardless of legal status and the way in which it is financed. However, unlike the US, corporate directors and senior managers are not personally liable for cartel conduct. Secondly, EU provisions only apply to cartel agreements that ‘may affect trade between Member States’, so EU law only applies to agreements that affect trade between two or more Member States. If the conduct applies to only a single such state, then only the national law of that country will apply. Third, as in the US, the European Commission, the Directorate General Competition (DG Comp) and the ECJ have effectively categorized hard-core cartels as ‘infringements by object’. This is somewhat like the categorization of hard-core cartels in the US as per se infringements, though, unlike in the US, parties accused of cartel conduct can attempt to justify it under Article 101(3).15 The ECJ has adopted a wide definition of the conduct that will be classified as an object offence. In the 1998 case European Night Services,16 the court appeared to adopt a relatively narrow classification of object offences with regard to agreements to include ‘obvious restrictions such as price-fixing, market sharing, or control of outlets’ but more recent cases such as T-Mobile17 and Allianz Hungaria Ciztosito Zrt18 appear to have widened the object category considerably. However, in Groupement des Cartes Bancaire,19 the ECJ appeared to retreat from this expansive definition of object offences and reaffirmed that restriction of competition by object should be limited to cases where there was a sufficient degree of harm to competition that an examination of anticompetitive effects would be superfluous. Fourth, penalties in the EU for cartel infringements are limited to civil fines imposed on undertakings, not individuals, and there are no criminal penalties for undertakings or individuals. The penalty regime is set out in Regulation 1/2003.20 Besides infringing agreements being void at law, DG Comp has the power to issue interim measures (injunctive relief) and to impose financial penalties. Article 23 of the Regulation provides a maximum civil fine of 10 per cent of the undertaking’s worldwide turnover in the preceding year. Note that the fine is the undertaking’s total turnover and is not limited to the affected product market or territory. So, for a global multinational company, the theoretical maximum fine would be enormous. In Australia, cartels are prohibited by the Competition and Consumer Act 2010 (Cth) (CCA), as amended. Following the 2015 Competition Policy Review (Harper Review),21 significant changes were made to the cartel regime in Australia. The Australian cartel provisions are complex and lengthy but were (to some extent) shortened and simplified as a result of the Harper recommendations. But it should be noted that the Australian statute is a ‘rules-based’ rather than ‘principles-based’ system, which partly explains the great length and complexity of the prohibitions. Part IV of the Act contains the relevant provisions.

14 See Case C-41/90 Höfner and Elser v Macrotron GmbH [1991] ECR I-1979; Case C180/98 Pavlov [2000] ECR I6451; Case C-309/99 Wouters v Algemene Raad [2002] ECR I-1577. 15 See Richard Whish and David Bailey, Competition Law (8th edn, OUP 2015) 127–9. 16 Case T-374/94 European Night Services [1998] ECR II-3141. 17 C-8/08 T-Mobile [2009] ECR I-4529. 18 C-32/11 Allianz Hungaria Ciztosito Zrt ECLI:​EU:​C2013:​160. 19 Case C-67/13 Groupement des Cartes Bancaire v Commission [1984] ECR 1679. 20 [2003] OJ L1/1. 21 Competition Policy Review, Final Report (March 2015) (Harper Review), http://​ competitionpolicyreview​.gov​.au/​final​-report/​.

120  Research handbook on methods and models of competition law Section 45AD of the CCA elaborately defines a ‘cartel provision’. Essentially, where the purpose or effect of a provision in a contract, arrangement or understanding is to fix, control or maintain the price of goods or services, it is a cartel provision if two or more of the parties to the contract, arrangement or understanding are, or are likely to be, in competition but for the agreement. Separately, where the purpose of a provision in a contract, arrangement or understanding is to prevent, restrict or limit production, or the productive capacity, of a party; or the supply or acquisition of goods or services; or amounts to an allocation of customers or markets or bid-rigging, it is a cartel provision if two or more parties to the contract, arrangement or understanding are, or are likely to be, in competition with each other. Sections 45AF and 45AG create criminal offences of a corporation ‘making’ or ‘giving effect’ to a ‘contract, arrangement or understanding’ that contains a cartel provision, with the requirement that the corporation knows or believes that the agreement contains a cartel provision. The maximum criminal penalties provided are a fine of AU$10 million or three times the total benefit obtained, or 10 per cent of the annual turnover of the corporation. Moreover, pursuant to section 78, a ‘person’ who attempts, induces, aids, abets, counsels or procures, or is directly or indirectly knowingly concerned in, a cartel offence is liable to a maximum of ten years’ imprisonment and/or a fine of AU$420,000. Sections 45AJ and 45AK then provide parallel civil prohibitions whereby, if a corporation makes or gives effect to a ‘contract, arrangement or understanding’ that contains a cartel provision, the corporation is liable to civil penalties that are elaborated in Part IV of the Act. Under section 76, the Australian Competition and Consumer Commission (ACCC) may take an action in the Federal Court to recover a pecuniary penalty against a corporation that has made or given effect to a cartel provision with a maximum civil penalty similar to the criminal financial fines mentioned above. Moreover, under section 76(1)(b), a person (for example, a director or employee) knowingly involved in the making of or giving effect to a cartel provision is liable to a civil penalty of up to AU$500,000. In addition, the court has a range of other enforcement powers, including the disqualification of corporate directors under section 86D for a period that the courts find appropriate. China, for these purposes, includes the mainland territory of China but excludes Hong Kong and Macau, which have their own separate competition regimes. The Anti-Monopoly Law (2008)22 (AML) prohibits cartel conduct. Article 3 defines prohibited ‘monopolistic conducts’ to include ‘monopolistic agreements among business operators’. Article 12 defines ‘business operator’ as a natural person, a legal person or any other organization that is engaged in the production of commodities or the provision of services. Article 13 provides for the detailed prohibition of anticompetitive agreements, which includes hard-core cartels. Articles 13–15 provide: Article 13 Any of the following monopoly agreements among the competing business operators shall be prohibited: (1) fixing or changing prices of commodities; (2) limiting the output or sales of commodities; 22 The Antimonopoly Law of the People’s Republic of China (AML), adopted 30 August 2008. See http://​english​.mofcom​.gov​.cn/​article/​policyrelease/​Businessregulations/​201303/​20130300045909​ .shtml.

Cartel prohibition and the search for deterrent penalties  121 (3) dividing the sales market or the raw material procurement market; (4) restricting the purchase of new technology or new facilities or the development of new technology or new products; (5) making boycott transactions; or (6) other monopoly agreements as determined by the Antimonopoly Authority under the State Council. For the purposes of this Law, ‘monopoly agreements’ refer to agreements, decisions or other concerted actions which eliminate or restrict competition. Article 14 Any of the following agreements among business operators and their trading parties are prohibited: (1) fixing the price of commodities for resale to a third party; (2) restricting the minimum price of commodities for resale to a third party; or (3) other monopoly agreements as determined by the Antimonopoly Authority under the State Council. Article 15 [Provides for an exemption regime similar to that of the EU.] Where a monopoly agreement is in any of the circumstances stipulated in Items 1 through 5 and is exempt from Articles 13 and 14 of this Law, the business operators must additionally prove that the agreement can enable consumers to share the interests derived from the agreement, and will not severely restrict the competition in relevant market.

The AML clearly adopts the EU structure of prohibition (Articles 13 and 14) and exemption (Article 15) if the stipulated conditions are proved to the satisfaction of the Antimonopoly Authority. The law does not adopt the per se prohibition of hard-core cartels evident in US jurisprudence or the ‘object or effect’ distinction found in EU law. This does raise the issue of whether it is necessary to prove the anticompetitive effect of hard-core cartel conduct under the AML. The types of agreement listed in Article 13 must be between ‘competitors’ and so the prohibition applies to horizontal relationships, while the Article 14 prohibitions refer to agreements between business operators and their trading partners and so are limited to vertical arrangements. The monopoly agreements set out in the Article are ‘agreements, decisions or other concerted actions which eliminate or restrict competition’, presumably in a relevant market, though this is not specifically stipulated in the Article. However, usefully, Article 12 does define ‘relevant market’, which refers to ‘the commodity scope or territorial scope within which the business operators compete against each other during a certain period of time for specific commodities or services’. Little clarity as to the precise legal standard that must be proved in cartel cases has emerged in China. Most cartel decisions by the Antimonopoly Authorities are brief. Usually, accused parties confess their guilt, allowing the authorities to simply recite the basic facts of the collusion, the fact that this conduct is prohibited, and the fact that competitive harm was caused in the relevant market. Very few cartel decisions by the Chinese enforcement agencies have been appealed to the People’s Court and so the precise legal nature of the prohibition of cartels is opaque. Article 46 of the AML provides the penalties applicable to conduct that breaches Article 14. First, the Anti-monopoly Authority may order the business operators to cease and desist from implementing their anticompetitive agreement. Second, the Authority can order confiscation of the illegal gains. Third, the Authority can impose a fine of between 1 per cent and 10 per cent

122  Research handbook on methods and models of competition law of the business operators’ sales revenue in the previous year. And fourth, where the monopoly agreement has not been implemented, a fine of up to 500,000 RMB shall be imposed. It should be noted that there are no criminal penalties for ‘business operators’ involved in cartel conduct and that no individual penalties are imposed on directors or employees who organized or operated the cartel. In this respect, too, the Chinese law very much follows EU substantive law. However, Article 50 does provide that ‘[w]here any loss was caused by a business operator’s monopolistic conducts to other entities and individuals, the business operator shall assume the civil liabilities’, thereby allowing a private right to damages – though there are none of the incentives to civil litigation that are found in US law. Thus, the substantive US law is brief yet extensive in coverage and provides both substantial criminal and civil penalties against business actors and individuals, with significant incentives to private plaintiffs. Both the EU and Chinese laws have similar substantive provisions; both provide for civil penalties against firms but only Australia imposes individual civil penalties. However, they diverge on criminal penalties for cartels, which currently are only available in Australia, and neither jurisdiction provides substantial incentives to private litigation.

III

INSTITUTIONAL AND PROCEDURAL ISSUES

In the US, public enforcement of the Sherman Act cartel prohibitions is undertaken exclusively by the Antitrust Division of the DOJ.23 Most US states also have local antitrust statutes that are enforced by state Attorneys General.24 The US follows a judicial enforcement model, with the DOJ acting as the complainant in civil cases and the prosecutor in criminal ones. Thus, high levels of procedural fairness are maintained and enforced by the federal judiciary. State-level cases follow a similar process. Cartel busting has long been a priority for the DOJ and the Department has been a forceful advocate of strengthening cartel enforcement using leniency programmes to reward a cartel participant who provides useful information and evidence against other cartel participants in return for immunity from civil or criminal action by the DOJ.25 Additionally, in order to incentivize guilty pleas in criminal cases, the DOJ has used deferred prosecution agreements26 in the vast majority of criminal cases, thereby ensuring successful prosecutions as a result of defendants agreeing to accept a criminal conviction in return for a reduced sentence. At least as important in US cartel enforcement is the role played by private litigation. The clear majority of antitrust cases in the US at the federal or state level are because of private civil litigation.27 An opt-out class action regime, treble damages, immunity from liability for

See, for example, Fashion Originators’ Guild v FTC, 312 US 457, 463–4 (1941). See Federal Trade Commission, ‘The Enforcers’, https://​www​.ftc​.gov/​tips​-advice/​competition​ -guidance/​guide​-antitrust​-laws/​enforcers. 25 Scott D Hammond (Deputy Assistant Attorney General, US DOJ), ‘The Evolution of Criminal Antitrust Enforcement over the Last Two Decades’ (24th Annual National Institute on White Collar Crime, 25 February 2010), https://​www​.justice​.gov/​atr/​file/​518241/​download. 26 See, for example, US v The Royal Bank of Scotland PLC, https://​www​.justice​.gov/​atr/​case​ -document/​deferred​-prosecution​-agreement/​. 27 In FY 2017, 655 new civil antitrust actions, both government and private, were filed in the federal district courts, with many more filed in state courts. See OECD, ‘Annual Report on Competition Policy Developments in the United States’ (DAF/COMP/AR(2018), 2017) 18, https://​one​.oecd​.org/​document/​ 23 24

Cartel prohibition and the search for deterrent penalties  123 legal costs for successful defendants, contingent attorney fee arrangements, and jury trials make the US litigation system by far the most attractive global forum for private litigation. Combined, the DOJ’s criminal and civil litigation with private enforcement create the world’s most effective anti-cartel regime. But, even with corporate and personal criminal and civil liability over decades, cartel activity in the US remains a significant problem.28 In the EU, the enforcement mechanism is radically different from that of the US. DG Comp, a unit of the EU Commission, has exclusive jurisdiction to implement EU competition law where conduct affects two or more Member States.29 It shares jurisdiction with national competition authorities (NCAs) of the EU’s Member States where anticompetitive conduct affects markets within the boundaries of a single Member State.30 DG Comp has power to investigate and make an administrative decision on whether conduct is anticompetitive and in breach of EU law. If an infraction is found, DG Comp also has power to impose very substantial civil penalties of up to 10 per cent of the undertaking’s total worldwide turnover for the period of the infraction.31 But there is no power to impose civil fines on individuals (other than for procedural infractions) and there is no power to impose criminal penalties at all against undertakings or individuals; the EU has no criminal competencies as this lies within the exclusive jurisdiction of the Member States. Adverse decisions by DG Comp are subject to full judicial oversight by the EU courts. The General Court has power to conduct a full review on the merits and application of law, with a further right to appeal to the ECJ on a point of law. National courts also have the right to refer matters of interpretation of EU law where they must apply EU law in cases before them.32 Individuals and undertakings have a private right of action against undertakings that breach EU competition law and cause them damage. This right is exercised via the court systems of individual Member States and not through the EU courts. Private litigation rates in the EU Member States have historically been very modest, but have been more common in the UK courts – though limited when compared to the US. Recognizing that private litigation is an important component of appropriate enforcement of competition law, the EU has adopted a specific Directive33 to facilitate greater private litigation through the national courts of the Member States. Class actions in some Member State jurisdictions are allowed but in many civilian jurisdictions there is no, or only limited, ‘discovery’ allowed of the opponent’s documents or data that could be used to prove the fact of transgressions or to confirm the motivation for particular conduct. Access to such factual documents or internal accounting DAF/​COMP/​AR(2018)18/​en/​pdf. See also United States Courts, ‘Table C-2A – U.S. District Courts – Civil Judicial Business’ (30 September 2017), https://​www​.uscourts​.gov/​statistics/​table/​c​-2a/​judicial​ -business/​2017/​09/​30. 28 The very fact that so many new cartel cases are brought in federal and state courts by the DOJ and private litigants shows that cartel activities that affect US markets remain a significant problem, particularly as only a very small number of cartels are actually uncovered. 29 See Council Regulation 2003/1 and Commission Notice on cooperation within the Network of Competition Authorities 2004/C101/03. 30 See Council Regulation 1/2003. 31 Ibid. See also European Commission, ‘Fines for Breaking EU Competition Law’ (November 2011), http://​ec​.europa​.eu/​competition/​cartels/​overview/​factsheet​_fines​_en​.pdf. 32 See TFEU, Art 267. 33 For a discussion of the consequences of impeding DG Comp investigations, see Maurits ter Haar, ‘Obstruction of Investigation in EU Competition Law: Issues and Developments in the European Commission’s Approach’ (2013) 36(2) World Competition: Law and Economics Review 285–­313.

124  Research handbook on methods and models of competition law records is essential to allow plaintiffs to prove and quantify damages claims, but the judicial systems of many Member States do not provide sufficient tools to allow plaintiffs access to such information. Moreover, no EU Member State provides litigation incentives – opt-out class actions, treble damages, immunity from paying the legal costs of successful defendants, or jury trials – comparable to the US system. Moreover, most EU Member States forbid attorney contingency fees. Unsurprisingly, private litigation to enforce competition law in most EU countries remains limited despite the EU’s attempts to make such litigation more attractive and more feasible. Litigation funding mechanisms via third party litigation funders have become more prominent in recent years where group legal actions have been brought for consequential damages caused by cartel activities,34 though again the impact of such initiatives remains marginal. In Australia, the ACCC has an exclusive right to bring civil penalty cases before the Federal Court to enforce the Australian competition law and to seek the application of civil penalties, disqualification orders and other orders that can be made against corporate and private individuals who are ‘knowingly involved’ in competition law infractions.35 Additionally, the ACCC can refer suitable cases to the Commonwealth Director of Public Prosecutions (CDPP) for criminal prosecution in appropriate cartel cases against corporate entities or private individuals, who can be imprisoned.36 Australia also has a direct right of private action, so that affected persons can take civil cases to prevent or remedy anticompetitive conduct that adversely affects them.37 Australia also allows class action civil law suits, which have been a growing phenomenon in recent years.38 Thus, private rights of action do provide a significant boost to the overall enforcement of competition law in Australia. But, the incentives for private enforcement are relatively low – no treble damages, no lawyer contingency fees,39 and no protection against liability for the legal costs of a successful defendant. However, as Australia is a common law jurisdiction, plaintiffs do have access to pre-trial discovery of relevant defendants’ documents and litigation funding arrangements are available in class action cases.40 Overall, private cartel litigation in Australia tends to be ‘follow-on’ actions after the ACCC has brought successful

34 See Commission Directive 2014/104/EC on harmonization of antitrust damages. By 2018, all Member States had introduced national laws to comply with the directive. See also 2013/396/EC Commission recommendation on common principles of injunctive and compensatory collective redress mechanisms. 35 CCA, s 75B(1). 36 Memorandum of Understanding between the Commonwealth Director of Public Prosecutions and the ACCC regarding Serious Cartel Conduct (15 August 2014), https://​www​.cdpp​.gov​.au/​sites/​default/​ files/​MR​-20140910​-MOU​-Serious​-Cartel​-Conduct​.pdf. 37 See, for example, CCA, s 77. 38 Class actions take place in the Federal Court and State Supreme Courts of Australia. Part IVA of the Federal Court of Australia Act 1976 (Cth) and Division 9.3 of the Federal Court Rules 2011 (Cth) provide a regime for commencing class actions in the Federal Court, which has been in place since March 1992. The Federal Court of Australia and the New South Wales Supreme Court drafted a joint protocol for the management of class action proceedings as a result of an influx of class actions to both courts in 2018. 39 See, for example, Legal Profession Uniform Law Application Act 2014 (Vic), Sch 1 (Legal Profession Uniform Law), s 183. 40 See Getting the Deal Through (August 2017), https://​ gettingthedealthrough​ .com/​ area/​ 27/​ jurisdiction/​5/​private​-antitrust​-litigation​-2017​-australia/​.

Cartel prohibition and the search for deterrent penalties  125 civil cases so that the judicial findings made and the evidence disclosed can be relied on in the subsequent private litigation. In China, the enforcement mechanism for the AML is a mixture of administrative investigation and the imposition of civil fines on ‘business operators’,41 as well as direct private rights of action in the civil courts for the recovery of damages.42 The administrative enforcement structure adopted under the AML was initially complex. An Antimonopoly Commission – effectively, a policy and coordination committee chaired by a senior minister – oversaw enforcement machinery with delegated investigation and sanctioning to three operational units situated within other administrative bodies – namely, the Prices Bureau of National Development and Reform Commission (NDRC), the Antimonopoly Unit of State Administration of Industry and Commerce (SAIC) and the Antimonopoly Bureau of Ministry of Commerce (MOFCOM). Enforcement responsibility for cartel conduct was split between NDRC and SAIC, the division being on the rather amorphous basis that NDRC investigated ‘price-related’ anticompetitive conduct and SAIC was tasked to deal with ‘non-price-related’ conduct. As noted in previous chapters, in March 2018 the National People’s Congress adopted a wide-scale administrative reform, and enforcement of the AML was moved to a new super-regulator – the State Administration for Market Regulation (SAMR).43 An obvious problem with siting the antimonopoly bureau within a much larger market-wide supervisory body is that the bureau does not have an independent budget or staff allocation and needs to compete with all the other SAMR bureaus for resources. China’s cartel decisions in the first decade of the enforcement of the AML have been significant but enforcement activity only became more intensive in 2013, when a range of state-owned enterprises, international cartels and foreign multinationals were investigated and punished for engaging in both horizontal and vertical monopolistic agreements. It should be noted that in China, no distinction is made between vertical price monopoly (resale price maintenance) and horizontal price-fixing – both, as discussed above, are effectively per se offences. As regards private litigation in China, at the time of writing there is no reported follow-on or original litigation in respect of horizontal cartel conduct. China, as a quasi-civilian jurisdiction, does not have extensive discovery rules to allow a plaintiff to access documents or data in the hands of the defendant and it can only be done in very restricted circumstances where the court itself undertakes an investigation of the parties’ claims. Thus, satisfying the burden of proof is difficult. However, in certain circumstances – particularly involving ‘hard core’ AML violations, including price-fixing – the Supreme People’s Court Rules on the application of the AML provide that in cases of group boycotts or collective agreements to prevent the development of new technology, the burden of proof is shifted to the defendant to show that its conduct has not had the effect of eliminating or restricting competition in the relevant market.44 There has been, however, a successful private action taken against Johnson & Johnson, the US pharmaceutical company, in the Shanghai People’s court for damages resulting from AML, Chs VI, VII. AML, Art 50. 43 See Noah A Brumfield, J Mark Gidley, Z Alex Zhang and Yi Ying, ‘China Merges Antitrust Enforcement Agencies into One, as Its Anti-Monopoly Law Approaches 10th Anniversary’ (White & Case, 29 March 2018), https://​www​.whitecase​.com/​publications/​alert/​china​-merges​-antitrust​ -enforcement​-agencies​-one​-its​-anti​-monopoly​-law​-approaches. 44 The SPC Rules on Several Issues Regarding the Application of Law to Civil Disputes Involving Anticompetitive Conduct in 2012, effective 1 June 2012 (SPC Rules), Art 8. 41 42

126  Research handbook on methods and models of competition law resale price maintenance, a vertical price monopoly case under the AML.45 There the burden and standard of proof were a significant issue. The plaintiff argued that as the conduct was effectively a per se matter, it was not necessary to prove any anticompetitive effect of the defendant’s conduct. The appeal court disagreed and held that the plaintiff is required to prove that the defendant has imposed a resale price maintenance restriction, that the restriction has had anticompetitive effects, and that the plaintiff suffered consequent damage. The court went on to hold that the plaintiff nevertheless would succeed as there was sufficient evidence to show that the defendant, while only having a 20 per cent share of the relevant markets, had maintained elevated price levels for 15 years through its influence on distributors. This formulation of the law does not appear to coincide with the view of the enforcement agencies. Again, further administrative decisions or case law may resolve this apparent conflict. This case may not be followed in relation to horizontal price-fixing cases, given the Supreme People’s Court Rules on AML litigation that ease the burden of proof on plaintiffs in appropriate cases. Another difficulty for plaintiffs in cartel cases is that, unfortunately, the status of government agency determinations of facts in cartel investigations and published in penalty decisions is unclear. It is arguable that determinations of facts established by the AML enforcement agencies should not have to be proved by private plaintiffs in appropriate cases, but there has not yet been any determination of this issue. A further negative feature of Chinese civil procedure law is that class actions are unknown, though disparate individual actions can be consolidated if based on the same facts.46 However, this is a ‘joint litigation’ mechanism that has been used in a small number of cases involving environmental pollution, securities fraud and certain other claims involving a public interest element.47 But, in such cases, each individual plaintiff must litigate, all plaintiffs and the court must consent to a representative case being taken on their behalf, any settlement must be agreed by all plaintiffs represented, and the settlement only applies to plaintiffs registered with the court as members of the joint litigation.48 Thus, the mechanism is complex and expensive. Moreover, attorney contingency fees are prohibited49 and damages are compensatory only. Consequently, the incentives for effective private anti-cartel enforcement in China are very weak. It is unlikely that private litigation will be a major complement to public enforcement in the foreseeable future. In summary, the US and Australia adopt a common law litigation model with regard to enforcement, but the US provides far greater incentives for private enforcement and collective action. The EU and China both adopt an administrative enforcement model for public enforcement, but with no role for criminal prosecution of antitrust infractions. The EU has tried to encourage greater private civil enforcement as a complement to public enforcement at EU and Member State levels but stand-alone actions remain relatively rare, with most cases being follow-on cases after a liability determination by DG Comp or a national agency, often by the commercial customers of cartel operators. Private enforcement must be undertaken in 45 Beijing Ruibang Yonghe Science & Commerce Co Ltd v Johnson & Johnson (Shanghai) Medical Equipment Co Ltd, Civil Judgment of Shanghai High People’s Court, 2012 Hu Gao Min San Zhong No 63, entered 1 August 2013. 46 SPC Rules, Art 6. 47 Law of Civil Procedure of the People’s Republic of China, Art 55. 48 Ibid, Arts 52, 53, 54. 49 Regulation of Legal Services Fee by China National Development and Reform Commission and Ministry of Justice, effective 1 December 2006, Art 12.

Cartel prohibition and the search for deterrent penalties  127 the national courts of Member States, not through EU courts, thus complicating litigation decisions – though recent legislation does allow private plaintiffs to forum shop (to some extent) between national jurisdictions when enforcing EU rights. Collective redress is largely ineffective in many Member States. The EU has only been able to recommend common principles but has not adopted uniform rules. In China, the courts have played a significant role in developing the jurisprudence of private litigation, but the absolute number of cases is small. China has none of the US-style incentives to litigate nor an effective collective redress system. It appears that the US system of encouraging private litigation, thus providing additional enforcement pressure to promote good behaviour by firms, is not well copied in other jurisdictions.

IV

OPTIMAL REMEDY REGIMES AND THE ROLE OF PERSONAL LIABILITY

An optimal anti-cartel penalty and remedy regime should seek to achieve several different but complementary objectives: first, the removal of unlawful profits as unjust enrichment; second, financial penalties to act as a punishment for the infraction; and third, a separate additional financial penalty to deter the infringer and others considering engaging in cartel conduct in the future. The public enforcement of anti-cartel laws also requires that the human actors directly engaged in the decisions to form and operate cartels must also be subject to individual penalties. Imposing ever-larger civil fines on undertakings is a necessary facet of cartel punishment and deterrence but corporate fines alone are insufficient to provide the necessary deterrent to individuals. Individual sanctions that will act as a deterrent to managers from engaging in cartel conduct through the agency of their corporate employers are also required to achieve optimal deterrence and prevent future cartel formation and operation. The precise nature of this individual remedy regime – civil or criminal fines, director disqualification, imprisonment and ancillary remedies – will be considered later in this section. Private enforcement by undertakings and individuals not only compensates direct and indirect victims for losses incurred, but also acts as an additional lever to promote corporate compliance with competition law. Finally, it should be remembered that an optimal remedy regime can only function if cartels are discovered, investigated and successfully prosecuted. Given their inherently secret nature, increasing detection rates is a vital precursor to achieving an optimal remedial system. In considering our four jurisdictions, only the US system comes close to the optimal regime paradigm outlined in the preceding paragraph. The federal courts have the full range of both civil and criminal powers under various statutes that can be imposed on both corporate and human cartel participants. The US has the reputation of having the most effective anti-cartel sanctions of any competition regime globally. Both corporations and individuals are subject to fines and treble civil damages actions by affected private plaintiffs. Additionally, since 2004, corporate executives involved in an anticompetitive conspiracy can be subject to a felony indictment and punished with up to ten years’ incarceration and a fine of up to US$1 million.50 50 15 USC § 1, as amended by Public Law 108-237, Title II, § 215(a), 118 Stat 668. The maximum sentence of imprisonment was increased from three to ten years and the maximum fine that could be imposed on an individual was increased to $1 million.

128  Research handbook on methods and models of competition law However, the development of the current comprehensive suite of cartel remedies and penalties took a very long time. As Werden explained, originally cartel activity in the US was criminalized by the Sherman Act but as a misdemeanour, not the more serious felony category of criminal offence. Moreover, until the late 1950s, criminal convictions were very rare.51 Even as late as 1967, Flynn52 opined that cartel offences were not crimes of ‘moral turpitude’ – in other words, they were not morally wrong and so deserving of severe punishment but were more akin to regulatory infractions or minor road traffic offences. Attitudes hardened in the mid-1970s, when Congress categorized cartel infringements as felonies and substantially increased the maximum sentence available from one to three years.53 In 1977, the DOJ issued sentencing guidelines that proposed a sentence of 18 months as the starting point for judicial consideration, with the term being increased or reduced depending on aggravating or mitigating circumstances.54 But the courts were initially reluctant to treat cartel offenders harshly.55 During the 1980s, the number of prosecutions and convictions increased, as did the severity of sentencing – including an increase in prison time served by individuals. This culminated in 1987, when new sentencing guidelines were issued56 which emphasized that harsher sentences should be imposed to increase the deterrent effect to prevent the creation of cartels and to incentivize the use of leniency programmes to allow whistle-blowers to escape the higher sentences to be imposed. The courts responded more positively, with average prison time imposed increasing substantially to 247 days from an average of only 44 days in the 1970s.57 Punitive fines also began to be imposed on corporate offenders by the early 1990s. The final turn of the sanctioning screw on individual cartel offenders occurred in 2004, when Congress increased the maximum term of imprisonment from three to ten years and the individual fine was increased to a maximum of US$1 million.58 In the last ten years, sanctions on individual cartel participants under federal law have fluctuated, as can been seen in Tables 6.1–6.3. As shown in Table 6.1, the number of criminal cartel cases has fluctuated from a low of 25 in 2007 to a high of 63 in 2011. It is not clear why this has been the case but, given the relatively small number of cases, this may be explained by the exigencies of individual circumstances, or it might possibly reflect the change in administration from Bush to Obama in January 2009. What is striking is the relatively small number of cases taken by way of criminal prosecution, despite the size of the US economy, the priority given by the DOJ to intensive cartel enforce51 Gregory J Werden, ‘Sanctioning Cartel Activity: Let the Punishment Fit the Crime’ (2009) 5(1) European Competition Journal 19. 52 John J Flynn, ‘Criminal Sanctions under State and Federal Antitrust Laws’ (1967) 45 Texas Law Review 1301, 1315. 53 Antitrust Penalties and Procedures Act, Public Law 93-528, § 3, 88 Stat 1708. 54 Guidelines for Sentencing: Recommendations in Felony Cases under the Sherman Act (24 February 1977). 55 See United States v Alton Box Board Co, 1977-1 Trade Cases (CCH) ¶ 61,336 (ND Ill 1977). This case was governed by the original misdemeanour penalty limit of one year’s imprisonment, but the judge’s comments reiterated that, at least at that stage, the judiciary was sceptical of the moral wickedness of cartel offences. 56 US Sentencing Commission, ‘Sentencing Guidelines and Policy Statements’ (13 April 1987), reprinted at 52 Federal Register 18,046 (1987). 57 J Clabault and M Block, Sherman Act Indictments: 1955–1980, Vol 2 (Federal Legal Press 1981), 529–47. 58 See First (n 13).

Cartel prohibition and the search for deterrent penalties  129 Table 6.1

US Department of Justice, criminal cartel prosecutions 2007–2016

Restraint of trade –criminal

2007

(Sherman §1)

2008

2009

2010

2011

2012

2013

2014

2015

2016

Filed

25

26

37

41

63

58

45

37

54

39

Won

19

25

29

28

37

82

37

45

40

30

Lost

1

3

2

0

0

0

0

0

1

1

Pending

41

39

45

42

69

43

50

42

55

63

Source: US DOJ, ‘Antitrust Division: Workload Statistics FY 2006–16’, 7, https://​www​.justice​.gov/​atr/​file/​788426/​ download.

Table 6.2

US Department of Justice, fines imposed in cartel cases 2007–2016

Fines imposed Total individual fines (US$000)

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

15,109

1,485

605

4,373

1,522

2,141

3,069

2,016

369

5,245

25

23

27

19

25

31

29

24

15

31

Number of individuals fined

Source: US DOJ, ‘Antitrust Division: Workload Statistics FY 2006–16’, 11, https://​www​.justice​.gov/​atr/​file/​ 788426/​download.

Table 6.3

US Department of Justice, sentences of incarceration in cartel cases 2007–2016

Incarceration Number of individuals sentenced Number of individuals sentenced to incarceration Average number of days of incarceration

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

39

31

44

37

39

55

39

35

15

36

34

19

35

29

21

45

28

21

12

22

923

754

726

898

502

747

750

787

402

330

Source: US DOJ, ‘Antitrust Division: Workload Statistics FY 2006–16’, 12, https://​www​.justice​.gov/​atr/​file/​ 788426/​download.

ment, and the aggressive promotion of the benefits of its leading leniency program. Even more striking are the second and third lines of Table 6.1. In each of the last ten years, the DOJ has had an astonishing success rate in criminal cartel cases. In five years in succession, the DOJ did not lose a single case – including the bumper year of 2012, when the score card shows 82/82 wins. This is a truly astonishing result. Again, as shown in Table 6.2, individual fines imposed in criminal cases have fluctuated widely, from a low of US$605,000 in 2009 to a high of US$15 million in 2007. Again, the most likely explanation for this variation is the specific facts of individual cases: nothing can be read into these raw statistics as to the severity or leniency of the federal judiciary in cartel cases. In Table 6.3, we note paradoxically that the average length of imprisonment per sentenced felon fell from 932 days in 2007 to 330 days in 2016. However, these averages do not necessarily reflect a softening of the federal judiciary’s hard line on cartel sentencing but may reflect the variable nature and characteristics of the particular cases decided in each year. What lessons can we then draw from America’s long relationship with individual penalties for cartel offences?

130  Research handbook on methods and models of competition law Acceptance by prosecutors, judges, academics, business, politicians and the general public that cartel offences are a serious antisocial form of harmful conduct took a very long time. The current penalty and remedy regime took 105 years to evolve and, for most of that period, criminal sanctions were an irrelevant or minor part of the antitrust armoury. This illustrates a trite fact: changing social attitudes takes a very long time and prolonged advocacy by ideologically committed champions of change over a very extended period. Even after the ideological battle about the harmful nature of cartels has been won, the number of criminal cases in the world’s largest economy remains relatively small. While both imprisonment terms and individual fines have increased significantly, the future trajectory of punishment – in terms of both the number of cases prosecuted and the severity of the penalties imposed – remains uncertain. This might be because the combined effect of the hard line taken in terms of individual punishment and a well-publicized leniency programme has genuinely reduced the number of operative cartels in the US. However, empirical economic research seems to suggest that at least 70 per cent of operative cartels remain undetected and so unpunished.59 If this research is correct, then the relative success for leniency and criminalization in deterring cartels – even in the US – is questionable, as cartel offenders may have a seven-in-ten chance of not being caught and, ex hypothesi, punished. Two of the most interesting facts discernible from the DOJ statistics are the relatively low number of prosecutions in the world’s largest economy, which has by reputation the hardest anti-cartel policy globally, and the remarkable success that the DOJ has in criminal cases that it does take to trial. Thus, it appears that the DOJ is very careful only to take those cases to court when it is sure or almost sure of a conviction. Most importantly, the statistics do not show the percentage of defendants who plead guilty to cartel charges, or the percentage of those guilty pleas induced through plea bargains. Plea bargains are an essential tool for the DOJ to secure guilty pleas and hence its remarkable success rate in criminal prosecutions. In 2008, O’Brian of the DOJ Antitrust Division confirmed that ‘[o]ver 90 percent of the hundreds of defendants charged with criminal cartel offenses during the last 20 years have admitted to the conduct and entered into plea agreements with the Division’.60 This practice has not changed since then and so is the key reason why the DOJ has been so successful in criminal cartel cases. Presumably, in cases where a plea bargain cannot be secured, prosecutions are unlikely to be taken – thus ensuring that success rates remain so astonishingly high. Defended cases formed not more than 20 per cent of criminal cases up to 2008 and the more recent 2016 statistics already discussed above appear to show that even fewer defended cases have since been brought to trial. To drive home the message, the then Deputy Attorney General, Sally Yates, issued a memo to DOJ prosecutors in which she emphasized this principle: ‘One of the most effective ways

59 OECD Secretariat, ‘Serial Offenders’ (October 2015). Cartel studies generally conclude that only about 10 to 30 per cent of all such conspiracies are discovered and punished (quoting from John M Connor, Price Fixing Overcharges: Revised 2nd Edition (27 April 2010). Available at SSRN: https://​ssrn​ .com/​abstract​=​1610262  or  http://​dx​.doi​.org/​10​.2139/​ssrn​.1610262). 60 See A O’Brian, ‘Cartel Settlements in the U.S. and EU: Similarities, Differences & Remaining Questions’ (13th Annual EU Competition, Law and Policy Workshop, Florence, 6 June 2008).

Cartel prohibition and the search for deterrent penalties  131 to combat corporate misconduct is by seeking accountability from the individuals who perpetrated the wrongdoing.’61 The DOJ is clearly wedded to its ‘gold standard’ of individual criminal sanctions supported by the essential tools of leniency agreements, including immunity from prosecution for the company and/or for corporate executives coupled with plea bargains for those other conspirators who do not benefit from leniency-derived immunity. As Snyder has stated, the DOJ views individual accountability as key to cartel deterrence and the threat of prison time is by far the most potent weapon to concentrate the minds of corporate executives who have instigated or been involved in a cartel conspiracy.62 Thus, the DOJ considers individual criminal responsibility an essential tool in its anti-cartel toolbox. It is wedded to its criminal sanctioning policy and eschews other types of personal liability as second-best alternatives and much less effective than imprisonment. In an article published in 2011 and later presented to an American Bar Association conference, a team of senior staff members from the DOJ set out the case for maintaining their criminal prosecution model. They are of the view that both theory and experience suggest rational decision-making by executives in cartel formation and operation and that the risk of imprisonment – rather than mere corporate or individual fines that would be reimbursed by employers – tends to weight decision-making against cartel participation. Imprisonment also helps encourage plea bargains and evidence gathering, which secures additional convictions in the same or other cartels. They also argue that director disqualification is not as potent a deterrent as imprisonment as it is a derivative of individual financial penalties in terms of lost income.63 Thus, the American position is clear: ●● Individual responsibility for cartel offences is a more effective deterrent than ever-higher corporate fines. ●● Prison time works, as it concentrates the minds of corporate executives on avoidance of involvement in cartels and in ensuring the adoption of effective corporate compliance far better than mere corporate or individual fines. ●● Disqualification as a director or manager of a firm is a poor alternative to a prison sentence. ●● Other jurisdictions should adopt the American model to enhance cartel deterrence. However, this position has several major flaws that make such a transplantation to other jurisdictions highly problematic: ●● Other jurisdictions will have the same problems of convincing lawmakers and other stakeholders of the moral wickedness of cartels and this may take decades to come to fruition, as it did in the US. ●● While the DOJ has civil and criminal prosecutorial authority, most other countries do not have a similar institutional architecture. Consequently, extensive coordination is needed

61 Sally Yates, ‘Individual Accountability for Corporate Wrongdoing’ (Memorandum, 9 September 2015), https://​www​.justice​.gov/​archives/​dag/​file/​769036/​download. 62 Brent Snyder, ‘Individual Accountability for Antitrust Crimes’ (Yale Global Antitrust Enforcement Conference, New Haven, Connecticut, 19 February 2016), https://​www​.justice​.gov/​opa/​speech/​deputy​ -assistant​-attorney​-general​-brent​-snyder​-delivers​-remarks​-yale​-global​-antitrust. 63 Gregory J Werden, Scott D Hammond and Belinda A Barnett, ‘Deterrence and Detection of Cartels: Using All the Tools and Sanctions’ (2011) 56 Antitrust Bulletin 207.

132  Research handbook on methods and models of competition law between the competition agency and the public prosecutor, who may be entirely ignorant of competition law and might afford cartel cases a very low priority given a case load of terrorist offences, homicides and other ‘traditional’ criminal offences. ●● Most other legal systems do not have the same prosecutorial tools as the DOJ. Most legal systems do not have a coordinated, seamless interface between leniency programmes and criminal plea-bargaining agreements. Thus, the two instruments and their intimate connection may not be available to competition agencies or public prosecutors in other countries. It is suggested that without the unique US systemic use of leniency and plea bargaining, the DOJ’s success in criminal prosecutions would be much lower than is currently the case. Consequently, it is suggested that in other countries, unless they can exactly replicate the circumstances extant in the US, reliance on individual criminal sanctions will be unsuccessful. In these circumstances, the option of harsh individual civil sanctions, as advocated in Section I of this chapter, as a viable alternative to personal criminal liability, appears to be a much more promising method of attaching individual responsibility to corporate insiders and to deter cartel formation. In the EU, the enforcement system is entirely administrative and civil in nature; moreover, as has been previously mentioned, only ‘undertakings’ are subject to civil remedies. Individuals cannot, unless the person is an undertaking, be subject to personal civil fines, and criminal penalties against individuals or undertakings are simply unavailable. Fundamentally, this is because the EU legal order does not provide for the creation of criminal offences, save in one limited case, and criminal jurisdiction is entirely within the competence of the domestic legal systems of the EU Member States, which vary very considerably. The position was succinctly explained by the EU’s Director General for competition:64 The EU enforcement system is, by contrast [with the US], an administrative one, built around financial sanctions against undertakings, not individuals. Fines against companies are exclusively set as a deterrent against cartels: there are no treble damages. The current legal framework of the European Union does not provide for criminal sanctions, and, in particular, custodial sanctions. This is what the debate about criminal sanctions usually refers to, imposed through a procedure involving a public prosecutor and a trial before a court … So a role for the EU in a criminal procedure is not a theoretical impossibility, although the current Treaty restricts it to the financial interests of the EU.

Consequently, the EU’s current penalty regime is seriously flawed, as personal liability cannot be imposed on corporate managers who instigate or operate cartels. In the case of Australia, prior to the 1989 OECD recommendation on toughening cartel sanctions, Australia had an extensive individual civil liability regime for anticompetitive agreements, including price-fixing and market-sharing,65 even though the number of individual civil sanctions for cartel infringements was relatively low.

64 Alexander Italianer (Director-General for Competition, European Commission), ‘Fighting Cartels in Europe and the US: Different Systems, Common Goals’ (Annual Conference of the International Bar Association, 9 October 2013), http://​ec​.europa​.eu/​competition/​speeches/​text/​sp2013​_09​_en​.pdf. 65 Prior to the 2009 introduction of criminal cartel prohibitions, section 76 of the then Trade Practices Act 1974 (Cth) contained pecuniary penalties of up to $50,000 for individuals and $250,000 for corporations in respect of each civil contravention of Part IV.

Cartel prohibition and the search for deterrent penalties  133 Progress to criminal liability for cartel conduct was slow66 but the CCA now criminalizes making a contract or arrangement, or arriving at an understanding that contains a cartel provision,67 and giving effect to a contract, arrangement or understanding that contains a cartel provision.68 The law makes cartel conduct a criminal offence and provides for imprisonment for up to ten years and individual fines of 2,000 penalty units (one penalty unit corresponded to AU$210 in 2017, so a maximum financial penalty of AU$420,000).69 The Australian experience with criminalization has been disappointing. In the period between the adoption of the criminal provisions in 2009 and 2017, only one criminal case was brought. This was a case involving a cartel in the ocean transport of motor vehicles to Australia between six shipping lines. Nippon Yusen Kabushiki Kaisha (NYK) pleaded guilty to a cartel offence and cooperated with the ACCC. A corporate fine of AU$25 million – the second-highest cartel penalty in Australian competition law history at that time – was imposed on 3 August 2017.70 (It should be noted that the highest fine at that time was also imposed in a civil proceeding.71) No individuals were charged, despite the judge noting that senior executives of NYK were knowingly involved in the cartel, as all their conduct was outside the jurisdiction.72 The fine reflected a 50 per cent discount on the appropriate penalty, as NYK pleaded guilty at the earliest opportunity and provided the ACCC with cooperation with regard to the other members of the cartel. Interestingly, Wigney J in his sentencing comments concluded by saying:73 The penalty imposed on NYK should send a powerful message to multinational corporations that conduct business in Australia that anti-competitive conduct will not be tolerated and will be dealt with harshly when it comes before this Court. That is so even where, as here, the decisions and conduct are engaged in overseas and as part of a global cartel. As has already been explained, but for NYK’s cooperation and willingness to facilitate the administration of justice, the fine that would have been imposed on NYK would have been substantially higher: as high as $50 million. That should serve as a clear and present warning to others who may have engaged in, or who may be planning to engage in, similar conduct in the future.

Australia does not have a formal plea-bargaining system. It does, however, have a well-established immunity and cooperation policy for cartel conduct, which covers both civil 66 See Trade Practices Review Committee, ‘Review of the Competition Provisions of the Trade Practices Act’ (Australian Department of Treasury, 31 January 2003) ch 11, http://​tpareview​.treasury​ .gov​.au/​content/​report​.asp. 67 See the Trade Practices Amendment (Cartel Conduct and Other Measures) Act 2009 (Cth). 68 CCA, s 44ZZRG. 69 Ibid, s 79. 70 Commonwealth Director of Public Prosecutions v Nippon Yusen Kabushiki Kaisha [2017] FCA 876, http://​www​.judgments​.fedcourt​.gov​.au/​judgments/​Judgments/​fca/​single/​2017/​2017fca0876/​ summary/​2017fca0876​-summary. 71 ACCC v Visy Industry Holding Pte Ltd (No 3) [2007] FCA 1617 (2 November 2007), https://​www​ .australiancompetitionlaw​.org/​cases/​visy2​.html. 72 See Commonwealth Director of Public Prosecutions v Nippon Yusen Kabushiki Kaisha [2017] FCA 876 at [62], [188], [191]. All conduct occurred overseas. No individuals concerned were citizens or permanent residents of Australia. Section 5 of the CCA provides that Part IV applies to conduct engaged in outside of Australia by bodies corporate incorporated in or carrying on a business in Australia. The judgment also noted that the managers involved in the cartel conduct could not have been compelled to attend ACCC interviews as they were permanently overseas. 73 Ibid, [300] (emphasis added).

134  Research handbook on methods and models of competition law and criminal cartels.74 Several new criminal cartel prosecutions were commenced in 2018, both against both corporate defendants and against individual managers allegedly involved in the offences.75 These cases are ongoing, so no conclusions can be drawn as yet. If the various defendants are convicted, the penalties imposed will be keenly analysed. With regard to the cartel provisions in the CCA, a 2015 Report into the CCA generally76 noted that the criminal cartel provisions were: ●● unnecessarily complex, making the law difficult to understand and comply with and did not provide business with sufficient clarity; and ●● framed too broadly and criminalized commercial conduct that ought not be characterized as cartel conduct, including joint venture activity and vertical arrangements between suppliers and their customers. Recommendations arising from the review were implemented by the Competition and Consumer Amendment (Competition Policy Reform) Act 2017 (Cth), which came into effect on 6 November 2017. It should be noted that the ACCC cannot itself undertake criminal prosecutions. Suitable cases are transferred to the Director of Public Prosecutions to undertake the criminal prosecutions in appropriate cases.77 Presumably, the complexity and opacity of the provisions, their overbroad nature, and a lack of cases for which the ACCC was confident that it could prove cartel conduct to the criminal standard provide the reasons why individual criminalization in Australia has, so far, been a failure. Interestingly, the Harper Panel did not critically consider whether prioritizing the use of existing civil individual sanctions and inviting the courts to impose the existing harsh civil penalties might not have been a better alternative, rather than persisting with a criminal regime that has proved so obviously unsuccessful. It is perhaps unfortunate that there was not a more critical analysis of the rationale for criminalization and consideration given to enhancing the severity of individual civil sanctions as a more viable alternative to criminalization and all the institutional complexities and difficulties of proof beyond reasonable doubt before a jury that criminalization inevitably entails in the Australian context. So, despite the NKY corporate guilty plea to a criminal cartel charge, it was only in 2018 that further prosecutions were launched, without verdicts to date, making it impossible to conclude that successful criminal prosecutions will become a greater feature of the Australian cartel enforcement landscape in the future.

74 ACCC, ‘ACCC Immunity and Cooperation Policy for Cartel Conduct’ (September 2014, updated 2018), https://​www​.accc​.gov​.au/​publications/​accc​-immunity​-cooperation​-policy​-for​-cartel​-conduct. 75 ACCC, ‘Criminal Cartel Proceedings Commenced against The County Care Group and Its Managers’ (Media Release, 15 February 2018), https://​www​.accc​.gov​.au/​media​-release/​criminal​-cartel​ -proceedings​-commenced​-against​-country​-care​-and​-its​-managers. A second group of prosecutions was launched in June 2018 against ANZ, Citibank and Deutsche Bank and several senior executives: see ACCC, ‘Criminal cartel charges laid against ANZ, Citigroup and Deutsche Bank’ (Media Release, 5 June 2018), https://​www​.accc​.gov​.au/​media​-release/​criminal​-cartel​-charges​-laid​-against​-anz​-citigroup​ -and​-deutsche​-bank. 76 Bruce Billson MP, ‘Review of Competition Policy’ (Media Release, 4 December 2013), http://​bfb​ .ministers​.treasury​.gov​.au/​media​-release/​014​-2013/​. 77 Memorandum of Understanding between the Commonwealth Director of Public Prosecutions and the ACCC (n 36).

Cartel prohibition and the search for deterrent penalties  135 Under the existing civil penalty provisions, section 77A of the CCA provides that a company must not indemnify a person (whether by agreement or by making a payment and whether directly or through an interposed entity) against any of the following liabilities incurred as an officer of the company: (a) civil liability; or (b) legal costs incurred in defending or resisting proceedings in which the person is found to have such a liability. Note that this prohibition on corporate indemnities of officers applies only to civil penalties and not to criminal fines. Section 77B provides that payment of such an indemnity is rendered void at law. Under section 79, a person who aids, abets, counsels or procures, induces, or is in any way directly or indirectly knowingly concerned in, or is a party to, or conspires with others to contravene, a cartel offence is taken to have contravened that provision and is punishable with a maximum penalty of ten years’ imprisonment and/or a maximum fine of (currently) AU$420,000. The Act goes on to provide a range of additional civil remedies and penalties for individuals involved in cartels in Australia: ●● under section 80, injunctions; ●● under section 82, damages; ●● under section 86C, non-punitive orders in civil or criminal cases – community service orders, probation, disclosure orders and an order to publish an advertisement in terms ordered by the court; ●● under section 86D, punitive orders disclosure and adverse publicity; ●● under section 86E, disqualification from managing a corporation for whatever period the court deems appropriate; and ●● under section 87, wide powers to order compensation or other orders rectifying harm caused by a person in contravention of the anti-cartel provisions. These wide-ranging penalty and remedial provisions for both public and private enforcement are not as widely used beyond the civil fines that the ACCC habitually seeks. Civil penalties in Australia are low by international standards, with the highest fine ever imposed being AU$46 million in 2018 in the Yazaki cartel case.78 Arguably, Australian civil penalties imposed on corporate entities are too low to achieve deterrence, if one accepts the orthodox position on the correct level of fines necessary to achieve corporate deterrence. Moreover, individual civil penalties are in theory significant, but a cursory examination of recent Australian cartel cases shows that even these are mild and unlikely to be the deterrent needed to ensure that corporate executives and their companies comply with the CCA. A recent OECD report confirmed the low levels of fines for anticompetitive activities in Australia generally by comparison with overseas jurisdictions, concluding: Despite Australia’s competition law system already being in line with international practices, it has particular characteristics, notably its method for applying sanctions. While in most regimes pecuniary penalties are set by reference to a detailed and publicly available methodology that focuses largely on the relevant sales of the infringing company’s product, in Australia the amount of pecuniary penalties is determined by the Federal Courts following an ‘instinctive synthesis’ of various factors. These differences do not prevent Australia from imposing substantial and deterrent sanctions for breaches

ACCC v Yazaki Corporation [2018] FCAFC 73.

78

136  Research handbook on methods and models of competition law of competition law. However, the maximum penalties that are imposed in Australia are nonetheless lower than in comparable jurisdictions for competition law infringements.79

To conclude the analysis of the Australian individual penalty regime for persons involved in cartels, it should be noted that the criminal regime introduced in 2009 as a direct result of the OECD recommendation in 1998 has, so far, singularly failed to provide a deterrent effect in this jurisdiction. After a slow start, poor drafting and complex administrative issues involving the ACCC and the CDPP meant that it took five years to negotiate and agree a memorandum of understanding.80 These factors combined to slow the selection of cases for criminal investigation, with the result that only a small number of criminal cases have been taken so far and several are still under way.81 The completed case was a follow-on action from an overseas-initiated investigation with a corporate defendant that cooperated and pleaded guilty, and even this case took five years to resolve. This very slow start of criminal prosecutions until 2017 betrays the timorous attitude of the ACCC and the CDPP in taking ‘difficult’ cartel prosecutions; it appears that neither agency wishes to risk defeat in a contested case before a jury. This is understandable, both from a resource allocation perspective and from a public relations standpoint. However, it underlines the difficulties of getting a criminal cartel prosecution system up and running in a new jurisdiction that does not have the advantages of a long history of cultural acclimatization to the moral wickedness of cartels, does not have a unified antitrust body for both criminal and civil enforcement, and, crucially, does not have a judicially binding plea-bargaining system whereby a personal defendant can be induced to plead guilty because they can be certain in advance of exactly what punishment they will receive. In Australia, the best that the ACCC or the CDPP can offer in a criminal prosecution is that they will ‘suggest’ a tariff to the sentencing judge but, beyond that, they can offer no certainty to the criminal defendant – unlike the position in the US. Given these impediments, it seems unlikely that criminal prosecutions of individuals in Australia will significantly increase soon. Finally, in respect of China, similarly to the EU, China’s AML remedies are limited to civil financial penalties and injunctive remedies solely against business operators.82 There is no provision in the AML that creates individual civil or criminal liability for cartel conduct. Consequently, in the first decade of the AML’s operation, no corporate managers have been held to account personally for forming or operating cartels in China.

V CONCLUSION This chapter has sought to survey the landscape of anti-cartel law in four jurisdictions – the US, the EU, Australia and China – and to compare the search in these jurisdictions for appropriate legislative provisions to prohibit cartels and to provide a system of penalties and other 79 See the OECD report comparing cartel penalties: OECD, Pecuniary Penalties for Competition Law Infringements in Australia (26 March 2018), http://​www​.oecd​.org/​fr/​concurrence/​pecuniary​-penalties​ -competition​-law​-infringements​-australia​-2018​.htm. 80 Memorandum of Understanding between the Commonwealth Director of Public Prosecutions and the ACCC (n 36). 81 CDPP v Country Care; CDPP v CFMMEU; CDPP v ANZ, Deutsche Bank, Citigroup, and others. 82 See AML 2008, Arts 47, 48.

Cartel prohibition and the search for deterrent penalties  137 remedies that will most appropriately punish and deter the formation and operation of cartels. The varying approaches to the substantive prohibitions show some diversity in approach, but that is not due to fundamentally different views of cartel conduct and the need for prohibition. Rather, it reflects the idiosyncrasies of the legal order of the jurisdiction, its drafting style, and other environmental factors. Where penalty regimes are concerned, this chapter has argued that while all are agreed that increasingly harsh penalties against cartels are needed, so too is a substantial increase in the rate of detection facilitated by tools such as leniency programmes. The 1998 OECD recommendation on toughening cartel penalties has generally led to increased civil corporate fines in each of the jurisdictions and to the adoption of criminalization in Australia – though not in the EU and China, for different reasons. However, the OECD recommendation did not only endorse higher civil corporate fines and criminalization. It states that ‘Member countries should ensure that their competition laws effectively halt and deter hard core cartels. In particular, their laws should provide for … effective sanctions, of a kind and at a level adequate to deter firms and individuals from participating in such cartels …’83 Consequently, this chapter argues that effective cartel sanctions ‘adequate to deter individuals and firms’ mean that criminalization in and of itself is insufficient to achieve the objective of adequate deterrence. The case of the US is unique for the reasons explained above in the effectiveness of its civil and criminal regime. Those features are difficult or impossible for other jurisdictions to replicate. This chapter has argued that the adoption and use of individual civil sanctions – fines, director disqualification and other appropriate measures – are a vital component in any effective anti-cartel remedy regime. Moreover, they avoid the difficulties of reliance on individual criminal sanctions that are constitutionally impossible in the EU, have proved ineffective in Australia (so far), and do not appear to be on the political agenda in China. This chapter advocates that the EU and China should adopt individual civil sanctions and that Australia should make much more extensive use of the impressive armoury of individual sanctions already available to the ACCC and the courts, in preference to relying on the more difficult to prove, and more unpredictable, outcomes of jury trials in defended criminal proceedings. If these recommendations were to be accepted, cartel deterrence would be enhanced in the EU, China and Australia, and consumers would reap the benefits.

83 OECD Council Recommendation Concerning Effective Action against Hard Core Cartels (25 March 1998) (emphasis added), http://​www​.oecd​.org/​daf/​competition/​2350130​.pdf.

7. Algorithm-driven collusive conduct Rob Nicholls

I INTRODUCTION Coordinated conduct, including price-fixing and other cartel-like behaviour, has traditionally been implemented by individuals in their own right or on behalf of corporations. Developments in the digital environment have created the potential for anticompetitive conduct to be facilitated by data-driven technologies. One potential facilitator is the use of algorithms, which are processes or rules to be followed in computerized problem-solving or calculations. Algorithmic-driven collusive conduct represents a new challenge to competition authorities. At the outset, an important issue for resolution in the analysis of the effects of algorithms on coordinated conduct is one of vocabulary. Similar terms with distinct but different meanings are used in relation to algorithms in computer science, economics and competition law. For example, the term ‘agent’ has a specific meaning in computer science which may be similar to the same term used in economics, particularly in the context of agent-based modelling. However, this is distinct from the concept of ‘agent’ in the legal context of principal and agent. In computer science and economics agent-based modelling, the agent is an autonomous entity that operates with bounded rationality. This is quite a useful approach to characterizing an agent, especially in the context of agents where actions are determined by artificial intelligence. Another way of dealing with the issue of conflicting definitions is to refer to agents once implemented as bots. This is short for robot. These bots operate in a way that is predetermined by the bounded rational instructions provided to it or, potentially, where the response of the bot varies over time in accordance with a changing set of rules. That is, the bot is either pre-programmed and remains with a set of program instructions, or is able to learn and to respond to learned behaviour. From a competition law and policy perspective, algorithmic-driven businesses have a risk of engaging in parallel conduct. Parallelism is not a problem per se. However, it can be an issue if the sector in question exhibits collusive oligopoly characteristics. There are three ways in which this could occur. The first is simple cartel conduct – when an algorithm is designed to facilitate collusion. The second occurs indirectly – that is, when an algorithm is not designed to facilitate collusion, but it is reasonable to assume that this will be a likely outcome. The third is accidental and occurs when an algorithm is not designed to facilitate collusion, but its implementation unpredictably leads to this outcome. In order to assess these risks, the chapter uses a typology created by Ariel Ezrachi and Maurice Stucke.1 This chapter explores collusion driven by algorithmic-driven conduct. It is important to note that the potential collusion is not purely horizontal. In a competition law world where cartel behaviour is the most heinous conduct, horizontal issues dominate. A key example of such



1

Ariel Ezrachi and Maurice E Stucke, Virtual Competition (Oxford University Press 2016).

138

Algorithm-driven collusive conduct  139 collusion is price-fixing. However, there is also vertical price-fixing collusion. The expression ‘vertical price-fixing’ is used in the EU. In other jurisdictions, such as Australia and the US, the term is ‘resale price maintenance’. Regardless of the descriptor, this is another form of collusion that can be facilitated by algorithmic-driven conduct. Having mentioned that there are both vertical and horizontal forms of collusion that can be algorithmically driven, clearly there is also the potential for some form of hybrid collusion based on both vertical and horizontal conduct. A well-documented example of this is a ‘hub-and-spoke’ conspiracy. In a simple form of hub-and-spoke conspiracy, a single information source can facilitate horizontal collusion, or the hub can be used to facilitate vertical collusion. For example, a trade association could act as an information hub, allowing members of that association to choose a single selling price without having to coordinate between themselves. However, horizontal, vertical and hybrid collusion do not require algorithmic-driven conduct. They might be facilitated by such conduct, but they are not necessarily driven by it. That is, algorithms can be used to facilitate conduct implemented by traditional communications such as a contract, arrangement or understanding, or they can drive that conduct. In particular, algorithms could be used to improve cartel stability in a traditional arrangement.2 Business applications of machine learning and artificial intelligence are increasing. The concept of sampling data (particularly consumer data) is increasingly being replaced by analysis of all of the available data. The reason for this change is that the cost of data capture, processing and analysis means that data sampling is no longer significantly cheaper than the analysis of the whole of the data population. This gives rise to the term ‘big data’.3 Simply collecting data does not provide the basis for a business. Instead, the data needs to be analysed and the mechanism by which that analysis occurs is the application of algorithms. Broadly, algorithms fall into two classes. The first class consists of those algorithms that transmit and exchange information. The second class comprises those algorithms that analyse data. Of course, in practice that analysis is not the final step in the process. What happens next is that the outputs of the analysis are used as the basis of some action and that action is usually automated. The form of that automation is usually algorithmic. For example, data analysis of the online shopping characteristics of a buyer can be used to recommend goods that are likely to appeal to that buyer. 2 JD Jaspers, ‘Managing Cartels: How Cartel Participants Create Stability in the Absence of Law’ (2017) 23 European Journal on Criminal Policy and Research 319 provides a review of the literature and draws a distinction between the economic and legal analyses of cartel stability. 3 There is a significant and growing body of literature on big data in the context of competition law and policy. Some recent examples providing useful insights include Simonetta Vezzoso, ‘Competition Policy in a World of Big Data’ in F Xavier Olleros and Majlinda Zhegu (eds), Research Handbook on Digital Transformations (Edward Elgar Publishing 2016); Dennis D Hirsch, ‘That’s Unfair! Or Is It? Big Data, Discrimination and the FTC’s Unfairness Authority’ (2015) 103 Kentucky Law Journal 345; Christopher Townley, Eric Morrison and Karen Yeung, ‘Big Data and Personalised Price Discrimination in EU Competition Law’ (2017) 36 Yearbook of European Law 683–748; Paul Lugard and Lee Roach, ‘The Era of “Big Data” and EU/U.S. Divergence for Refusals to Deal’ (2017) 31 Antitrust 58; Ania Thiemann and Pedro Gonzaga, ‘Big Data: Bringing Competition Policy to the Digital Era. A Report for the OECD’ (DAF/COMP(2016)14, OECD 2016); Autorité de la concurrence and Bundeskartellamt, ‘Competition Law and Data’ (2016); Canadian Competition Bureau, ‘Big Data and Innovation: Implications for Competition Policy in Canada: Draft Discussion Paper’ (2017); Harlan Kroll et al, ‘Accountable Algorithms’ (2017) 165(3) University of Pennsylvania Law Review 633.

140  Research handbook on methods and models of competition law It is easy to see simple applications of such analysis. Consider a merchant who is supplying consumer goods using Amazon Marketplace. The merchant might decide that he or she wants to have the most aggressive pricing but still make a sensible margin. The approach here would be to collect information and then analyse it with a simple pair of rules. The first rule would be that the price must always be above the marginal cost. The second rule would be that the price must be below the lowest price offered by a competitor provided that the first rule is met. This pair of rules could be implemented very readily, and the effect of the rules is likely to be that a competitive outcome is achieved and, if all merchants applied the same rules, commodity pricing would be the outcome. However, a variation of the rules would lead to the perverse outcome demonstrated in the pricing of the book The Making of a Fly on Amazon.4 As has been well documented, in this case there were two merchants. One wanted a price premium over the lowest price offered and the other wanted a discount to the highest price offered. It did not take long for one of the merchants to offer a $23 million book. This was an extreme outcome driven by poorly designed application of algorithmic conduct. The two merchants did not collude to drive up prices and, consequently, there was no cartel conduct. However, the effect was certainly not consumer welfare maximizing. This chapter reviews the issues that flow from the application of algorithms to business practice. It does so by examining the characteristics of algorithmic coordination. It then moves on to the complications that flow from platform businesses that are algorithm based. It analyses the potential for collusive conduct issues associated with algorithmic-driven business and the interaction between competition law and algorithmic-driven conduct. Finally, a competition law analysis looks at some of the limits of the law using the US, Europe and Australia as examples of implementation, and conclusions are drawn.

II

ALGORITHMIC COORDINATION

One of the important aspects of using big data in real time is the way in which businesses, including platform operators, respond to information that is available from other businesses. Typically, this response is algorithmically driven. Control of an algorithm would normally be expected to be in the hands of the operator. However, the operator may decide to automate this process. Such automation might use information collection agents with boundedly rational instructions. The programmed bounded rationality has the potential to lead to the information collection agents acting in a way that responds to information signalling but is not anticipated or controlled by the operator. Effects associated with complex adaptive systems such as ‘emergence’ of unpredicted patterns are possible.5 The term ‘agent’ in this context is an automatic process (sometimes known as a ‘bot’). The distinction between a robotic agent and an agent

4 Olivia Solon, ‘How a Book about Flies Came to Be Priced $24 Million on Amazon’, Wired Magazine (27 April 2011), https://​www​.wired​.com/​2011/​04/​amazon​-flies​-24​-million/​. 5 This has been examined in the context of the financial system: Richard Bookstaber, ‘Using Agent-Based Models for Analyzing Threats to Financial Stability’ (Office of Financial Research, US Department of Treasury, Working Paper 0003, 21 December 2012); Richard Bookstaber, The End of Theory: Financial Crises, the Failure of Economics, and the Sweep of Human Interaction (Princeton University Press 2017).

Algorithm-driven collusive conduct  141 with a human principal is important and the balance of the chapter uses the term ‘bot’ when referring to an automated agent. A

Potential Threat

The potential threat posed by algorithmic coordination has been considered by a number of international organizations. The European Competition Commissioner has raised the issue in respect of regulators.6 The Organisation for Economic Co-operation and Development (OECD) held a roundtable on ‘Algorithms and Collusion’ in June 2017.7 In addition to the background briefing from the Secretariat8 and a commissioned article from Ezrachi and Stucke,9 there were contributions from the EU,10 Italy,11 the Russian Federation,12 Singapore,13 Ukraine,14 the UK,15 the US16 and the Business and Industry Advisory Committee (BIAC) of the OECD.17 The OECD has also published a post-roundtable report,18 as well as making available a helpful bibliography prepared by the European Commission.19 Following the OECD, the International Competition Network has also considered some of the issues.20 In addition to

6 Margrethe Vestager, ‘Algorithms and Competition’ (Speech at the Bundeskartellamt 18th Conference on Competition, Berlin, 16 March 2017, European Commission 2017), https://​ec​.europa​ .eu/​commission/​commissioners/​2014​-2019/​vestager/​announcements/​bundeskartellamt​-18th​-conference​ -competition​-berlin​-16​-march​-2017​_en. 7 OECD, ‘Algorithms and Collusion – OECD’ (2017), http://​www​.oecd​.org/​competition/​algorithms​ -and​-collusion​.htm. 8 OECD, ‘OECD Directorate for Financial and Enterprise Affairs Competition Committee: Algorithms and Collusion – Background Note by the Secretariat’ (2017). 9 Ariel Ezrachi and Maurice E Stucke, ‘OECD Directorate for Financial and Enterprise Affairs Competition Committee: Algorithms and Collusion – Note by Ariel Ezrachi & Maurice E. Stucke’ (2017). 10 European Union, ‘OECD Directorate for Financial and Enterprise Affairs Competition Committee: Algorithms and Collusion – Note from the European Union’ (2017). 11 Italy, ‘OECD Directorate for Financial and Enterprise Affairs Competition Committee: Algorithms and Collusion – Note from Italy’ (2017). 12 Russian Federation, ‘OECD Directorate for Financial and Enterprise Affairs Competition Committee: Algorithms and Collusion – Note by the Russian Federation’ (2017). 13 Singapore, ‘OECD Directorate for Financial and Enterprise Affairs Competition Committee: Algorithms and Collusion – Note from Singapore’ (2017). 14 Ukraine, ‘OECD Directorate for Financial and Enterprise Affairs Competition Committee: Algorithms and Collusion – Note by Ukraine’ (2017). 15 United Kingdom, ‘OECD Directorate for Financial and Enterprise Affairs Competition Committee: Algorithms and Collusion – Note from the United Kingdom’ (2017). 16 United States, ‘OECD Directorate for Financial and Enterprise Affairs Competition Committee: Algorithms and Collusion – Note by the United States’ (2017). 17 BIAC, ‘OECD Directorate for Financial and Enterprise Affairs Competition Committee: Algorithms and Collusion – Note from BIAC’ (2017). 18 OECD, Algorithms and Collusion: Competition Policy in the Digital Age (2017). 19 Cyril Ritter, ‘Bibliography on Antitrust and Algorithms’ (7 June 2017). Available at SSRN: https://​ ssrn​.com/​abstract​=​2982397. 20 Chris Jenkins, ‘Algorithms, Personalised Pricing and Coordination’ (2019), http://​ ec​ .europa​ .eu/​competition/​cartels/​icn/​index​_en​.html; Jehanne Richet, ‘Some Economic Considerations about Algorithms and Tacit Collusion’ (2019); Szilvia Szekely, ‘Pricing Algorithms and EU Competition Law’ (2019); Randall Hofley and Anita Banicevic, ‘Algorithms and Canadian Cartel Law Enforcement’ (2019); Antonio Capobianco, ‘Digital Cartels and Algorithms’ (2019).

142  Research handbook on methods and models of competition law Table 7.1

Ezrachi and Stucke typology of algorithmic coordination

Scenario

Description

Messenger

Humans agree to collude and use computers to execute their will.

Hub-and-spoke

The use of a single pricing algorithm to determine the market price charged by numerous users.

Predictable agent

A world where pricing algorithms act as predictable agents and continually monitor and adjust to each other’s prices and market data.

Digital eye

Computers, in learning by doing, determine independently the means to optimize profit.

the internationally coordinated regulatory work, there have been many commentaries.21 The level of concern about algorithmic collusion appears to be lower in the US than in the EU.22 Some of this is based on the assertion by regulators that the use of an algorithm to coordinate conduct in a market is covered by the present law.23 However, this is far from a settled argument as this chapter demonstrates. The primary issue for examination is whether there is a business model for the use of algorithms in setting prices. There are obvious and readily appreciable examples from business that are based on algorithmic conduct in the sharing economy (for example, Uber and Airbnb) and social media (for example, Facebook and YouTube). There are also examples from relatively small businesses. Work at Northeastern University has demonstrated that merchants on Amazon Marketplace using algorithmic pricing have greater sales.24 Determining the nature of algorithmic collusion and the rationale of why it is important will be considered next. One way of doing this is to apply the analytical framework offered by Ezrachi and Stucke25 and set out in Table 7.1. The effect of this approach to analysis is that it adds one additional element to the potential sources of collusion set out in the Introduction – that is, the enhancement to parallelism offered by predictable agents that leads to ‘tacit collusion on steroids’.26 This element has been partially challenged, with Ittoo and Petit arguing that the ‘algorithmic tacit collusion conjecture should not presently be taken as a given’.27 However, the same authors concede ‘that regulators may 21 See, for example, the 2017 Spring Volume 2 of the Antitrust Chronicle, entitled ‘The Algorithms Have Landed’; Nicolas Petit, ‘Antitrust and Artificial Intelligence: A Research Agenda’ (2017) 8 Journal of European Competition Law & Practice 361; Ezrachi and Stucke, Virtual Competition (n 1); Michal S Gal, ‘Algorithmic-Facilitated Coordination: Market And Legal Solutions’ (2017) 2 CPI Antitrust Chronicle 22. 22 United States (n 16). 23 Canadian Competition Bureau (n 3); Australian Competition and Consumer Commission, ‘The ACCC’s Approach to Colluding Robots: Speech by Rod Sims’ (2017), https://​www​.accc​.gov​.au/​speech/​ the​-accc’s​-approach​-to​-colluding​-robots; Maureen K Ohlhausen, ‘Should We Fear the Things That Go Beep in the Night? Some Initial Thoughts on the Intersection of Antitrust Law and Algorithmic Pricing’ (Remarks from the Concurrences Antitrust in the Financial Sector Conference, New York, 23 May 2017, Federal Trade Commission 2017), https://​www​.ftc​.gov/​public​-statements/​2017/​05/​should​-we​-fear​ -things​-go​-beep​-night​-some​-initial​-thoughts​-intersection; Vestager (n 6). 24 Le Chen, Alan Mislove and Christo Wilson, ‘An Empirical Analysis of Algorithmic Pricing on Amazon Marketplace’ (Proceedings of the 25th International World Wide Web Conference, Northeastern University 2016). 25 Ezrachi and Stucke, Virtual Competition (n 1), 36. 26 Ibid. 27 Ashwin Ittoo and Nicolas Petit, ‘Algorithmic Pricing Agents and Tacit Collusion: A Technological Perspective’ in Hervé Jacquemin and Alexandre de Streel (eds), L’intelligence artificielle et le droit (Larcier 2017), 241–56.

Algorithm-driven collusive conduct  143 be right to be vigilant’.28 Deng also calls for regulator attention, while arguing that algorithmic collusion has not yet occurred.29 In this context, it is interesting to note that economic analysis suggests that algorithmic collusion is an equilibrium outcome in some situations.30 It is helpful to consider how the different types of conduct could occur. The main rationale for this approach is to assess whether the scenarios proposed are distinct from the ‘human’ means of engaging in prohibited collusive conduct, or just another type of human intervention. One way of perceiving the tools that compare prices is to think about a price comparison website. In order to be able to compare prices, the price comparison website needs to be able to collect prices from potential suppliers. In principle, the price comparison website could simply ‘scrape’ prices from other websites in order to be able to acquire the data that it needs. In practice, it is more useful to have an autonomous agent continually monitoring websites for price changes. This is likely to be facilitated by the increasing use of an application programming interface (API). This provides the ‘hooks’ by which application software can call for information. In many cases, these data sets can be ‘live’. APIs tend to be used for business-to-business transactions. This contrasts with the business-to-consumer experience, where consumers will typically look at web-based outputs rather than using machines to review data. The benefit to using an API is that the data underlying the application can be accessed readily. However, an API also changes the way in which data providers decide whether to provide an application programming interface or not. If a data provider decides that it will provide an API, there is a choice between whether that interface is presented publicly or privately. In either case, the API needs to have mechanisms by which authentication of a user can be determined. If access to the data is restricted in any way, an authorization mechanism will be required. Clearly, there is the potential that an API might be made available to facilitate collusion and that authorization could limit access to colluding parties to the exclusion of others. Indeed, such exclusion could be used as a punishment for cartel defection.31 Marshall and Marx have proposed a number of ways in which cartel enforcement can be optimized by cartelists.32 More recently, Marshall has proposed the use of external consultants (as evidenced in the EU case against AC-Treuhand) as a cartel stability mechanism.33 Consistency of prices using an API is another solution. This distinction between authentication and authorization is critical to understanding the scope and application of an API. In particular, it may well be that only authentication is required for public APIs. That is, the determination that any specific application is allowed to interface the API is on the basis that it has undergone some form of pre-authentication. This is because it is necessary to ensure that the user uses the data in a way that the underlying

Ibid, 1. Ai Deng, ‘When Machines Learn to Collude: Lessons from a Recent Research Study on Artificial Intelligence’ (2017). 30 Timo Klein, ‘056/VII Tinbergen Institute Discussion Paper Autonomous Algorithmic Collusion: Q-Learning under Sequential Pricing’ (30 August 2017). Available at SSRN: https://​ssrn​.com/​abstract​=​ 3029662. 31 For a discussion on retaliation, see Robert C Marshall and Leslie M Marx, The Economics of Collusion: Cartels and Bidding Rings (MIT Press 2012) 54. 32 Ibid. 33 Robert C Marshall, ‘Unobserved Collusion: Warning Signs and Concerns’ (2017) 5(3) Journal of Antitrust Enforcement 329; Lukas Solek, ‘Passive Participation in Anticompetitive Agreements’ (2017) 8(1) Journal of European Competition Law & Practice 15. 28 29

144  Research handbook on methods and models of competition law data provider deems appropriate. For example, if there is a public API to data that includes personally identifying information, only applications that will use that information in a way consistent with appropriate local privacy policy will have the relevant authentication to access the data. Authorization and authentication must both occur in the case of a private API. An inefficient alternative is data providers engaging in a process known as ‘screen scraping’. Essentially, an intermediary can take information from a web page designed for consumers to create a data set, to which they offer an ad hoc API. The problem with this type of arrangement is that the underlying provider of the data firstly may not authorize screen scraping and secondly may well change the way that the data are presented so that the screen-scraping intermediary needs to rearrange its data and potentially its API from time to time. The bot needs to have a set of rules or instructions as to what to do with prices it obtains and how to monitor changes. It makes sense to have this agent implemented with a degree of intelligence and a limited instruction set. That is, the bots will be boundedly rational. In order to ensure that these bots are efficient, they will be driven by an algorithmic approach. Initially, such algorithms are likely to be based on efficient search and data capture. However, the efficiency of algorithms can be improved either by incremental improvements or by radical changes. In developing a typology for innovation in the ICT sector, Bauer and Shim34 draw a distinction between radical innovation (which implies ‘that many aspects of a process, a product, or of the competencies of participants in the innovation system are affected’35) and incremental innovation. They argue that innovation can be divided into modular innovations (which have low levels of coordination between the modular innovation and the remainder of the system) and coupled innovations (which are highly integrated and require technical and economic connection between multiple layers of the system). In this analysis, the form of innovation that captures supernormal profits is radical and modular. The approach to the creation of bots is based on machine learning, that is, the mechanisms by which an algorithm can be developed. In general, machine learning has three elements: ●● supervised learning; ●● unsupervised learning; and ●● reinforcement. Supervised learning uses sample data to train an algorithm and test data to find out how successful the training has been by running the algorithm. Unsupervised learning directs a machine to find patterns. Reinforcement learning is the adaptation of an algorithm in response to data found in the wild. One way of developing algorithms that have effective supervised learning characteristics is to use evolutionary algorithms.36 There are two issues that flow from the use of algorithmic-driven bots for price discovery. The first is that the operation of the bot is generally independent of the business that created it. The second is that the operation of the bot may not be understood by the business that created it, especially if evolutionary approaches were used. One example of the way in which the risk of accidental collusion could occur arises 34 Johannes M Bauer and Woohyun Shim, Regulation and Innovation Behavior in Telecommunications (Michigan State University 2012). 35 Ibid, 9. 36 See, for example, Melanie Mitchell, An Introduction to Genetic Algorithms (MIT Press 1996); Melanie Mitchell, ‘Complex Systems: Network Thinking’ (2006) 170 Artificial Intelligence 1194.

Algorithm-driven collusive conduct  145 in the context of algorithms, where the mechanism by which the algorithm functions is not understood – that is, where the algorithm has become a ‘black box’ to the business that created it. This might seem paradoxical when significant effort is invested in improving algorithmic efficiency. However, this is precisely the rationale behind evolutionary algorithms.37 In an evolutionary or genetic algorithm, there is a set of candidate solutions to a problem. These are applied to a function to create a set of potential solutions. Some solutions are better than others. The test is similar to the evolutionary test for biological fitness. The process starts with known candidate solutions. These are changed through manipulations inspired by biological processes including point mutation (random variations of some parts of the algorithm) and crossing over (generating new algorithms by merging parts of existing ones). Then, the fitness of the new pool of candidates is computed and the algorithms with the highest fitness are allowed to survive in the next generation. The process is iterated until only solutions with large fitness survive. The effect is that the algorithmic solution becomes a black box. The knowledge of its success is the only parameter that is certain. The source code for the algorithm is not useful to determine its function. It would be hard to show that the application of the algorithm met fiduciary obligations or otherwise as it is not clear that a director or officer could be certain that the algorithm was or was not acting in accordance with those obligations. That is, the source code for a black box algorithm does not reveal how that algorithm operates but merely that it has instructions to operate. As reported by Harrington,38 the first successful coordination in repeated ‘prisoner’s dilemma’ was solved using an evolutionary or genetic algorithm.39 In this case, the solution was also a black box algorithm. The ethical problems that bot trading creates in the financial services sector are reasonably well understood.40 The application to antitrust law is a little more recent.41 The issue is made more complex by the potential for the complex adaptive systems effect of ‘emergence’ as mentioned above. This effect is the sign of complex outcomes from a population of boundedly rational actors.42 Analysis of these effects has been applied to both financial regulation43 and antitrust.44 The problem arises when two bots determine that it is more efficient to swap prices 37 Ugo Pagallo, The Laws of Robots: Crimes, Contracts, and Torts (Springer 2013) Vol 10; Samir Chopra and Laurence F White, A Legal Theory for Autonomous Artificial Agents (University of Michigan Press 2011). 38 Joseph E Harrington, ‘Developing Competition Law for Collusion by Autonomous Artificial Agents’ (2018) 14(3) Journal of Competition Law & Economics 331–63. 39 John H Miller, ‘The Coevolution of Automata in the Repeated Prisoner’s Dilemma’ (1996) 29 Journal of Economic Behavior & Organization 87; Philip Hingston and Graham Kendall, ‘Learning versus Evolution in Iterated Prisoner’s Dilemma’ (Proceedings of the Congress on Evolutionary Computation, IEEE Cat No 04TH8753, 2004). 40 See, for example, Michael P Wellman and Uday Rajan, ‘Ethical Issues for Autonomous Trading Agents’ (2017) 27(4) Minds and Machines 1. 41 Ezrachi and Stucke, Virtual Competition (n 1), 34. 42 See, for example, Melanie Mitchell, Complexity: A Guided Tour (Oxford University Press 2009). 43 Bookstaber, The End of Theory (n 5); Prasanna Gai, Andrew Haldane and Sujit Kapadia, ‘Complexity, Concentration and Contagion’ (2011) 58 Journal of Monetary Economics 453; Andrew Haldane, Rethinking the Financial Network (Bank of England 2009); Sreekala Kochugovindan and Nicolaas J Vriend, ‘Is the Study of Complex Adaptive Systems Going to Solve the Mystery of Adam Smith’s “Invisible Hand”?’ (1998) 3 Independent Review 53. 44 Avishalom Tor, ‘Understanding Behavioral Antitrust’ (2014) 92 Texas Law Review 573; Avishalom Tor, ‘Boundedly Rational Entrepreneurs and Antitrust’ (2016) 61 Antitrust Bulletin 520.

146  Research handbook on methods and models of competition law than to separately acquire those prices. The logical extension of this efficiency arises when the bots agree that they will both have the same price to increase efficiency. That is, the bot engages in the cartel conduct of price-fixing. The criticism of this potential type of collusion is that there may be no opportunity for bots to exchange information to create this reinforced learning. However, there has been economic analysis showing that such collusion occurs ‘in the wild’.45 Ariel Ezrachi and Maurice Stucke consider the ways in which algorithmic tacit collusion might operate.46 They argue that ‘one would expect it in markets with several important characteristics’:47 ‘Algorithmic tacit collusion would likely arise in concentrated markets involving homogenous products where the algorithms can monitor to a sufficient degree the pricing and other keys terms of sale … and once deviation is detected, a credible deterrent mechanism exists.’ Ezrachi and Stucke argue that the third characteristic48 is the one described in the EU Merger Guidelines that ‘the reactions of outsiders, such as current and future competitors not participating in the coordination, as well as customers, should not be able to jeopardise the results expected from the coordination’.49 In effect, the argument is that the most likely market conditions for algorithmic tacit collusion are similar to those in which human collusion has occurred in the past. The human style collusion has the potential to occur between bots and this is the ‘digital eye’ scenario set out in Table 7.1. The concept here is that boundedly rational bots with a shareholder wealth-maximizing instruction set could meet in cyberspace and collude to fix prices. As has been shown, this scenario has been rejected by some. However, there is certainly a risk of unintended collusion. Part of the issue that Ezrachi and Stucke identify with this scenario is that it may not be detected by the principals of the bots, let alone a competition regulator. However, as the next section demonstrates, there is a higher potential for simple collusion in multisided markets.

III

PLATFORM COMPLICATIONS

A

Overview of Platforms

A platform is a market with more than one ‘side’. There is an extensive literature on two-sided or double-sided markets.50 One ‘bricks-and-mortar’ example of a two-sided market is the 45 Emilio Calvano, Giacomo Calzolari, Vincenzo Denicolò and Sergio Pastorello, ‘Artificial Intelligence, Algorithmic Pricing and Collusion’ (2019) 55 Review of Industrial Organization 1; Matthew Kassel, ‘Beware Algorithms That Could Collude on Prices; In a Study, Two Pricing Algorithms Learned on Their Own to Raise Prices Together to Unfairly High Levels’, Wall Street Journal (online) (1 April 2019). 46 Ariel Ezrachi and Maurice E Stucke, ‘Two Artificial Neural Networks Meet in an Online Hub and Change the Future (of Competition, Market Dynamics and Society)’ (1 July 2017) Oxford Legal Studies Research Paper No 24/2017; University of Tennessee Legal Studies Research Paper No 323. Available at SSRN: https://​ssrn​.com/​abstract​=​2949434; Ezrachi and Stucke, Virtual Competition (n 1). 47 Ezrachi and Stucke, Virtual Competition (n 1), 3. 48 Ibid, 4. 49 EC Merger Guidelines, para 41. 50 See, for example, Daniel F Spulber, Market Microstructure: Intermediaries and the Theory of the Firm (Cambridge University Press 1999) ch 3; Bernard Caillaud and Bruno Jullien, ‘Chicken & Egg:

Algorithm-driven collusive conduct  147 shopping mall.51 Here, the agents on one side are shoppers. These consumers benefit from the aggregation of major stores and supermarkets with a range of speciality stores and food outlets in a single place. The agents on the other side are the stores. The store owners benefit from the aggregation of a range of shoppers. The intermediary or platform is the shopping mall operator. It extracts occupancy fees from the store owners that reflect the value of the shopper aggregation and the value of the infrastructure of the mall beyond the floor area of the store. In general, the shopping mall operator does not transact with the shopper side of the market except to provide information. That is, the mall operator does not run a store. Rysman52 offers a simple but clear definition of a two-sided market: ‘Broadly speaking, a two-sided market is one in which 1) two sets of agents interact through an intermediary or platform, and 2) the decisions of each set of agents affects the outcomes of the other set of agents, typically through an externality.’ B

Types of Two-sided Market and Competition Issues

The shopping mall example is one form of a two-sided market and was chosen as a comparison because it has physical form rather than logical form. A useful distinction between types of two-sided markets is offered by Filistrucchi et al.53 This separates two-sided markets into two-sided transaction markets and two-sided non-transaction markets. The potential competition law issues associated with platforms have been recognized by the OECD,54 the UK House of Lords,55 the French and German competition authorities,56 the

Competition among Intermediation Service Providers’ (2003) 34 RAND Journal of Economics 309; Mark Armstrong, ‘Competition in Two-Sided Markets’ (2006) 37 RAND Journal of Economics 668; Jean-Charles Rochet and Jean Tirole, ‘Platform Competition in Two-Sided Markets’ (2003) 1 Journal of the European Economic Association 990; Jean-Charles Rochet and Jean Tirole, ‘Two-Sided Markets: A Progress Report’ (2006) 37 RAND Journal of Economics 645; William A Brock, ‘Contestable Markets and the Theory of Industry Structure: A Review Article’ (1983) 91 Journal of Political Economy 1055. 51 Thomas Eisenmann, Geoffrey Parker and Marshall W Van Alstyne, ‘Strategies for Two-Sided Markets’ (2006) 84 Harvard Business Review 92; Andrei Hagiu and Julian Wright, ‘Multi-Sided Platforms’ (2015) 43 International Journal of Industrial Organization 162; Johan Frishammar, Javier Cenamor, Harald Cavalli-Björkman, Emma Hernell and Johan Carlsson, ‘Digital Strategies for Two-Sided Markets: A Case Study of Shopping Malls’ (2018) 108 Decision Support Systems 34. 52 Marc Rysman, ‘The Economics of Two-Sided Markets’ (2009) 23 Journal of Economic Perspectives 125, 125. 53 Lapo Filistrucchi, Damien Geradin, Eric Van Damme and Pauline Affeldt, ‘Market Definition in Two-Sided Markets: Theory and Practice’ (2014) 10(2) Journal of Competition Law and Economics 293. 54 Ania Thiemann and Pedro Gonzaga, ‘Big Data: Bringing Competition Policy to the Digital Era. A Report for the OECD’ (2016), https://​one​.oecd​.org/​document/​DAF/​COMP(2016)14/​en/​pdf. 55 House of Lords, ‘Online Platforms and the Digital Single Market’ (2016), https://​publications​ .parliament​.uk/​pa/​ld201516/​ldselect/​ldeucom/​129/​129​.pdf. 56 Autorité de la concurrence and Bundeskartellamt, ‘Competition Law and Data’ (2016), http://​ www​.autoritedelaconcurrence​.fr/​doc/​reportcompe​titionlawanddatafinal​.pdf.

148  Research handbook on methods and models of competition law Dutch competition authority,57 the European Commission58 and the EU.59 These issues are solely associated with electronic, rather than physical, platforms. One important element is the transparency (or opacity) of the operation of a platform. Another element is based on the concept of ‘tipping’. In a competition law context, ‘tipping’ occurs when a competitive market is replaced by a market with monopoly characteristics as a result either of a merger or of the exit of players.60 In the platform space, the tipping phenomenon, as the ‘winner-takes-all’ or ‘winner-takes-most’ characteristic, is clear from the significant market share of each of Facebook in social media advertising and Google in search advertising.61 C

Amplification of Algorithmic Hazards

The most important characteristic of multisided markets from an antitrust perspective is that they are not simply a set of transactions with players on each side. Multisided market literature demonstrates that multisided markets have network externalities and that benefits accrue to each side of the market.62 The effect of this is the need for competition authorities to deal with multisided markets holistically and not as a series of conventional markets. This means that the usual competition law measurement and enforcement tools may not be appropriate to multisided markets and this has been the focus of substantial research. Much of 57 Netherlands Authority for Competition and Markets (ACM), ‘Taking a Closer Look at Online Video Platforms’ (2016), https://​www​.acm​.nl/​en/​publications/​publication/​16342/​Taking​-a​-closer​-look​ -at​-online​-video​-platforms/​. 58 European Commission, ‘Online Platforms and the Digital Single Market Opportunities and Challenges for Europe’ (2016), http://​eur​-lex​.europa​.eu/​legal​-content/​EN/​TXT/​PDF/​?uri​=​CELEX:​ 52016DC0288​&​from​=​EN. 59 European Data Protection Supervisor, ‘EDPS Opinion on Coherent Enforcement of Fundamental Rights in the Age of Big Data’ (2016), https://​edps​.europa​.eu/​sites/​edp/​files/​publication/​16​-09​-23​ _bigdata​_opinion​_en​.pdf. 60 David Evans and Richard Schmalensee, ‘The Antitrust Analysis of Multisided Platform Businesses’ in Roger D Blair and D Daniel Sokol (eds), The Oxford Handbook of International Antitrust Economics, Volume 1 (Oxford University Press 2015); Stephan Kreifels, ‘Digital Platforms and Competition Law’ (2016) 5(1) Journal of European Consumer and Market Law 33; Gaurav Jakhu and Payal Malik, ‘Dilemma in Antitrust Enforcement: How Use of Economics Can Guide Enforcement Rules in Multi-Sided Markets’ (2017) 5 Journal of Antitrust Enforcement 260; Marios C Iacovides and Jakob Jeanrond, ‘Overcoming Methodological Challenges in the Application of Competition Law to Digital Platforms – a Swedish Perspective’ (2018) 6(3) Journal of Antitrust Enforcement 437; Ernst Fehr and Klaus M Schmidt, ‘A Theory of Fairness, Competition, and Cooperation’ (1999) 114 Quarterly Journal of Economics 817; OECD, ‘Rethinking Antitrust Tools for Multi-Sided Platforms 2018 – OECD’ (2018). 61 The role of the platform in facilitating winner-takes-all in algorithmic collusion is discussed in Paolo Siciliani, ‘Tackling Algorithmic-Facilitated Tacit Collusion in a Proportionate Way’ (2019) 10 Journal of European Competition Law & Practice 31. 62 See, for example, Eisenmann, Parker and Van Alstyne (n 51); Geoffrey G Parker and Marshall W Van Alstyne, ‘Two-Sided Network Effects: A Theory of Information Product Design’ (2005) 51 Management Science 1494; Jean-Charles Rochet and Jean Tirole, ‘Platform Competition in Two-Sided Markets’ (2003) 1 Journal of the European Economic Association 990; Armstrong (n 50); Rochet and Tirole, ‘Platform Competition in Two-Sided Markets’ (n 50); Rochet and Tirole, ‘Two-Sided Markets: A Progress Report’ (n 50); Hagiu and Wright (n 51); Evans and Schmalensee (n 60); Jean Charles Rochet and Jean Tirole, ‘Tying in Two-Sided Markets and the Honor All Cards Rule’ (2008) 26 International Journal of Industrial Organization 1333; Kalina S Staykova and Jan Damsgaard, ‘A Typology of Multi-Sided Platforms: The Core and the Periphery’ (ECIS 2015 Completed Research Papers, Paper 174, 2015); Filistrucchi et al (n 53).

Algorithm-driven collusive conduct  149 the focus has been on the potential for (nearly perfect) price discrimination at the consumer level.63 However, there is also a significant risk of collusion by suppliers on the upstream sides of the platforms. The antitrust risk on the upstream is partly driven by algorithm-based conduct on the downstream. This was identified by the US Department of Justice in its action against Trod Ltd, in respect of collusive pricing of posters.64 However, this case involved simply the use of an algorithm to give effect to a standard prohibited collusive practice. That is, it was an example of the ‘messenger’ scenario from Table 7.1. There is also an increased risk that each of the ‘hub-and-spoke’ and ‘predictable agent’ scenarios will be used in a platform environment. The ‘hub-and-spoke’ scenario arises when a common algorithm is used by multiple players. In a perfectly competitive world, this appears to be an unlikely outcome as the choice of algorithm would be likely to represent a strategic competitive advantage. However, in a low-margin competitive environment such as eBay, AliExpress or Amazon Marketplace, the use of any algorithmic help is useful. As a result, there is a significant number of firms that offer ‘repricing’ software designed to provide competitive pricing and ensure that their clients’ products are promoted by the platform (for example, in the case of Amazon Marketplace, that the product is in the ‘BuyBox’). Although there are a number of such intermediaries, the practical effect is that only very limited numbers of algorithms are used by many merchants on the most popular platforms. These common tools also lead to the risk of the ‘predictable agent’ scenario. The purpose of the repricing tools is to create a competitive offering from the merchant. Essentially, it also acts as a tool to discover rivals’ costs on the basis that the minimum acceptable margin for merchants is likely to be comparable. The effect of the use of algorithms in a selling space where the products are identical is that the best consumer price is based on the private value of the second-most-competitive seller. This may not be problematic. Efficient auction design aims to get to this price in order to avoid buyer’s remorse.

63 See, for example, Townley, Morrison and Yeung (n 3); Michal S Gal and Niva Elkin-Koren, ‘Algorithmic Consumers’ (2016) 30 Harvard Journal of Law and Technology 309; Maurice E Stucke and Ariel Ezrachi, ‘Alexa et al., What Are You Doing with My Data?’ (2018) 5 Critical Analysis of Law 148; David Lazer, ‘The Rise of the Social Algorithm’ (2015) 348 Science 1090; Michal S Gal, ‘Algorithmic Challenges to Autonomous Choice’ (2018) 25(1) Michigan Technology Law Review 59; Frederik Zuiderveen Borgesius and Joost Poort, ‘Online Price Discrimination and EU Data Privacy Law’ (2017) 40 Journal of Consumer Policy 347; Ezrachi and Stucke, ‘Two Artificial Neural Networks’ (n 46); Ramsi Woodcock, ‘The Power of the Bargaining Robot’ (2017) 2 CPI Antitrust Chronicle 40; Gal (n 21); Bill Baer, Sonia Kuester Pfaffenroth and Vesselina H Musick, ‘Pricing Algorithms: The Antitrust Implications’ (2018). Available at: https://​www​.arnoldporter​.com/​en/​perspectives/​publications/​2018/​04/​ pricing​-algorithms​-the​-antitrust​-implications. 64 Department of Justice, ‘Online Retailer Pleads Guilty for Fixing Prices of Wall Posters’ (Media Release, 11 August 2016), https://​www​.justice​.gov/​opa/​pr/​online​-retailer​-pleads​-guilty​-fixing​-prices​ -wall​-posters.

150  Research handbook on methods and models of competition law D

Algorithmic Amplification of Hub-and-spoke Conspiracies

Falls and Saravia propose a three-pronged taxonomy of hub-and-spoke conspiracies.65 The categories are vertical, horizontal and mixed. These are categorized according to the type of harm caused and/or the type of market power that derives from each situation, as set out below. The first class of conspiracy is the increase of market power or reduction of competition at the level of the vertical party. With this type of hub-and-spoke conspiracy, a distributor typically acts as the hub that facilitates the conduct by instigating agreements with manufacturers to take action against other rival distributors.66 It is within this class that the anticompetitive incentives of each party, particularly the distributors, are highlighted. Not only is the majority of competition removed from the distributor level, but those within the manufacturer level are more likely to be compliant with the distributor’s agreement either to avoid exclusionary action established by the distributor or to ensure a split of the distributor’s supernormal profits.67 While there is a horizontal agreement between the manufacturers, the (vertical) harm occurs at the distributor level (such that other distributors are inhibited from purchasing or on-selling goods provided by the manufacturers). The second class of hub-and-spoke conspiracies differs from its prior counterpart as harm is shifted towards the horizontal party members. With a horizontal hub-and-spoke conspiracy, the vertical player (in this example, the distributor) is less likely to be the instigator or organizer of the conspiracy.68 This is because benefits of a horizontal arrangement will lie with the vertical parties (the retailers). The vertical party acts as facilitator and relays information to form the horizontal agreement between the retailers.69 In this instance, the harm occurs at the horizontal level, with the retail consumers being the group that suffers the greatest from collusive action such as fixed or increased prices across the industry. In a hybrid arrangement, the first and second classes of hub-and-spoke conspiracy operate concurrently, with harm at both the vertical and horizontal levels. It is reasonably easy to see that algorithmic-driven conduct could amplify the effects of a hub-and-spoke conspiracy. The speed of information flows, especially if facilitated by a number of APIs, could improve conspiracy efficiency.

65 As well as the cited article, there is an extensive literature on hub-and-spoke conspiracies. See, for example, Craig G Falls and Celeste C Saravia, ‘Analyzing Incentives and Liability in “Hub-and-Spoke” Conspiracies’ (2015) 19 Distribution 9–23; Gian Luca Zampa and Paolo Buccirossi, ‘Hub and Spoke Practices: Law and Economics of the New Antitrust Frontier’ (2013) 9 Competition Law International 91; Rhonda L Smith and Arlen Duke, ‘Information Exchange, Hub and Spoke Arrangements and Collusion’ (2015) 43 Australian Business Law Review 81; Nicolas Sahuguet and Alexis Walckiers, ‘A Theory of Hub-and-Spoke Collusion’ (2017) 53 International Journal of Industrial Organization 353; Okeoghene Odudu, ‘Indirect Information Exchange: The Constituent Elements of Hub and Spoke Collusion’ (2011) 7 European Competition Journal 205. 66 Falls and Saravia (n 65), 10–11. 67 Ibid, 10. 68 Ibid, 11–12. 69 Ibid.

Algorithm-driven collusive conduct  151 E

Perfect Price Discrimination

Price discrimination occurs when a merchant can determine the price that a buyer is prepared to pay, without disclosing the price at which the merchant is prepared to sell. This also has analogies to auction theory but where the seller reveals its public value in order to persuade the buyer to reveal their private value.70 The use of algorithms could also lead to something close to perfect price discrimination – that is, where each buyer pays a different (but close) price, again depending on their willingness to pay. Here algorithmic conduct could have one of four outcomes. The first is that there is no regulatory intervention and consumers pay what they are willing to pay. In effect, the absence of intervention is justified by the presumption that there is no market failure as price discrimination is treated by competition authorities as a normal part of vibrant competition. From a reputational risk perspective, the issue with perfect price discrimination is that if customers discover that they could have paid less, they may desert the merchant. The second outcome is that arbitrageurs are likely to enter. Intermediaries who algorithmically appear to have a lower willingness to pay could offer a service to customers and ensure that the customer pays no more than the amount that they are willing to pay. The primary reputational issue here is that the customer relationship moves from the merchant to the intermediary. The secondary issue is that the margin taken by the intermediaries is likely to converge on a fixed amount, as occurred with two major hotel platforms until European Commission intervention.71 A third issue in reputational risk is the risk of the merchant engaging in actual (and potentially illegal) discrimination. For example, if a merchant’s pricing algorithm charges more to women with an Italian surname, that is likely to raise an issue. That is, if the price discrimination discriminates against an identifiable group, there is likely to be a social media backlash, at least within the group that is being discriminated against. This is likely to lead to reputational harm, but not competition law risk. The amplification of the effects of trading errors by social media is a risk that businesses cannot afford to take. A fourth issue may be another reason why major platforms, such as Amazon, do not directly engage in price discrimination. Such platforms have an information advantage. Amazon links merchants to buyers on the Amazon platform, as well as selling products directly to consumers. Amazon makes a market as well as trading in that market. When Amazon sells, it makes a margin. When merchants sell, Amazon receives a percentage of the selling price. Amazon also acquires enough knowledge about individual consumers’ willingness to pay to be able to price discriminate. However, Amazon can also acquire knowledge about the aggregate price elasticity of demand for any product that is offered on the platform. This elasticity curve is likely to be more valuable to the platform than the value of price discrimination. It could be offered as a service for merchants or be used by the platform to decide on the product mix that is offered.

70 Martin Cave and Rob Nicholls, ‘The Use of Spectrum Auctions to Attain Multiple Objectives: Policy Implications’ (2017) 41 Telecommunications Policy 367. 71 For a retrospective on the effects of intervention, see Andrea Mantovani, Claudio Piga and Carlo Reggiani, ‘On the Economic Effects of Price Parity Clauses – What Do We Know Three Years Later?’ (2018) 9(10) Journal of European Competition Law & Practice l.

152  Research handbook on methods and models of competition law The effect is that engaging in algorithm-driven price discrimination may have higher reputational risk issues than gathering value from the information provided by the same algorithms.

IV

COMPETITION AND ANTITRUST LAW ANALYSIS

A

Concertation Not Contract

This section has a focus on concerted action or concerted practices, rather than on contracts. The rationale is that the use of algorithms to implement cartel conduct would be prohibited under antitrust or competition law in almost any jurisdiction that has such a statute. That is, the use of an algorithm to implement prohibited cartel conduct is as likely to be prohibited as the use of a person to implement the same conduct. That is, the ‘messenger’ scenario set out in Table 7.1 is most likely to be prohibited.72 In contrast, each of the ‘hub-and-spoke’ and ‘predictable agent’ scenarios has the potential to be implemented without an agreement – that is, where there may be a meeting of the minds of the parties but in the absence of any specific or identifiable commitment. Jurisdictions take different approaches to this conduct. B

The Australian Position

In its report on competition law and policy in Australia in 2015, the Harper Panel concluded that the existing provisions on price signalling and information exchanges in the Competition and Consumer Act 2010 (Cth) (CCA) were not adequate and needed to be replaced.73 A key concern was that the current provisions were unlikely to capture some types of inappropriate coordinated conduct between competitors, such as information exchanges about commercial strategy, consumers or prices.74 The Panel highlighted how this came about due to the perception that under the current law some level of ‘obligation’, ‘agreement’ or ‘commitment’ between competitors was required to establish unlawful collaborative conduct, and that mere evidence of a pattern of price signalling or information exchange was not enough to constitute a breach.75 Under the CCA, persons (including corporations) are prohibited from entering into a contract or arrangement, or arriving at an understanding, that has the purpose or effect of substantially lessening competition.76 The Harper Panel proposed an extension of the current law to include the concept of ‘concerted practices’ to remedy this concern.77 This recommendation was accepted and section 45 of the CCA was extended as follows: ‘A corporation must not: … engage with one or more persons in a concerted practice that has the purpose, or has or

72 See Rob Nicholls and Brent Fisse, ‘Concerted Practices and Algorithmic Coordination: Does the New Australian Law Compute?’ (2018) 28 Competition and Consumer Law Journal 26. 73 Ian Harper, Peter Anderson, Su McCluskey and Michael O’Bryan, Competition Policy Review: Final Report (2015). 74 Ibid, 369–70. 75 Harper et al (n 73). 76 CCA, s 45(2). 77 Harper et al (n 73), 9.

Algorithm-driven collusive conduct  153 is likely to have the effect, of substantially lessening competition.’78 The purpose of these changes is to ensure that conduct between competitors that has the purpose or effect of substantially lessening competition is captured irrespective of whether such conduct falls short of constituting an express contract, arrangement or understanding between those competitors, which was already prohibited. Introducing the concept of ‘concerted practices’ into Australian competition law demonstrates an increased willingness by the government to align Australian competition law with the laws of the UK, the EU and the US.79 However, given that the concept of ‘concerted practices’ is new to the Australian legal landscape, issues arise as to how it will be defined and applied in Australia. The concept is not explicitly defined in the CCA. The Explanatory Memorandum states that the word ‘concerted’ has a clear and practical meaning. It also states that a concerted practice may exist in addition to, or ancillary to, a contract, arrangement or understanding.80 Problems raised by the lack of clarity here have been analysed.81 If the change in Australian law removes the need for a person to be involved in the creation of a concerted practice, it is useful to understand how different international jurisdictions have dealt with the concept (or ‘concerted actions’, as labelled under US antitrust law). Two relevant jurisdictions include the US and the EU, both of which have long histories of prohibiting ‘concerted actions’ or ‘concerted practices’ respectively.82 This section examines the US and the EU approaches to concerted practices, tracing how the concept has developed in these jurisdictions. C

Concerted Actions under US Antitrust Law

The US concept of concerted actions (or concerted practices) is based in section 1 of the Sherman Antitrust Act of 1890, which prohibits ‘every contract, combination in the form of trust or otherwise, or conspiracy, in restraint of trade or commerce among the several States, or with foreign nations’.83 In general, US courts have interpreted the words ‘contract’, ‘combination’ and ‘conspiracy’ to mean ‘agreement’.84 The term ‘agreement’, however, is a broad and ill-defined concept in US antitrust law and thus the courts have applied section 1 to a broad

CCA, s 45(1)(c) (emphasis added). Harper et al (n 73), 369. 80 Explanatory Memorandum – Competition and Consumer Amendment (Competition Policy Review) Bill 2017 (Cth), 33. 81 Rob Nicholls and Deniz Kayis, ‘Concerted Practices Contested: Evidentiary Thresholds’ (2017) 25 Competition and Consumer Law Journal 125; Caitlin Davies and Luke Wainscoat, ‘Not Quite a Cartel: Applying the New Concerted Practices Prohibition’ (2017) 25 Competition and Consumer Law Journal 173; Lindsay Foster and Hanna Kaci, ‘Concerted Practices: A Contravention without a Definition’ (2018) 26 Competition and Consumer Law Journal 1. 82 Australian Competition and Consumer Commission, Framework for Concerted Practice Guidelines (September 2016) 2–3; Harper et al (n 73). 83 US 15 § 1. 84 Federico Ghezzi and Mariateresa Maggiolino, ‘Bridging EU Concerted Practices with U.S. Concerted Actions’ (2014) 10 Journal of Competition Law & Economics 647, 649. 78 79

154  Research handbook on methods and models of competition law spectrum of arrangements other than formal and express agreements.85 Most relevantly, US courts have been willing to apply antitrust prohibitions to the following situations: ●● where competitors have not entered into an express anticompetitive agreement but rather there is only circumstantial evidence, such as parallel market behaviour, which implies that inappropriate coordinated conduct among competitors has occurred; ●● where a corporation has unilaterally disclosed sensitive information to the market in either private or public and, in turn, invited its competitors to collude; ●● where competitors have arranged an information exchange of market-sensitive information, such as information concerning commercial strategy, consumers or prices; and ●● in instances where hub-and-spoke conspiracies or horizontal collusion has occurred. Through prohibiting the above types of actions under the Sherman Act, the doctrine of concerted actions has further developed in the US. This is discussed below. 1

Using circumstantial evidence and parallel behaviour to establish a concerted action Modern judicial efforts to define the concept of ‘concerted actions’ under section 1 of the Sherman Act stem from five key Supreme Court cases. These cases imply that a concerted action between competitors can result from means other than direct assurances or explicit agreements between parties. The cases are Interstate Circuit, Inc v United States;86 United States v Paramount Pictures, Inc;87 American Tobacco Co v United States;88 Theatre Enterprises, Inc v Paramount Film Distributing (Theatre Enterprises);89 and First National Bank v Cities Services Co.90 The cases were foundational to establishing the concept of concerted actions in US antitrust law. In particular, these cases create three key principles: first, that a concerted action does not require an express agreement; second, that circumstantial evidence can lead to an inference of inappropriate coordinated conduct; and third, that if competitors’ actions are merely consciously parallel and economically justified, an agreement will not exist. Subsequent Supreme Court cases have adapted the above principles into the ‘plus factors’ approach. Specifically, the Supreme Court has extended the third principle above, which was first introduced in Theatre Enterprises. The two most relevant cases to understanding the development of this new approach include Monsanto Co v Spray-Rite Service Corporation (Monsanto)91 and Matsushita Electrical Industrial Co v Zenith Radio Corporation (Matsushita).92 In Monsanto, the Supreme Court held:

85 Gregory J Werden, ‘Economic Evidence on the Existence of Collusion: Reconciling Antitrust Law with Oligopoly Theory’ (2004) 71 Antitrust Law Journal 719. 86 306 US 208 (1939). 87 334 US 131 (1948). 88 328 US 781 (1946). 89 346 US 537 (1954). 90 391 US 253. 91 465 US 752 (1984). 92 475 US 574 (1986).

Algorithm-driven collusive conduct  155 The correct standard is that there must be evidence that tends to exclude the possibility of independent action by the [parties]. That is, there must be direct or circumstantial evidence that reasonably tends to prove that [the parties] had a conscious commitment to a common scheme designed to achieve an unlawful objective.93

The Court emphasized that the burden to establish the fact of an unlawful agreement is on the plaintiff. In Matsushita, the Supreme Court developed the modern ‘plus factors’ formula that is now applied to cases where parties rely on circumstantial evidence to establish a concerted action that breaches section 1 of the Sherman Act. The Court held that ‘antitrust law limits the range of permissible inferences from ambiguous evidence in a Section 1 case … [and highlighted that] conduct as consistent with permissible competition as with illegal conspiracy does not, standing alone, support an inference of antitrust conspiracy’.94 The Court went on to further clarify that a plaintiff must present evidence that ‘tends to exclude the possibility’ that the alleged conspirators acted.95 The decision in Matsushita instructs courts to assess the economic justifications of the defendant’s actions when considering circumstantial evidence and parallel behaviour. US courts, including lower courts, have applied the rule in Matsushita to mean that plaintiffs must not only produce evidence of conscious parallelism (or parallel market behaviour among competitors), but also produce extra evidence that makes the possibility of inappropriate coordinated behaviour more likely than not to establish a concerted action in violation of the Sherman Act.96 Such pieces of extra evidence are known as ‘plus factors’.97 There is an extensive and growing literature on plus factors.98 2 Unilateral disclosures: attempts to monopolize and invitations to collude The section above discussed situations in which parallel market behaviour by competitors might be captured by the prohibition in section 1 of the Sherman Act. This section discusses another form of concerted action – invitations to collude. In general, an invitation to collude occurs where a corporation or its representative unilaterally discloses sensitive information such as price, or about consumers or corporate strategy, to the market in an attempt to solicit

465 US 752 at 768 (1984). 475 US 574 at 588 (1986). 95 Ibid. 96 See, for example, Bell Atlantic Corp v Twombly, 550 US 544 at 554 (2007): ‘The inadequacy of showing parallel conduct or interdependence, without more, mirrors the ambiguity of the behavior: consistent with conspiracy, but just as much in line with a wide swath of rational and competitive business strategy unilaterally prompted by common perceptions of the market.’ 97 See In re Musical Instruments and Equipment Antitrust Litigation (Guitar Center), 798 F 3d 1186 at 1194 (9th Cir 2015): ‘[plus factors are] economic actions and outcomes that are largely inconsistent with unilateral conduct but largely consistent with explicitly coordinated action’. 98 Gal and Elkin-Koren (n 63); Joseph E Harrington Jr, ‘Posted Pricing as a Plus Factor’ (2011) 7 Journal of Competition Law and Economics 1; Smith and Duke (n 65); Barak Orbach, ‘Hub-and-Spoke Conspiracies’ (2016) 15 Antitrust Source 1; Ghezzi and Maggiolino (n 84); William H Page, ‘Communication and Concerted Action’ (2007) 38 Loyola University Chicago Law Journal 405; William E Kovacic, Robert C Marshall, Leslie M Marx and Halbert L White, ‘Plus Factors and Agreement in Antitrust Law’ (2011) 110 Michigan Law Review 393. 93 94

156  Research handbook on methods and models of competition law its competitors to enter into a horizontal price-fixing or market allocation agreement.99 Such disclosures are unlawful where the agreement entered into by the competitors does not have any countervailing procompetitive benefit.100 Invitations to collude are prohibited under section 2 of the Sherman Act (‘section 2’) and section 5 of the Federal Trade Commission Act (1914) (‘section 5’). Section 2 prohibits any acts that ‘monopolize, or attempt to monopolize, or combine or conspire with any other person or persons, to monopolize any part of the trade or commerce’. In contrast, section 5 is more general in nature and provides that any ‘unfair methods of competition in or affecting commerce … are unlawful’. Section 5 applies to situations of anticompetitive behaviour even where it is difficult to establish the existence of an agreement between the parties. Invitations to collude may flow from private communications between competitors in which one party unilaterally discloses sensitive information to the other, or from public announcements in which a corporation unilaterally discloses sensitive information to the market in general. The position of US antitrust law concerning private disclosures can be illustrated by three key cases. In each case, the Federal Trade Commission (FTC) prosecuted the respondent corporations on the basis that they had breached section 5 following complaints they received from competitors. The first case is Quality Trailer Products Corporation.101 The second case is AE Clevite, Inc.102 The third case is In the Matter of YKK (USA) Inc.103 In contrast, where there has been a public disclosure of information, it is more difficult to identify what course of action authorities should take. A public announcement by a corporation is a public data source that is available to all their competitors and thus it is more difficult to establish an invitation to collude. The following cases provide two examples of where public disclosures may be captured by section 5. The first is In the Matter of Valassis Communications, Inc.104 In that case, the CEO of the respondent company, Valassis Communications, announced the company’s intention to raise the price of newspaper advertising inserts in a public call with analysts. Evidence produced by the FTC suggested that the respondent’s CEO was aware that its main competitor, News America, would be listening to the call. The FTC held that there was no reasonable business justification for the respondent to disclose pricing information on the call and thus the CEO’s conduct constituted a breach of section 5. The second case is In the Matter of U-Haul International, Inc and AMERCO.105 In that case, managers from the respondent company, U-Haul, approached Budget, one of U-Haul’s main competitors, and asked them to match U-Haul’s higher prices. At the same time, U-Haul’s CEO stated that the company was showing price leadership on a public investor conference call. The FTC held that U-Haul’s public and private disclosures breached section 5, as they created a significant risk of anticompetitive behaviour.

99 Albertina Albors-Llorens, ‘Horizontal Agreements and Concerted Practices in EC Competition Law: Unlawful and Legitimate Contacts between Competitors’ (2006) 51 Antitrust Bulletin 837, 23. 100 Ibid. 101 115 FTC 944 (1992). 102 58 Fed Reg 35,459 (FTC 1 July 1993). 103 116 FTC 628 (1993). 104 141 FTC (C-4160) (2006). 105 FTC File No 081-0157 (2010).

Algorithm-driven collusive conduct  157 3 Arrangements to exchange information Another form of concerted action includes where counterparts arrange to exchange sensitive information. Such arrangements may be in violation of section 1 of the Sherman Act. The position regarding arrangements to exchange information under US antitrust law was summarized in United States v Container Corporation of America.106 In that case, the Supreme Court stated that the dissemination of price information is not itself a per se violation of the Sherman Act, but such conduct would be unlawful where it ‘had an anticompetitive effect in the industry, chilling the vigor of price competition’.107 The Court confirmed that under US antitrust law, an arrangement to exchange information is not an offence per se; rather, it will only be an offence where such an arrangement has an anticompetitive purpose. Further to this, the FTC and the Department of Justice have both stated that whether the arrangement is anticompetitive depends on the nature of the information exchanged.108 4 Hub-and-spoke conspiracies An example of a horizontal hub-and-spoke conspiracy was the e-books case.109 Prior to the formation of Apple as a competitor within the industry, Amazon held the dominant market share in terms of selling e-books through its Kindle platform. The agreement that Amazon shared with the publishers was similar to a ‘wholesale distribution model’, where Amazon would resell the e-books to its consumers at its own prices. The issue with this model was that Amazon set all its pricing of e-books at $9.99, raising fears among publishers that such a low price would increase cannibalization of sales for hard copy books (averaging around $26 per book).110 However, with the introduction of Apple and its iPad, Apple created a new arrangement with publishers whereby Apple acted as an agent and instead each publisher was given control to set its own prices for e-books. Publishers threatened to withdraw from Amazon’s Kindle agreement and subsequently Amazon also converted to the agency model. Ultimately, this move towards an agency model left consumers paying a much higher price for e-books. This matter was heard in the US and the UK, with both courts agreeing that there was the existence of a concerted practice.111 The third class of hub-and-spoke conspiracies is a hybrid of the two previously discussed classes. The case United States v Apple Inc demonstrates how it is also possible for the e-books case to be classified as a hybrid hub-and-spoke case.112 While Apple did facilitate a horizontal 393 US 333 (1969). Ibid, 337. 108 See US Department of Justice and Federal Trade Commission, Statements of Antitrust Enforcement Policy in Health Care, Statement 6: Enforcement Policy on Provider Participation in Exchanges of Price and Cost Information (August 1996 revision), http://​www​.ftc​.gov/​bc/​healthcare/​industryguide/​policy/​ statement6​.pdf. See also US Department of Justice and Federal Trade Commission, Antitrust Guidelines for Collaborations Among Competitors (April 2000) 6, http://​www​.ftc​.gov/​os/​2000/​04/​ftcdojguidelines​ .pdf: ‘The Agencies recognized that sharing of information among competitors may be precompetitive and is often reasonably necessary to archive the precompetitive benefits of certain collaboration; for example, sharing certain technology, know-how, or other intellectual property may be essential to achieve the precompetitive benefits of an R&D collaboration.’ 109 Case COMP/39.847/E-Books [2013]. 110 Ibid, [20]–[25]. 111 E-Books 2013 (n 109), [87]–[93]; United States v Apple Inc, No 12 Civ 2826 (SDNY 10 July 2013), 48. 112 Ibid. 106 107

158  Research handbook on methods and models of competition law agreement, with the harm being associated at the horizontal level, the agreements that Apple had with the publisher included a ‘most-favoured-nation’ clause. Such a clause meant that publishers were required to offer Apple the lowest set retail price when compared to other competing retailers, effectively forcing competitor retailers to comply with the agency model or completely lose access to e-book products.113 This also meant that Apple was gaining a significant financial advantage over the other distributors (suggesting that harm occurred on both levels). D

Concerted Practices under European Union Law

The US was the first country to introduce antitrust laws and the development of the EU’s competition law framework has been greatly influenced by the American experience. Article 101 of the Treaty on the Functioning of the European Union (TFEU) prohibits ‘all agreements, decisions by associations of undertakings and concerted practices’ which have an anticompetitive purpose or effect.114 Unlike section 1 of the Sherman Act, the term ‘concerted practice’ is explicitly used in Article 101 of the TFEU.115 The purpose of this was to ensure that the prohibition applies to all types of inappropriate coordinated conduct, not just express agreements between competitors.116 As explained by the European Commission (the Commission) in its decision in Re Polypropylene:117 The object of the Treaty in creating a separate concept of concerted practice is to forestall the possibility of undertakings evading the application of Article [101] by colluding in an anti-competitive manner falling short of a definite agreement by (for example) informing each other in advance of the attitude each intends to adopt, so that each could regulate its commercial conduct in the knowledge that its competitors would behave in the same way …118

The concept of ‘concerted practices’ is not defined in the TFEU. To determine how it has been applied in the EU, it is necessary to consider the case law on concerted practices. Such decisions are discussed in the section below. 1 The definition of concerted practices under EU law The first time the European Court of Justice considered the definition of ‘concerted practices’ was in the case ICI v Commissioner (Dyestuffs).119 Dyestuffs concerned an investigation of ten dyestuffs producers, who controlled approximately 80 per cent of the European dyestuffs market. The Commission claimed that the producers had engaged in a concerted practice 113 Brent Fisse, ‘Facilitating Practices, Vertical Restraints and Most Favoured Customers’ (2016) 44 Australian Business Law Review 325. 114 Treaty on the Functioning of the European Union (2009), Art 101(1). 115 The insertion of ‘concerted practices’ in Article 101 was greatly influenced by the development of the ‘concerted actions’ in US antitrust law. See the Opinion of Advocate General Mayras in Case 48/69, ICI v Commissioner [1972] ECR 619 at 669–670. See also the Opinion of Advocate General Vesterdorf in Case T-1/89, Rhone-Poulenc v Commissioner [1991] ECR 11-867 at 11-927. 116 Albors-Llorens (n 99), 1. 117 Decision of the Commission in Re Polypropylene [Decision IV/31.149, 1986 OJ L 230]. 118 Ibid, [87]. 119 ICI v Commissioner [1972] ECR 619.

Algorithm-driven collusive conduct  159 aimed at fixing prices. In making this claim, the Commission pointed to three separate occasions between 1964 and 1967 on which the producers had uniformly raised prices. They pointed to evidence that representatives from the defendant producers had all met before each price increase to contend that they were a result of concerted practice.120 In the Commission’s Opinion to the Court, Advocate General Mayras argued that a concerted practice could be distinguished from an anticompetitive agreement as it could not be separated from the broader effects it had on the market. That is, a concerted practice could only be established when such conduct had an identifiable anticompetitive effect.121 In its decision, the Court provided a broader definition of concerted practices, stating that it was ‘a form of co-ordination between undertakings which, without having reached the stage where an agreement properly so-called has been concluded, knowingly substitutes practical cooperation between them for the risks of competition’.122 The Court further explained that ‘by its very nature, then, a concerted practice does not have all the elements of a contract but may inter alia arise out of co-ordination which becomes apparent from the behaviour of the participants’.123 The Court also highlighted that although the behaviour of the parties was a crucial element to establishing a concerted practice, its impact on the market although relevant was not an explicit requirement. Three years later, the Court reconsidered the issue of concerted practices in the case Suiker Unie v Commissioner (Suiker).124 In that case, the Commission claimed that the main sugar producers and distributors in the EU had engaged in a concerted practice to protect certain markets in Italy, Germany and the Netherlands. The Commission pointed to evidence that the defendants had exchanged information on their future strategies in these markets to argue that a concerted practice had occurred. However, there was no evidence of a concrete plan or agreement between the defendants to engage in anticompetitive behaviour. In its decision, the Court rejected the defendants’ argument that for a concerted practice to be established, evidence of a premeditated plan to lessen competition was required. Rather, the Court explained that all that was required to establish a breach of Article 101 was any direct or indirect contact between operators, the object or effect whereof is either to influence the conduct on the market of an actual or potential competitor or to disclose to such a competitor the course of conduct which they themselves have decided to adopt or contemplate adopting in the market.125

The Court further explained that in order for the defendants’ behaviour not to be captured by Article 101, they would need to demonstrate that their actions were independent and not based on knowledge of their competitors’ future behaviour.126

Ibid, 679. Ibid, 671. 122 Ibid, [64]. 123 Ibid, [65]. 124 Joined Cases 40–48, 50, 54–56, 111, 113 & 114/73 [1975] ECR 1663. 125 Ibid, [174]. 126 Ibid, [175]. 120 121

160  Research handbook on methods and models of competition law Dyestuffs and Suiker were foundational to establishing the concept of concerted practices in EU competition law. In particular, these cases highlight that the definition of concerted practice consists of two elements:127 ●● the existence of direct or indirect reciprocal contacts between undertakings aimed at knowingly removing uncertainty as to future market behaviour; and ●● subsequent behaviour in the market pursuant to the contact between the undertakings. Subsequent cases have introduced a third causality element to the above definition. A good example of such a case is Huls v Commission (Huls).128 In that case, the Commission claimed that several polypropylene producers had engaged in a concerted practice aimed at fixing prices. In making this claim, the Commission pointed to evidence of regular meetings between the producers in which they discussed target prices and developed a system to split demand in the market. In its decision, the Court stated that there were three requirements for a concerted practice: ‘(a) the undertakings (businesses) concerting with each other; (b) subsequent conduct on the market; and (c) a relationship of cause and effect between the two’.129 In relation to this third element, the Court went on further to explain that once it can be established that there had been contact between undertakings aimed at removing uncertainty as to future market behaviour, there is a presumption that ‘the undertakings taking part in the concerted action and remaining active on the market take account of the information exchanged with their competitors for the purposes of determining their conduct on that market’. Huls lowers the standard required by Dyestuffs and Suiker to establish a concerted practice. This is because the words of the Court tip the balance of the test in favour of the first element outlined above, meaning that a concerted practice may be inferred simply from contact between competitors and does not require the Commission to demonstrate that such contact produced any anticompetitive effects as this will be presumed. 2

Using circumstantial evidence and parallel behaviour to establish a concerted practice As discussed above, US authorities have often relied on circumstantial evidence such as parallel market behaviour to establish the existence of a concerted practice. This is because documentary evidence, such as contracts between colluding parties, is difficult to obtain as the parties are careful to not leave a trail.130 EU authorities face similar issues. This section examines whether, under EU competition law, parallel behaviour by competitors would be sufficient to establish a concerted practice if it were the only evidence that such conduct had occurred. The first case that considered this issue was Dyestuffs. In that case, the defendant producers argued that their uniform price increases were a result of the oligopolistic nature of the European dyestuffs market. In its decision, the Court made clear that while parallel behaviour was relevant, it was not conclusive in establishing the existence of a concerted practice. Albors-Llorens notes that while the Court stated that parallel behaviour would only provide ‘strong evidence’ and that consideration of the relevant market’s specific features was more 129 130 127 128

Albors-Llorens (n 99). Case C-199/92P [1999] ECR 1-4287. Ibid, [161]. Albors-Llorens (n 99), 851.

Algorithm-driven collusive conduct  161 relevant, it only superficially followed this standard. That is, she contends that the Court in Dyestuffs only superficially considered the characteristics of the dyestuffs market in making its decision and instead treated the evidence that the defendants acted in a similar manner as conclusive. This in turn indicates ‘that concerted practice could exist through a combination of parallelism and a deviation from normal market practices’.131 Three years later, the Court reconsidered the issue of parallel conduct in Suiker. In its decision, the Court stated that ‘intelligent adaptions to the existing and anticipated conduct of competitors’ were legitimate.132 The Court’s words have been interpreted to mean that conscious parallelism in an oligopoly is not an offence per se. That is, evidence of parallel behaviour by competitors in an oligopolistic market is not in itself enough to establish a concerted practice.133 Ten years later, the Court decisively clarified its position regarding parallel conduct in Ahlstrom Osakeyhtio v Commission (Woodpulp II).134 In that case, the Commission claimed that a group of US and Scandinavian wood pulp producers had engaged in a concerted practice aimed at fixing prices by following a system in which they all gave advance quarterly price announcements to their customers. Because of these announcements, prices across the European wood pulp market increased uniformly. In its decision, the Court applied the standard established in Dyestuffs and Suiker, stating that ‘parallel conduct cannot be regarded as furnishing proof of concertation unless concertation constitutes the only plausible explanation for such conduct’.135 Moreover, unlike in Dyestuffs, the Court based its finding that the price announcements did not constitute a concertation on evidence of the wood pulp market forwarded by a panel of expert economists who contended that such behaviour was regular within the industry. Albors-Llorens contends that the decision in Woodpulp II is significant for two reasons:136 First, it clarified that, in the absence of evidence of contacts between market participants, parallel behavior will be a proof of collusion only where there is no other plausible explanation for such parallelism. This is a very high standard of proof. Secondly, the judgment made it very clear that, for [Article 101] to apply to oligopolistic markets, the competition authorities need to prove that the line that separates behaviour resulting from market structure and intelligent adaptation to market conditions from collusion has been unmistakably crossed by the undertakings in question.

3 Unlawful and legitimate exchanges of information As discussed above, information exchanges between competitors may constitute a concerted practice in certain circumstances. There is a fine line between legitimate exchanges of information and unlawful exchanges. Each of the 1968 Notice on Cooperation Agreements137 and the Commission’s 7th Report on Competition Policy138 provided guidance. The effect is that Ibid, 853–4. Joined Cases 40–48, 50, 54–56, 111, 113 & 114/73 [1975] ECR 1663. 133 See the Commission decision in Zinc Producers Group [1984] OJ L 220 27, [75]–[76]. 134 Joined Cases G-89/85, G-104/85, G-114/85, G-116/85, G-117/85 and G-125/85 to G-129/85 [1993] ECR I-01307. Albors-Llorens (n 99). 135 Joined Cases G-89/85, G-104/85, G-114/85, G-116/85, G-117/85 and G-125/85 to G-129/85, [126] (emphasis added). 136 Albors-Llorens (n 99), 854. 137 Notice on Cooperation Agreements [1968] OJ C 75 3. 138 Seventh Report on Competition Policy (1977), Pt 7. 131 132

162  Research handbook on methods and models of competition law the Commission will consider three key criteria when determining whether an information exchange between counterparts is legitimate or unlawful. These are: ●● the nature of the information being exchanged; ●● the structure of the market in which the counterparts participate; and ●● whether the information exchange is likely to improve transparency within this market. The first two of these criteria appear to be consistent with the position under US antitrust law, discussed above. However, the third criterion provides an interesting context as to the scope of prohibited conduct in the EU. Both Aalborg Portland v Commission139 and Boel v Commission140 relate to situations where the defendant has attended a meeting in which other competitors have disclosed sensitive information and they have claimed that they did not know that anticompetitive information would be discussed in these meetings or that they did not agree to take part in the collusive behaviour agreed by the other competitors at these meetings. In these cases, it was held that in order to not be liable under Article 101, the defendants would need to publicly distance themselves from what was discussed at the meetings or inform authorities of the nature of the meetings.141 4 Hub-and-spoke conspiracies under EU law In the EU, as set out above, the prohibition on concerted practices has been applied to both horizontal and vertical arrangements. Hub-and-spoke conspiracies occur where there is no direct communication between competitors. However, this is a relatively new area of EU law. Most of the cases have come from the UK. In this context, it is important to note that law in the UK is meant to be applied in a manner that is consistent with the EU and thus the UK decisions may have ramifications and application in broader EU law. The commonality of the approach is created by Chapter 1 of the Competition Act 1998 (UK) having almost identical wording to Article 101 of the TFEU. It prohibits ‘any agreements between undertakings, decisions by associations of undertakings or concerted practices which (a) may affect trade within the United Kingdom and (b) have as their object or effect the prevention, restriction or distortion of competition within the United Kingdom’. Three cases provide examples of the EU approach. The first is Musique Diffusion Françoise v Commission,142 which is an example of indirect contact between competitors. In this case, the supplier received a complaint from a French distributor concerning the influx of imports from the UK and Germany. The supplier informed British and German distributors of the French distributor’s complaint. The Court held that the supplier was trying to pressure the British and German distributors to stop exporting their products to the French market.143

139 Joined Cases C-204/00P, C-205/00P, C-211/00P, C-213/00P, C-217/00P and C-219/00P [2004] ECR I-00123. 140 See Case T-142/89 [1995] ECR 11-867. 141 Aalborg at [82], [84]; Boel at [89]. 142 Case 100/80 [1983] ECR 1825. 143 Ibid, [72]–[79].

Algorithm-driven collusive conduct  163 The second case, JJB Sports/All Sports v Office of Fair Trading,144 related to price-fixing of replica football shirts in the UK market. The defendants were the retailers JJB Sports and All Sports. The defendants complained to the manufacturer that another retailer, Sports Soccer, was discounting the price of replica football jerseys. The manufacturer subsequently exerted pressure on Sports Soccer to raise its price. It was held that the defendant retailers and the manufacturer had engaged in a concerted practice aimed at fixing prices. It was further held that general complaints by retailers to a manufacturer would not be unlawful. However, complaints in which a specific competitor is named are unlawful.145 The Court explained that if one retailer A privately discloses to a supplier B its future pricing intentions in circumstances where it is reasonably foreseeable that B may make use of that information to influence market conditions, and then B passes that pricing information on to a competing retailer C, then in our view A, B, and C are all to be regarded on those facts as parties to a concerted practice.146

In Argos & Littlewoods v Office of Fair Trading,147 Argos and Littlewoods, both toy and games retailers, disclosed future pricing information to their supplier, Hasbro, which in return informed them of its pricing expectations. There was no direct contact between Argos and Littlewoods. However, the Court took a view that was similar to JJB Sports/All Sports. It held that there had been indirect contact through the third-party supplier. As a consequence, it held that the retailers and the supplier had engaged in a concerted practice aimed at fixing prices.148

V

DISCUSSION AND CONCLUSIONS

A

The Issues

There are two critical issues that need to be addressed arising from the foregoing. The first is whether algorithmic-driven conduct will lead to the predicted ‘tacit collusion on steroids’.149 The second is whether competition law and policy need to respond to changes in business models. This chapter has examined some of the specific issues that flow from the ‘hub-and-spoke’ and ‘predictable agent’ scenarios set out in Table 7.1. The basis of this is that the ‘messenger’ scenario is already prohibited by antitrust law and that the ‘digital eye’ scenario may not yet have occurred. However, what the chapter has demonstrated is that the potential risk of automated collusion rises in the context of multisided markets or platforms. It is also important to realize that some of the distinctive characteristics of multisided markets mean that conduct which would be prohibited in traditional markets is procompetitive in multisided markets.150 In particular, the

Case 1021/1/1/03 and 1022/1/1/03, 2004 CAT 17. Ibid, [665]. 146 Ibid, [659]. 147 Case 1014 and 1015/1/1/03. 148 Ibid, [701]–[704]. 149 Ezrachi and Stucke, Virtual Competition (n 1). 150 See, for example, Jakhu and Malik (n 60); Rysman (n 52); OECD, ‘Rethinking Antitrust Tools for Multi-Sided Platforms 2018’ (n 60); Gal and Elkin-Koren (n 63); Harrington (n 38). 144 145

164  Research handbook on methods and models of competition law assumption that tipping always occurs in platforms does not lead to ‘winner-takes-all’, even if the outcome is ‘winner-takes-most’.151 But there remains a problem. If consumers receive the second-best price as a result of algorithmic pricing, this is a poorer outcome than receiving the best price from the lowest-price merchant. The effect is that algorithmic coordination allows the lowest-price merchant to extract rent from the consumer. Whether this amounts to an antitrust issue depends on the degree of market power held by that merchant. In particular, if the leading merchant is also the platform operator (the market maker of the multisided market), then there is potential for the rent to be as a result of prohibited collusive conduct or as a result of prohibited unilateral conduct. It suggests that the Falls and Saravia approach to hub-and-spoke conspiracies may need to be extended to address this area.152 One of the criticisms of the change in the Australian law to address concerted practices is that there is no clarity about whether the harm is vertical or horizontal. In this context, perhaps it is a strength as it does not restrict the Australian Competition and Consumer Commission in its enforcement work.153 There is also an opportunity for competition regulators to use algorithmic-based tools for cartel detection. This is a form of regtech extended from its usual application in the financial services sector.154 This is already in place for cartel detection from open sources.155 However, there is also a risk of chilling a competitive environment by prematurely deciding that there is collusion in algorithmic conduct.156 B

Some Conclusions and Prescriptions

In Verizon Communications, Inc v Law Offices of Curtis V Trinko,157 the Supreme Court stated that cartels remain ‘the supreme evil of antitrust’. That is, price-fixing by natural persons or corporations is considered to be heinous.

151 David S Evans and Richard Schmalensee, ‘Debunking the “Network Effects” Bogeyman’ (2017) 40 Regulation 35. 152 Falls and Saravia (n 65). 153 A recent Australian High Court case (Australian Competition and Consumer Commission v Flight Centre Travel Group Ltd (2016) 261 CLR 203; [2016] HCA 49) has already determined that an agent and its principal can engage in price-fixing. See, for analysis, Brent Fisse, ‘The High Court Decision in ACCC v Flight Centre – Crash Landings Ahead?’ (2017) 45 Australian Business Law Review 61; Davies and Wainscoat (n 81). 154 Douglas W Arner, Janos Nathan Barberis and Ross P Buckley, ‘FinTech, RegTech and the Reconceptualization of Financial Regulation’ (2017) 37(3) Northwestern Journal of International Law & Business 371; Douglas W Arner, Janos Nathan Barberis and Ross P Buckley, ‘The Evolution of Fintech: A New Post-Crisis Paradigm?’ (University of Hong Kong Faculty of Law Research Paper No 2015/047, 2015). 155 See, for example, Martin Huber and David Imhof, ‘Machine Learning with Screens for Detecting Bid-Rigging Cartels’ (2019) 65 International Journal of Industrial Organization 277; Pal Vadász, Andras Benczúr, Geza Füzesi and Sándor Munk, ‘Identifying Illegal Cartel Activities from Open Sources’ in Babak Akhgar, P Saskia Bayerl and Fraser Sampson (eds), Open Source Intelligence Investigation: Advanced Sciences and Technologies for Security Applications (2016). 156 Salil K Mehra, ‘Antitrust and the Robo-Seller: Competition in the Time of Algorithms’ (Springer 2016) 100 Minnesota Law Review 1323; Salil K Mehra, ‘Robo-Seller Prosecutions and Antitrust’s Error-Cost Framework’ (2017) 2 CPI Antitrust Chronicle 36. 157 540 US 398 (2004).

Algorithm-driven collusive conduct  165 There is a problem of nexus between the creators of bots and their longer-term relationship. If the creators regard the bot’s conduct as being wrong, they would no longer condone the action and would stop the bots from tacitly colluding. That is, if the bots are agents of a firm, they would not be permitted to engage in conduct that would place the firm at legal risk. However, they do not seem to have done so. This creates a situation where there is an assumption, whether tacit or express, that bots are autonomous actors. In effect, ethical consistency requires that bots are not agents of the firm if there is any prospect that they might engage in price-fixing. The problem facing antitrust regulators is whether autonomous actors engaged in algorithmic tacit collusion are in breach of the law. The problem facing competition law policy-makers is whether autonomous actors engaged in algorithmic tacit collusion are causing any anticompetitive harm that needs to be prohibited. This issue is critical in the antitrust space, with approaches such as applying the essential facilities doctrine being suggested158 and whole journal issues devoted to the topic.159 On the other hand, there is significantly less concern with the effects of algorithmic-driven conduct in the corporate law context. One reason for this may be that the competition law is more responsive to changes, potentially a reflection of the way that jurisprudence in competition law has evolved. As we have seen, the Australian law, consistent with the approaches of the US and the EU, currently requires that there is a ‘contract, arrangement or understanding’ for there to be price-fixing conduct. Australian jurisprudence means that these terms require the ‘meeting of the minds’ mentioned above, but also ‘commitment’. Even if bots are found to be agents of firms, it is not clear that algorithmic tacit collusion would meet either of these two tests. As Australian law and jurisprudence have not diverted significantly from US or EU law, then the same issue may arise in other jurisdictions. In the EU, there is a concerted practices defence known as the ‘oligopoly defence’.160 The defence asserts that market structure leads to similar conduct, and that the conduct is not to be treated as a concerted practice. This defence effectively eliminates action for tacit collusion and, along with the vertical restraint limitation discussed below, significantly limits the

Lugard and Roach (n 3). John M Newman, ‘Complex Antitrust Harm in Platform Markets’ (2017) 2 CPI Antitrust Chronicle 52; Frank Pasquale, ‘When Antitrust Becomes Pro-Trust: The Digital Deformation of U.S. Competition Policy’ (2017) 2 CPI Antitrust Chronicle 46; Woodcock (n 63); Mehra, ‘Robo-Seller Prosecutions and Antitrust’s Error-Cost Framework’ (n 156); Dylan I Ballard and Amar S Naik, ‘Algorithms, Artificial Intelligence and Joint Conduct’ (2017) 2 CPI Antitrust Chronicle 29; Gal (n 21); Ariel Ezrachi and Maurice E Stucke, ‘Looking Up in the Data-Driven Economy’ (2017) 2 CPI Antitrust Chronicle 16; Paul Johnson, ‘Should We Be Concerned That Data and Algorithms Will Soften Competition?’ (2017) 2 CPI Antitrust Chronicle 10. 160 Alina Kaczorowska-Ireland, European Union Law (4th edn, Routledge 2016) 996; Barry Rodger and Angus MacCulloch, Competition Law and Policy in the EU and UK (Routledge 2014); Nicolas Petit, ‘The Future of the Court of Justice in EU Competition Law’ in Court of Justice of the European Union (ed), The Court of Justice and the Construction of Europe: Analyses and Perspectives on Sixty Years of Case-law (TMC Asser Press 2013) 417; Nicolas Petit, ‘The Oligopoly Problem in EU Competition Law’ in Ioannis Lianos and Damien Geradin (eds), Handbook on European Competition Law: Substantive Aspects (Edward Elgar Publishing 2013) 289; George A Hay, ‘Horizontal Agreements: Concept and Proof’ (2006) 41 Antitrust Bulletin 877, 888. 158 159

166  Research handbook on methods and models of competition law operation of the prohibition against concerted practices. It would be reasonable to consider eliminating this defence in the context of multisided markets. As set out above, a 2017 amendment to the Australian law prohibits a firm from engaging ‘in a concerted practice that has the purpose, or has or is likely to have the effect, of substantially lessening competition’. That is, the amended law eliminates the need for the firms to have ‘commitment’ and merely a ‘meeting of the minds’. Australia might become one of the first jurisdictions to have the flexibility to deal with algorithmic tacit collusion under competition law. However, the legislation was not drafted to have an effect on automated price-fixing.161 It was intended to deal with hub-and-spoke conspiracies and the shared use of a data source. Indeed, the Explanatory Memorandum162 provides examples of such conduct. Competition law usually assumes that efficient market mechanisms are beneficial to consumers and only act to prohibit restraints that have an adverse effect on competition. The challenge is to ensure that changes to the law are measured, appropriate and timely. There is a significant risk if the law tries to ‘keep up’ with technology, this creates a significant risk.163 The debate mentioned earlier in this section is healthy. However, over-reaction or a premature response brings with it risks of creating inefficiency. There is also a problem of over-enforcement. This is the ‘coincidence or conspiracy’ matter. The issue from an antitrust regulator perspective is that detection of cartels is difficult.164 The risk is that conspiracies will be found where coincidences should have been noted. Any amendment to the law needs to address this problem and it is not clear that the Australian legislative drafting meets this objective.

Nicholls and Fisse (n 72). Explanatory Memorandum (n 80). 163 Lyria Bennett Moses, ‘Agents of Change: How the Law Copes with Technological Change’ (2011) 20 Griffith Law Review 763. 164 Caron Y Beaton-Wells, ‘Immunity for Cartel Conduct: Revolution or Religion? An Australian Case Study’ (2014) 2 Journal of Antitrust Enforcement 126; D Daniel Sokol, ‘Cartels, Corporate Compliance, and What Practitioners Really Think about Enforcement’ (2012) 78 Antitrust Law Journal 201; Yassine Lefouili and Catherine Roux, ‘Leniency Programs for Multimarket Firms: The Effect of Amnesty Plus on Cartel Formation’ (2012) 30 International Journal of Industrial Organization 624; Michael O’Kane, ‘International Cartel Criminalisation and Leniency: Recent Lessons from the UK and Global Comparisons’ (2012) 8 Competition Law International 55; Margaret C Levenstein and Valerie Y Suslow, ‘What Determines Cartel Success?’ (2006) 44 Journal of Economic Literature 1; Caron Y Beaton-Wells and Brent Fisse, Australian Cartel Regulation (Cambridge University Press 2011). 161 162

8. Vertical agreements under EU competition law: proposals for pushing Article 101 analysis, and the modernization process, to a logical conclusion Miguel de la Mano and Alison Jones

I INTRODUCTION The analysis of vertical agreements under Article 101 of the Treaty on the Functioning of the European Union (TFEU) has been controversial since the Article’s first enforcement. The problems created by the initial tendency of the European Commission (the Commission) to interpret the concept of a restriction of competition set out in Article 101(1) broadly, and to rely on Article 101(3), the legal exception to Article 101(1),1 as the main vehicle for authorizing vertical agreements, are well known. The resultant difficulties eventually led the Commission to recognize that modernization of the system was required. Although the case law of the Court of Justice (the Court of Justice of the European Union is now comprised of the Court of Justice and the General Court) always displayed a more nuanced approach to that initially adopted by the Commission, and although Commission policy towards vertical agreements has advanced considerably since modernization, this chapter argues that difficulties still persist. Further, as the Commission (following the launch of its e-commerce sector inquiry2) and a number of national competition authorities (NCAs) are taking a renewed interest in vertical practices, especially in relation to online markets and online selling, it is important that these problems be resolved. This chapter commences in Section II by introducing the difficulties that resulted from the Commission’s initial appraisal of vertical agreements and describing how the modernization proposals sought to resolve them. It observes that in spite of the reform and the Commission’s 1 Article 101 divides substantive analysis between Article 101(1), which prohibits agreements and so on between undertakings that have as their object or effect the restriction of competition, and Article 101(3), which provides a legal exception from the Article 101(1) prohibition for agreements that satisfy its four conditions: see, for example, Alison Jones and Brenda Sufrin, EU Competition Law: Text, Cases, and Materials (6th edn, Oxford University Press 2016) chs 3, 4, 11. 2 See European Union, ‘Sector Inquiry into E-Commerce’ (10 May 2017), https://​ec​.europa​.eu/​ competition/​antitrust/​sector​_inquiries​_e​_commerce​.html. The inquiry was launched in May 2015. The Commission published Commission Staff Working Document: Preliminary Report on the E-Commerce Sector Inquiry on 15 September 2016 and the Report from the Commission to the Council and the European Parliament: Final Report on the E-Commerce Sector Inquiry (Commission’s Final Report), which summarizes the main findings of the e-commerce sector inquiry and incorporates comments submitted by stakeholders during the public consultation, on 10 May 2017. On 2 February 2017 the Commission opened three investigations in relation to aspects of e-commerce: European Commission, ‘Antitrust: Commission Opens Three Investigations into Suspected Anticompetitive Practices in E-Commerce’ (2 February 2017), http://​europa​.eu/​rapid/​press​-release​_IP​-17​-201​_en​.htm.

167

168  Research handbook on methods and models of competition law promise of a more ‘effects-based’ approach based on a consumer welfare objective, antitrust analysis has not evolved in the way anticipated. Rather, in Section III, it is argued that the current legal framework still fails to reflect adequately the economic logic of vertical restraints. First, it continues to rely heavily on broad presumptions of illegality that are difficult to rebut and are, as applied, not justified by economic theory, evidence or experience. This creates a risk of Type I errors (or false positives, arising from an overinclusive rule that sometimes condemns conduct that is competitive, benign or beneficial) and is likely to deter procompetitive distribution arrangements. Second, the dearth of decided cases means that, despite significant advances in the economic and legal assessment of vertical mergers and abuse of dominance (which often involve the analysis of similar competitive effects), a transparent structure for analysing and balancing the competitive harms and benefits of vertical arrangements has not developed. This uncertainty has been compounded by some disparities in approach to enforcement emerging at the national level. Section IV consequently proposes an approach which focuses on the underpinning objectives of the rules as opposed to historic categories of analysis (such as distinctions between intrabrand and interbrand restraints3 and the application of presumptions of illegality that are practically difficult to rebut). It argues that the questions of when, and why, a vertical agreement is treated as restrictive by object and/or presumptively illegal need clarification, explanation and refinement; it is not justifiable to apply a presumption of illegality against vertical restraints as extensively and rigidly as is currently done.4 Further, it suggests that a clearer framework is required for the assessment of vertical restraints to determine whether they (a) restrict competition and, if so, (b) satisfy the conditions of Article 101(3). This should help to reduce uncertainty and diverging enforcement practices across EU Member States. The chapter suggests how these changes might be effected and an administrable system governing vertical restraints put in place. It argues that this requires the Commission to bring more vertical agreement cases that may be reviewed by the EU courts, especially some complex effects cases,5 including both infringement and non-infringement decisions, that clearly delineate the boundaries of Article 101.6 Commitments decisions7 could also be used, 3 Intrabrand restrictions restrain competition (price or non-price competition) between distributors of a particular brand. Interbrand restraints restrain competition between manufacturers and their competitors (see further Section III). This chapter does not deal with the licensing of digital content. 4 Similar considerations apply in the assessment of exclusivity rebates under Article 102 of the TFEU. However, in Case T-286/09, Intel v European Commission (General Court of the European Union, 12 June 2014), the General Court advanced a three-category approach to exclusive dealing and rebates: (1) exclusive dealing obligations and exclusivity (or loyalty or fidelity) rebates are presumptively abusive, subject to objective justification; (2) quantity rebates are presumptively legal; and (3) other conditional rebates are unlawful if it is shown by a consideration of all the circumstances that they are capable of restricting competition. The Court of Justice set aside the General Court’s judgment: Case C-413/14 P, Intel v European Commission (Court of Justice of the European Communities, 6 September 2017). See n 179. 5 See n 150 and text. 6 See Wouter PJ Wils, ‘The Relationship between Public Antitrust Enforcement and Private Actions for Damages’ (2009) 32 World Competition 3 (stressing the importance of effective enforcement of the competition laws in ensuring not only deterrence, compensation and remediation, but also clarification of the law) and Joshua D Wright and Douglas H Ginsburg, ‘The Economic Analysis of Antitrust Consents’ (2018) 46 European Journal of Law and Economics 245. 7 Commitments decisions under Regulation 1/2003, Article 9 of the TFEU involve only a decision not to pursue a preliminary finding of an infringement on the basis that commitments are given by the parties to change their behaviour. They do not involve a finding of infringement or non-infringement.

Vertical agreements under EU competition law  169 and monitored, to collect empirical data on the effects of contract changes on competition in practice. It also argues that this change of approach should be reflected in the revision of the EU regime governing vertical agreements.8

II

MODERNIZATION: EVOLUTION IN ARTICLE 101 ANALYSIS

A

The Initial Approach

The Commission’s early approach to Article 101, interpreting the concept of a restriction of competition set out in Article 101(1) broadly to encompass many restraints on firms’ economic freedom and ability to engage in cross-border trade, led to conceptual and practical problems. It also created a central role both for Article 101(3) in the Article 101 framework9 and for the Commission in its enforcement (national courts and NCAs could not apply Article 101(3)). As Carles Esteva Mosso, Deputy Director-General for the Directorate-General for Competition at the Commission, explains, under this ‘previous, more form-based approach to the interpretation of “restriction of competition”, a large number of agreements were considered to be caught by the test of Article 101(1) and required exemption under Article 101(3)’.10 Although the Commission sought to resolve the bottle-neck problem11 through various techniques – including the development of the de minimis principle,12 the use of comfort letters and the introduction of block exemption regulations (BERs) exempting certain categories of agreements from the Article 101(1) prohibition – these were imperfect solutions and trenchant criticism of this approach mounted. In particular, complainants asserted that the ‘anaemic’13 analysis conducted by the Commission under Article 101(1), combined with the rigidity of

8 In its Final E-Commerce Report (n 2), the Commission states that its enforcement actions and the data gathered during the e-commerce sector inquiry will serve as critical inputs into the future review of the 2010 regime before the expiry of the block exemption regulation governing vertical agreements (see Commission Regulation (EU) No 330/2010 on the application of Article 101(3) of the Treaty on the Functioning of the European Union to categories of vertical agreements and concerted practices [2010] OJ L102/1) in May 2022. The Commission launched that review in October 2018, see especially Consultation February–May 2019: European Commission, ‘EU Competition Rules on Vertical Agreements – Evaluation’, https://​ec​.europa​.eu/​info/​law/​better​-regulation/​initiatives/​ares​-2018​ -5068981/​public​-consultation​_en. To track the development of the review, see https://​ec​.europa​.eu/​ competition/​consultations/​2018​_vber/​index​_en​.html. 9 See, for example, Goodyear Italiana-Euram [1975] OJ L38/10; Campari [1978] OJ L70/69, para 7; Section II.A; n 152 and text. 10 Carles Esteva Mosso, ‘The Contribution of Merger Control to the Definition of Harm to Competition’ (Brussels, Global Competition Law Centre Conference, 1 February 2016) 2. 11 Until 2004, parties fearful that their agreement might infringe Article 101(1) had to either provide notification of their agreement to the Commission (which had the exclusive right to grant individual exemptions but had limited resources to grant such exemptions in practice: Regulation No 17 First Regulation implementing Articles 85 and 86 of the Treaty [1962] OJ 204/62, Article 9(1)) or draft it to fall within a block exemption. 12 The Court of Justice has confirmed that Article 101(1) applies only to agreements that both restrict competition and affect trade appreciably: see Case 5/69, Völk v Vervaecke [1969] ECR 295. 13 See Barry E Hawk, ‘System Failure: Vertical Restraints and EC Competition Law’ (1995) 32 Common Market Law Review 973.

170  Research handbook on methods and models of competition law the early BERs and the Commission’s inability to deal with all agreements notified to it for exemption, generated legal uncertainty, led to legal formalisms and analysis of agreements by category, risked the condemnation and deterrence of innocuous agreements or legitimate business practices, and, crucially, eliminated ‘what should be the heart of the matter and antitrust (i.e. economics/law) substantive analysis of a particular agreement or practice, i.e. its competitive harms and benefits)’.14 B

The Modernized Regime

1 Background and components of the modernized system The pressure eventually led the Commission to recognize that change was required. The modernization programme commenced in January 1997 with the adoption by the Commission of a Green Paper on Vertical Restraints,15 which included an economic analysis of the impact of vertical restraints on competition. Its conclusions, and the debate which followed, led to a new regime centred around a more economic, flexible, overarching BER covering all vertical restraints concerning intermediate and final goods and services that met its requirements (1999),16 the publication of guidelines on the appraisal of vertical restraints (2000),17 and the abandonment of the authorization and notification system and the removal, by Council Regulation 1/2003 (in 2004),18 of the Commission’s exclusive jurisdiction over Article 101(3). The latter developments paved the way for self-analysis by firms of their agreements and a fuller role for NCAs and the national courts in the enforcement process. The BER, now Regulation 330/201019 (Verticals Regulation), and the Vertical Guidelines (the Guidelines),20 were revised and replaced in 2010, but are currently under review. Collectively, these changes marked a major shift in Commission policy and a recognition that appraisal of vertical agreements should not be based primarily on their content and form but should also take account of their competitive effects. The Commission thus promised more realistic analysis of agreements under Article 101 based on a consumer welfare objective,21 utilizing Article 101(1) to identify anticompetitive effects in terms of parameters of

14 Ibid, 2.5: the author complained that the notification and authorization system set up by Regulation 17 [1959–1962] OJ 204/62, Article 9(1) had failed. 15 European Commission, ‘Green Paper on Vertical Restraints in EC Competition Policy’ COM(1996) 721, 22 January 1997, 721. 16 Commission Regulation (EC) No 2790/1999 on the application of Article 81(3) of the Treaty to categories of vertical agreements and concerted practices [1999] OJ L336/21. 17 Guidelines on Vertical Restraints [2000] OJ C291/1. 18 Council Regulation (EC) No 1/2003 of 16 December 2002 on the implementation of the rules on competition laid down in Articles 81 and 82 of the Treaty [2003] OJ L1/1 (Council Regulation). The Commission’s determination to follow a ‘more economic’ approach was also epitomized by the creation of the post of Chief Competition Economist in July 2003. 19 Commission Regulation (EU) No 330/2010 [2010] OJ L102/1. 20 Guidelines on Vertical Restraints [2010] OJ C130/10 (Guidelines). 21 Guidelines on the application of Article 81(3) (now 101(3)) (Article 101(3) Guidelines) [2004] OJ C101/97, paras 13, 42. Although the case law of the EU Court has not fully endorsed the approach and places greater focus on the need to protect not only the interests of consumer but also the structure of the market and the process of competition, the welfare of consumers certainly forms at least part of the EU competition law fabric: see, for example, Case C-501/06 P, GlaxoSmithKline Services Unlimited v Commission [2009] ECR I-9291, paras 61–3.

Vertical agreements under EU competition law  171 competition and Article 101(3) to weigh countervailing efficiencies. In the Guidelines, the Commission recognizes that, tested against this benchmark, many vertical agreements will not restrict competition at all: rather, competition concerns can arise only if there is insufficient competition or some degree of market power at the supplier or buyer level (or both).22 As stated in the Guidelines: ‘When a company has no market power, it can only try to increase its profits by optimizing its manufacturing or distribution processes.’23 Although the Guidelines, for the most part, take on board lessons from the economic literature and recognize that reductions in intrabrand competition are generally likely to be problematic only where interbrand competition is weak, this chapter argues that, in the detail, they fail to reflect sufficiently these broader statements on policy, and that: (1) by better aligning complementary activities (that is, manufacturing and distribution), vertical restraints tend to generate efficiencies24 that benefit parties to the agreement and end customers; (2) interbrand competition is a core driver of economic efficiency and consumer welfare; (3) firms compete for downstream customers not only on price but on quality, service, availability, after-sales support and new product development; and (4) vertical restraints that facilitate discrimination across different customer segments may, in certain circumstances, stimulate competition and ultimately increase overall welfare. In our view, the net result of these omissions is that, even after the significant modernization improvements, the Commission’s system still has problems, especially as it continues to concern itself acutely with certain (mainly intrabrand) restraints and, in some scenarios, effectively disregards efficiencies associated with them. Consequently, there is still analysis of agreements by category (see Section II.B.2) and a scarcity of modern guidance on how antitrust analysis of vertical agreements ought to be conducted (see Section II.B.3). In Section III we argue that as this current system fails to reflect fully the economic logic of vertical restraints, it risks the condemnation and deterrence of innocuous or legitimate business practices, generates legal uncertainty, and fails to elucidate how substantive antitrust analysis is to be conducted. 2 A category of vertical restraints is presumed to be incompatible with Article 101 There remains a category of vertical restraints that are strongly discouraged, which the Commission describes as presumed to be incompatible with Article 101. These restraints are ordinarily found (a) to restrict competition by object, and so are ‘assumed’ to restrict competition within the meaning of Article 101(1); and (b) to be unlikely to satisfy the Article 101(3) conditions25 – consequently, they cannot benefit from the Verticals Regulation (they are designated by the Commission as hard-core restraints). The Commission thus applies a

Guidelines, [6]. Ibid, [106]. 24 Contra ibid, [6]. 25 See, for example, Luc Peeperkorn, ‘Defining “By Object” Restrictions’ (2015) 3 Concurrences 40; Alison Jones, ‘Left Behind by Modernisation? Restrictions by Object under Article 101(1)’ (2010) 6 European Competition Journal 649. 22 23

172  Research handbook on methods and models of competition law ‘double presumption’ of illegality to hard-core restrictions, considering that such restraints are presumed to infringe Article 101(1) and not to satisfy the conditions of Article 101(3).26 (a) By object restraints are assumed to restrict competition appreciably Under Article 101, an agreement whose object or ‘precise purpose’27 ‘reveals in itself a sufficient degree of harm to competition law’28 or a ‘sufficiently deleterious impact on competition’29 is assumed to restrict competition appreciably30 and to infringe Article 101(1).31 In such cases, no assessment of anticompetitive effects is required or permitted.32 Case law stresses that to identify an agreement incorporating object restrictions, ‘regard must be had inter alia to the content of its provisions, the objectives it seeks to ascertain, and the economic and legal context of which it forms part’.33 Over the years, jurisprudence has clarified that agreements containing ‘established’ clauses are liable, in principle,34 to be found to pursue a restrictive objective. These include not only horizontal cartel agreements,35 but also certain restraints in vertical agreements, including those that: ●● involve minimum resale price maintenance (RPM – where the supplier specifies the resale price of the product at a fixed or minimum level);36 ●● confer absolute territorial protection (ATP) on a distributor or are otherwise aimed at partitioning national markets – ‘in principle, agreements aimed at prohibiting or limiting parallel trade have as their object the prevention of competition’;37 or

Guidelines, para 47; see nn 56–57 and text. Taking account of its clauses and the legal and economic context in which it operates (see further Section IV), Case C-501/06 P, GlaxoSmithKline [2009] ECR I-9291, para 58. 28 Case C-67/13 P, Groupement des cartes bancaires v Commission (CB) (General Court of the European Union, 11 September 2014) para 57. 29 Case 56/65, STM [1966] ECR 235, 249. 30 See Case C-226/11, Expedia Inc v Autorité de la concurrence (Court of Justice of the European Union, 13 December 2012); and see the Commission’s Notice on agreements of minor importance which do not appreciably restrict competition under Article 101(1). Contra Case 5/69, Völk v Vervaecke [1969] ECR 295. 31 Article 101(1) prohibits an agreement if either its object or its effect is to restrict competition: Case 56/65, STM [1966] ECR 235, 249. 32 See, for example, Case C-286/13 P, Dole Food and Dole Fresh Fruit Europe v Commission (General Court, 19 March 2015), paras 111–35; Article 101(3) Guidelines, para 22. 33 Case C-501/06 P, GlaxoSmithKline [2009] ECR I-9291, para 58. 34 See n 41 and text. 35 Horizontal collusion to fix prices, limit output, share markets or rig bids or to reduce capacity. See, for example, Case C-209/07, BIDS [2008] ECR I-8637. See also Case C-67/13 P, CB (n 28) and Case C-8/08, T-Mobile Netherlands BV v Raad van bestuur van de Nederlandse Mededingingsautoriteit [2009] ECR I-4529, paras 36–43. 36 See n 85 and text; Asus (Case AT/40.465, 26 September 2018). 37 Case C-501/06 P, GlaxoSmithKline (n 21), para 59. See also, for example, Cases 56 and 58/64, Établissements Consten SàRL & Grundig-Verkaufs-GmbH v Commission (Consten and Grundig) [1966] ECR 299 (see n 42 and text); Case C-403, Premier League Ltd v QC Leisure and Case 429/08, Murphy v Media Protection Services Ltd [2011] ECR I-09083. 26 27

Vertical agreements under EU competition law  173 ●● ban online selling (which reduces the ability of a distributor to sell outside its territory – see Pierre Fabre v Président de l’Autorité de la concurrence38) and, it seems, certain selective distribution systems39 (SDSs) (an SDS restricts the number or type of dealers and prohibits sales from authorized to non-authorized distributors). In spite of the flexible nature of the characterization exercise described above,40 the cases dealing with such restraints in distribution agreements have, save in the most exceptional circumstances,41 refused to contemplate the possibility that the context of a case supports a finding that the overarching objective is not to restrict competition but to enhance efficiency of the supply chain to the benefit of the parties and end customers.42 This category of restraints is thus ordinarily found to restrict competition by object irrespective of (1) the rationale for the incorporation of the restraint; (2) the intensity of interbrand competition; and/or (3) the degree of market power of the parties (in contradiction to broader statements about its importance made by the Commission in the Vertical Guidelines43). The approach adopted in these cases contrasts starkly, and is not easy to reconcile, with that taken in other judgments of the EU courts – for example, in Delimitis v Henninger Bräu.44 In this case, discussed further below, the Court of Justice accepted that as an obligation imposed on the café proprietor to purchase most of its beer requirements from a brewer entailed advantages for both the supplier and the reseller, the purpose/object of the agreement could not be said to restrict competition. Rather, its effects had to be considered. As the category of object restraints is not closed, a number of cases arising at the national level have raised the question of whether other vertical restraints, such as price relationship

38 See Case C-439/09, Pierre Fabre v Président de l’Autorité de la concurrence [2011] ECR-SC I-09419; n 84 and text; Case AT-40/428, Guess (17 December 2018) (in which a €40 million fine was imposed for geo-blocking, including through a prohibition of selling online without specific authorization). 39 Ibid (holding that SDSs restrict competition by object unless objectively justified); contra case C-230/16, Coty Germany GmbH v Parfümerie Akzente GmbH (Court of Justice of the European Union, 6 December 2017), where the Court did not state that an SDS restricts competition by object, but rather stated that an SDS is compatible with Article 101 insofar as it complies with certain conditions. 40 See Case C-501/06 P, GlaxoSmithKline [2009] ECR I-9291. 41 The two cases both involved licensing of intellectual property rights: Case 27/87, Erauw-Jacquéry Sprl v La Hesbignonne Société Coopérative [1988] ECR 1919; Case 62/79, Coditel v Ciné Vog Films (Coditel I) [1980] ECR 881; but see now the e-commerce reports (n 2) (expressing concern about geo-blocking in licences of digital content). This chapter does not deal with the licensing of digital content: see n 3. 42 In, for example, the early Cases 56 and 58/64, Consten and Grundig [1966] ECR 299, the Court of Justice rejected the parties’ argument that an agreement, incorporating clauses that sheltered the exclusive distributor from all intrabrand competition in France and conferring ATP upon it, was necessary to prevent free-riding and to encourage interbrand competition: see Section III below. Instead, focusing on the resulting isolation of the French market, the Court held that clauses that result in the segregation of a national market, and/or in maintaining separate national markets, were liable to restrict competition by object. 43 See Guidelines, para 6 and text. 44 Case C-234/89 [1991] ECR I-935. See also Vertical Guidelines, para 61.

174  Research handbook on methods and models of competition law agreements,45 most favoured nation clauses (MFNs),46 or clauses preventing members of an SDS selling on certain online marketplaces47 or using price comparison engines may additionally be classified as restrictive of competition by object. Some judgments of the Court of Justice may have encouraged such claims by utilizing language suggesting that object characterization might be appropriate to restraints that are likely to eliminate or seriously weaken competition48 or have the potential to have ‘a negative impact on competition’.49 In Groupement des cartes bancaires v Commission (CB),50 however, the Court of Justice stressed that because a finding that an agreement restricts competition by object exempts the Commission (or other claimant) from its ordinary burden of demonstrating a restriction of competition, the category of object restrictions is a restrictive one51 and must be confined to agreements that obviously harm the proper functioning of competition.52 Further, in Coty Germany GmbH v Parfümerie Akzente GmbH,53 the Court of Justice made it clear that a contractual clause, incorporated in a selective distribution system for luxury goods, prohibiting dealers from distributing products via third-party platforms did not restrict competition so long as it was designed to preserve the luxury image and prestige of the goods, was laid down uniformly and applied in a non-discriminatory fashion, and was proportionate in the light of the objective pursued. (b) ‘By object’ restraints are presumed not to satisfy the conditions of Article 101(3) It is frequently argued that the object category in the EU is distinct from the per se rule in the US. A finding of per se illegality does not allow any justifications for the conduct to be raised, while in contrast a finding that an agreement is restrictive of competition by object does not prevent an argument that the agreement satisfies the four criteria54 of Article 101(3).55 As, however, the EU Commission has adopted the view that vertical restraints falling in the object category, like horizontal cartel activity, constitute ‘hard-core restraints’ that are presumed

45 See, for example, the decision of the (then) Office of Fair Trading (OFT) in Tobacco (Competition Appeal Tribunal, Case CE/2596-03, 15 April 2010), which was quashed in Case Nos 1160–1165/1/1/10, Imperial Tobacco Group plc v OFT [2011] CAT 41. 46 See n 154 and text. 47 See Case C-230/16, Coty Germany GmbH v Parfümerie Akzente GmbH (Court of Justice of the European Union, 6 December 2017); see, for example, the Federal Cartel Office’s cases involving Adidas (June 2014) and Asics (August 2015). 48 Case C-32/11, Allianz Hungária Biztosító Zrt, Generali-Providencia Biztosító Zrt v Gazdasági Versenyhivatal (European Court of Justice, 14 March 2013). 49 Case C-8/08, T-Mobile [2009] ECR I-4529. 50 Case C-67/13 P (General Court of the European Union, 11 September 2014). 51 It thus held the General Court had erred in its ruling that the category was not to be interpreted restrictively. 52 See further Section IV. 53 Case C-230/16 (Court of Justice of the European Union, 6 December 2017). 54 That is, that the agreement achieves specified benefits (including ‘improving the production or distribution of goods’ or ‘promoting technical or economic progress’); that a fair share of those benefits is passed on to consumers; that the agreement does not contain any indispensable restraints; and that it does not eliminate competition in respect of a substantial part of the products in question: see Case C-501/06 P, GlaxoSmithKline [2009] ECR I-9291; see, for example, Guidelines, para 225. 55 At first sight, it thus seems closer to situations in the US where the courts adopt ‘truncated’ analysis: see, for example, National Collegiate Athletic Association v Board of Regents of the University of Oklahoma, 466 US 85 (1984).

Vertical agreements under EU competition law  175 not to satisfy the conditions of Article 101(3)56 (and so are not covered by the Verticals Regulation57), this distinction is arguably more theoretical than real. Indeed, there is extremely limited guidance as to when agreements incorporating object restraints may exceptionally satisfy the four onerous conditions of Article 101(3) in practice58 and it seems unclear how parties can prove that the beneficial effects of their agreement offset anticompetitive effects that have been assumed, not established.59 Because serious consequences may follow if an infringement is uncovered,60 parties generally avoid incorporating these object/hard-core restraints in their agreements and having to advance efficiency justifications for them even if they have no market power and perceive such restraints to be helpful, or indispensable, to the efficient distribution of their products or services.61 The design of Article 101 thus makes it extremely difficult for parties to an agreement to rebut the presumption of illegality applied to object/hard-core restraints, meaning that such restraints are generally perceived by business to be de facto illegal.62 3 Other agreements: safe harbour and individual analysis The Verticals Regulation provides a broad exemption for agreements involving parties that are considered unlikely to have market power – proxied by use of a 30 per cent market share

56 The Commission’s view is that these hard-core restraints: (a) generally fail to create objective economic benefits; (b) do not benefit consumers; and/or (c) are unlikely to be considered indispensable to the attainment of any efficiencies created by the agreement in question (efficiencies generated can ordinarily be achieved by less restrictive means – the indispensability criterion thus seems to incorporate a ‘sliding scale’, meaning that it is harder for more severe restraints to satisfy it: see Jonathan Faull and Ali Nikpay (eds), The EU Law of Competition (3rd edn, Oxford University Press 2014) 3.494). 57 Although, therefore, the foundational premise of the Verticals Regulation is that, in the absence of market power, vertical agreements can be presumed to be compatible with Article 101, it does not apply, irrespective of the parties’ market positions, to agreements incorporating ‘hard-core restrictions’ set out in Regulation 330/2010, Article 4. These include RPM (but not maximum or recommended resale prices); subject to four exceptions, restrictions concerning the territory into which, or the customers to whom, the buyer may sell (although one exception allows restrictions on active sales into the exclusive territory or to an exclusive customer group reserved to the supplier or another buyer, it does not permit restrictions on passive sales – bans on online selling are typically treated as precluding passive sales and as object/hard-core restraints according to Guidelines, para 51); restrictions on active or passive selling by authorized retail distributors in an SDS to end-users (or other authorized distributors); and certain restrictions on the sale of components as spare parts by the manufacturer. 58 It is now rare for competition agencies to conduct detailed analysis of when a vertical agreement may satisfy the conditions of Article 101(3): see Section II.B.2. 59 The pass-on condition of Article 101(3) essentially requires a balancing of the anticompetitive and procompetitive effects of the agreement. 60 Where an EU competition agency uncovers agreements incorporating object restrictions, fines may ensue: see, for example, Commission decisions in Case IV/35.733, Commission v Volkswagen [1998] OJ L124/60; Case T-62/98, Volkswagen v Commission [2000] ECR II-2707 (fine reduced); aff’d Case C-338/00 P, Volkswagen v Commission [2003] ECR I-9189; Cases IV/35.706 and IV/36.321, Nintendo [2003] OJ L255/33; aff’d (but fines reduced); Case T-13/03, Nintendo and Nintendo of Europe v Commission [2009] ECR II-975; COMP/39.154, Apple/iTunes (18 March 2008); COMP/37.975, Yamaha (16 July 2003). 61 Firms sometimes fear that the mere tabling of efficiency claims may be considered as an admission that the agreement in question contains severe restraints and is consequently unjustifiable. 62 In Sections III and IV, we argue that this approach is not justified by economic theory or experience and so requires change.

176  Research handbook on methods and models of competition law threshold63 – and which, as explained above, do not contain hard-core restraints.64 As the Regulation is directly applicable, and its benefit can only be withdrawn prospectively,65 it provides legal certainty and operates as an important mechanism for authorizing a large group of vertical agreements whose efficiencies are presumed to offset any anticompetitive effects that might arise. Where conditions of the Regulation are not satisfied, however, individual examination is required to determine whether an agreement is caught by Article 101(1) (has restrictive effects) and, if so, whether each of the four conditions under Article 101(3) is met. Under the modernized regime, this has become a question for ‘self-assessment’. As the Commission has not, post-modernization, adopted either an infringement decision involving an analysis of the restrictive effects of a vertical agreement, or a non-infringement decision, and has not published a ‘guidance letter’, firms must rely on jurisprudence that is relatively sparse (and old) and the Vertical Guidelines for such guidance. Some of this jurisprudence is not easy to reconcile with the consumer welfare approach advocated in the Vertical Guidelines. The 1966 case of STM66 clarifies that an assessment of whether an agreement has in fact restricted competition to an appreciable extent requires an appraisal of the agreement in its market context and in the light of the competition that would occur if the agreement in question had not been made. Some judgments of the EU courts – for example, the Court of Justice’s judgment in Delimitis67 – support the view that, consistent with economic analysis, the inquiry must focus on the important question of whether or not the agreement, alone or in conjunction with a network of similar agreements, would likely have an appreciable impact on the parameters of competition and allow the parties to exercise market power. The Court stressed the importance of assessing whether the agreement (a beer supply agreement obliging a café proprietor to purchase most of its beer requirements from the brewer) appreciably contributed to a foreclosure of access to the market. This required a definition of the relevant market and an assessment of whether there was a concrete possibility for new competitors to penetrate that market or for existing competitors to expand. The case thus (1) advocates an ‘effects-based’ approach to assess the competitive effects of vertical restraints under Article 101(1); and (2) establishes the importance of interbrand competition to the assessment. Nonetheless, in cases dealing with intrabrand restraints, greater weight has been attached to the importance of the structure of competition and undistorted competition in all market segments (including the distributor level) than to their impact on interbrand competition. This case law reflects a greater suspicion of restraints on rivalry between a supplier’s dealers than

63 Arguably, however, market shares provide a poor proxy for market power, especially in differentiated product markets. Indeed, the approach adopted in relation to vertical restraints is removed from the more nuanced analysis applied in relation to mergers: see especially Section IV.C below. 64 A category of ‘excluded’ restraints does not benefit from the Verticals Regulation (see Article 5), but does not necessarily prevent the agreement from benefiting from the block exemptions: see Jones and Sufrin (n 1), ch 11. 65 Council Regulation, Art 29(1)(2); Verticals Regulation, recital 15 (see also Article 6). The Commission, or an NCA, can withdraw its benefit from the agreement, which is valid and compatible until then. 66 Case 56/65, STM [1966] ECR 235, 249. 67 Case C-234/89, Delimitis v Henninger Bräu [1991] ECR I-935.

Vertical agreements under EU competition law  177 on restraints between a supplier and its competitors, and considers the former to be restrictive of competition unless necessary: ●● to facilitate the penetration of a new market by an undertaking through the prevention of free-riding on a dealer’s marketing and promotion efforts (see, for example, Société Technique Minière v Maschinenbau Ulm GmbH (STM)68); ●● to encourage non-price competition between dealers (see, for example, Metro-SB-Grossmärkte GmbH v Commission (Metro 1)69); or ●● to ensure the commercial success of a franchise agreement (Pronuptia de Paris GmbH v Pronuptia de Paris Irmgard Schillgallis (Pronuptia)70). In other words, in contrast to Delimitis (and the approach taken by the Commission when appraising the impact of vertical mergers – see Section IV below), these cases do not typically inquire into whether the agreement has had, or is likely to have, substantial or significant anticompetitive effects, or whether it affects ‘actual or potential competition to such an extent that on the relevant market negative effects on prices, output, innovation or the variety or quality of goods and services can be expected with a reasonable degree of probability’.71 Rather, it characterizes intrabrand restraints as restrictive of competition, unless objectively necessary and proportionate, irrespective of whether or not the parties have market power. Accordingly, this approach seems to conflate object and effect72 and Article 101(1) and 101(3) analysis,73 as well as missing a critical step in the appraisal – namely, the identification of the agreement’s likely impact on competition.74 4 What policy drives the current approach? Rather than an articulated concern about diminished interbrand competition or a high probability of negative effects on prices,75 the jurisprudence dealing with intrabrand restraints seems to be influenced, partly at least, by an inherent anxiety about restraints on rivalry between dealers

Case 56/65, STM [1966] ECR 235. Case 26/76, Metro 1 [1977] ECR 1875, para 21. But see also Case C-439/09, Pierre Fabre GmbH (Court of Justice of the European Union, 6 December 2017), n 39 and text. 70 Case 161/84, Pronuptia [1986] ECR 353. 71 Article 101(3) Guidelines, paras 24–6. 72 The objective necessity function thus seems to perform a combined classification and truncated analysis function. If the restraints are not objectively necessary, they are assumed to restrict competition (they are restrictive of competition by object). If they are objectively necessary, however, they do not restrict competition (by object or effect). The cases do not provide clear guidance, however, as to how it can be determined whether a restraint is objectively necessary to an agreement, especially as, as has been seen, RPM and ATP are almost never considered to be required to achieve a legitimate purpose such as the penetration of a new market and the elimination of free-riding. 73 The appraisal does seem to involve some loose form of balancing, or at least consideration, of the agreement’s potential harms and benefits, so duplicating, or at least overlapping, with the analysis that the Commission states in its Guidelines is reserved for Article 101(3). Indeed, in some cases, the balancing conducted under Article 101(1) closely resembles that conducted in free movement cases under the ‘EU rule of reason’: see, for example, Cases C-403 and 429/08, Murphy [2011] ECR I-09083 and Giorgio Monti, ‘Restraints on Selective Distribution Agreements’ (2013) 36 World Competition 489. 74 See, for example, Monti (n 73). 75 See Case C-67/13 P; see also discussion at nn 183–189 and text. 68 69

178  Research handbook on methods and models of competition law (given the ‘considerable impact of distribution costs on the aggregate cost price’76), particularly those that impact on the single market, through prohibiting or limiting the opportunities for parallel or cross-border trade77 and perpetuating price differences78 between Member States. The internal market objective thus ‘adds an extra dimension to the analysis of vertical restraints’79 and means that arguments that have been found persuasive and decisive in cases such as Delimits have been disregarded in cases involving intrabrand restraints, making cross-border resale and price arbitrage more difficult.80 Indeed, the Commission has stressed that companies should not ordinarily be allowed to create barriers to trade between Member States, to seal off territories ‘hermetically …, making interpenetration of national markets impossible’81 and so contribute to the cost of ‘not realising’ the EU’s single market objectives.82 The Court’s judgment in Pierre Fabre,83 confirming that online selling constitutes a form of passive (not active) selling that can be restricted only in exceptional circumstances, also supports the Commission’s view that, generally, every distributor should be allowed to use the internet to sell its products: ‘the promotion of online sales is extremely important for the internal market in Europe because it broadens the market, improves the choices for customers, and generally speaking, enhances competition’.84 The rationale behind the cases holding that a complete elimination of intrabrand price competition between dealers is restrictive of competition by object85 is not as clearly spelt out in the Court’s judgments, but may be driven by concerns that price restraints can lead to segmented markets and increased scope for price discrimination or differentiation between Member States.86 The Vertical Guidelines also seek to provide an economic justification for the approach to RPM. For example, the Guidelines state that RPM should be characterized as a hard-core restraint (subject to the double presumption of illegality) because it can restrict competition in a number of ways, including through facilitating collusion between suppliers and enhancing price transparency; by eliminating intrabrand price competition; by softening competition between suppliers and retailers; by preventing distributors from lowering sale prices; by lowering pressure on manufacturers’ margins; by foreclosing competitors; and Cases 56 and 58/64, Consten and Grundig [1966] ECR 299, 342–3. Case C-501/06 P, GlaxoSmithKline [2009] ECR I-9291, para 59. 78 See Guidelines, para 224. 79 Green Paper on Vertical Restraints, para 70; see European Union, ‘Sector Inquiry into E-Commerce’ (n 2); see also Zsolt Patki, ‘The Cost of Non-Europe in the Single Market: “Cecchini Revisited” – An Overview of the Potential Economic Gains from Further Completion of the European Single Market’ (European Parliamentary Research Service, September 2014). 80 Case C-403, Premier League and Case 429/08, Murphy [2011] ECR I-09083. 81 Nintendo [2004] OJ L255/33, para 338. 82 Green Paper on Vertical Restraints, para 78: ‘[t]he elimination of barriers to trade may not achieve its objective if producers and/or distributors introduce practices contrary to integration’. 83 Case C-439/09, Pierre Fabre v Président de l’Autorité de la concurrence [2011] ECR-SC I-09419. 84 Interview with Dr Alexander Italianer, Director General for Competition, European Commission (theantitrustsource, April 2011) 1, 6; Guidelines, paras 52–4. 85 See especially Case 161/84, Pronuptia de Paris v Schillgallis [1986] ECR 353, para 25; Case 26/76, Metro-SB and Case 243/83, SA Binon & Cie v SA Agence et Messageries de la Presse [1985] ECR 2015, para 44. 86 Unless the supplier is dominant, charging different prices to different customers, based on location, customer characteristics, or the distribution channel, is usually a manifestation of competition working well. It may enhance market integration since it incentivizes market expansion and entry in new markets even if customers have a lower willingness to pay or if competition is already intense. 76 77

Vertical agreements under EU competition law  179 by reducing dynamism and innovation at the distribution level.87 Indeed, in 2018,88 in four separate decisions condemning strategies pursued individually by consumer electronics manufacturers to ensure that retailers within SDSs, selling online, adhered to resale prices, the Commission held that RPM ‘may be considered so likely to have negative effects, in particular on the price, choice, quantity or quality of the goods and services, that it may be considered redundant, for the purposes of applying Article 101(1) … to prove that they have actual effects on the market’ – the conduct, by its very nature, restricts competition.89 The close monitoring of retailers’ resale prices through software meant that the manufacturers could avoid price erosion across, potentially, the entire (online) retail network. Section III argues, however, that although these may constitute valid theories of harm in some cases, the impact of RPM depends on the overall market circumstances.

III

PROBLEMS STEMMING FROM THE CURRENT FRAMEWORK

A Overview Although the modernization programme made great leaps, Section II above revealed that the general apprehension about intrabrand restraints means that analysis of such restraints has not developed to focus on the theory of harm, interbrand competition or the efficiencies that might be generated by vertical agreements through aligning the parties’ complementary activities. In Section III.B below, we argue that this approach fails to acknowledge that, by aligning supplier and dealer incentives, vertical intrabrand restraints frequently generate efficiencies that restrictions on intrabrand competition will not necessarily weaken interbrand competition (but rather may strengthen it), that problems stemming from weak interbrand competition cannot generally be solved by increasing intrabrand competition, and that consumer demand depends not just on price but on other factors (such as service, quality and image).90 Although anticompetitive theories for such restraints undoubtedly exist, they need specific circumstances to arise. Section III.C maintains that, consequently, the hard-line approach to object/hard-core creates a danger of Type I errors. Further, the existing jurisprudence dealing with other restraints does not provide a satisfactory legal framework for assessing the competitive benefits and harms of vertical restraints. EU policy towards such restraints may consequently be harming consumer welfare and efficiency, a core common objective of both the competition and the internal market rules.91

Guidelines, para 224. See, for example, Asus (Case AT/40.465, 26 September 2018). 89 Ibid, paras 105–7. 90 See, for example, Edward Iacobucci and Ralph A Winter, ‘European Law on Selective Distribution and Internet Sales: An Economic Perspective’ (2016) 81(1) Antitrust Law Journal 47. 91 See Article 101(3) Guidelines, para 13: ‘the creation and preservation of an open single market promotes an efficient allocation of resources throughout the Community for the benefit of consumers’. 87 88

180  Research handbook on methods and models of competition law B

The Economic Logic of Vertical Restraints

1 Aligning supplier and dealer incentives The widespread adoption of vertical restraints (where legal), by suppliers and distributors that do not have market power suggests that in many cases the objective pursued is to resolve supply chain inefficiencies rather than to restrict competition. Some older surveys indicate that the most common restraint is RPM (when not prohibited by law).92 A more recent report prepared by Oxera and Accent93 demonstrates that in the UK, businesses with or without market power employ a wide range of vertical restraints,94 the most common being selective and exclusive distribution agreements (which often exclude online retailers or platforms). In addition, many manufacturers specify recommended retail prices to their retailers, in some cases accompanied by recommended ranges of discounts.95 The Commission’s e-commerce sector inquiry also reveals that suppliers are now increasingly using a selective distribution model.96 This raises the question of why firms, in particular those with little or no market power, would wish to incorporate such vertical restraints in an agreement, especially if the manufacturer wishes to sell as much of its product as possible. The key answer is that restraints within a supply chain provide the opportunity to eliminate production or distribution inefficiencies in the chain, to deal with externalities and align suppliers’ and retailers’ incentives. Another way to achieve these aims, and internalize conflicting objectives, is through vertical integration (either organically or through merger).97 Vertical restraints, however, offer the prospect of achieving these benefits through contracting, while maintaining independence to contract with multiple parties.98 This facilitates greater specialization and the opportunity to reap economies

92 See, for example, Frederic M Scherer and David Ross, Industrial Market Structure and Economic Performance (3rd edn, Houghton Mifflin 1990) 549; Thomas R Overstreet Jr, Resale Price Maintenance: Economic Theories and Empirical Evidence (Bureau of Economic Staff Report to the Fair Trade Commission, November 1983) 153, 155. 93 Oxera Consulting LLP and Accent, Vertical Restraints: New Evidence from a Business Survey (Report, 24 March 2016). 94 Presumably chosen to comply with EU competition law rules. 95 Some manufacturers merely specify RRP, while others take active measures to ensure that the retailers price at or close to the RRP (in terms of level of prices or ranges of discounts). 96 See Section III.B.1. 97 Consistent with mainstream economic thinking, the Commission recognizes in its Guidelines on the assessment of non-horizontal mergers under the Council Regulation on the control of concentrations between undertakings [2008] OJ C265/6 that, in the absence of market power, vertical integration does not generally harm competition or consumers; and, even if one or both merging parties enjoy significant market power and there is a risk that the merger will result in foreclosure or facilitate collusion, efficiencies resulting from the alignment of supplier/dealer incentives can offset anticompetitive effects and ultimately benefit consumers: see Section IV.C. 98 As a result, the risks of anticompetitive effects arising from vertical restraints are lower than in the case of vertical integration. In addition, a vertical merger is typically irreversible. In contrast, vertical restraints can be renegotiated as demand or supply circumstances change, promoting dynamic competition. Irreversibility can exacerbate anticompetitive foreclosure effects – for example, if it facilitates the merging parties to commit not to supply an input to a competing downstream dealer by adapting the input to the specific requirements of its integrated downstream division.

Vertical agreements under EU competition law  181 of scale, scope and learning effects at each level of the supply chain. In turn, this leads to greater social welfare and more dynamic competition.99 Although vertical restraints may harm competition in certain circumstances,100 in many situations, suppliers have no incentive to restrict the scope of their distribution networks.101 On the contrary, suppliers ordinarily want dealers to take actions that would expand demand and steal customers from their competitors (thereby promoting interbrand competition), for example through lowering retail prices, investing to reduce the costs of distributing or selling a given supplier’s specific brand, or taking measures that improve the product’s or service’s quality or attractiveness for consumers. For example, retailers can raise consumer demand for a product and expand sales by offering services that improve information, convenience or quality for end customers. These activities are, however, costly. Although the benefits of retailers’ services accrue to both the retailer and the supplier (who makes a positive margin when retail sales expand), the costs of service may be borne solely by the retailer. It follows that, when choosing the optimal amount of service effort, the retailer will only take into account its own benefits and will ignore the benefits of increased sales for the upstream supplier from the expansion of retail sales. Consequently, retailers, in the absence of vertical restraints, are likely to underinvest in activities that would increase demand for the supplier’s brand. Economists have considered whether the supplier cannot simply compensate for the retailer’s relationship-specific investments. However, in practice, these investments may not be directly observable or verifiable by the supplier and cannot be enforced contractually. Vertical restraints thus provide the opportunity to ensure that both contractual parties invest optimally in the relationship. The procompetitive effects of vertical restraints thus rest on two propositions. First, a simple price contract may leave retailers with inadequate incentives to provide sales and service effort in their various dimensions. Second, contracts with vertical restraints

99 Different vertical contracting practices can sometimes be used for addressing a given inefficiency in vertical relationships. In practice, however, depending on the fine details of the situation at hand, perceived risks and market conditions, in some scenarios one particular vertical restraint is likely to be optimal. 100 See Section III.B.3. 101 See, for example, Pauline M Ippolito, ‘Resale Price Maintenance: Empirical Evidence from Litigation’ (1991) 34 Journal of Law & Economics 263; Phillip L Hersch, ‘The Effects of Resale Price Maintenance on Shareholder Wealth: The Consequences of Schwegmann’ (1994) 42 Journal of Industrial Economics 205; James C Cooper, Luke M Froeb, Dan O’Brien and Michael G Vita, ‘Vertical Antitrust Policy as a Problem of Inference’ (2005) 23 International Journal of Industrial Organization 639; Pauline M Ippolito and Thomas R Overstreet Jr, ‘Resale Price Maintenances: An Economic Assessment of the Federal Trade Commission’s Case against the Corning Glass Works’ (1996) 39 Journal of Law & Economics 285; Shantanu Dutta, Jan B Heide and Mark Bergen, ‘Vertical Territorial Restrictions and Public Policy: Theories and Industry Evidence’ (1999) 63 Journal of Marketing 121, 122 (‘our results suggest that efficiency arguments should play an important role in the public policy debate on vertical restraints’); Francine Lafontaine and Margaret Slade, ‘Exclusive Contracts and Vertical Restraints: Empirical Evidence and Public Policy’ in Paolo Buccirossi (ed), Handbook of Antitrust Economics (MIT Press 2008); Vertical Restraints: New Evidence from a Business Survey (n 93); Dennis W Carlton and Judith A Chevalier, ‘Free Riding and Sales Strategies for the Internet’ (2001) 49 Journal of Industrial Economics 441; but cf Alexander MacKay and David Aron Smith, ‘The Empirical Effects of Minimum Resale Price Maintenance’ (16 June 2014), http://​alexandermackay​.org/​ files/​The​%20Empirical​%20Effects​%20of​%20MRPM​.pdf; see nn 174–175 and text.

182  Research handbook on methods and models of competition law can restore or at least enhance and align those incentives.102 Vertical restraints may therefore be intended to replicate contractually the efficiency-enhancing effects of a vertical merger and to achieve an efficient price and output level and be necessary: ●● to internalize pricing inefficiencies and eliminate a double-marginalization problem;103 ●● to encourage relationship-specific investments (for example, in brand image)104 by avoiding hold-up by the party with relatively greater bargaining power; and/or ●● to ensure that retailers have optimal incentives to invest in services necessary to encourage consumer demand for a product, to launch a new product,105 to signal the quality of a product,106 or to build brand reputation without risk of other retailers free-riding on their efforts.

102 See, for example, Lester G Telser, ‘Why Should Manufacturers Want Fair Trade?’ (1960) 3 Journal of Law & Economics 86; Richard A Posner, ‘The Next Step in the Antitrust Treatment of Restricted Distribution: Per Se Legality’ (1981) 48 University of Chicago Law Review 1; Frank H Easterbrook, ‘Vertical Arrangements and the Rule of Reason’ (1984) 53 Antitrust Law Journal 135; Barry Nalebuff, Bundling, Tying, and Portfolio Effects (DTI Economics Paper No 1, 2003); Benjamin Klein and Kevin M Murphy, ‘Vertical Restraints as Contract Enforcement Mechanisms’ (2008) 31 Journal of Law & Economics 265. 103 Various types of vertical restraints, including maximum RPM (that is, price ceilings) and two-part tariffs (by introducing alternative means of sharing the profits), can be used to increase efficiency by removing the coordination failure in the more exceptional circumstance that occurs when a manufacturer and its retailer both enjoy market power. In the absence of such restraints, both parties will exploit their market power, ignoring the impact of this price increase on the other party’s profit, leading to retail prices that are not only above costs but also above the desirable level for the vertical structure as a whole. 104 Contracts are generally incomplete and future actions by either party may not be observable. This means it is not possible to cover all contingencies that may occur or to condition all relevant decisions to be taken by each party (on price, quantity, product characteristics, etc) on verifiable variables (including, possibly, announcements by the parties, for example concerning their valuations, costs, etc). Moreover, there are many investments which lose most of their value outside a particular relationship because they are tailored and dedicated to a particular partner (for example, a franchise devotes important investments to carry and promote a particular brand or a firm may design its machinery to work with a particular intermediate good or input). In such cases, the risk that the relationship may break down will generally lead to an underinvestment problem. If a retailer fears that its promotion efforts to establish a brand’s image might next year benefit another shop located in the same area and carrying the same brand, it will think twice before investing heavily in such an activity. Likewise, a manufacturer will be deterred from investing in assets which might improve a distributor’s performance if the latter is likely to switch to other brands. Vertical restraints, for example exclusive territories and exclusive dealing, mitigate the risk of such opportunistic behaviour (for example, a firm getting out of the relationship after the partner has made specific investments in it) and the resulting underinvestment. 105 When a manufacturer requests a distributor to carry a new product line, it exposes the distributor to a certain amount of risk, for example with respect to the necessary market development effort. By deploying territorial restrictions or using an SDS, the manufacturer may be able to reduce this risk and ensure that the risk is allocated optimally between suppliers and distributors. This increases the chances that new products will be developed in the first place. 106 Customers may have only imperfect information regarding the quality or other unobservable characteristics of a product, especially in the case of new products. In that case, the identity or specific actions of the retailer may provide a useful signal. For example, a supplier may prefer to distribute through a store which has a reputation for stocking high-quality products. This implicit certification activity by retailers involves costs (for example, locating the store in an upmarket district of town and employing suitable assistants). Hence, retailers may be unwilling to stock products if they are being sold at lower prices by outlets, for example discount outlets or internet distributors, which would undermine the exclusivity and

Vertical agreements under EU competition law  183 For example, retailers can raise consumer demand for a product and expand sales by offering services that improve information, convenience or quality for end customers. A rich literature explains that a number of vertical restraints, including RPM, exclusive dealing, territorial restrictions or selective distribution, may be used as mechanisms to allow suppliers to elicit optimal service levels (for example, pre-sale services, quality certification and the building of brand reputation).107 The risk of free-riding is particularly high where retailers cannot separate aspects of retail service that build demand for the manufacturer’s product from other retail activities and/or they cannot ‘sell’ the former to consumers or the manufacturer on a stand-alone basis. Often, transaction costs as well as economies of scope in retailing prevent the separation and sale of brand-specific retail services. Vertical restraints may also be designed to prevent retailers from lowering retail prices in circumstances where they simply attract customers from competing retailers, but without increasing output or the brand manufacturer’s profits.108 The importance of vertical restraints as a way to align supplier and dealer incentives has increased substantially in the internet age, where e-commerce has rapidly transformed distribution and retailing methods. Numerous products are now sold online or through platforms, and price comparison websites exist to facilitate consumers’ buying choices. The growth of e-commerce not only facilitates market access,109 it also increases price transparency and lowers search costs, enabling customers instantaneously to obtain and compare product and price information online, and to switch swiftly from one channel (online/offline) to another. Nonetheless, e-commerce also creates new challenges to ensure that manufacturers’ and retailers’ incentives remain aligned. In-store services are becoming less important as e-commerce develops and consumers invest time online to learn about the product. In some situations, however, consumers continue to rely on retail services – for example in the case of complex products, experience goods or one-off purchases of durable goods. In such cases the risk of free-riding can be exacerbated by e-commerce110 – a consumer can simply visit

quality image they are promoting. This argument may justify distributing only through ‘selected’ stores, which meet specified quality criteria, and refusing to supply the product to discount stores, supermarkets or certain internet distributors. 107 See, for example, nn 101–102. In their absence, retailers may choose a service level which is lower than optimal for both the supplier and end consumers. This is because when choosing the optimal amount of service effort, the retailer will only take into account its own benefits and will ignore the benefits of increased sales for the upstream supplier from the expansion of retail sales. Although the benefits of retailers’ services accrue to both the retailer and the supplier (who makes a positive margin when retail sales expand), the costs of service may be borne solely by the retailer. In addition, retailer efforts may not be directly observable or verifiable by the supplier and cannot be enforced contractually. 108 Although the manufacturer may be indifferent to where the purchase takes place, each individual retailer might be discouraged from investing in promotional activities. This typically hurts both the manufacturer and also consumer welfare. 109 For example, smaller retailers may, with limited investment efforts, become visible and sell products to a large customer base and in multiple Member States through marketplaces. 110 See, for example, Vertical Restraints: New Evidence from a Business Survey (n 93) (finding that vertical restraints have offered ways to manage the competitive impact of increased e-commerce in many sectors, in particular by preventing free-riding and maintaining the pre-sale and after-sales service quality and protecting brand image); Dennis W Carlton and Judith A Chevalier, ‘Free Riding and Sales Strategies for the Internet’ (2001) 49 Journal of Industrial Economics 441.

184  Research handbook on methods and models of competition law a brick-and-mortar store to ask questions, to try or test a product, and then make the purchase online.111 Although internet distributors may also provide extensive product information and customer reviews on their websites,112 two factors make free-riding by internet retailers a potentially greater problem than free-riding in the reverse direction. First, much of the effort of the brick-and-mortar retailers takes the form of a per customer cost, while online retailers are more likely to incur fixed costs in providing support. Second, because the promotional effort of a brick-and-mortar retailer consists of personal interaction between customers and sales consultants, it is much more difficult to verify it directly.113 The Commission’s Final Report on e-commerce114 acknowledges that although price is considered as the key competitive factor among retailers, the importance of quality and brand image considerations is stressed by manufacturers as the most important feature of interbrand competition. Thus, the report confirms that e-commerce presents renewed challenges for manufacturers to ensure that they can keep control over the image, reputation and positioning of their brand in the mid to long term and maintain a consistent quality of pre- and after-sales service. These developments and challenges are fuelling new distribution practices and restraints relating to internet sales, for example online RPM, dual pricing practices, MFNs, territorial and/or cross-border sales restrictions (geo-blocking) and marketplace (platform) bans, restrictions on the use of price comparison tools, and the exclusion of pure online players from the distribution network.115 Although much remains to be learned about the competitive effects of such arrangements, it is clearly possible that these mechanisms seek to protect brand image, combat counterfeiting, limit the effect of free-riding, and encourage investment in the provision of customer services. For example, a manufacturer that constrains the percentage of sales that can be made via the internet may simply be adjusting the mix of its price and non-price demand variables (image investment, for example), just as if it were adjusting these variables directly. The Commission’s Final Report recognizes that e-commerce is having a profound effect on manufacturers’ distribution strategies: ●● First, more manufacturers are integrating vertically into online distribution and so are competing with their independent distributors. E-commerce (1) significantly lowers the transaction costs of reaching a wide customer base; (2) facilitates adapting and communicating new prices to customers in response to changing competitive conditions; and (3) gives the manufacturer greater control over brand image. According to the Commission’s Final Report, 64 per cent of manufacturers have launched their own websites within the

Commission’s Final Report (n 2), para 11. For example, a consumer may compare smartphone prices and characteristics in an online retailer but then purchase the preferred brand in a nearby physical shop. 113 In addition, the manufacturer could use fixed fees to compensate online retailers who provide information or promotion-laden sites. Fixed fees cannot be effectively used to compensate brick-and-mortar retailers for the effort they dedicate to each customer. 114 Commission’s Final Report (n 2). 115 See n 154 and text. 111 112

Vertical agreements under EU competition law  185

●●

●● ●● ●●

last ten years and 3 per cent reported that they took the decision to fully integrate distribution activities.116 Second, SDSs are used more widely. Both the number of SDSs and the use of selection criteria have increased significantly over the last ten years. These allow manufacturers to control distribution more effectively by, in particular, (1) maintaining a coherent brand image across offline and online sales; (2) avoiding free-riding between the two channels; and (3) offering a certain degree of protection against the sale of counterfeit products. Third, it acknowledges the increasing use being made of the vertical restraints outlined above117 that allow greater control over the distribution of products. Fourth, manufacturers are increasingly pursuing a multi-channel strategy, using both offline and online distribution mechanisms. Fifth, sellers may sell, and purchasers may buy, through brick-and-mortar stores, manufacturer or retailer websites, third-party platforms or marketplaces (such as Amazon or eBay), and/or price comparison websites.

For differentiated products, in particular durable goods that represent an important share of consumer expenditure, manufacturers want to ensure that technical features, innovations and other important product characteristics are appropriately explained and presented to consumers. The environment for such services can be created both in brick-and-mortar shops and online. Consumers nowadays make use of both channels and switch between channels to collect information, compare offers and ultimately make a purchase. An increased proportion of sales is generated via the manufacturers’ own retail activities, both online and offline. These developments suggest that the relative importance of intrabrand versus interbrand competition is evolving. Manufacturers are seeking to tighten their control over the reputation of their brands by integrating vertically into online sales and using SDSs. These market developments tend to reduce intrabrand competition but intensify interbrand competition. 2 Even ‘object/hard-core’ restraints have procompetitive potential Procompetitive justifications may be the driving economic motivation even for restraints on intrabrand competition, identified in the EU as object or hard-core restraints. Indeed, in many cases, a supplier is unlikely, unilaterally, to want to shield its dealers from intrabrand competition through territorial or price restraints unless it wishes to encourage sales and service efforts by dealers. Otherwise, the effect would be to reduce sales in the downstream market and reduce supplier profits.118 For example, SDSs and a number of vertical restraints may be required to deal with the problem that retail promotional activities confer a positive externality on both suppliers and other retailers. Without such restraints, a given retailer will likely invest too little in promotion, both because other retailers and the supplier will free-ride on its efforts and because it can itself 116 Vertical integration is not always an efficient option for businesses (especially small businesses) that want to expand into new markets and reach a wide customer base, however. Further, firms are likely to benefit from the support of online distribution platforms. It would typically make no economic sense for a small supplier to vertically integrate into distribution and set up its own e-commerce platform. 117 See n 154 and text. 118 Although, therefore, dealers are protected from intrabrand competition, their ability to raise retail prices or skimp in the provision of services will be constrained if competition at the manufacturer level is sufficiently strong.

186  Research handbook on methods and models of competition law free-ride on the efforts of others. Another way to induce retailers to provide optimal levels of service, sales effort, or advertising for the vertical chain and to ensure an optimal density number and distribution of retailers is to establish a minimum sales price. In particular, RPM may incentivize distributors to increase promotion efforts (for example, to introduce a new brand or to enter a new market) or to provide services at the point of sale (for example, for complex and/or high-quality products)119 where it is too difficult or costly for a manufacturer to exactly specify in a contract what a distributor should do to increase sales. With retail price competition for the product eliminated, retailers compete with each other for sales by offering valuable retail service to consumers. Retailers may use RPM-protected margins to safeguard investment in promotion and retail services store reputation. Exclusive rights of sale in a given territory to a single dealer, or a group of dealers, may also allow the manufacturer to ensure that dealers in different areas do not free-ride on another dealer’s investment. Territorial exclusivity also induces the optimal level of investment in the relationship with the supplier and may allow the manufacturer to certify its product quality by granting territorial exclusivity to a retailer with the desired reputation. Territorial exclusivity can be reinforced through contractual provisions preventing other retailers and/or the manufacturer from selling into it – whether through active or passive selling. The dichotomy drawn in the EU between active and passive selling (or qualified and absolute territorial protection)120 detracts from the core issue which should be how these restraints impact on competition. From an economic perspective the form of the ‘sale’ makes little difference. Absent supply chain inefficiencies, a supplier ordinarily benefits from competition between retailers and is likely to want them to compete and rely on all available distribution channels, including the internet. Indeed the Commission’s Final Report in its e-commerce sector inquiry121 acknowledges that there may be procompetitive motives for online sales restrictions. 3 Theories of harm Despite their propensity to produce efficiencies, there is concern that efficiency justifications may sometimes be exaggerated and may only apply in specific circumstances. Further, it is uncontroversial that vertical restraints may have horizontal effects by restricting intrabrand competition or interbrand competition, creating conditions in which the intensity of competition is reduced, and collusion (upstream or downstream) is facilitated, or anticompetitive foreclosure occurs. Further, if alterations resulting from vertical restraints ‘fail to reflect the preferences of infra-marginal consumers, the interests of consumers in general may not be served’.122 Economic concerns about RPM, for example, derive from the fact that it gives the manufacturer power over retail price, eliminates intrabrand price competition between retailers, and sets a price floor on resale prices – that is to say, without RPM, prices would be lower. In addition, theory indicates that RPM123 may facilitate collusion – either at supplier or dealer 119 See, for example, the Australian Competition and Consumer Commission (ACCC) determination in Tooltechnic Systems (Aust) Pty Ltd (Authorisation No A91433, 5 December 2014). 120 See n 57 and text. 121 See Commission’s Final Report (n 2). 122 William S Comanor, ‘Vertical Price-Fixing, Vertical Market Restrictions, and the New Antitrust Policy’ (1985) 98 Harvard Law Review 983, 991; see, for example, Tooltechnic (n 119). 123 See, for example, Resale Price Maintenance (n 92); Robert Pitofsky, ‘In Defense of Discounters: The No-Frills Case for a Per Se Rule against Vertical Price Fixing’ (1983) 71 Georgetown Law Journal

Vertical agreements under EU competition law  187 level – or lead to foreclosure of suppliers,124 or innovative and discounting retailers.125 For example, RPM may make manufacturer collusion more stable by making it easier to detect deviations from a collusive strategy and reducing incentives for deviation;126 retailers may induce a manufacturer to ‘impose’ RPM, to support their collusion, and to prevent deviations from the strategy;127 a supplier may use RPM to exclude competitors from the market by offering retailers a higher margin as a ‘payment’ for their willingness not to deal with competitors; or a retailer may demand RPM from its supplier(s) to exclude discounting competitors or new entrants from their market. Similarly, territorial restraints or restraints on online selling may have anticompetitive consequences through exclusion or collusion: a retailer that is well established may obtain territorial exclusivity and prevent competing retailers (or online retailers) from gaining access to a key good; such restraints may facilitate collusion between suppliers (through market sharing) and/or soften upstream competition between them,128 or facilitate collusion at the retail level.129

1487; Comanor (n 122); Daniel O’Brien and Greg Shaffer, ‘Vertical Control with Bilateral Contracts’ (1992) 23 RAND Journal of Economics 299; Patrick Rey and Thibaud Vergé, Resale Price Maintenance and Horizontal Cartel (CMPO Discussion Paper 02/047, July 2004); Patrick Rey and Thibaud Vergé, ‘The Economics of Vertical Restraints’ in Paolo Buccirossi (ed), Handbook of Antitrust Economics (MIT Press 2008); Bruno Jullien and Patrick Rey, ‘Resale Price Maintenance and Collusion’ (2007) 38 RAND Journal of Economics 983; amici curiae brief filed by 37 states supporting the respondent before the Supreme Court in Leegin Creative Leather Products v Diaz, 33 Cal Rptr 3d 139 (Cal 2005); Luc Peeperkorn, ‘RPM and Its Alleged Efficiencies’ (2008) 4 European Competition Journal 1; Kenneth G Elzinga and David E Mills, ‘The Economics of Resale Price Maintenance’ (2008) Issues in Competition Law and Policy 1841; Greg Shaffer, ‘Anticompetitive Effects of RPM (Resale Price Maintenance) Agreements in Fragmented Markets’ (Report, Office of Fair Trade, February 2013); John Asker and Heski Bar-Isaac, ‘Exclusionary Minimum Resale Price Maintenance’ (NBER Working Paper No 16564, December 2010); and Amelia Fletcher and Morten Hviid, ‘Broad Retail Price MFN Clauses: Are They RPM “at Its Worst”’ (2016) 81 Antitrust Law Journal 1; Guidelines, para 224. 124 Customer foreclosure takes place when the vertical restraint is likely to foreclose upstream rivals by restricting their access to a sufficient customer base. 125 Input foreclosure takes place when vertical restraint is likely to raise the costs of downstream rivals by restricting their access to an important input. 126 Cartels are less stable if cartel members cannot observe wholesale prices and retail prices fluctuate – it is hard to distinguish between fluctuations due to cost changes in downstream markets and those due to individual deviations by cartel members. RPM eliminates this retail price deviation and thus creates stability. 127 See, for example, Case 1022/1/1/03, JJB Sports plc v Office of Fair Trading [2004] CAT 17. 128 Suppose, for example, that competing manufacturers distribute their products through distinct retail networks. If they maintain strong intrabrand competition within their retail networks, then the retail price of each product will closely reflect the evolution of the wholesale price for that product; as a result, the situation resembles one of direct face-to-face competition between the manufacturers. If, instead, manufacturers reduce intrabrand competition by, for example, assigning exclusive territories to their distributors, these distributors will have more freedom for setting their prices: typically, the retail price for one product will then (at least partially) respond to increases in rival manufacturers’ wholesale prices, thereby encouraging the retailers to raise their prices. In sum, assigning exclusive territories to retailers could be an effective way for a manufacturer to commit itself to a ‘friendlier’ behaviour vis-à-vis its rivals, which may encourage that rival to charge higher prices. 129 In an extreme case, with a monopoly manufacturer and a set of retailers who, absent the territorial restrictions, would otherwise compete, territorial restrictions could enforce a market division arrangement similar to explicit collusion between the retailers.

188  Research handbook on methods and models of competition law When considering theories of harm it is important, however, to recall the crucial differences between horizontal130 and vertical restraints. In particular, it has been seen that the latter may align incentives between firms at different levels of the supply chain, increase dealer services and demand for the supplier’s product and strengthen interbrand competition, so enhancing the ability and incentives of the parties to compete. Horizontal restraints, in contrast, align incentives between firms that compete with each other at the same level of the supply chain and often in the same market and consequently tend directly to restrict interbrand competition between rival firms and increase market power. Further, in assessing the correct mechanism for antitrust analysis, the likelihood of competitive harm should be considered.131 This will depend on the circumstances of the case. Take, for example, the case of a supplier whose brand has 5 per cent of the downstream market and which imposes RPM on its distributors to try to encourage the provision of services that will expand demand for its product. Even if RPM eliminates intrabrand price competition between all retailers, that specific brand still competes with other brands owned by rival suppliers. If there are, for example, 19 other suppliers, each accounting for 5 per cent of the market, no supplier is likely to profit directly from unilaterally using RPM to increase prices: existing customers can switch to any of the other 19 brands. In such circumstances, the price restraint would not generally injure competition or consumers. On the contrary, in the absence of market power, in all likelihood RPM is intended to achieve one of the procompetitive objectives discussed above.132 Even in the case of ATP and RPM, therefore, theories of harm may require careful analysis. Whether such anticompetitive effects are possible, or merely theoretical, depends upon the specific circumstances of the case, the degree of interbrand competition, the nature and scope of the restraint in question, whether restraints are dealer or retailer led, whether similar restraints are used by other firms operating in the market, and the availability of alternative means to internalize vertical inefficiencies, including vertical integration.133 For example, RPM will only work as a foreclosure strategy by a supplier if the retailers bound to the manufacturer via RPM agreements comprise a sufficiently large share of the relevant market. The theory of harm is untenable where upstream competitors and new entrants

But see n 136 and text. Despite the risk that horizontal restraints will increase market power, competition agencies and courts generally recognize that, with the exception of hard-core cartels, horizontal restraints (like horizontal mergers) may generate offsetting efficiencies, which should be weighed against any anticompetitive effects identified: see nn 179–185 and text. 132 A policy that prohibits a single supplier with 5 per cent market share from unilaterally resorting to whatever vertical restraint it considers most appropriate to achieve any such efficiencies may therefore chill interbrand competition and hurt consumers: see Sections III.C and IV. Indeed, assume that the other 19 suppliers in this example are vertically integrated and the unintegrated supplier decides that, rather than using RPM, it will acquire all its dedicated distributors to control retail pricing. It seems unlikely that any competition authority would challenge, let alone prohibit, such a merger, irrespective of whether the supplier claims efficiencies. 133 Vertical Restraints: New Evidence from a Business Survey (n 93), noting that the impact on consumers from having vertical restraints would depend on the type of restraint, the position of the relevant parties and the market context. In some cases, removing restrictions would be likely to lead to some short-term price reduction and may increase availability for consumers, but this could be at the cost of lower quality (or lower perceived quality) and service, and lower availability in the longer term. See also, for example, Tooltechnic (Authorisation No A91433, 5 December 2014) (n 119). 130 131

Vertical agreements under EU competition law  189 retain access to the market via competing retailers or alternative channels of distribution. Similarly, input foreclosure also is plausible only if the supplier has a significant degree of market power in the input market and the input represents a significant cost factor or is an otherwise critical component for dealers. Similarly, the use of RPM does not automatically imply that collusion can be sustainable, let alone that RPM itself induces collusion. A well-known set of particular circumstances needs to be present for collusion to emerge and to remain sustainable. Indeed, economic theory indicates that collusion will generally be implausible in industries that do not feature a tight oligopoly with high barriers to entry and that certain factors may make collusion more or less likely and stable. For example, collusion is easier to establish in the case of firms competing with homogenous products (where there is limited non-price competition).134 Further, a level of market power upstream and/or downstream is generally required for collusion to be sustainable, and the intrabrand restraints must facilitate firms reaching an understanding on the terms of coordination and monitoring and punishing deviations. Where collusion could be monitored and enforced through other more effective and less conspicuous means, it would seem unlikely that RPM would be chosen. Another antitrust concern is that although e-commerce is an important driver of price transparency and price competition, certain online practices or restraints relating to online selling may raise competition concerns. For example, in its Final E-Commerce Report, the Commission noted that price-tracking software, which is becoming more prevalent and sophisticated, is allowing both retailers and manufacturers to track online retail prices.135 This creates a possible risk that the software can be used as a mechanism for facilitating or strengthening collusion between retailers or enabling manufacturers to monitor and enforce retailers’ compliance with their particular pricing policies (especially if manufacturers have integrated into retailing).136 In addition, evidence indicates that although price comparison sites and

134 The collusion mechanism must be robust to countervailing buyer power or new entry and expansion by non-coordinating firms. A retailer collusion story is not convincing in many scenarios. Under this theory, retailers induce a manufacturer to ‘impose’ RPM, to support their collusion and to preclude deviations for the strategy. Retailers (possessing monopsony power) thereby delegate both the implementation and the enforcement of the cartel to the manufacturer. One problem with this theory is that such collusion may be undermined by non-price rivalry between dealers. Another is that the manufacturer would be worse off under the agreement and its sales would be reduced. Where they are used as a mechanism to facilitate explicit (as opposed to tacit) collusion, cartel enforcement may follow. 135 Half of responding retailers stated that they track the online prices of competitors. Of these, 67 per cent used automatic software programs. Of these software users, 78 per cent would then adjust their own prices to those of their competitors. The Commission found that most retailers adjusted their prices manually, but a ‘significant number’ used both manual and automatic price adjustments, while about 8 per cent used only automatic adjustments. Some manufacturers were also engaging in tracking the online retail prices of their products sold by distributors (approximately 30 per cent did so systematically and others on a more specific basis) and 38 per cent of these manufacturers used price-tracking software. 136 See, for example, Case AT/40.465, 26 September 2018. Rather than providing perfect competition so that retailers can, and do, respond unilaterally and competitively to pricing developments in the marketplace, such software may increase transparency and, consequently, reduce the incentive to deviate from a collusive strategy and make it easier and quicker to detect those that do deviate: Final E-Commerce Report (n 2), 175–7; see, for example, Ariel Ezrachi and Maurice E Stucke, Virtual Competition: The Promise and Perils of the Algorithm-Driven Economy (Harvard University Press 2016). See also, for example, the Decision of the Competition and Markets Authority: Online Sales of Posters and Frames (Case 50223, 12 August 2016) (UK) (in this case, an agreement between online sellers not to undercut

190  Research handbook on methods and models of competition law online marketplaces may at first sight enhance transparency and competition,137 recent theory and practice indicate that the reality may be more complex. Not only may price comparison sites reduce incentives for consumers to shop around, but suppliers may pay or incentivize price comparison sites to place their offerings high in the ranking or use such sites as an advertising platform. They may therefore lead to customers paying higher prices.138 C

A Continued Risk of Errors and Lack of Clarity as to How Full Article 101 Analysis Is To Be Conducted?

1 Overinclusive rules in the EU? Most antitrust systems rely on rules (or presumptions139) of illegality to condemn certain behaviour. Such rules serve important ends, particularly the attainment of procedural economy and the clear prohibition, and deterrence, of patently anticompetitive behaviour. In such circumstances, the administrative savings may outweigh the cost of small false positives and exceed the efficiencies that can be derived from moving to a more comprehensive antitrust analysis. It is crucial, however, if such rules are not to sacrifice economic benefits generated by practices with ambiguous competitive effects, that they should be founded upon a sound understanding, from theory, evidence and experience, that the practice in question typically imposes harm. They should be reserved for conduct which is manifestly anticompetitive, which is most unlikely to produce offsetting benefits or redeeming efficiencies. The mere theoretical possibility that a particular agreement may restrict competition should not suffice. This is an important point, which emphasizes the link between antitrust rules and standards140 and between ‘restrictions by object’ and ‘restrictions by effect’ in EU competition law, and makes it clear that they should not reflect two totally separated analytical approaches but should be joined by unifying concepts. In each case the analysis (the decision to place an agreement in the object category or to appraise the restrictive effects) must be referable to a theory of harm and, in particular, to the question of whether the agreement can be expected to lead to a situation with market-wide increased price/reduced output, compared with a no-agreement counterfactual. Without such a theory, the crucial question of whether the agreement is restrictive or enhances economic efficiency cannot be addressed.

each other’s prices was implemented using automated repricing software, which the parties configured to give effect to the illegal cartel). 137 Savvy consumers can use them to hunt down the best available deal; firms, worried about losing customers, feel an obligation to improve their offerings all the time: see, for example, Competition & Markets Authority, Digital Comparison Tools Market Study (Final Report, 26 September 2016). 138 See, for example, David Ronayne, Price Comparison Websites (Working Paper, October 2015). However, to determine the overall impact on consumers, the procompetitive and anticompetitive effects of increased price transparency need to be balanced: see Sections III.C and IV. 139 Although no conduct is prohibited per se in the EU, object/hard-core restraints are perceived to be virtually illegal per se and to be strongly discouraged: see nn 56–60 and text. 140 See, for example, Isaac Ehrlich and Richard A Posner, ‘An Economic Analysis of Legal Rulemaking’ (1974) 3 Journal of Legal Studies 257, 258 (using the term ‘standard’ to refer to ‘a general criterion of social choice’, such as a mandate to promote ‘competition’ and the term ‘rule’ to refer to a more precise statement that circumscribes the assessment of factors relevant to a decision according to the standard); Louis Kaplow, ‘A Model of Optimal Complexity of Legal Rules’ (1995) 11 Journal of Law, Economics & Organization 150.

Vertical agreements under EU competition law  191 In accordance with this reasoning, many systems apply rules or presumptions of illegality to condemn horizontal cartel arrangements. In CB,141 for example, the Court of Justice explained that such agreements may ‘be considered so likely to have negative effects, in particular on the price, quantity or quality of the goods and services, that it may be considered redundant, for the purposes of applying Article [101(1)], to prove that they have actual effects on the market’ (emphasis added).142 The discussion in Section III.B, however, suggests that as the effect of vertical restraints is complex, ambiguous and contested – they may either increase or decrease efficiency, depending on the context of the case – economic theory does not support general conclusions that vertical restraints are either ‘good’ or ‘bad’ for welfare. As a matter of pure economics, therefore, it is arguably not justified (as it is for cartel agreements) to assume harm to competition and/or to presume that they have no redeeming efficiencies or involve restraints unnecessary (or disproportionate) to the achievement of efficiencies. In short, the double presumption of illegality lacks a solid foundation in economic principles.143 If this is correct, the strict stance against vertical restraints creates a risk of Type I errors (which are not offset by administrative savings) and is likely to be deterring agreements that may result in efficiencies in distribution and/or help firms to penetrate new markets. Although the cost to society of such deterrence is unseen and, hence, difficult to quantify, economic theory warns that it may not be insubstantial.144 Suppliers may prefer to avoid the risk of an antitrust infringement by finding other mechanisms to achieve their objectives – for example, by choosing not to release products in some EU markets where pricing would need to be lower (if this created a risk of undermining the firm’s position in higher-priced markets)145 or by vertically integrating instead. Such solutions might forgo the use of independent distributors, result in reduced intrabrand and interbrand competition, and/or result in a reduction in e-commerce.146 Indeed, it has been seen that the Commission’s Report confirms that manufacturers are seeking tighter control over distribution through vertical integration and/or the conclusion of SDS.147 It seems possible, therefore, that a system designed to promote and protect intrabrand competition may in fact be undermining it.

141 Case C-67/13 P (General Court of the European Union, 11 September 2014). See also discussion at nn 183–189 and text. 142 Ibid, para 51. 143 Iacobucci and Winter (n 90), 63; see, for example, Margaret E Slade, ‘The Effects of Vertical Restraints: An Evidence Based Approach’ in Konkurrensverket, Swedish Competition Authority (ed), The Pros and Cons of Vertical Restraints (2008). 144 See Frédéric Bastiat, Selected Essays on Political Economy (trans Seymour Cain, The Foundation for Economic Education, 1995). 145 See, for example, COMP/30.228, Distillers (Red Label) [1983] OJ C245/3. 146 If suppliers are driven towards vertical integration and ownership of their online distribution channel, the ability and incentives of independent distributors/retailers to enter or remain in the market are limited. Further, the ability for small manufacturers to access more efficient or specialized distributors is reduced, meaning that they may also have to vertically integrate online and sell directly through their own webpage or an online marketplace. But this often entails very high costs and is liable to harm consumers. 147 See, for example, Cases 56 and 58/64, Consten and Grundig [1966] ECR 299 (some years after the decision, Grundig acquired Consten and a number of its exclusive dealers: see Valentine Korah and Denis O’Sullivan, Distribution Agreements under the EC Competition Rules (Hart Publishing 2002) 62).

192  Research handbook on methods and models of competition law The risk of Type I errors may be particularly acute if the rules developed in the pre-internet age are simply extended into the online world. It is still relatively early days for online commerce, which has injected doses of high adrenaline into retail distribution across the EU. The interaction of online and offline sales is leading to creative destruction, trial and error, and all kinds of experimentation with new business models (for example, through mobile apps) and supply chain practices that infuse intense dynamic competition and stimulate innovation. The rigid approach to vertical restraints may, therefore, be constraining such dynamism and market participants from entering into vertical arrangements that stimulate interbrand competition and counteract concentration of market power or collusion. The new realities of a more globalized, technology-driven and digitalized competitive environment may suggest that an approach that allows for some consideration of case-specific facts seems crucial. Although it is recognized that the Commission has wider e-commerce policy objectives and wishes to respond to consumer frustrations about impediments to their making cross-border transactions via the internet,148 care must be taken not to do so at the expense of efficiency – also an overarching objective of the internal market project.149 2 Lack of guidance on individual analysis Because of the substantial reliance on both the category of object restraints by the Commission, and NCAs, in their infringement decisions,150 reliance on the de minimis principle and the Verticals Regulation (by firms), and the paucity of Article 267 references from national courts,151 the question of how individual analysis of vertical agreements is to be conducted under Article 101(1)(3) arises relatively rarely before the EU courts. There has therefore been no steady flow of cases before the courts that has allowed them to develop and hone the case law in this area. Consequently, some lack of clarity continues to shroud questions about how the parts of Article 101 work together, and concerns manifest in the jurisprudence about restraints on rivalry and internal market integration prevail over analysis based on an assessment of the impact of the conduct on economic efficiency and the welfare of consumers in the EU.152 The result is a shortage of guidance on how assessment is to be conducted in relation to the new vertical models of distribution and practices that have been emerging online (subjects

148 See Regulation (EU) 2018/302 on addressing unjustified geo-blocking and other forms of discrimination based on customers’ nationality, place of residence or place of establishment within the internal market and amending Regulations (EC) No 2006/2004 and (EU) 2017/2394 and Directive 2009/22/EC [2018] OJ L60I/1. 149 See Article 101(3) Guidelines, para 13: ‘the creation and preservation of an open single market promotes an efficient allocation of resources throughout the Community for the benefit of consumer’. 150 The last infringement decision adopted by the Commission in relation to a vertical agreement under Article 101 was in 2005: COMP/36.623, Peugeot [2006] OJ L173/20. 151 Case C-345/14, SIA ‘Maxima Latvija’ v Konkurences padome (Court of Justice of the European Union, 26 November 2015); Case C-230/16, Coty (Court of Justice of the European Union, 6 December 2017). 152 The Verticals Regulation, although creating valuable legal certainty, also clouds the relationship between its two parts. Indeed, the Commission’s Vertical Guidelines offer no explanation as to why such a broad overarching block exemption is necessary if, as seems probable, many (or even most?) agreements satisfying its conditions are, given the parties’ lack of market power and the absence of ‘severe’ restraints within the agreement, unlikely to have a restrictive effect on competition (and so to infringe Article 101(1)).

Vertical agreements under EU competition law  193 of the Commission’s e-commerce sector inquiry), including agency arrangements.153 This problem is compounded by the fact that it was the NCAs, not the Commission, that initially took the lead in this area and that they sometimes struggled to act consistently with one another and, so, to adopt a uniform interpretation of EU competition law.154 These perceived differences in the interpretation and application of EU law across jurisdictions create challenges for business and may even deter firms wishing to expand sales across Member States.

IV

SOME PROPOSALS FOR CHANGE: AN EVOLUTION IN ARTICLE 101 APPRAISAL

A

The Need to Develop a More Flexible and Nuanced Framework for Vertical Agreements

The parts above have demonstrated that, in spite of improvements in relation to the working and operation of the Verticals Regulation, a number of the problems identified in relation to the pre-modernized framework persist, including: (1) overly broad application of Article 101(1) and, in particular, overreliance on presumptions of illegality through dependence on, and seeking to expand, the object category; (2) legal uncertainty about how common contractual provisions should be analysed, especially those being used on online markets; and (3) reliance on block exemptions (to reduce legal uncertainty). This approach continues to eliminate and de-emphasize ‘what should be the heart of’ the system155 – substantive (economic) analysis of the overall competitive effects of a given vertical agreement. Change thus seems to be required to realize the Commission’s modernization aspirations and to ensure that the Article 101 analytical framework better reflects the economics of vertical restraints and that the different categories of analysis are linked by uniform concepts underpinning EU competition law. 153 Agency agreements fall outside Article 101(1) entirely in certain circumstances (even if the agreement contains hard-core restraints), broadly where the agent bears no risks (that is, all commercial risks are covered by the principal): see Guidelines, paras 14–15; Case C-217/05, Confederación Española de Empresarios de Estaciones de Servicio (CEES) v Compañia Española de Petróleos SA [2006] ECR I-11987. In spite of the major importance characterization of the agreement has for competition law assessment, the question of when an agency agreement exists is not easy to determine, especially where suppliers sell online via platforms following an agency model. The issue did not have to be decided in, for example, COMP/39.847, e-books [2013] OJ C378/25 and AT/40.153, Amazon (28 July 2017). 154 For example, with regard to MFNs used in relation to online hotel booking, the German NCA adopted decisions finding both that HRS’s and Booking.com’s MFNs infringed competition law (both German and EU), continuing proceedings against the latter, even after other NCAs closed investigations: see generally Edurne Navarro Varona and Aarón Hernández Canales, ‘Online Hotel Booking’ (Competition Policy International Antitrust Chronicle, May 2015). The Commission and relevant NCAs formed a working group to evaluate and monitor the effects of the different solutions adopted and to consider if there is a need to look at this sector again. See also n 60 and text. 155 See Barry E Hawk, ‘System Failure: Vertical Restraints and EC Competition Law’ (1995) 32 Common Market Law Review 973, 2.5: the author complained that the notification and authorization system set up by Regulation 17 [1959–1962] OJ Spec Ed 87, Article 9(1) had failed.

194  Research handbook on methods and models of competition law This chapter thus proposes both (1) an evolution in approach to the assessment of when an agreement is restrictive of competition by object; and (2) an analytical framework, based on mainstream economic thinking, for the individual assessment of agreements under Article 101. The latter will play a more significant role if the scope of the object category is curtailed and would push the modernization process to its logical conclusion. B

Refining the Object Category

1 Is refinement really required? It has been argued in Section III that the approach to object restraints may be deterring procompetitive or competitively neutral agreements. Techniques could be adopted to mitigate the harshness of the current approach and to reduce the risk of false positives, for example through: (1) careful prioritization of cases (public enforcement agencies could decide not to pursue a case unless the practice seems likely to produce significant anticompetitive effects); (2) greater individual use of Article 101(3) to ‘except’ agreements generating efficiencies from the Article 101(1) prohibition, in particular through the Commission adopting non-infringement decisions under Article 10 of Regulation 1/2003;156 or (3) removal of the list of hard-core restraints from the Verticals Regulation.157 It is submitted, however, that none of these options are ideal. First, some NCAs are understood to prioritize vertical cases in the way suggested – for example, deciding not to bring RPM cases where no anticompetitive effects appear to exist.158 It is submitted, however, that this approach is effectively tantamount to acceptance that the rule is wrong.159 As it deviates from that rule it also leads to uncertainty which is exacerbated by the fact that it is certainly not followed by all NCAs, some of which are not able to prioritize and/or have taken robust action against RPM, territorial restraints and restraints on online selling. Firms may therefore feel compelled to follow the strictest approach adopted or create different commercial strategies across Europe to the detriment of the single market. Further, as prioritization does not impact on the law, parties to such agreement remain exposed to private litigation alleging an infringement of Article 101. Second, we acknowledge that a flurry of non-infringement decisions providing clear guidance on the application of Article 101(3) to agreements incorporating ‘object’ restraints seems

156 See, for example, David Bailey, ‘Reinvigorating the Role of Article 101(3) under Regulation 1/2003’ (2016) 81 Antitrust Law Journal 111. Although commitments decisions (see n 7) have been used to deal with agreements that the Commission might have ‘exempted’ in the past, they do not involve a finding of infringement or non-infringement. 157 See n 57 and text. 158 See, for example, the Netherlands Competition Authority strategy and enforcement priorities with regard to vertical agreements: Authority for Consumers & Markets, ‘ACM’s Strategy and Enforcement Priorities with Regard to Vertical Agreement’ (20 April 2015) (the ACM assumes that vertical restraints are generally procompetitive in the absence of market power, even those incorporating hard-core restraints). 159 The legal uncertainty and difficulties generated by the Commission’s publication of a set of prioritization guidelines in relation to Article 102 are well known.

Vertical agreements under EU competition law  195 unlikely. Not only are NCAs not permitted to adopt non-infringement decisions,160 but the Commission’s internal incentives are not designed to dedicate resources to adopting this kind of decision. Further, it has been seen that a particularly acute problem is that it seems difficult to use Article 101(3) (as currently drafted and interpreted) effectively to weigh the beneficial effects of an agreement against its anticompetitive ones, in circumstances where the latter have not been assessed in any way but simply assumed.161 Thirdly, removing the list of hard-core restraints would lead to a perceived rule of virtual per se legality for these practices where parties’ market shares do not exceed 30 per cent and make it more difficult to sanction conduct that, in certain circumstances, could facilitate collusion between suppliers or dealers or have foreclosure effects that are not offset by redeeming efficiencies. This might arguably create a risk of Type II errors – an underinclusive rule that sometimes tolerates conduct that in fact is competitively destructive. Finally, and crucially, it is submitted that these techniques do not address the root of the problem, namely that these vertical restraints may not harm competition at all and may offer scope for efficiencies, suggesting that it is not always appropriate to characterize them as restrictions of competition by object in the first place. 2 Identifying object restrictions – the importance of the characterization process In the US, since it has been accepted that the Sherman Antitrust Act of 1890 is a consumer welfare prescription,162 there has been a concern that categories of antitrust analysis adopted when interpreting section 1 of the Sherman Act should not become conclusions, displacing the fact-specific analysis on which antitrust law is focused.163 In the context of vertical agreements, this sentiment has led to a move away from reliance on per se rules. In three landmark cases dealing with vertical intrabrand restraints, Continental TV, Inc v GTE Sylvania,164 State Oil v Khan165 and Leegin Creative Leather Products Inc v PSKS, Inc166 – the Supreme Court overturned previous rulings holding that customer and territorial restraints,167 maximum RPM,168 and minimum RPM169 were illegal per se. It held that as (1) per se rules are appropriate only for restraints that always, or almost always, tend to restrict competition, have manifestly anticompetitive effects, and lack any redeeming virtue; and (2) the market impact of vertical intrabrand restrictions are complex because of their ability simultaneously to reduce intrabrand competition and to stimulate interbrand competition and bring about distribution efficiencies, the competing effects resulting from such a restraint should be analysed under the rule of reason. Indeed, in Leegin,170 the majority accepted that because the economics literature was Case C-375/09, Tele2Polska [2011] ECR I-3055. See Section II.B.2. 162 See, for example, Reiter v Sonotone Corp, 442 US 330, 343 (1979). 163 Mark A Lemley and Christopher R Leslie, ‘Categorical Analysis in Antitrust Jurisprudence’ (2008) 93 Iowa Law Review 1207. 164 433 US 36 (1977). 165 522 US 3 (1997). 166 551 US 887 (2007). 167 United States v Arnold, Schwinn & Co, 388 US 365, 376 (1967). 168 See Keifer-Stewart Co v Joseph E. Seagram & Sons Inc, 340 US 211 (1951); Albrecht v Herald Co, 390 US 145 (1968). 169 Dr Miles Medical Co v John D Park & Sons Co, 220 US 373 (1911). 170 551 US 887 (2007) (Justice Kennedy wrote for the majority). For a fuller analysis of the case, see, for example, Alison Jones, ‘Resale Price Maintenance: A Debate about Competition Policy in Europe?’ 160 161

196  Research handbook on methods and models of competition law ‘replete’ with procompetitive justifications for a manufacturer’s use of RPM, rule of reason was required. The arguments set out so far in this chapter could provide support for a similar change of approach in the EU – and for recognition that neither theory nor experience suggests that these vertical restraints reveal in themselves a sufficiently obvious harm to competition or to the internal market. The reality, however, is that it seems unlikely that the EU courts could be persuaded that such a reversal is required.171 Rather, a long line of cases indicates that although EU courts have not, as was done in CB,172 articulated concerns that such agreements pose a high probability of negative effects on prices and so on,173 they have demonstrated anxiety that these practices may harm competition through limiting important competition/ rivalry between dealers and suppliers and, crucially, erecting private barriers to cross-border trade inimical to the internal market objective. Further, the Commission’s view seems, like the dissenting minority in Leegin,174 to be that efficiency arguments may not exist in reality as frequently as theory suggests (for example, how often does the opportunity for free-riding really arise?) and that RPM at least does create a high risk of softening competition, collusion and foreclosure. In the absence of convincing empirical data suggesting otherwise, therefore, it is unlikely to be convinced that a wholesale change is necessary.175 An abandonment of this starting position on ATP, RPM and bans on online selling might therefore be argued to create Type II error risks, given the budget limits constraining the activities of ECN members.176 The starting position on ATP, RPM and bans on online selling thus appears entrenched and seems unlikely to be overturned. Although there is no formal system of precedent in the EU, the Court of Justice generally strives for consistency,177 preferring to develop the law on a more incremental basis. It has been seen, nonetheless, that the case law is clear that agreements should be held to be restrictive of competition by object only where an analysis of the purpose of the agreement, taking account of both its clauses and the context in which it operates, reveals a sufficiently deleterious impact on competition.178 The category of object restraints is not therefore as (2009) 5 European Competition Journal 479. 171 See, for example, Jones, Left Behind by Modernisation (n 25); but see the judgment of the Court of Justice in Intel v Commission (Court of Justice of the European Communities, Case C-413/14 P, 6 September 2017). 172 CB (General Court of the European Union, Case C-67/13 P, 11 September 2014), para 51. 173 But see Asus (Case AT/40.465, 26 September 2018), paras 105–7. 174 Justice Breyer, writing for the minority, noted that US history provided empirical support for the view that RPM led to considerably higher retail prices and that economists generally concurred on this point (see also MacKay and Smith (n 101)). In contrast, he found no satisfactory answer to the question as to when and how often benefits, such as the prevention of free-riding, were likely to occur and considered that courts would have extreme difficulty in identifying instances in which benefits were likely to outweigh harm. As antitrust law should be informed by economics but could not always replicate economists’ (sometimes conflicting) views, he concluded that it would sometimes be acceptable to provide a rule of per se unlawfulness to a business practice that could at times produce benefits. See also, for example, Peeperkorn (n 123); Pitofsky (n 123). 175 See n 101. 176 See, for example, Ramsi Woodcock, ‘Per Se in Itself: How Bans Reduce Error in Antitrust’ (11 January 2017), https://​papers​.ssrn​.com/​sol3/​papers​.cfm​?abstract​_id​=​2896453. 177 See, for example, Anthony Arnull, ‘Owning up to Fallibility: Precedent and the Court of Justice’ (1993) 30 Common Market Law Review 247. 178   Case 56/65, STM [1966] ECR 235, 249.

Vertical agreements under EU competition law  197 simple as constituting a definitive and clear list. The list of ‘established’ vertical restraints simply provides an illustration of agreements/restraints whose purpose, or objective, may in many contexts be considered to be highly likely to be sufficiently deleterious – the purpose and the context of the agreement could, however, make it clear that this assumption cannot be made. In CB,179 the Court of Justice stressed that the category of object restraints is inappropriate for cases where a more detailed market analysis is required to assess the impact of the agreement, for example cases involving complex measures or where experience180 with the restraint is (in that case the horizontal cooperation was argued to be designed to ensure the success of the carte bleu system, in particular through combating free-riding and balancing issuing/acquisition activities).181 In these latter situations, negative effects cannot be considered so likely to make assessment of effects redundant.182 In line with CB, it seems that horizontal cooperation agreements with the potential to have mixed effects on competition – even those containing price or output restraints – do not fall within the object category183 unless they do not truly concern, for example, joint research and development, production or joint purchasing, but serve as a tool to engage in a disguised cartel.184 Logic, and case law, require that a similar ‘characterization’ exercise be carried out in relation to vertical restraints, meaning that RPM and territorial restraints plausibly necessary to the pursuit of a legitimate procompetitive objective should not be found to restrict competition by object.185 In Murphy,186 the Court of Justice accepted that a broadcasting licensing agreement containing territorial limitations aimed at partitioning national markets would not be regarded as restrictive by object where other circumstances falling within its economic and legal context justified the finding that such an agreement is not liable to impair competition,

179 CB (General Court of the European Union, Case C-67/13 P, 11 September 2014); see also, for example, Intel v Commission (Court of Justice of the European Communities, Case C-413/14 P, 6 September 2017) (the Court of Justice clarified previous case law finding exclusivity rebates to be abusive of competition (Case 85/76, Hoffmann-La Roche v Commission [1979] ECR 461), holding that if a dominant firm submits, on the basis of supporting evidence, that its exclusivity rebates are not capable of restricting competition and the Commission carries out an analysis of their likely effects, the General Court must examine arguments seeking to call into question the Commission’s analysis even where the Commission argues that the rebates at issue are, by their very nature, capable of restricting competition). 180 The importance of experience was also stressed by the Court of Justice in Case C-286/13 P, Dole (General Court, 19 March 2015), para 115; see also the Opinion of Wathelet AG in Case C-373/14, Toshiba Corp v Commission (Court of Justice of the European Union, 20 January 2016). 181 The Court of Justice referred the matter back to the General Court, which found that the agreement infringed Article 101 – the free-riding arguments had not been substantiated: Case T-491/07 RENV, CB v Commission (Court of Justice of the European Union, 30 June 2016). 182 See also Case C-345/14, Maxima Latvija (Court of Justice of the European Union, 26 November 2015), para 22; Case C-309/99, Wouters v Algemene Raad van de Nederlandse Orde van Advocaten [2002] ECR I-1577, paras 106–10. 183 See, for example, Guidelines on the applicability of Article 101 of the Treaty on the Functioning of the European Union to horizontal co-operation agreements [2011] OJ C11/1, paras 128, 160–1, 205–6. 184 See n 183; see also the Commission’s analysis in COMP/39.226, Lundbeck (19 June 2013), upheld in Case T-472/13, H Lundbeck and Lundbeck v Commission [2013] OJ C325/76; COMP/39.612, Périndopril (Servier) (9 July 2014), on appeal in Case T-691/14, Servier SAS v Commission (General Court of the European Union, 12 December 2018). 185 This should be a distinct exercise to the objective necessity analysis currently advocated in the cases: see nn 68–74 and text. 186 Cases C-403 and 429/08, Murphy [2011] ECR I-09083, para 139.

198  Research handbook on methods and models of competition law and in Pierre Fabre187 the Court of Justice held that an SDS incorporating a ban, or de facto ban, on internet selling necessarily restricted competition (and restricted competition by object) unless objectively justified (it constituted a proportionate measure to achieve a legitimate aim188). New restraints should also not be added to the object category unless theory or experience justifies a finding that the clauses and context reveal a high probability of anticompetitive effects. The problem of an overly expansive object category could therefore be addressed through a careful characterization process – that is, a close examination of the agreement’s purpose, content and context before it is determined whether object or effects analysis is appropriate. This may appear a minor and obvious development following CB, but its effect, if followed, would be radical. To date there is no case involving an ordinary vertical agreement (rather than an intellectual property licensing agreement) in which the Commission (or Court) has been prepared to accept on the facts that either ATP or RPM is not restrictive of competition by object.189 Rather, in practice, the jurisprudence generally just presumes that such restraints are not justified and inevitably lead to consumer harm. Decision-makers need, therefore, to become more willing to accept that object categorization is not appropriate where the parties raise a plausible efficiency justification for RPM or territorial/customer restraints, bringing the practice into line with the approach set out by the Court of Justice in CB. If there is a convincing efficiency story, the claimant should be required to demonstrate actual or likely restrictive effects before the parties are required to provide a robust justification of the efficiencies within the Article 101(3) forum. C

A Clearer Framework for Effects Analysis

1 Developing the legal framework Concern about the open-textured nature of full antitrust analysis has often led decision-makers to shy away from adopting it and an anxiety that it will become tantamount to a rule of per se legality (given the difficulty it presents for claimants) leading to Type II errors.190 The EU administrative system, however, provides a flexible forum for the Commission to develop an administrable and workable framework for an assessment of vertical restraints which need not be an expensive, excessively complex or time-consuming task. Indeed, it seems crucial to the evolution required that the Commission, which retains a central role in EU competition law enforcement, should bring a series of carefully chosen cases of wider EU rel-

Case C-439/09, Pierre Fabre GmbH (Court of Justice of the European Union, 6 December 2017). The Court of Justice held that the aim of maintaining a prestigious product image was not a legitimate purpose and could not be relied upon to provide any objective justification for such a clause; see also n 39. 189 See n 41. 190 Herbert Hovenkamp, The Antitrust Enterprise: Principle and Execution (Harvard University Press 2005) 8; Stephen Calkins, ‘California Dental Association: Not a Quick Look But Not the Full Monty’ (2000) 67 Antitrust Law Journal 495; Douglas H Ginsburg, ‘Vertical Restraints: De Facto Legality under the Rule of Reason’ (1991) 60 Antitrust Law Journal 67; Michael A Carrier, ‘The Rule of Reason: An Empirical Update for the 21st Century’ (2009) 16 George Mason Law Review 827; D Daniel Sokol, ‘The Transformation of Vertical Restraints: Per Se Illegality, the Rule of Reason, and Per Se Legality’ (2014) 79 Antitrust Law Journal 1003. 187 188

Vertical agreements under EU competition law  199 evance and with a high precedent potential. If advanced in this way through decisions that are reviewable on appeal, the law could evolve and provide greater clarity of structure for NCAs and national courts to follow in subsequent cases. Indeed, some systems have developed structured forms of analysis often employing burden-shifting frameworks to organize the evaluation of theories of harm and supporting evidence (in agreements, mergers and dominance cases). In the US, in particular, there has been a detectable movement away from analysis based on the classification of conduct into specific categories, and a movement towards a more concept-based approach, which focuses on market power, the theory of harm, proof of adverse effects, and assessment of countervailing efficiencies.191 If the Commission were to bring a number of carefully targeted effect cases in this way, the perception of effects analysis as unmanageable and disorderly could be quelled, and resources could be concentrated on developing the Article 101 structure – rather than the limits of the ‘by object’ category – to determine how: ●● anticompetitive effects in terms of parameters of competition can be identified under Article 101(1) (departing from the objective necessity test set out in some cases); and ●● procompetitive effects identified under Article 101(3) can be balanced against the anticompetitive effects identified under Article 101(1). Further, if the Commission adopts commitments decisions it should consider putting in place mechanisms that measure how changes made impact on markets in practice. More broadly, these steps could feed into the review of the 2010 vertical regime and the opportunity to ensure that the guidelines and block exemption are updated and brought more closely in tune with economic thinking and modern business realities. 2 Assessing and balancing anticompetitive and procompetitive effects In merger cases, the Directorate-General for Competition routinely conducts analysis which requires identification of the likely impact of a merger on the main parameters of competition (likely anticompetitive effects through foreclosure or coordinated effects and countervailing factors) and efficiencies. In the context of Article 101, an analogous assessment192 can be developed to determine whether (1) the conditions exist either for one of the parties to the vertical agreement to exercise market power or for effective collusion (by suppliers or retailers) to be sustained; (2) the agreement in some way maintains, enhances or facilitates the exercise of that market power; and, if so, (3) the restraints are indispensable to achieving efficiencies that align incentives and stimulate interbrand competition and that are likely to be passed on to consumers and offset the harm suffered (Article 101(3)). The enquiry under Article 101(1) should therefore focus first on the question of whether the agreement is likely to enable or facilitate the exercise of market power – not on the objective necessity of the restraints.193 Where a vertical structure faces vigorous competition from both other brands and other retailers, there is less potential for any type of vertical restraint to reduce competition or economic efficiency. As strong competition at each level fosters efficiency of 191 See, for example, in the US, United States v VISA USA, Inc, 344 F 3d 229, 238 (2d Cir 2003) (agreements); United States v Microsoft Corp 253 F 2d 34 (DC Cir 2001) (per curiam) (monopolization); United States v Baker Hughes Inc 908 F 2d 981 (DC Cir 1990) (merger). 192 Arguably, a merger assessment is more complex, since it is typically forward looking. 193 See Section III.B.3.

200  Research handbook on methods and models of competition law both supply and distribution services, minimal analysis should be required in cases involving firms with small market shares that operate in upstream and downstream markets that are not concentrated. Development of the internet, in particular, has expanded the opportunity for interbrand competition in many markets. In contrast, where vertical restraints are incorporated into agreements between firms operating in less competitive markets194 (oligopolistic ones or where one of the firms is dominant), there is greater risk that they may be used to reduce competition or economic efficiency and so closer antitrust scrutiny is consequently warranted. In certain circumstances there may be evidence of actual harmful effects resulting from an implemented agreement. In other cases, anticompetitive effects will have to be established through indirect evidence, economic reasoning, and an assessment of indicators of market performance and the economic and legal context in which the agreement takes place. In conducting such analysis, it makes sense to benchmark it against that conducted in relation to vertical mergers – where there is vertical integration between the supplier and dealer(s). Like vertical mergers, vertical restraints can eliminate intrabrand competition without directly impacting on (interbrand) competition between two rivals competing in the same relevant market. Unlike full integration, vertical restraints typically allow (even promote) a certain degree of intrabrand competition (for example, on service), so logic suggests that restraints should not be treated more severely under Article 101 than vertical mergers under the EU merger rules. Thus, the same concepts and methodological approaches should be used, taking account of the different ex ante and ex post nature of the review. This would eliminate any ‘concentration privilege’ and ensure that both mergers and agreements are only considered anticompetitive if they are likely to result in consumer harm.195 As in the case of vertical mergers, the first step must be to assess whether the use of vertical intrabrand restraints is likely to result in input or customer foreclosure or coordination. Here the factors discussed in Section III.B.3 are crucial. This approach would resemble the methodology proposed by (then) Assistant Attorney General Christine Varney196 for US courts to appraise RPM under the rule of reason, post Leegin.197 She suggested that a plaintiff seeking to establish that RPM infringes section 1 of the Sherman Act would have to demonstrate one of four theories of harm:

Competition in the market as a whole is what matters here, and not intrabrand competition as such. See, for example, Esteva Mosso (n 10). 196 Christine A Varney, ‘Antitrust Federalism: Enhancing Federal/State Cooperation’ (Remarks for the National Association of Attorneys General, Columbia Law School, 7 October 2009); see also Case 1022/1/1/03, JJB Sports plc v Office of Fair Trading [2004] CAT 17; Iacobucci and Winter (n 90), 63; Federal Trade Commission, ‘FTC Modifies Order in Nine West Resale Price Maintenance Case’ (5 May 2008), http://​www​.ftc​.gov/​opa/​2008/​05/​ninewest​.shtm. 197 In Leegin 33 Cal Rptr 3d 139 (Cal 2005), Justice Kennedy stated that three factors would be particularly important to the inquiry of anticompetitive effects in RPM cases: (1) the number of manufacturers making use of the practice (careful scrutiny would be necessary if many competing manufacturers adopted the practice); (2) the source of the restraint (harm to competition would be more likely to arise if RPM were introduced as a result of retailer pressure, perhaps to facilitate a retailer cartel or supporting a dominant, inefficient retailer – in contrast, a manufacturer’s interest in efficient distribution would ordinarily be aligned with that of the consumer); and (3) whether the manufacturer or retailer has market power (‘[i]f a retailer lacks market power, manufacturers likely can sell their goods through rival retailers … And if a manufacturer lacks market power, there is less likelihood it can use the practice to keep competitors away from distribution outlets’: Leegin Creative Leather Products v PSKS, 551 US 877, 898 (2007). 194 195

Vertical agreements under EU competition law  201 (1) RPM used by manufacturers to identify cheating on a price-fixing agreement; (2) RPM used to organize a retailer cartel by coercing manufacturers to eliminate price cutting; (3) RPM used by a dominant retailer to protect it from retailers with better distribution systems and lower cost structures, therefore forestalling innovation and distribution; or (4) RPM used by a manufacturer with market power to give retailers an incentive not to sell the products of small rivals or new entrants. Further, she noted, reflecting concerns of the Leegin court and economists, that particularly careful scrutiny of RPM would be merited if it was used widely by manufacturers or if the retailer, rather than the manufacturer, was the source of the restraint (RPM is more likely to be anticompetitive when it results from retailer coercion).198 Where likely restrictive effects are established, the burden shifts to the parties to establish procompetitive effects resulting from the agreement under Article 101(3) – for example, that by internalizing double mark-ups, preventing free-riding, encouraging investment in customer services, permitting a cost-effective alternative to service contracts, facilitating market entry for new firms and brands, or otherwise aligning the incentives of the parties, the agreement is likely to enhance the ability of the firms to act procompetitively for the benefit of consumers, thereby counteracting the adverse effects on competition that the agreement might otherwise have.199 A particularly difficult hurdle for those relying on Article 101(3) in the EU to overcome, however, is the ‘indispensability’ criterion, which requires a showing that the restrictions are indispensable to the achievement of the efficiencies, while allowing consumers a fair share of them. Although it is clear that restrictions are not indispensable, if efficiencies can be achieved by other practicable or less restrictive means or if they are not reasonably necessary to produce the efficiencies there is limited guidance on how this condition applies in practice.200

V CONCLUSIONS Although there is relatively little empirical evidence, such evidence as there is, together with modern economic theory, does not support arguments either that vertical intrabrand restraints are inherently suspicious or that they always improve efficiency and enhance welfare. Rather, the economic literature reviewed in this chapter indicates that such restraints tend to produce efficiencies and that, although they can give rise to anticompetitive outcomes, this is likely only in certain circumstances, in particular where market power exists either in the upstream or downstream markets. It has also been seen that vertical restraints are being used to encourage online distributors to compete on non-price dimensions of competition. E-commerce is 198 See Case 1022/1/1/03, JJB Sports plc v Office of Fair Trading [2004] CAT 17; Iacobucci and Winter (n 90), 63. 199 See, for example, the Commission’s Guidelines on the assessment of non-horizontal mergers under the Council Regulation on the control of concentrations between undertakings: [2008] OJ C265/6, para 52. 200 Case C-68/12, Protimonopolný úrad Slovenskej republiky v Slovenská sporiteľňa a.s. (Court of Justice of the European Union, 7 February 2013); see nn 54–56 and text. On this point, see the analysis of the ACCC in Tooltechnic (n 119).

202  Research handbook on methods and models of competition law offering new opportunities for firms, small or large, to enter and expand into new markets, to introduce new products and innovative services, and to test new ideas and markets at a fraction of the transaction costs of the past. It is also transforming the retail experience, giving consumers access, with low search costs, to a wide array of online distributors across a wider geographic scope. It has been seen, nonetheless, that although vertical restraints on interbrand competition frequently fall outside of Article 101(1) and/or benefit from the safe harbour of the Verticals Regulation, a combination of factors has led to intrabrand restraints being treated harshly. Many such restraints are simply assumed to restrict competition and are excluded from the benefit of the Verticals Regulation, meaning that their incorporation in a vertical agreement remains extremely risky, even for firms which hold very limited market power and in markets where interbrand competition is intense. Others are subject to an uncertain appraisal under Article 101, which focuses more closely on the necessity of the restraints contained within the agreement than its overall impact on competition. The current regime appears therefore to be out of kilter with the economics underpinning vertical restraints and the Commission’s public statements about competition law and policy. The approach is also at odds, and hard to reconcile, with merger policy, where it tends to be assumed that vertical mergers offer scope for efficiencies, and where a range of relevant factors, not just a contractual clause in isolation, are analysed to determine the likely impact of such mergers on the main parameters of competition (price, output, quality, innovation, etc.). Such mergers are not condemned without the establishment of a robust theory of likely unilateral or coordinated effects. The approach may, therefore, be deterring procompetitive agreements and forcing firms either to integrate vertically or to forgo certain means of distribution entirely. The strict policy that was designed to eliminate barriers to undistorted competition in an integrated market may in fact be harming competition and deterring new and innovative online distribution practices. This chapter has considered how best to align EU law with mainstream economic thinking. Economics could be relied upon to support the view that (a) vertical intrabrand restraints should never be found to restrict competition by object, or even (b) that all vertical restraints, including price and territorial restraints, should be presumptively legal when incorporated within an agreement between parties that do not have market power. It has been seen, however, that the EU policy towards RPM, ATP and bans on online selling is entrenched, and acute concern has been articulated about the possible risks to competition and the internal market objective that result from reduced intrabrand competition, market partitioning and price discrimination between Member States. Further, expansion of the law in this way might create unacceptable risks of Type II errors. Consequently, this chapter calls for a more gradual, but nonetheless significant, change to be developed in the jurisprudence through (1) the adoption of a more realistic approach to categorization – that is, a greater willingness to consider plausible efficiency justifications before finding that a vertical restraint is restrictive of competition by object, especially where restraints are being applied in new online contexts; and (2) clearer guidance as to how agreements incorporating intrabrand restraints are to be analysed under Article 101. It has been argued that such evolution is likely to be dependent upon the Commission being prepared to take more vertical cases, which could be used to develop and clarify the law in this important area through reasoned decisions and an increased flow of cases and appeals before the EU courts.

Vertical agreements under EU competition law  203 If the Commission, EU courts, NCAs and national courts become more willing to analyse economic theory, experience and context, both when deciding whether a vertical agreement should be characterized as restrictive of competition by object and/or when assessing its mixed effects, the seeds for the development of a more consistent and coherent framework for antitrust analysis will be planted. In addition, in line with modernization ambitions, this would represent a further move away from analysis based on historic categories towards a more concept-based approach, focusing on market power, the theory of harm, proof of adverse effects and assessment of countervailing efficiencies. Such an approach would provide the ideal foundation for the forthcoming reform of the Guidelines and Verticals Regulation, and allow the regime to be revised in a way that will provide lucid guidance for firms competing online and offline at all levels of the supply chain.

9. Unilateral conduct analysis: focus on harm in multiple guises Rhonda L. Smith and Deborah Healey

I

INTRODUCTION AND BACKGROUND

This chapter explores the economic concepts behind the enforcement of competition law provisions aimed at unilateral anticompetitive conduct and the application of those provisions to specific types of conduct in multiple jurisdictions. It focuses on the methodology employed to determine whether conduct is harmful to market competition. A

Philosophies Affect the Application of the Law

The philosophy behind competition policy and law and the approach to enforcement differs significantly between countries and helps to explain the sometimes significantly different approach to assessing unilateral conduct. In the US, the Chicago School dominated antitrust thinking from the late 1960s. Its central tenet was that markets work best without regulatory intervention. This reflected the assumption that barriers to entry are low and, where they are significant, they tend to be the result of government intervention. Unsurprisingly, as Chicago School-influenced economists and lawyers joined the US Department of Justice (DOJ) and the Federal Trade Commission (FTC) and were appointed to the courts, the approach to market conduct other than horizontal agreements was essentially that of laissez-faire. So long as there was a business rationale for the conduct, including – or perhaps especially – unilateral conduct, it was likely to be judged efficiency enhancing rather than anticompetitive. By way of contrast, the philosophical underpinnings of European competition law, and much of that in Asia (for example, Taiwan, Hong Kong and Japan), reflect a belief in fairness, as is evident from the policy objective of the law. From its origins in the Treaty of Rome – reinforced in the Treaty of Lisbon 2009 – the EU has had fairness as a stated objective of its competition law.1 This means that the aim is to achieve efficiency through competitive markets and then ensure a fair distribution of the wealth from those markets,2 recognizing that producers incur risks that must be rewarded to provide incentives for innovation. Unlike that of the Chicago School, this approach regards transfers between consumers and producers as a relevant aspect of competition law. Thus, Article 101(3) of the Treaty on the Functioning of the EU (TFEU)3 states that ‘any agreement …, any decision … or any 1 It should be noted that ‘fairness’ has various interpretations within EU law. See, for example, Alfonso Lamadrid de Pablo, ‘Competition Law as Fairness’ (2017) 8 Journal of European Competition Law & Practice 147, 147–8. 2 Doris Hildebrand, ‘The Equality and Social Fairness Objectives in EU Competition Law: The European School of Thought’ (2017) 1 Concurrences Review 41. 3 Treaty on the Functioning of the European Union (Treaty No 2008/C 115/01, signed 13 December 2007) (TFEU).

204

Unilateral conduct analysis: focus on harm in multiple guises  205 concerted practice …, which contributes to improving the production or distribution of goods or to promoting technical or economic progress, while allowing consumers a fair share of the resulting benefit … may not be prohibited even if it is anti-competitive’,4 that is, if the resulting benefit/wealth gain is redistributed fairly and on an equal footing between market participants. Article 102 states that the conduct of firms that possess substantial market power will be an abuse of that power if it has the effect of ‘directly or indirectly imposing unfair purchase or selling prices or other unfair trading conditions’. This may apply to business-to-business transactions, as well as to sales to consumers. For example, the European Commission (EC) found that Google had abused its market dominance as a search engine by giving its own comparison shopping service preferential treatment over competing comparative shopping services. As a consequence ‘[i]t denied other companies the chance to compete on the merits and to innovate. And most importantly, it denied European consumers a genuine choice of services and the full benefits of innovation.’5 B

The Rationale for Competition Law

The rationale for competition law is discussed at length in other parts of this volume. In short, competition policy aims to protect and promote competition because it results in efficient market outcomes and increased economic welfare,6 assuming there are no significant market failures. However, market power enables a firm to influence market outcomes by restricting output and raising prices or otherwise adversely altering the terms of trade.7 This causes increased production cost due to under-exploited economies of scale (production inefficiency). Increased prices cause some to demand more of less preferred products, causing a movement of resources to enable increased production (allocative inefficiency) represented by the deadweight loss. However, the effect of market power on dynamic efficiency is less clear. Schumpeter argued that market power was necessary to provide the incentive for innovation, while Arrow argued that it was competition that was necessary.8 It is now generally agreed that innovation is most likely where firms possess some market power but also face competition.9 4 The agreement must not afford the parties the possibility of eliminating competition in respect of a substantial part of the products in question. 5 European Commission, ‘Antitrust: Commission Fines Google €2.42 Billion for Abusing Dominance as Search Engine by Giving Illegal Advantage to Own Comparison Shopping Service’ (Media Release, 27 June 2017), http://​europa​.eu/​rapid/​press​-release​_IP​-17​-1784​_en​.htm. 6 That is, the unit cost of production is minimized (production efficiency) and resources are allocated to the products that consumers value most and for which they are prepared to pay at least the cost of production (allocative efficiency). Individual firms are price takers earning just enough profit (normal profit) to retain them in the market. Dynamic efficiency refers to the speed and extent of change by firms in response to changed market conditions. 7 Price is quoted for a unit of product of given quality and so, even if the nominal price remains unchanged, a change in quality is an effective change in price. Other terms and conditions associated with supply may include delivery, after-sales service and credit terms. These too may change and hence change the effective price faced by consumers. 8 Joseph Schumpeter, Capitalism, Socialism and Democracy (Harper 1942); Kenneth J Arrow, ‘Economic Welfare and the Allocation of Resources to Invention’ in National Bureau for Economic Research, The Rate and Direction of Inventive Activity: Economic and Social Factors (Princeton University Press 1962) 467–92. 9 Jonathan Baker, ‘Beyond Schumpeter vs. Arrow: How Antitrust Fosters Innovation’ (2007) 73 Antitrust Law Journal 575.

206  Research handbook on methods and models of competition law C

Competition Law and Substantial Market Power

Market power is the threshold requirement for the application of provisions condemning unilateral conduct in all jurisdictions. It is the ability profitably to maintain prices above the cost of supply in the long run – the ability to give less and charge more. Market power is a matter of degree and different jurisdictions set the threshold for problematic market power at different levels – monopoly, dominance or substantial market power (SMP). However, as the FTC states, ‘[c]ourts do not require a literal monopoly before applying rules for single firm conduct; that term is used as shorthand for a firm with significant and durable market power’.10 For convenience, mostly in this chapter the term ‘SMP’, or simply ‘market power’, is used to refer to all three levels. Competition law does not prohibit mere possession of SMP. To do so would take away the incentive to compete vigorously – market power can be seen as the prize for such conduct. If markets are competitive, firms will only acquire market power by being more efficient than their rivals – lower-cost producers and/or suppliers of better products. This is the type of conduct that competition law values. Such market power is unlikely to last sufficiently long to do much harm to consumers as other firms will strive to replace the firm with market power by out-competing it. Competition law is not directed towards market power that is transitory, only to that which persists in the long term. To maximize profits, a firm possessing SMP will restrict output and raise prices above the competitive level. This will not contravene competition law. Likewise, such firms can respond to competition from rivals by adding value to their products and/or by finding ways to reduce the cost of production, sometimes referred to as ‘competition on the merits’. However, they cannot engage in unilateral conduct that in some jurisdictions has the purpose or intent of damaging the competitive process (not simply competitors) and/or, in other jurisdictions, that has an anticompetitive effect – conduct that is entry deterring or exclusionary. Thus: Over and over, the courts face the huge fire-breathing dragons that are incessantly dragged up before them. But then each dragon sits down, sticks its thumb in its mouth, and says, ‘It is true that you people don’t want dragons roaming your streets, but I couldn’t help being a dragon, and I am really a very nice dragon. The crime is dragonizing – not being a dragon – and while I admit being a dragon, you can’t prove I acted like one. In fact, I have always acted like a little woolly lamb.11

Thus, ‘it is acceptable to be a dragon, but not to dragonize’:12 Because it is in the nature of dragons to dragonize, it has never been easy to say when a monopolist is going too far. The standard is also a little different in many other countries, where the law says that a dominant firm may not abuse its dominance. So what is a dominant firm? And, for that matter, what is a monopolist?13

10 Federal Trade Commission, ‘Guide to Antitrust Laws: Single Firm Conduct’, https://​www​.ftc​.gov/​ tips​-advice/​competition​-guidance/​guide​-antitrust​-laws/​single​-firm​-conduct. 11 Martin Shapiro, Law and Politics in the Supreme Court (The Free Press 1964) 268. 12 A Foer, ‘E-Commerce Meets Antitrust: A Primer’ (2001) 1 Journal of Public Policy & Marketing 20, 51. 13 Ibid 51.

Unilateral conduct analysis: focus on harm in multiple guises  207 Although the possession of market power is not of itself a breach of competition law, both the US and the EU refer to the ‘special responsibility’ of firms with SMP. In Google Shopping, for example, the EC referred to ‘the special responsibility of the dominant undertaking’.14 In the EU, this may reflect a focus on form rather than effect when assessing unilateral conduct. In the US, with its laissez-faire philosophy, this is unexpected. In this chapter, the key processes for assessing unilateral conduct in the US, the EU and China are discussed in Section II. Section III addresses general issues relating to the methodology employed by competition authorities and the courts when assessing unilateral conduct. One of the critical issues currently engaging the competition community is how to address competition issues, including unilateral conduct, in digital markets. Section IV discusses some issues associated with platform businesses.

II PROCESS A Legislation In the US, the primary statutory provision for addressing anticompetitive unilateral conduct is the Sherman Act 1890. In addition, each state has antitrust provisions. Some closely follow section 2 of the Sherman Act; others have laws that prohibit unfair or deceptive practices. Many states have specific provisions relating to certain industries. Section 2 of the Sherman Act states: Every person who shall monopolize, or attempt to monopolize, or combine or conspire with any other person or persons, to monopolize any part of the trade or commerce among the several States, or with foreign nations, shall be deemed guilty of a felony …

The aim of section 2 is to preserve a competitive environment that encourages economic growth. To this end, it creates three offences: monopolization, attempts to monopolize, and conspiracy to monopolize.15 The first of these is described as ‘the core section 2 offense’.16 It requires establishing ‘(1) the possession of monopoly power in the relevant market and (2) the wilful acquisition or maintenance of that power as distinguished from growth or development as a consequence of a superior product, business acumen, or historic accident’.17 Attempted monopolization requires proof of engagement in predatory or anticompetitive conduct with the specific intent to monopolize and that there is a dangerous probability of achieving monopoly power.18 Market power less than monopoly may be sufficient for this.

Google Search (Shopping) (Case AT.39740) [2017] 345. ‘Single-Firm Conduct Under Section 2 of the Sherman Act’ in US Department of Justice, Competition and Monopoly (2008, updated 2015) 5. 16 William F Adkinson Jr, Karen L Grimm and Christopher N Bryan, ‘Enforcement of Section 2 of the Sherman Act: Theory and Practice’ (Working Paper, 3 November 2008) 2. 17 United States v Grinnell Corp, 384 US 563 (1966) 570–71, quoted in Antitrust Modernization Commission, Reports and Recommendations (April 2007) ch 2 (Modernization Commission Report). 18 Modernization Commission Report (n 17), CH 2. 14 15

208  Research handbook on methods and models of competition law In addition, section 5 of the Federal Trade Commission Act (FTC Act) prohibits ‘unfair methods of competition’.19 Both sections are general prohibitions against anticompetitive unilateral conduct but they leave interpretation of the terms used and what constitutes the prohibited conduct to the courts. This approach differs in some significant respects from Article 102, the corresponding EU provision. As noted, the EU has fairness as a stated objective of its competition law. A unique aspect of EU competition law, including Article 102, is to ensure a system of undistorted competition in the internal market. Article 102 of the TFEU20 specifically provides that ‘[a]ny abuse by one or more undertakings of a dominant position within the internal market or in a substantial part of it shall be prohibited as incompatible with the internal market in so far as it may affect trade between Member States’. Unlike the general prohibition in section 2 of the Sherman Act, Article 102 identifies ‘abuse’ as the prohibited conduct of dominant firms. Article 102 specifies particular types of conduct – exclusionary, exploitative or discriminatory – that are prohibited, as in its application does section 2 of the Sherman Act. Identification of conduct that falls into one of these categories is a key issue. The basic competition provisions contained in the TFEU are supported by regulations containing general rules relating to matters such as the investigative powers of the EC, and regulations dealing with particular types of conduct or specific sectors. Article 102(a) contains a distinctive feature of EU competition law. Together with Article 54 of the European Economic Area Agreement, it provides for violation due to excessive pricing.21 This provision is concerned with harm to consumers, not exclusionary conduct and the competitive process. It implicitly assumes either that markets will not self-correct in response to excessive pricing/terms, or that they will not do so quickly enough to prevent significant harm to consumers, which contrasts sharply with the US approach. The US and European approaches to competition law, including as it relates to unilateral conduct, form templates for newer competition laws. Historical events or ties have also influenced legislation. Thus, Japan’s first competition law was a replica of the US Sherman Act, introduced during American occupation after the Second World War and directed at a similar issue to that Act (reducing the power of the conglomerates, the zaibatsu). Correspondingly, competition law in India and in Hong Kong during the period of British rule resembled British law. However, as time went on, newly enacted competition laws began to reflect particular local issues and concerns. This is evident in China’s Anti-Monopoly Law 2008 (AML). The AML specifies the framework for addressing anticompetitive conduct, including unilateral conduct. It prohibits a business from abusing a dominant market position. Dominance is defined in the AML as the ability to control the price, volume or other trade conditions of a product in a relevant market, or to impede market entry. Chapter 3 of the AML states that a dominant business operator ‘is prohibited from engaging in certain types of conduct that constitute an abuse of dominance’. Article 17 provides an inclusive list of such conduct, divided into exploitative and exclusionary conduct. The structure of the AML with respect to unilateral conduct is quite similar to that of Article 102 of the TFEU – for example, it is based on dominance, allows for collective dominance, 19 This contrasts with ‘unfairness’ in the EU context, which relates primarily to distribution and not business methods. 20 Ex Art 82 EC, ex Art 86 EC. 21 Article 102(a) applies not only to excessive pricing, but also to reductions in product quality.

Unilateral conduct analysis: focus on harm in multiple guises  209 and lists prohibited conduct. However, it is more prescriptive than Article 102. The AML provides factors to be assessed in determining dominance, including quite a detailed discussion of the role of market shares. Like the EU, it lists the types of conduct regarded as abusive and includes excessive pricing. The important distinction, however, is the presumptive thresholds for dominance contained in Article 19. Dominance is presumed where the relevant market share of one undertaking accounts for one-half or more of the relevant market; the joint market share of two undertakings as a whole accounts for two-thirds or more in the relevant market; or the aggregate market share of three undertakings as a whole accounts for three-quarters or more of the relevant market. These presumptions can be rebutted. B Exemptions Generally, within the competition law of a country or region, there is provision for exemption in certain circumstances. In both the US and China, for example, conduct is exempt if there is a legitimate business rationale for it. In the US, section 2 of the Sherman Act is violated only if the company tries to maintain or acquire a monopoly through unreasonable methods. For the courts, a key factor in determining what is unreasonable is whether the practice has a legitimate business justification.’22 In China, however, a legitimate business rationale does not exempt excessive pricing and buying products at unfairly low prices. Given this, the ‘business rationale’ exemption includes conduct adopted to realize normal operations and to achieve normal benefit, or if the conduct has positive effects on the efficiency of the economy, public interests and economic development.23 Since the inception of the then Trade Practices Act 1974 (Cth), Australian law has provided an administrative route for exemption. In general terms, this provides that conduct that has a net public benefit can be exempt from the conduct prohibitions of the Act. However, until November 2017, exemptions were not available in relation to unilateral conduct (section 46 of the Competition and Consumer Act 2010 (Cth) (CCA)). Following an amendment to the CCA, section 46 prohibits a firm that possesses SMP from engaging in conduct that has the purpose, effect or likely effect of substantially lessening competition.24 In part to align it with the treatment of other prohibited conduct, the amendments also provided for authorization of anticompetitive unilateral conduct that resulted in a net public benefit. In the EU, however, there are no statutory defences under Article 102. Nevertheless, the European Court of Justice (ECJ) has stressed that a dominant firm may seek to justify behaviour that would otherwise constitute abuse, either by arguing that the behaviour is objectively justifiable or by showing that any resulting negative consequences are outweighed by the greater efficiencies it promotes.25 Faella found, however, that where efficiency justifications have been explicitly considered in EC abuse of dominance cases, they were rejected.26 Federal Trade Commission (n 10). SAIC and NDRC Rules. See, for example, Ziqing Zheng, ‘Antimonopoly and Unilateral Conduct 2018’, Global Competition Review (1 August 2018) paras 31–2, https://​globalcompetitionreview​.com/​ jurisdiction/​1005113/​china. 24 Previously, the prohibition has applied to the taking advantage (use) of SMP for an anticompetitive purpose. 25 Case C-209/10 Post Danmark AS v Konkurrencerådet [2012] ECR I-172, [41]–[42]. 26 Gianluca Faella, ‘The Efficient Abuse: Reflections on the EU, Italian and UK Experience’ (2016) 1 Competition Law and Policy Debate 33. 22 23

210  Research handbook on methods and models of competition law In some countries, there are other exemptions. In relation to the US, Calkins observes: Unfortunately, there is a very long list of exemptions and immunities from the U.S. antitrust laws … Statutory exemptions include ones for agricultural cooperatives, insurance, labor, sports broadcasting, and various forms of communication, energy, and transportation. The most important court-made exemptions are those for petitioning government (known as the Noerr-Pennington doctrine), and for government action itself and the results thereof (known as the state action doctrine).27

In China, state-owned enterprises (SOEs) are not theoretically exempt from competition law and it has been stated, for example, that SOEs must notify mergers in the same way as do private companies. However, various discretions contained in the AML provide scope for flexibility in respect of SOEs.28 According to Cheng, the current policy in relation to SOEs is to (1) dominate and expand in key industries affecting national security and the ‘commanding heights’ of the national economy, (2) compete on equal footing with non-state enterprises in competitive industries, (3) compete mainly among SOEs in natural monopoly industries such as telecom and electricity, and (4) serve as the main instrument of industrial policy: develop pillar industries such as steel and automobiles and create national champions to compete internationally.29

C

Enforcement – Courts versus Administrative Decision-making

In the US,30 the DOJ and the FTC have joint responsibility for enforcing competition laws. Although the Antitrust Modernization Commission recognized that their roles overlap and that there are disadvantages from this dual system of enforcement,31 it did not recommend a structural change given ‘the significant costs and disruption of moving to a single-agency system at this point in time would likely exceed the benefits’.32 In China, until 2018, three agencies (the Ministry of Commerce, the National Development and Reform Commission (NDRC) and the State Administration for Industry and Commerce (SAIC)) were jointly responsible for the enforcement of the AML. Both SAIC and NDRC were responsible for unilateral conduct and there was considerable overlap between them. However, during 2018,33 the three agencies were moved into a new government agency, the State Administration for Market Regulation. Within this, the Anti-Monopoly Bureau is responsible for the enforcement of the AML, including abuse of dominance. Equivalent agen-

27 Stephen Calkins, ‘Competition Law in the United States’ in Vinod Dhall (ed), Competition Law Today: Concepts, Issues, and the Law in Practice (Oxford University Press 2007) 36. 28 See, for example, the objects clause, Article 7 (potential carve-out for SOEs in significant industries), and various provisions that emphasize the ‘national interest’ in their application. 29 Thomas Cheng, ‘Roundtable on Competition Neutrality’ (Note to OECD Competition Committee Meeting No DAF/COMP/WD(2015)49, 11 June 2015) 5. 30 For details regarding the US, see Modernization Commission Report (n 17), ch 2. 31 Potentially these include additional cost, duplication of effort, uncertainty and inconsistency. However, the Commission found that the agencies coordinated their operations and these disadvantages were not significant. 32 Modernization Commission Report (n 17), 129–30. 33 Cheng Liu and Audrey Li, ‘2018 in Review – the Beginning of a New Era of AML Enforcement’, King & Wood Mallesons (9 January 2019), https://​www​.kwm​.com/​en/​cn/​knowledge/​insights/​2018​-in​ -review​-the​-beginning​-of​-a​-new​-era​-of​-aml​-enforcement​-20190109.

Unilateral conduct analysis: focus on harm in multiple guises  211 cies have been set up at the provincial level. These changes are intended to increase efficiency and consistency in enforcement. In the EU, the EC has primary responsibility for the enforcement of competition law. Within the EC, the Directorate-General for Competition has responsibility for ‘elaborating’ policy, investigating alleged anticompetitive conduct, and enforcement action. Since 2004, the national competition authorities (NCAs) of EU Member States can apply EU competition rules directly, although the European Competition Network determines which authority is best placed to undertake an investigation. Under Regulation 1/2003, the EC is responsible for ensuring that the NCAs apply EU competition rules in a uniform manner. However, Dunne observes that divergences exist with respect to many aspects of the decentralized enforcement framework.34 She is critical of the EC for its failure to articulate why harmonization is necessary, exploring this via an analysis of fining practices, and concludes: ‘The key concerns in devising a converged approach, therefore, are to balance consistency with sufficient flexibility, and to reflect a EU-wide consensus on fining practices, not merely a topdown “re-centralisation” of sanctions.’35 There are two models for the enforcement of competition law, as well as a hybrid of the two. The first model is one in which conduct that may contravene the competition law is investigated by the competition authority, which, subject to its finding, prosecutes the parties concerned in court. This is the approach of the DOJ. Antitrust cases, including those relating to section 2 of the Sherman Act, are heard initially by the appropriate District Court. Appeals from this court are heard by the Court of Appeals (Circuit Court) and then may be appealed to the Supreme Court.36 In other countries, such as Australia and New Zealand, the competition authority must also prosecute allegedly anticompetitive conduct in court. In some cases, such as in Hong Kong, following an investigation of the alleged conduct, adjudication is by a specialist tribunal rather than a court. The second model is where the competition authority is empowered to investigate and then make an administrative decision concerning allegedly anticompetitive conduct. These administrative decisions then may be subject to judicial review. This is the model applicable in the EU. The EC can issue prohibition decisions, impose fines and enter into binding agreements with companies. Its decisions can be appealed, first to the Court of First Instance (the General Court) and then to the ECJ.37 Corresponding arrangements apply to NCAs, but decisions by a superior court in a member country may be referred to the ECJ. As in the EU, in China the competition authorities make administrative decisions. Article 53 of the AML provides that companies that disagree with a decision may apply for administrative review or file an administrative lawsuit. A party may challenge a finding of abuse by law enforcement agencies in the following cases: (1) where the main evidence (in support of such act) is insufficient; (2) where incorrect laws and regulations are applied; (3) where statutory procedures are violated; (4) where such act is performed in excess of authority; or (5) where official powers are abused.

34 Niamh Dunne, ‘Convergence in Competition Fining Practices in the EU’ (2016) 53 Common Market Law Review 453. 35 Ibid, 2, 31, 36 State attorneys general may bring a civil action in relation to alleged anticompetitive unilateral conduct. 37 Decisions by a superior court in a member country may be referred to the ECJ.

212  Research handbook on methods and models of competition law In the US, the FTC operates something of a hybrid of the two models. It enforces the antitrust laws through section 5 of the FTC Act, which prohibits unfair methods of competition through action in the federal courts. It can investigate possible contraventions and ‘may enforce Section 5 through internal administrative litigation (known as Part III proceedings) before an administrative law judge whose decision is subject to review by the five FTC Commissioners and then a federal court of appeals’.38 In January 2017, for example, the FTC filed a case in federal court against Qualcomm, alleging that it had used its monopoly position in baseband chips for mobile phones to impose anticompetitive licensing terms for standard-essential patents and that this restricted the ability of its competitors in baseband chips to compete.39 Each of these enforcement systems has advantages and disadvantages. One of the advantages of the competition authorities making administrative decisions concerning anticompetitive conduct is their expertise and hence ability to address complex economic matters, contrasting with the generalist nature of courts. However, a key concern is that the competition authority acts as investigator, prosecutor, jury and sentencing judge.40 Court assessment is expensive, notoriously unreliable when a jury is involved, and protracted – although appeals from administrative decisions may be even more protracted. D

Private Enforcement

Frequently, there is provision for the private enforcement of competition laws, although the extent of public versus private enforcement varies significantly between jurisdictions. In the US, most antitrust cases are brought by businesses and individuals who have been injured by anticompetitive conduct, often as a class action. Private litigants can sue in federal court seeking (treble) damages, injunctive relief and reasonable legal fees. The former, together with the asymmetric cost rule, tends to make ‘private actions for damages an instrument of deterrence and punishment’.41 Private parties can also seek court orders preventing anticompetitive conduct (injunctive relief) or bring a case under state antitrust laws. As noted by Adkinson et al, ‘the Agencies play an important role in private enforcement through their amicus participation in the Supreme Court’.42 There is no right of private action under the FTC Act. The balance between private and public enforcement in the EU is the reverse of that in the US. Lack of provision for treble damages may help to explain why there is relatively little private enforcement.43 Recently, attempts have been made to encourage private enforcement, partly to ease the burden on the EC. Nevertheless, these measures ‘are designed to create an

38 Modernization Commission Report (n 17), 129 (footnotes omitted). Typically, states investigating a matter arising under the federal antitrust laws will jointly investigate with either the DOJ or the FTC, or they may conduct a separate investigation. 39 FTC v Qualcomm, Case 5:17-cv-00220, 17 January 2017. 40 ‘Prosecutor, Judge and Jury – Enforcement of Competition Law in Europe Is Unjust and Must Change’, The Economist (18 February 2010). 41 Wouter PJ Wils, ‘The Relationship between Public Antitrust Enforcement and Private Actions for Damages’ (2009) 32 World Competition 3, 17 (emphasis added) and see nn 62–73, http://​ssrn​.com/​author​ =​456087. See also the discussion of the objective of penalties below. 42 Adkinson et al (n 16), 8. 43 Other reasons include the relative newness of the provisions, but perhaps more importantly a concern about having the conduct stopped rather than seeking compensation, which is costly and uncertain.

Unilateral conduct analysis: focus on harm in multiple guises  213 effective system of private enforcement by means of damages actions that complements, but does not replace or jeopardise, public enforcement’.44 Article 50 of China’s AML provides for private action through the intermediate people’s courts by companies and individuals seeking damages due to anticompetitive conduct. In 2012, the Supreme People’s Court provided guidance to the courts as to how to handle private antitrust claims. However, perhaps unsurprisingly, private enforcement of abuse of dominance cases has been relatively insignificant.45 What is the right balance between public and private litigation? Better access to justice for those damaged by anticompetitive conduct increases the probability of successful detection and prosecution. Merrett and Smith provide an assessment of private competition law litigation in Australia and conclude that it is woefully deficient.46 Suggestions for encouraging more private litigation include: ●● reducing the duration and complexity of proceedings – what can be done to get matters through the Court process more quickly? ●● managing the risk of costs – is there scope to allow relief from the usual approach to costs where private proceedings benefit the general public? ●● improving the utility of remedies – can we devise remedies which both encourage litigation and provide effective relief to victims of anticompetitive conduct?47 E Penalties Assuming that the role of penalties is to deter anticompetitive conduct, penalties should be sufficient to eliminate benefit from the conduct, taking into account the probability of the conduct being detected (and successfully prosecuted where conduct is assessed by a court). Historically, courts have been reluctant to impose maximum penalties and consequently they have failed to deter adequately. Recently, in many jurisdictions, the maximum allowable penalty has increased significantly.48 In 2018, contravention of section 2 of the Sherman Act carried a penalty of up to US$10,000,000 for a corporation, or US$350,000 for a person, or imprisonment not exceeding three years, or both, at the discretion of the court.49 In the EU, having established an abuse of dominance, the EC has power, pursuant to Article 23 of Regulation 1/2003, to impose a fine of up to 10 per cent of the annual turnover of the

44 European Commission, White Paper on Damages Actions for Breach of the EC Antitrust Rules (COM(2008)165 of 2 April 2008). 45 For a detailed analysis, see Alexandr Svetlicinii, ‘Private Litigation under China’s Anti-Monopoly Law: Empirical Evidence and Procedural Developments’ (2017) 7 KLRI Journal of Law and Legislation 163. 46 Alexandra Merrett and Rhonda Smith, ‘The Public Benefits of Private Litigation’ (2014) 19 State of Competition 1. 47 Ibid. 48 Unlike in some other jurisdictions, the Sherman Act refers to ‘punishment’; others refer to deterrence. 49 Federal Trade Commission, ‘Guide to Antitrust Laws: The Antitrust Laws’, https://​www​.ftc​.gov/​ tips​-advice/​competition​-guidance/​guide​-antitrust​-laws/​antitrust​-laws.

214  Research handbook on methods and models of competition law Table 9.1

Highest EU fines

Company

Year

Fine (€)

Google Android

2018

5 billion

Google Shopping

2017

2.7 billion

Intel

2009

1.45 billion

Qualcomm

2018

1.2 billion

Microsoft

2004

794 million

Servier

2014

582 million

Telefonica

2007

207 million

company/companies concerned.50 Table 9.1 lists the highest EU fines, all of which relate to abuse of dominance.51 Similarly, Article 47 of China’s AML provides that where a business operator has violated the provisions of the law by abusing its dominant market position, the antitrust enforcement agency shall order the business operator to stop the illegal act and confiscate the illegal income. A fine of 1 to 10 per cent of the sales revenue of the preceding year shall be imposed. The exact amount of the fine depends on factors such as the nature, extent, duration and circumstances of the illegal behaviour. The rules provide that such fines could be reduced or waived if the business stops its conduct.52 In 2015, Qualcomm was fined US$975 million (€863 million) for abuse of its dominance in wireless technology by excessive pricing.53 The following year, Tetra Pak was fined US$97 million for abusing its dominant position in the market for paper-based aseptic packaging equipment, technology services for such equipment, and paper-based aseptic packaging materials.54 Then, in 2017, two pharmaceutical firms, Second Pharma Co, Ltd of Zhejiang Province and Tianjin Handewei Pharmaceuticals Co, Ltd, were fined 443,900 yuan for charging excessively high prices and for refusal to deal.55 Competition law may provide for a range of other penalties, and these tend to be more variable across jurisdictions. For example, the US DOJ can seek criminal sanctions for offences under section 2 of the Sherman Act56 – a penalty not available to the FTC, in the EU or in China. Both the US and the EU57 statutes provide for divestiture. Other remedies may be behavioural, such as the issue of cease-and-desist orders, or a requirement to deal with rivals on certain terms, or equitable relief.

As at September 2018. Ana Zarzalejos, ‘The 7 Biggest Fines the EU Has Ever Imposed against Giant Companies’, Business Insider (19 July 2018), https://​www​.businessinsider​.com/​the​-7​-biggest​-fines​-the​-eu​-has​-ever​ -imposed​-against​-giant​-corporations​-2018​-7/​?r​=​AU​&​IR​=​T/​#6​-servier​-fined​-582​-million​-in​-2014​-6. 52 Zhenguo Wu, ‘Perspectives on the Chinese Anti-Monopoly Law’ (2008) 75 Antitrust Law Journal 73, 111. 53 NDRC Administrative Penalty Decision [2015] No 1, http://​ www​ .ndrc​ .gov​ .cn/​ zwfwzx/​ xzcf/​ 201503/​t20150302​_754177​.html. 54 SAIC Administrative Penalty Decision [2016] No 1, http://​www​.competitionlaw​.cn/​info/​1025/​ 23864​.htm. 55 NDRC Administrative Penalty Decision [2017] No 2, http://​ www​ .ndrc​ .gov​ .cn/​ xzcf/​ 201708/​ t20170815​_861782​.html. 56 Although this power has not been used in practice. 57 Council Regulation (EC) 1/2003 of 16 December 2002 on the implementation of rules on competition laid down in Articles 81 and 82 of the Treaty [2003] OJ L 1, Art 7. 50 51

Unilateral conduct analysis: focus on harm in multiple guises  215

III METHODOLOGY For its conduct to be susceptible to the competition law prohibitions relating to unilateral conduct, a firm must possess the requisite degree of market power and then conduct must be found to be anticompetitive. Several policy decisions (which may be embodied in the law) must be made at this point. The first is whether the conduct will be assessed based on its form or its effect. The second concerns the inevitable risk of an erroneous decision and the nature of acceptable risk. Less efficient firms will fail to survive in competitive markets in the long run. This is not a competition concern; rather, it is the market at work to ensure survival of the fittest. This contrasts with anticompetitive conduct that may cause firms to fail, even when they are efficient. A significant issue for competition authorities and the courts is to correctly distinguish between the two. Various tests developed to analyse this issue are discussed later. A

Establishing Market Power

For unilateral conduct to raise competition concerns, the firm concerned must possess SMP. Competition authorities generally provide guidance as to how they will assess market power.58 The usual approach when assessing a firm’s market power is to consider the factors that would enable it to act somewhat independently of market conditions. The changes that result in a substantial lessening of competition are caused by the same factors as those that confer SMP. These factors are discussed in detail in Chapter 11 and only a brief summary of the indicators of SMP is provided below. Most assessments of market power begin by considering the market share of the firm under investigation, as well as its share relative to other firms in the relevant market. Market shares are often used to provide a ‘safe harbour’, such that firms with low market shares may not be considered further. The US FTC, for example, states that ‘[c]ourts look at the firm’s market share, but typically do not find monopoly power if the firm (or a group of firms acting in concert) has less than 50 per cent of the sales of a particular product or service within a certain geographic area’.59 However, market shares are not a good indicator of market power – a small market share may be used strategically, while a firm without current competitors may lack market power if barriers to entry and/or expansion are low. Evidence of entry does not necessarily mean that barriers to entry are low. What is important is the likelihood of entry, whether it will be an effective constraint on the incumbent firm(s), and whether it will be sufficiently timely to prevent significant harm to consumers.60 Even if barriers to entry are high, a firm will not possess SMP if buyers possess sufficient countervailing power. A challenge for assessing

58 See, for example, Commission Communication (EC) 2009/C of 24 February 2009 on Guidance from the Commission’s enforcement priorities in applying Article 82 of the EC Treaty to abusive exclusionary conduct by dominant undertakings [2009] OJ C 45; US Department of Justice, Competition and Monopoly: Single Firm Conduct Under Section 2 of the Sherman Act (2008), https://​www​.justice​.gov/​ atr/​competition​-and​-monopoly​-single​-firm​-conduct​-under​-section​-2​-sherman​-act. 59 Federal Trade Commission, ‘Monopolization Defined’, https://​www​.ftc​.gov/​tips​-advice/​ competition​-guidance/​guide​-antitrust​-laws/​single​-firm​-conduct/​monopolization​-defined. 60 This is referred to as the LET test. See FTC and DOJ, Horizontal Merger Guidelines (19 August 2010) 27ff, https://​www​.justice​.gov/​atr/​file/​810276/​download.

216  Research handbook on methods and models of competition law SMP arises when a firm’s key patent is expiring: at what point does its market power cease to be substantial?61 If the exercise of SMP enables a firm to earn supra normal profit, this might suggest that high profits are indicative of that power. However, in United Brands,62 an early EU case, the ECJ dismissed profitability as a useful indicator of SMP.63 In 2003, a report commissioned by the then UK Office of Fair Trading64 noted difficulties with its use, including: ●● it is not well established what profitability analysis should be measuring; and ●● the choice of various measures and interpretation issues. Despite these issues, the report found that profitability analysis should be seen as one among a number of complementary economic indicators and techniques that can be used together in a competition policy analysis. Niels, Jenkins and Kavanagh note that the EC conducted a profitability analysis as part of its review of the extent of competition in the payment cards and retail banking industries in the EU.65 Nevertheless, the US DOJ has cautioned against over-reliance on evidence of high profits, price-cost margins and demand elasticity as evidence of market power.66 Market power may be leveraged from one market into another, often through tying arrangements.67 In Tetra Pak II, the ECJ referred to and accepted ‘associative links’ between two markets in assessing whether Tetra Pak was dominant in the relevant market.68 A claim of leveraging requires establishing: ●● the existence of a separate horizontal or vertical market to which the market power can be transferred; ●● close connection between non-dominated and dominated markets; and ●● that the abusive conduct must stem from the dominant position of an undertaking.69

61 See, for example, Australian Competition and Consumer Commission v Pfizer Australia Pty Ltd (2015) 323 ALR 429. 62 United Brands Company v Commission of the European Communities [1978] ECR 207. 63 Prior to this, in 1956 in the cellophane case, the court had accepted that profitability was an indicator of market power. 64 Oxera, ‘Assessing Profitability in Competition Policy Analysis’ (Office of Fair Trading, Economic Discussion Paper 6, July 2003) 1.3, https://​www​.oxera​.com/​wp​-content/​uploads/​media/​oxera​_library/​ downloads/​ reports/​ oft​ -assessing​ -profitability​ .pdf. See also Gunnar Niels, Helen Jenkins and James Kavanagh, Economics for Competition Lawyers (2nd edn, Oxford University Press 2016) ch 3. 65 European Commission, Sector Inquiry under Article 17 Regulation 1/2003 on Retail Banking: Interim Report I Payment Cards (12 April 2006), http://​ec​.europa​.eu/​competition/​sectors/​financial​ _services/​inquiries/​interim​_report​_1​.pdf; European Commission, Sector Inquiry under Article 17 Regulation 1/2003 on Retail Banking: Interim Report II Current Accounts and Related Services (17 July 2006), http://​ec​.europa​.eu/​competition/​sectors/​financial​_services/​inquiries/​interim​_report​_2​.pdf. 66 US Department of Justice, Single Firm Conduct under Section 2 of the Sherman Act (2008) 28. 67 For many years, Chicago School economists argued that leveraging by tying was not economically rational as the amount of economic rent that can be extracted from a supply chain is finite. So, if it is all extracted at one level in the supply chain, it cannot be added to by linking this to another level in the supply chain. However, this criticism is invalid if the monopolist cannot extract all of the consumer surplus in the market in which it has market power. 68 Case C-333/94 PECJ, Tetra Pak International SA v Commission of the European Communities [1996] ECR I-5951, [2], [31]. 69 Eduardo Cunha, ‘Leveraging Abuses’, Freie Universitat Berlin (17 June 2017), https://​wikis​.fu​ -berlin​.de/​display/​oncomment/​Leveraging+abuses.

Unilateral conduct analysis: focus on harm in multiple guises  217 Leveraging was the basis for establishing market power in the Microsoft cases on both sides of the Atlantic. For example, in the EU, the Commission’s basic argument was that Microsoft extended its market power from PC operating systems (where Windows controlled over 95 per cent of the market) into a complementary market – that of the operating systems for work group servers.70 In June 2017, in its Google Shopping decision, the EC stated that ‘Google has leveraged its market dominance in general internet search into a separate market, comparison shopping’.71 Of course, leveraging per se is not anticompetitive; rather, it is the conduct it enables that may raise concerns, as recognized by the US Supreme Court in Trinko.72 B

Form-based versus Effects-based Approach

To distinguish between legitimate and illegitimate unilateral conduct, assessment may be based on the form of conduct or its effects. The focus of most jurisdictions is on effects. In some, such as Australia, assessment is based on either the purpose or the effect of the conduct.73 However, traditionally, Europe adopted a form-based approach, which meant: ‘Once a company was found to be dominant, its “special responsibility” not to impair competition meant that it could not engage in certain forms of behaviour, such as pricing below variable cost, tying products, or offering loyalty rebates.’74 If the alleged anticompetitive unilateral conduct fell into one of these categories, it was, in effect, per se illegal. This approach has been widely criticized. It is argued that conduct that fails to have an adverse effect on the competitive process should not be of concern. More importantly, it fails to recognize that conduct – especially unilateral conduct – may have different effects in different circumstances, damaging competition in some cases but not others. In 2008, the EC issued a guidance document which introduced a more effects-based approach.75 This had only mixed acceptance in the courts, but in the Intel decision the ECJ rejected the findings of the General Court, which had disregarded the EC’s analysis of effects, and it endorsed an effects-based approach.76

70 Christos Genakos, Kai Uwe Kühn and John Van Reenen, ‘The European Commission versus Microsoft: Competition Policy in High-Tech Industries’, CentrePiece (Summer 2007) 3. 71 Google Search (Shopping) (Case AT.39740) Commission Decision [2018] OJ C 9/11. 72 Verizon Communications Inc v Law Offices of Curtis V Trinko, 540 US 398 (2004). 73 Prior to November 2017, it had regard only to purpose. 74 Oxera, ‘The Latest Intel from Luxembourg: An Advocate General Prefers Effects Analysis in Abuse Cases’ (November 2016), https://​www​.oxera​.com/​agenda/​the​-latest​-intel​-from​-luxembourg​-an​ -advocate​-general​-prefers​-effects​-analysis​-in​-abuse​-cases/​. 75 Commission Communication (EC) of 24 February 2009 on Guidance on the Commission’s enforcement priorities in applying Article 82 of the EC Treaty to abusive exclusionary conduct by dominant undertakings [2009] OJ C 45. 76 Case C‑413/14 P Intel Corporation Inc v European Commission, Commission Decision 1/2003 [2017] ECR I-632 [119].

218  Research handbook on methods and models of competition law C

Type I and Type II Errors

Having established that a firm whose unilateral conduct may be anticompetitive possesses the requisite degree of market power, the next step for the competition authority and/or the court is to determine whether the conduct actually is anticompetitive. But distinguishing between conduct that is procompetitive or competitively neutral and anticompetitive conduct frequently is not easy. A form-based approach may be subject to error because types of conduct are categorized incorrectly – for example, all loyalty discounts may be prohibited – or because particular conduct is wrongly considered to belong in a particular category. However, there is no less risk that conduct will be incorrectly assessed when an effects test is applied. Often the effect of conduct depends on particular circumstances, such as the extent of a discount and the period for which it is offered. A test that aims to avoid overlooking anticompetitive conduct is more likely to incorrectly conclude that legitimate conduct is anticompetitive (Type I error), while a test that focuses on avoiding incorrectly concluding that conduct is anticompetitive is more likely to falsely identify anticompetitive conduct as legitimate conduct (Type II error). Reducing one type of error increases the risk of the other. Consequently, there is a trade-off between them. Generally, policymakers aim to minimize Type I error,77 reflecting the Chicago School belief that markets are self-correcting. Devlin and Jacobs are critical of these generalizations.78 D Tests Economists and lawyers have sought to develop tests that help distinguish between competitive and anticompetitive conduct. Each has positive and negative aspects, making a particular test more appropriate in some circumstances than in others. Perhaps conceptually the simplest of these tests is the comparison of the market situation with the alleged conduct (the factual) and without it (the counterfactual), the difference being the effect of the conduct on competition. Perhaps the primary benefit of a counterfactual is that it is relatively simple and generally does not require (much) additional factual information. However, it has been criticized primarily because it is often unrealistic or speculative, and there may be several possible counterfactuals whose outcomes may be inconsistent. Four other commonly used tests are explained briefly in Table 9.2 and their advantages and disadvantages are identified. Although used in other jurisdictions, until relatively recently, the form-based approach made them of little relevance in the EU, except in relation to claims of predation and anticompetitive vertical agreements.

77 An exception is where the law treats certain types of conduct (for example, cartels) as per se illegal. In these cases, the view, at least implicitly, is that the probability that the conduct is anticompetitive is so high that the risk of a Type I error is so low that it can be ignored. However, unilateral conduct is generally not treated as per se, the exception until relatively recently being the treatment of tying in the US. 78 Alan Devlin and Michael Jacobs, ‘Antitrust Error’ (2010) 52 William & Mary Law Review 75, 79–81.

Unilateral conduct analysis: focus on harm in multiple guises  219 Table 9.2

Commonly used tests

Test

Description

Comments

Profit sacrifice

Conduct is considered unlawful if

- Criticized for a focus on the short run where competition policy

short-term profit is sacrificed for higher focus is on the long run long-term profit.

- Requires knowledge of competitive price - May be under- or over-inclusive – conduct may be harmful to competition but not involve profit sacrifice and vice versa The US DOJ has explicitly rejected the use of this test in relation to section 2 of the Sherman Act.a

No economic sense

Conduct should only be condemned if it - May be over-inclusive – for example, successful R&D which is economically rational only because it excludes previously efficient rivals tends to eliminate or lessen competition. - Difficult to apply when the relevant conduct has both positive and negative effects The US DOJ has consistently applied this test in cases under section 2 of the Sherman Act.b

Equally efficient firm Conduct should be unlawful only if it

- Too lenient – allows exclusion of new firms that are not yet

would be likely to exclude a firm which equally efficient but could have become so if they had time to grow is at least as efficient as the dominant

- Exclusion of an equally efficient firm that is not an effective

firm.

competitive constraint has few, if any, implications for competition - Use in relation to non-price conduct poses problem for determining the counterfactual Applied to alleged price discrimination in Post Danmark.c

Price squeeze or

Occurs where the difference between

This test has been a feature of telecommunications cases.

margin squeeze

the wholesale price of an input and the

In the EU, price/margin squeeze is treated as a stand-alone abuse.

downstream price of the output fails to

In the US, it is subject to regulation rather than section 2 of the

allow a reasonable (profit) margin.

Sherman Act.d

Notes: a US Department of Justice, Competition and Monopoly: Single-Firm Conduct Under Section 2 of the Sherman Act, ch 3, https://​www​.justice​.gov/​atr/​competition​-and​-monopoly​-single​-firm​-conduct​-under​-section​-2​ -sherman​-act​-chapter​-3; b Gregory Werden, ‘The “No Economic Sense” Test for Exclusionary Conduct’ (2006) 31(2) Journal of Corporation Law 293; c Case C-209/10 Post Danmark AS v Konkurrencerådet [2012] ECR I-172, [38]–[40]; d Germain Gaudin and Despoina Mantzari, ‘Margin Squeeze: An Above-Cost Predatory Pricing Approach’ (Düsseldorf Institute for Competition Economics, Discussion Paper 207, January 2016), http://​www​ .dice​.hhu​.de/​fileadmin/​redaktion/​Fakultaeten/​Wirtschaftswissenschaftliche​_Fakultaet/​DICE/​Discussion​_Paper/​207​ _Gaudin​_Mantzari​.pdf.

E

The Anatomy of a Unilateral Conduct Case

Predatory pricing is a form of unilateral anticompetitive conduct that illustrates well the methodology employed to assess such conduct, as well as the not-insignificant differences of approach between jurisdictions, reflecting differences both in legislative provisions and in philosophy.

Box 9.1

CASE STUDY: PREDATORY PRICING

Predatory pricing is a form of exclusionary pricing conduct. It occurs when a firm temporarily lowers its price substantially in a relevant market to tame aggressive rivals, to foreclose rivals and/or to deter entry, after which the firm can raise its price above the competitive level. When investigating a claim of predatory pricing, it must be established that

220  Research handbook on methods and models of competition law the alleged predator possesses SMP for its pricing conduct to raise competition concerns. Its market power may be because it faces few actual rivals and/or due to high barriers to entry. While any firm can choose to reduce its prices, in order to reduce the market price below the equilibrium price and sustain the reduction it is necessary for there to be significant excess capacity, as shown in the diagram below. If this does not exist, the predator will need to create it prior to, or simultaneously with, its pricing conduct.

From a competition law perspective, the challenge is to determine whether observed price cutting is predatory and so anticompetitive, or competition at work and procompetitive. Low or reduced prices are not necessarily predatory. The legal tests for predatory pricing differ between jurisdictions, but generally require showing that the reduced price is below some measure of cost – a temporary profit sacrifice. Since the decision of the US Supreme Court in Brooke Group in 1993,79 the relevant measure of cost is average variable cost or average avoidable cost80 of rivals. In the EU, following the AKZO case,81 the cost test was satisfied if price was below the dominant firm’s average variable cost or, if part of a plan to eliminate competitors, if price was greater than average cost but less than total cost. However, in its 2008 Guideline, the EC changed these cost standards to average avoidable cost and long run incremental average cost – this change was reflected in the ECJ decision in Post Danmark in 2012.82 The second limb of the test for predatory pricing in Brooke Group is the requirement that the alleged predator should be able to recover (recoup) its losses in that market and within a relatively short period. Yet, the losses incurred through predation can be viewed as an investment in entry deterrence and as such are unrecoverable sunk costs. This, together Brooke Group Ltd v Brown & Williamson Tobacco Corp, 509 US 209 (1993). In the long run, if prices are below the economic cost of supply, the firm should exit the market, subject to certain qualifications. 81 AKZO Chemie BV v Commission (Case 62/86 R) [1986] ECR I-00359. 82 Case C-209/10 Post Danmark AS v Konkurrencerådet [2012] ECR I-172, [44]. 79 80

Unilateral conduct analysis: focus on harm in multiple guises  221 with the difficulty of establishing the firm’s costs, means that the US standard for predatory pricing is very high. This is deliberate, making the likelihood of successful prosecution low indeed. The Supreme Court opined that predatory pricing was rare, and even more rarely successful, and consumers should be allowed the benefit of lower prices. This position was reaffirmed in 2009 in linkLine.83 The requirement for recoupment has not been adopted in the EU. It was rejected in AKZO and this was reaffirmed in 2009 in Wannadoo.84 The benefit to the predator may be simply the ability in future to price above the competitive level, and/or the reputational effect that deters future entry into this or some other market, as recognized in the 2008 EC Guideline. Other jurisdictions tend to require satisfaction of a cost-based test but may or may not require demonstration of the likelihood of recoupment.

IV

THE DIGITAL ENVIRONMENT

The digital revolution has challenged competition law analysis. Traditional approaches need to be adjusted to accommodate new or altered business models, and concerns that they may result in competition for, rather than in, the market. There are issues associated with market definition, as well as with how a firm’s market power should be assessed. As the business landscape can alter rapidly, it is less clear what constitutes transitory market power and when markets are contestable. Discussion of these questions is ongoing, but what is clear is that competition analyses applied to these new industries need to be highly fact specific. In this section, one feature of the digital economy is discussed: platform businesses. A

Platform Businesses

Directly or indirectly, a platform business provides access for one group of platform customers to another group of platform customers – this has been described as a ‘matchmaker’ role.85 Absent the platform, transaction costs or other factors would prevent or limit this connection between parties. A platform can ‘decrease search costs and reduce deadweight loss by brokering exchanges that never would have happened without it’.86 Most media businesses are traditional platform businesses. However, platform businesses often have multiple customer groups, offer a menu of products to exploit the synergies available on the platform, and are continuously evolving – indeed, platform businesses compete by innovating. One type of platform business is based on sharing – for example, Uber and Airbnb. A second type includes businesses such as Google and Facebook, which can be

Pacific Bell Telephone Co v linkLine Communications Inc, 555 US 438 (2009). France Télécom SA v Commission of the European Communities (Case C-202/07 P) [2009] ECR I-214, [69]–[72]. 85 David S Evans and Richard Schmalensee, Matchmakers: The New Economics of Multisided Platforms (Harvard Business Review Press, 2016). 86 Alec Stapp and Ryan Hagemann, ‘Hearings on Competition and Consumer Protection in the 21st Century’ (Submission to FTC, 20 August 2018) 2, https://​www​.ftc​.gov/​system/​files/​documents/​public​ _comments/​2018/​08/​ftc​-2018​-0050​-d​-0027​-155052​.pdf. 83 84

222  Research handbook on methods and models of competition law loosely described as being information based. A third type comprises platforms that enable direct transactions, such as a stock exchanges and credit card issuers.87 Indirect network effects are a distinctive feature of platform businesses. They result from the value that one group of customers places on access to the platform being positively related to the size of another group of platform customers.88 Indirect network effects may be very significant for one group of customers (for example, advertisers) but insignificant for other customers on the same platform (for example, readers/viewers). Given this, a platform operator charges different prices to different customer groups. Customers whose network effects are relatively small are charged a low, zero or negative price, while customers with greater network effects are charged a higher price. Consequently, prices for particular customer groups are not reflective of the cost of supply, although in the long run it must be true for the platform business as a whole. B

Market Definition

The question of the relevant market for a competition analysis for platform businesses tends to focus on whether only one side of the platform forms the market, or whether the relevant market should include other sides as well. Having identified the conduct that gives rise to the inquiry and the product it relates to, the next step is to identify substitute products that would constrain a small but significant, non-transitory increase in price (SSNIP). The difficulty here is that if the price for a customer group is zero, a SSNIP has no meaning. If the SSNIP is applied to the positive price charged by the platform (and there may be more than one), this prejudges the one/two-sided question in favour of a single market. An alternative approach is to begin by considering that the service supplied by the platform is access of one group of customers to another, although the form that the service takes differs between customer groups. Given this, the next question is whether the conduct at issue occurs in relation to a particular group of customers but is constrained by the reaction of another group of customers – that is, whether there are significant indirect network effects. Platform businesses are divided into transaction platforms and non-transaction platforms. The latter include traditional media businesses where there is no direct transaction between advertisers and readers, for example, and network effects for readers are small. In other words, if the publisher raised advertising rates, there would be little response from readers. This lack of constraint justifies ignoring responses from readers and focusing on the response of advertisers alone to a SSNIP. It should be noted, however, that the problem is asymmetric. If the price increase applied to newspapers, loss of readers would be of concern to advertisers and market definition would include the advertising side of the market. The approach is different for transaction platforms. Here, different customer groups interact directly, facilitated by the platform. In this case, network effects are likely to be significant 87 An interesting issue concerns online trading. To the extent that a retailer, for example, creates an online facility for its customers, in principle this is no different from its bricks and mortar operations. However, if a business is set up to bring together parties that want to sell various products, and parties interested in buying such products, then this is a platform that makes a profit from facilitating these transactions but not participating in them – a business such as eBay. 88 This contrasts with direct network effects, which occur when customers obtain more value from a product the greater the number of other consumers of the product. Telephone use is an obvious example.

Unilateral conduct analysis: focus on harm in multiple guises  223 on both sides of the platform. One group of customers may act in a way that alters the pricing experienced by the other group – for example by offering a subsidy. In Amex, merchants offered rewards to cardholders who settled a transaction using a card other than an Amex card.89 This interaction alters the pricing structure previously determined by the platform operator; in Amex, this included points from the loyalty programme. Less card usage by cardholders meant less demand by merchants for accreditation. So, responses of both (all) customer groups must be taken into account in assessing constraints on the decision-making of the platform operator. The usual guide to market definition for platform markets is: ●● non-transaction platform – define the market based on the customer group affected by the conduct; and ●● transaction platform – define the market by including all customer groups. However, even for a non-transaction platform, directly or indirectly, conduct affects pricing. This is because the price set by the platform operator for one customer group takes into account the price to be paid by other customer groups and vice versa. This means that the price charged to one group cannot be understood without also considering the others.90 So all customer groups should be included in the market. Having determined the conduct and the type of market, inquiries about close substitutes to which customers can turn in response to a SSNIP are required It is not only the product dimension of markets that raises issues in platform businesses: Defining the geographic dimension of a two-sided platform suffers the same problems associated with the HMT as the product dimension. To these it adds spatial issues. The HMT specifies that in order to identify the market, starting from the location of the firm whose conduct raises the competition concern, the SSNIP is applied to successively enlarged areas until it is profitable because the sources of substitute supply have been included within the market. This tends to imply contiguous or proximate areas of competition.91 Yet in globalised markets clearly this need not be the case.92

To date, few of the cases considered by the courts have assessed the geographic dimension of a market involving platform businesses. A notable exception is a Chinese case, Qihoo 360 v Tencent.93 The products of Tencent include instant messaging, microblogging and online games. These products are free, funded by the sale of advertising. Qihoo 360 was the leading provider of internet security software, while Tencent was the leading provider of instant

Ohio v American Express Co, 585 US __ (2018). ‘France: Google Wins Court Decision vs Evermaps’, Competition Policy International (29 November 2015), https://​www​.competitionpo​licyinternational​.com/​france​-google​-wins​-court​-decision​ -vs​-evermaps/​. 91 For a critical appraisal of this approach, see Steven E Crane and Patrick J Welch, ‘The Problem of Geographic Market Definition: Geographic Proximity vs Economic Significance’ (1991) 19 Atlantic Economic Journal 12. 92 Rhonda L Smith and Arlen Duke, ‘The Geographic Dimension of Markets: Some Observations’ (2017) 25 Competition and Consumer Law Journal 1, 20. 93 For a more detailed description of this case, see David Evans and Vanessa Zhang, ‘Qihoo 360 v Tencent: First Antitrust Decision by the Supreme Court’, Competition Policy International (21 October 2014), https://​www​.competitionpo​licyinternational​.com/​qihoo​-360​-v​-tencent​-first​-antitrust​-decision​-by​ -the​-supreme​-court. 89 90

224  Research handbook on methods and models of competition law messaging. The alleged anticompetitive conduct concerned the circumstances surrounding Tencent’s introduction of security software. The geographic dimension of the relevant market was not considered at first instance. However, on appeal to the Supreme People’s Court, Qihoo 360 claimed that the geographic dimension of the market was limited to mainland China and Tencent argued that it was global. The court found that ‘the operators and users of instant messaging services, however supplied, are not limited to those based in mainland China. Due to the openness and interoperability of the internet, operators and users are not confined by national borders.’94 Users could be located outside of mainland China and suppliers could provide service from anywhere in the world. Determining the geographic dimension of the market in this case was relatively straightforward. However, the court provided little guidance as to how this task should be approached in more complex cases. C

Market Power

A threshold issue in competition inquiries involving unilateral conduct is to establish the requisite degree of market power. Traditionally, market power has been defined as the ability to raise price above the competitive level without losing so much business that the increase is not profit maximizing. This implicitly assumes that competition within the relevant market is in terms of price (or at least the terms of trade).95 However, given the way in which pricing is structured by platform businesses, this is usually not the basis for competition. Instead, competition is based on innovation.96 Given this, market power might be defined in terms of effort to innovate. As has traditionally been the case, whether a firm possesses SMP can be established by reference to factors that confer or indicate market power. When discussing these earlier in this chapter, market shares and market concentration were dismissed as poor indicators. Rather, the key indicator of market power is the height of barriers to entry. Although conditions that may create these barriers are similar to those in traditional markets, for platform businesses the factors that create barriers, as well as the relative importance of traditional factors, may be different. Bamberger and Lobel97 suggest that for platform businesses that focus on access to information and search and online marketplaces, sunk costs associated with entry are high – a great deal of specialist infrastructure is required. High fixed costs also provide a strong incentive to expand the customer base to exploit economies of scale, thereby raising the bar in terms of the ability of entrants to be equally cost effective. However, for platform businesses that focus on the service economy,98 sunk costs are much lower, as this requires an app but not extensive infrastructure. The losses incurred by an entrant while market share is low and average fixed costs are high represent a sunk cost in traditional markets.99 For platform busi Ibid 17–18. As price is specified for a product of given quality, quality is included in this description. 96 Given the probability of success for innovation, this tends to mean a higher cost structure than if competition were to be on price. 97 Kenneth A Bamberger and Orly Lobel, ‘Platform Market Power’ (2017) 32 Berkeley Technology Law Journal 1051. 98 These platforms enable offline interactions by connecting users online (for example, credit cards): ibid 1054. 99 For a discussion of this issue, see Rhonda L Smith and David K Round, ‘Temporal Considerations and Entry: Sunk Costs, Scale Economies and Strategic Behaviour’ (2009) 17 Competition and Consumer Law Journal 176. 94 95

Unilateral conduct analysis: focus on harm in multiple guises  225 nesses, however, this cost is exacerbated where indirect network effects are significant, given the need to build critical mass on both sides of the platform quickly, as explained below. D

Barriers to Entry

1 Network effects Indirect network effects have the potential to confer a significant first mover advantage on the platform operator, thereby creating an entry barrier. This is because of the need to quickly and simultaneously build a critical mass of the various types of platform customers. Later entrants must either win customers away from an incumbent or offer a differentiated platform product, often making entry more difficult and costly. In addition, Bamberger and Lobel observe that the presence of a dominant platform makes it harder (and more costly) for an entrant to attract venture capital to facilitate entry.100 Nevertheless, several factors may reduce any market power derived from network effects. These include: ●● consumers may multi-home; ●● where network effects are local rather than global, users prefer platforms offering more personalized services – a differentiated product;101 ●● network effects may be negative – for example, due to poorer service as the user base grows; and ●● in platform markets with different technologies, the disruptive effect may be procompetitive. To determine the significance of indirect network effects requires careful interrogation of the factual information. 2 Data Another means of deterring market entry is to tie up essential inputs so that they are unavailable to entrants, making entry more costly or impossible. The success of a platform business depends on the amount of attention it can attract from users and how well consumer needs are matched to the offerings of the business. This, in large part, depends on the amount and the quality of data available. There has been considerable debate as to whether access to data can act as a barrier to entry and so confer market power, given that data is a public good and so cannot be locked away. However, it may be the sunk cost incurred to quickly amass data and process it that represents the barrier to entry. Data that are short lived and provide precise details of consumer intentions are highly valuable and difficult to replicate, making them a more effective exclusionary device.102 3 Switching costs Just as potential entrants may be deprived of access to essential inputs, they may also be deprived of access to customers. Factors that may result in lock-in include the direct costs of

Conversely, this differentiation may raise switching costs: Bamberger and Lobel (n 97), 1085. Stapp and Hagemann (n 90), 3–4. 102 Oxera, Market Power in Digital Markets (Report, Prepared for European Commission, 30 September 2018) ss 3, 7. 100 101

226  Research handbook on methods and models of competition law switching, time taken to learn and ramp up a new system, brand loyalty (including trust), and inertia (avoidance of choice and uncertainty by not switching).103 Businesses may be unable or less willing to switch from one platform to another if their customers multi-home (see below). As discussed by Bamberger and Lobel, other switching costs include technical and contractual restrictions on transferring content from the platform on which it was developed to that of a rival, access to data on the original platform that feeds into the service provided, enabling it to be customized, and access to the platform’s rating and review system, which encourages businesses to stay on the platform where they have an established reputation.104 4 Customer responses – single and multi-homing In traditional markets, market power may be constrained by the countervailing power of one or more consumers.105 Countervailing power is unlikely in relation to platform businesses, though customer reactions to an exercise of market power may nonetheless provide a constraint. Businesses and consumers may engage with one platform only (single homing) or with several platforms (multi-homing). Generally, multi-homing by businesses constrains platform operators who compete to attract them. If customers on one side of a platform single home, then platforms compete aggressively for these customers. But the businesses supplying them have little bargaining power with the platform operator because they need to be on numerous platforms and so cannot bargain based on the risk of defection to another platform. Multi-homing is affected by the extent of differentiation between platforms and by switching costs. In Epyx, a UK case, it was the strong preference for single homing on one side of the platform, as well as conduct by the platform operator, which was the basis for a finding that Epyx possessed SMP.106 Oxera suggests that consumers on platform markets may constrain the market power of a platform because of concerns about how their data is used.107 The significance of this depends on how transparent data use is and the ability to respond, as well as the switching costs involved. Some empirical evidence is cited which suggests that consumers generally are less concerned about their data being used than about whether they benefit from its use. There can be no doubt that technological change is challenging all aspects of competition analysis, including that involving unilateral conduct. However, while traditional approaches to analysis may need to be adapted, they do not need to be discarded. Perhaps the most chal-

103 Bamberger and Lobel (n 97) note that lack of user lock-in and low switching cost caused the FTC to drop its case against Google relating to favouring its own information over that of others in search results in 2013: 1083. 104 Ibid, 1067. 105 See Kate Collyer, Hugh Mullan and Natalie Timan, ‘Measuring Market Power in Multi-Sided Markets’, Competition Policy International (15 September 2017), https://​ www​ .competitionpo​ licyinternational​.com/​wp​-content/​uploads/​2017/​09/​CPI​-Collyer​-Mullan​-Timan​.pdf. 106 Competition and Markets Authority, ‘Vehicles Service, Maintenance, and Repair Platforms Investigation’ (9 September 2014), https://​www​.gov​.uk/​cma​-cases/​investigation​-into​-the​-supply​-of​ -vehicle​-service​-maintenance​-and​-repair​-platforms​-in​-the​-uk. 107 Oxera, Market Power in Digital Platforms (Report Prepared for the European Commission, 30 September 2018), https://​www​.oxera​.com/​wp​-content/​uploads/​2018/​10/​Market​-power​-in​-digital​ -platforms​.pdf.

Unilateral conduct analysis: focus on harm in multiple guises  227 lenging aspects of the new environment is the highly fact-intensive nature of the necessary enquiries.

V CONCLUSION This chapter has explored the approaches of different jurisdictions to the assessment of anticompetitive unilateral conduct. It has considered the philosophical and economic underpinnings of the problem and illustrated these different approaches in relation to problematic categories of conduct. It has also outlined methodologies for assessing levels of market power and anticompetitive impact. Finally, it has focused on current controversies relating to digital market conduct analysis, many of which are not yet settled.

10. Mergers Rhonda L. Smith

I INTRODUCTION Mergers, acquisitions and takeovers, called ‘concentrations’ in some jurisdictions (hereafter ‘mergers’),1 are continuous occurrences in the business world.2 They may be motivated by a variety of factors, including growth, efficiency, tax advantages and to facilitate exit. Consequently, mergers are often efficiency enhancing – they result in more efficient production and/or distribution and they may facilitate increased innovation. Most mergers raise no competition concerns – their objective is not anticompetitive and neither is their effect. However, a small percentage of mergers do raise concerns initially. These may either disappear on closer examination or are addressed in a way that allows the merger to proceed. An even smaller percentage of mergers are found to be anticompetitive and are stopped or proceed to court.3 This group often seems much larger than it is because it generally involves large companies that are well able to use the media to draw attention to their thwarted transaction. The resources that competition agencies devote to merger investigations may appear disproportionately large. However, mergers have the potential to change market structure in a way that cannot be reversed subsequently – often referred to as the inability to unscramble the egg. Mergers take various forms. A horizontal merger (the main focus of this chapter) occurs when the merger parties operate in the same market; that is, they are competitors. Possible 1 A ‘merger’ involves the shareholders of two companies becoming the shareholders of a new merged company. An ‘acquisition’ occurs when one company acquires or purchases a shareholding in, or the assets of, another company. A takeover occurs when one company acquires another. This may be a ‘friendly’ or agreed takeover, or it may be a hostile takeover. The latter refers to a situation where the purchase has not been approved or recommended by the board of the target. The acquisition may be partial, and treatment of partial acquisitions varies between jurisdictions. Generally, the issue is whether the merger confers control, directly or indirectly, on the acquirer. For a discussion of partial acquisitions, see Organisation for Economic Co-operation and Development (OECD), Directorate for Financial and Enterprise Affairs Competition Committee, ‘The Concept of a Merger Transaction’ (Note by the Secretariat, 18 June 2013) 9–17, http://​www​.oecd​.org/​officialdocuments/​publicdisplaydocumentpdf/​ ?cote​=​DAF/​COMP/​WP3(2013)5​&​docLanguage​=​En. See also Torben Stühmeier, ‘Competition and Corporate Control in Partial Ownership Acquisitions’ (Discussion Paper No 211, Düsseldorf Institute for Competition Economics 2016), http://​www​.dice​.hhu​.de/​fileadmin/​redaktion/​Fakultaeten/​ Wirtschaftswissenschaftliche​_Fakultaet/​DICE/​Discussion​_Paper/​211​_Stuehmeier​.pdf. 2 For a discussion of the reasons for mergers, see Dennis C Mueller, ‘Efficiency Versus Market Power through Mergers’ in Manfred Neumann and Jurgen Weigand (eds), The International Handbook of Competition (2nd edn, Edward Elgar Publishing 2004). 3 For example, in the EU, as of the end of 2017, some 6,805 mergers had been notified to the Directorate-General for Competition (DG Comp) under Council Regulation (EC) No 139/2004 of 20 January 2004 on the control of concentrations between undertakings (the EC Merger Regulation) [2004] OJ L24/1 (European Union Merger Regulation) since it originally took effect in 1990. Of these, only 265 cases went to a Phase II investigation and only 29 were blocked: European Commission, ‘Final Decisions – June 96’ (28 February 2019), http://​ec​.europa​.eu/​competition/​mergers/​statistics​.pdf.

228

Mergers  229 consequences of such mergers are discussed below. Vertical mergers occur when the merger parties operate at different levels in the supply chain – for example, when a flour miller acquires a bakery. Such vertical links may be used strategically, in relation to access to essential inputs or by ‘locking in’ customers, thereby limiting potential demand for the products of rivals and entrants. So, the competition concern is whether the merger results in an increase in market power. Mergers involving parties that are neither competitors nor vertically related are referred to as conglomerate mergers. These are only likely to attract scrutiny if the parties supply products that are weak substitutes or complements, such as when a manufacturer of carbonated beverages acquires a fruit juice plant. This may enable leveraging of market power from one market to another, as well as strategic conduct such as tying or bundling of products. Section II discusses how mergers may have anticompetitive consequences and therefore why competition authorities are concerned about them. The balance of the chapter discusses the treatment of mergers in three jurisdictions – the US, the EU and Australia. Section III compares and discusses the merger control process. Section IV explains the processes used to analyse the competition effects in the three jurisdictions. Some brief conclusions about the treatment of mergers under competition law are drawn in Section V.

II

HORIZONTAL MERGERS AND ANTICOMPETITIVE EFFECTS

Horizontal mergers may be anticompetitive because they increase the likelihood of collusion and/or because they confer or increase the unilateral power of the merged entity. A

Increased Risk of Collusion

Historically, this was the initial concern about the likely effects of mergers. It arises mainly in relation to horizontal mergers but may be an indirect effect of other types of mergers. Potentially a merger may enable or encourage post-merger coordinated interaction among firms in the relevant market, coordinated interaction being ‘conduct by multiple firms that is profitable for each of them only as a result of the accommodating reactions of the others’.4 This includes explicit agreement as to how firms will compete or refrain from competing,5 as well as tacit collusion and parallel accommodating conduct. The conditions required for coordination include a relatively small number of market participants, high barriers to entry and transparent pricing – it is also helpful, but not necessary, for the product to be relatively homogeneous. Thus, to the extent that a merger reduces the number of firms in the market and may reduce the extent of differentiation of both firms and products,6 the risk of coordinated conduct may be increased. Such a weakening of competition harms consumers, who will face

4 US Department of Justice and Federal Trade Commission, ‘Horizontal Merger Guidelines’ (19 August 2010) 24, https://​www​.ftc​.gov/​sites/​default/​files/​attachments/​merger​-review/​100819hmg​.pdf (US Horizontal Merger Guidelines 2010). 5 Ibid. 6 Coordinated effects are more likely and more direct in the case of a horizontal merger where a firm acquires a particularly vigorous or aggressive rival, leaving only smaller, weaker players in the market.

230  Research handbook on methods and models of competition law higher prices and less favourable trading terms, poorer-quality products and possibly reduced service. B

Unilateral Market Power

A merger may increase unilateral market power or substantially lessen competition simply by reducing the number of competitors in the market, assuming that barriers to entry into the relevant market are high. This is because an effective competitive constraint has been removed. Indeed, that may be the primary purpose of the merger. In the case of a vertical merger, this may occur if the merger results in the acquisition of a firm that supplies an essential input or that attracts particularly profitable customers. This may foreclose the market or raise rivals’ costs, making them less competitive and less of a constraint. The merged firm is therefore able to increase prices, lower quality, lessen innovation and thus increase profits. C

Differentiated Products

In markets with differentiated products, unilateral effects may result from a merger even when a number of alternative suppliers remain. This may occur where firms supplying differentiated products do not uniformly constrain each other. To illustrate, assume that the firms in a market that produce differentiated products are distributed along a line such that the distance between them is indicative of the competitive constraint each applies to the others – the greater the distance between them, the weaker the competitive constraint. See Figure 10.1 (the numbers denote the firms).

Figure 10.1

Firms producing differentiated products

A merger is proposed between Firms 2 and 3 (for example, between Coca Cola and Pepsi). Prior to the merger, Firm 2 is the closest constraint on Firm 3 and vice versa. If Firm 2 were to raise its price or reduce quality, at least some of its customers would consider substituting the product of Firm 3. Likewise, if Firm 3 engaged in such conduct, its customers would look to Firm 2 for supply. Post merger, although four other firms remain in the market, the constraint on the merged entity would be weakened – 1 and 4 are now the closest constraints on the merged entity and each is a less effective constraint than 2 on 3 and 3 on 2 pre merger. In summary, the conditions necessary for unilateral effects with differentiated products are that the products of the two firms proposing to merge are the closest substitutes for each other for a substantial proportion of consumers, and that remaining constraints are not sufficiently strong to prevent anticompetitive conduct.7 However, whether the merger is likely to be anti7 See Szabolcs Lorincz, ‘Unilateral Effects in Differentiated Product Markets’ (International Competition Network Merger Working Group Teleseminar, 15 November 2011) 5. In addition to

Mergers  231 competitive also depends on the dynamic responses in the market – that is, whether existing firms are able to alter their products to be more competitive and whether entry is likely. D

Leveraging Market Power Via a Merger

As business models change and become more complex, strategic behaviour8 may play an increasingly significant role in analysing competition effects, including those from a merger. Vertical and conglomerate mergers may enable power to be leveraged from one market to another and this may damage the competitive process. In the case of a vertical merger, acquiring a key industry supplier, for example, not only may secure supplies for the downstream acquirer, but may also enable it to restrict supply to competitors who need access to the same supplier. Whether and, if so, in what circumstances a conglomerate merger may result in competition concerns due to leveraging is controversial and caused major disagreements between the US and the EU in the early 2000s when, for example, the GE/Honeywell merger was not opposed in the US but was found to be anticompetitive in the EU.9

III

THE MERGER PROCESS

A

Legislative Provisions

The world’s first competition law, the Sherman Antitrust Act of 1890,10 did not contain a specific mergers provision. Instead, depending on their likely effects, mergers were addressed under section 1 or section 2 of that Act. It was not until the passing of the Clayton Antitrust Act of 191411 that a provision relating to the acquisition of stock (shares) was introduced. The Clayton Act prohibits mergers if in ‘any line of commerce or any activity affecting commerce in any section of the country, the effect of such acquisition[s] may be substantially to lessen competition, or to tend to create a monopoly’.12 The interpretation of ‘substantial lessening of competition’ is discussed in the next section. In 1950, the Celler-Kefauver Act13 extended the coverage of the Clayton Act to asset acquisitions and acquisitions involving firms that were not direct competitors. Surprisingly, the EU did not introduce a specific mergers provision until 1990. Under this, the creation of dominance alone may have been sufficient to strike down a merger. However, dominance is no longer required under the current provision, the European Union Merger Regulation, Article 2(2)–(3) of which prohibits mergers where the ‘concentration … would

unilateral effects associated with differentiated products, three other sources of unilateral effects are distinguished in the US Horizontal Merger Guidelines 2010 (n 4), 20–4. 8 Strategic behaviour is conduct intended to provide a firm with a long-term commercial advantage. 9 For a discussion of the effects of conglomerate mergers in the Australian context, see Rhonda L Smith and Adrian Rahman, ‘Conglomerate Mergers and the ACCC’s Merger Guidelines’ (2006) 14(1) Competition and Consumer Law Journal 20. 10 15 USC §§ 1–7. 11 15 USC §§ 12–27, 29 USC §§ 52–3 (Clayton Act). 12 Clayton Act, s 7, quoted in US Horizontal Merger Guidelines 2010 (n 4), 1. The Federal Trade Commission Act of 1914, 15 USC §§ 41–58, s 5, may also be relevant. 13 Pub L 81-899, 64 Stat 1125.

232  Research handbook on methods and models of competition law significantly impede effective competition in the common market or in a substantial part of it, in particular as a result of the creation or strengthening of a dominant position’.14 By reversing the question to be considered, it becomes a ‘significant impediment to competition’ test and so focuses on the possibly harmful effects of the merger, rather than on whether the merged entity will become dominant. This is more aligned to the US approach. Australia’s first effective competition law, the Trade Practices Act 1974 (Cth), like the Clayton Act, included a prohibition on mergers that were likely to substantially lessen competition.15 This provision attracted strong criticism from the outset. Business interests argued that it prevented mergers needed to provide Australian businesses with sufficient scale to compete internationally. Consequently, at election time, the major political parties not infrequently agreed to a review of the competition law, particularly the test for mergers, should they win office. Following such an inquiry, in 1977 the test was changed from the initial substantial lessening of competition test to a dominance test.16 However, in a relatively small economy with many oligopolistic markets, the ability to successfully prosecute mergers where at least one other firm of reasonable size would remain post merger was very difficult.17 Thus, in 1993, shortly after the EU introduced its merger provision based on dominance, the mergers test in Australia was changed back to one referenced to a substantial lessening of competition. Once a merger occurs, it is very difficult to then separate the businesses again if later the merger is found to be anticompetitive. Thus, both the US and the EU have mandatory pre-merger notification which prevents the merger proceeding until it has been assessed.18 This creates a heavy workload for the competition authorities, especially as the number of mergers occurring increases over time. To address this, thresholds for notification are used to identify mergers that do not need to be notified. In the US, the thresholds are non-jurisdictional, are based on the size of the transaction, and, depending on the size of the transaction, may also take account of the parties’ turnover or assets.19 In the EU, thresholds are based solely on

14 Council Regulation (EC) No 139/2004 of 20 January 2004 on the control of concentrations between undertakings (the EC Merger Regulation) [2004] OJ L24/1. 15 Trade Practices Act 1974 (Cth), s 50(1). 16 Trade Practices Amendment Act 1977 (Cth), s 27. 17 Motta argues that the use of dominance works fairly well where a merger increases the risk of collusion. However, he regards it as a poorer fit when the outcome from a merger is to increase the firm’s unilateral market power: Massimo Motta, Competition Policy: Theory and Practice (Cambridge University Press 2004) 271. To address this issue, the European Commission accepted the possibility of joint dominance in the context of a proposed merger between Nestlé and Perrier, and gave rise to the peculiarly European approach. However, in 2002, in a merger involving UK-based suppliers of packaged tours, Airtours, the EU Court of First Instance (now the General Court) revisited the concept of collective dominance. The Court did not rule that the concept of collective (or joint) dominance was invalid, but it raised substantially the evidentiary threshold required to prove it. It required (i) that sufficient market transparency existed for each member of the dominant monopoly to monitor each other’s conduct; (ii) means to retaliate against departures from the common policy; and (iii) that the reaction of current and future competitors, and consumers, would not disturb the common policy: Case T-342/99, Airtours Plc v Commission of the European Communities [2002] ECR II-02085. 18 In the US, this is provided for under the Hart-Scott-Rodino Antitrust Improvements Act of 1976, Pub L 94-435, Stat 1383. 19 OECD, Directorate for Financial and Enterprise Affairs Competition Committee, ‘Local Nexus and Jurisdictional Thresholds in Merger Control’ (Background Paper by the Secretariat, 14–15 June 2016) 9, http://​www​.oecd​.org/​officialdocuments/​publicdisplaydocumentpdf/​?cote​=​DAF/​COMP/​ WP3(2016)4​&​docLanguage​=​En. In the US, from January 2017, merger transactions valued at less

Mergers  233 turnover.20 Yet it is questionable whether there is a close relationship between the threshold factors and the effect of the proposed merger on competition.21 Unlike the US and the EU, Australia has no statutory requirement for pre-merger notification.22 However, it would be unusual, although not unheard of, for the parties to not notify the Australian Competition and Consumer Commission (ACCC) of a major merger,23 possibly due to the penalties that can be applied if the merger is subsequently found to be anticompetitive.24 B

The Role of the Competition Authorities and the Courts

In the US, mergers that raise competition concerns are assessed either by the Antitrust Division of the Department of Justice (DOJ) or by the Federal Trade Commission (FTC). The two operate cooperatively and which authority undertakes the investigative task appears to depend on which has the greatest interest in the matter, previous experience with the industry, and available investigative capacity. There does not appear to be any evidence that merger parties engage in gaming to ensure that one authority rather than the other investigates a particular merger. Following an investigation of a proposed merger, the DOJ or the FTC issues an administrative decision. If this is unfavourable, the merger may be abandoned. However, the decision may be appealed. In the case of the FTC, this will be to the US Court of Appeals for the District of Columbia, while the orders of the Federal District Court in DOJ cases may be appealed to the US Court of Appeals for the relevant circuit. Very few merger cases have been appealed from there to the Supreme Court.25 than $80.8 million are exempt: Ngoc Hulbig, ‘FTC Announces 2017 Thresholds for Merger Control Filings under the HSR Act and Interlocking Directorates under the Clayton Act’, Mondaq (30 January 2017), http://​www​.mondaq​.com/​unitedstates/​x/​564108/​Antitrust+Competition/​FTC+Announ​ ces+2017+Threshol​ds+for+Merger+Con​trol+Filings+Unde​r+the+HSR+Act+and​+Interlocking+Dir​ ectorates+Under+the+Clayton+Act. 20 In the EU, notification is required for mergers that transcend national borders and which have an annual combined business turnover that exceeds a worldwide turnover of €5 billion and exceeds a Community-wide turnover of €250 million: OECD, ‘Local Nexus and Jurisdictional Thresholds’ (n 19); OECD, Directorate for Financial and Enterprise Affairs, ‘Local Nexus and Jurisdictional Thresholds in Merger Control from the EU Perspective’ (Presentation by DG Comp, June 2016), https://​www​ .slideshare​.net/​OECD​-DAF/​local​-nexus​-and​-jurisdictional​-thresholds​-in​-merger​-control​-from​-the​-eu​ -perspective​-dg​-comp​-june​-2016​-oecd​-discussion). 21 Julie Clarke, ‘Current Issues in Merger Law’ in John Duns, Arlen Duke and Brendan Sweeney (eds), Comparative Competition Law (Edward Elgar Publishing 2015) 198. 22 Nevertheless, the Australian Competition and Consumer Commission (ACCC) encourages notification if the merger involves products that are supplementary or complementary, or where the merged entities will have a market share that exceeds 20 per cent: ACCC, ‘Informal Merger Review Process Guidelines’ (September 2013) 7–8, https://​www​.accc​.gov​.au/​system/​files/​D17​-156292​%20Informal​ %20Merger​%20Review​%20Process​%20Guidelines​%20​-​%20updated​%20November​%202017​_0​.PDF. 23 Most notably, in 1988, Australian Meat Holdings acquired two meat processing plants in Queensland from Borthwicks. Australian Meat Holdings was later found to have contravened the merger provision of the competition law. It was fined and ordered to divest the plants. See Trade Practices Commission v Australian Meat Holdings Pty Ltd (1988) 83 ALR 299. 24 These penalties are discussed below. 25 There have been just two cases in the last six years. In 2013, the US Supreme Court ruled in favour of the FTC when Phoebe Putney Health System sought to acquire Palmyra Medical Center: Fed Trade Comm’n v Phoebe Putney Health Sys, Inc, 568 US 216 (2013). Then, in 2017, the US Court of Appeals for the District of Columbia blocked a bid by health insurer Anthem Inc’s proposed acquisition of Cigna

234  Research handbook on methods and models of competition law With limited exceptions, within the EU, the European Commission (EC) has exclusive jurisdiction over ‘concentrations’. So, as in the US, the initial investigation of a merger is undertaken by the competition authority. However, distinctively, the Directorate-General for Competition has the power to make determinative decisions on mergers – that is, it can block a merger and can accept remedies. These decisions can be challenged in the courts – at first instance, in the General Court and then in the European Court of Justice – but litigation plays a smaller role than in the US. In Australia, the process is more complex. Currently, the ACCC is responsible for the enforcement of the merger provision of the Competition and Consumer Act 2010 (Cth) (CCA). However, its role varies according to whether the merger is assessed informally or, until November 2017, via a formal clearance (see Box 10.1 below). Only in relation to the latter was an administrative decision made. This was appealable to the Federal Court of Australia to be heard initially by a single judge, by a full bench of the court (three judges) on further appeal, and ultimately by the High Court of Australia. In the alternative, the parties may seek an exemption from the merger provision of the CCA – known as an authorization – if they can show that there will be a net public benefit from the merger. However, before the changes in November 2017, this was assessed by the Australian Competition Tribunal.26 The Tribunal’s decision could be appealed to the Full Federal Court, but only on points of law. Since November 2017, the ACCC has been responsible for assessing merger authorization applications.

Box 10.1 ALTERNATIVE MERGER ASSESSMENT PROCESSES IN AUSTRALIA Informal ‘Clearance’ The informal clearance process27 is the approach normally followed by merging parties and initially involves the parties providing the ACCC with information relevant to the merger and its likely effect on competition. If the proposal raises competition concerns, additional information may be sourced from interested parties. The Commission then announces whether, on the information available to it, it considers that the merger would be likely to substantially lessen competition in the relevant market. There is no administrative decision that can be challenged either by the parties to the merger or by others. If the parties propose to proceed despite the ACCC indicating that the merger would be anticompetitive, the Commission can seek an injunction to stop it until it can be challenged in court. Corp: United States of America v Anthem Inc and Cigna Corporation, USCA 17-5024 (2017). Other proposed mergers, such as the second attempt by Staples to acquire Office Depot in 2016, while much publicized, have not progressed to the Supreme Court: In re Staples, Inc and Office Depot, Inc, Federal Trade Commission, Docket No 9367 (19 May 2016). 26 Until 2008, applications for authorization of mergers were subject to an administrative decision of the ACCC, which could be appealed to the Australian Competition Tribunal (formerly the Trade Practice Tribunal). 27 A detailed description of the ACCC’s processes is contained in ACCC, ‘Informal Merger Review Process Guidelines’ (September 2013), https://​www​.accc​.gov​.au/​system/​files/​D17​-156292​%20Informal​ %20Merger​%20Review​%20Process​%20Guidelines​%20​-​%20updated​%20November​%202017​_0​.PDF. The role of the Tribunal is discussed further in Section III.

Mergers  235 Alternatively, following an adverse response from the Commission, the parties may seek a declaration from the court that the merger is not anticompetitive. Formal Clearance Although the informal process generally works well for less contentious mergers, by the early 2000s there was an increasing concern that, where mergers were contentious, there was a lack of transparency which resulted in uncertainty, and the process was not timely, especially as potentially the merger could be challenged in court by interested third parties even if the ACCC did not challenge it. Consequently, a voluntary formal clearance process was introduced in 2007. The ACCC had 40 days within which to decide whether to grant clearance. If clearance was not granted, or was subject to conditions to which the parties objected, the parties could seek a review by the Australian Competition Tribunal, which had 30 days within which to make a decision (extendable to 90 days at the discretion of the Tribunal). This process was never used. As the Business Council of Australia noted, ‘it is unduly complicated by strict technical formal requirements for a compliant application’.28 In addition, generally businesses are satisfied with the flexibility provided by the informal process. Nevertheless, in 2015, the Competition Policy Review (the Harper Review) concluded that the existence of a formal merger clearance option serves a useful purpose, even if it is seldom used, since it provides a time-limited,29 accessible alternative to the ACCC’s informal clearance process … [However,] the process needs reform to remove unnecessary restrictions and requirements that may have acted as a deterrent to its use.30

The government accepted this recommendation in December 2015 and changes were enacted in October 2017. As a result, the formal merger review process was discontinued and became part of the authorization process. Responsibility for assessing applications for the authorization of mergers shifted from the Australian Competition Tribunal to the ACCC. Parties are required to give an undertaking not to complete the merger while it is under review. C

Time Frame for Decisions

The time frame for decisions in relation to mergers is critical, not simply because they delay the business benefitting from the merger, but because the proposed merger has been made public, allowing competitors to identify strategies in response. Consequently, protracted decision-making may ‘kill’ the proposal. Given this, all three jurisdictions under discussion impose strict time limits. The EU, for example, splits the investigation of notified mergers into Phase I and Phase II. During the former, it is determined whether the merger has an EU dimension and whether it

28 Business Council of Australia, ‘Submission to Competition Policy Review’ (Summary Report, July 2014) 16, http://​legacy​.bca​.com​.au/​publications/​submission​-to​-the​-competition​-policy​-review. 29 There is a statutory limit of 40 business days, although this may be extended by up to 20 days. 30 Ian Harper, Peter Anderson, Su Mccluskey and Michael O’Bryan QC, ‘Competition Policy Review Final Report’ (March 2015) 328, https://​treasury​.gov​.au/​publication/​p2015​-cpr​-final​-report.

236  Research handbook on methods and models of competition law raises competition concerns. In general, this phase is limited to 25 working days and the parties may submit remedies but must do so within 20 working days of notification. The Commission proceeds to Phase II if any competition issues are not redressed in Phase I. This phase involves in-depth market and economic analysis and the Commission usually issues a Statement of Objections to the parties, which have a right of response. The second phase is limited to 90 working days. Undertakings may be offered to overcome the competition issues at between 55 and 65 working days. If undertakings do not eliminate the competition issue and the merger has not been withdrawn, it will be prohibited. Where appeal to the court is allowed, fast-track processes are increasingly being adopted for hearings. In Australia, for example, a merger not addressed via the authorization process may be adjudicated by the court if the ACCC believes it to be anticompetitive. In 2007, a fast-track process was introduced by the Federal Court to shorten timelines and reduce the costs of litigation, including in competition cases. This is achieved in part by using case summaries rather than formal pleadings and by limiting access to discovery. In addition, parties are allocated a fixed amount of time for submissions, evidence-in-chief and cross-examination. In relation to authorization applications for mergers, the Tribunal is subject to a statutory timeline of three months, although under certain circumstances this may be extended by a further three months. D

Information Requirements

Parties are required to provide a considerable amount of information when completing mandatory notification forms, but, irrespective of whether there is mandatory pre-merger notification, investigation of a proposed merger inevitably requires provision of, or access to, information and this may result in significant costs for the parties – including the time of senior staff and management. The information required differs significantly between jurisdictions, which may be a major issue for the parties in global acquisitions, adding significantly to costs, as well as the time taken to obtain clearance. The information required initially may be limited, but it is likely to increase if a more detailed investigation is undertaken. ‘Information requirements have expanded and grown in complexity, owing in part to the development of digital technologies, as well as to advances in economic analysis, which depend upon extensive data.’31 It has been suggested that the EU typically requires more information than the US in the early stages of an inquiry but, as the investigation proceeds, the US requires increasing amounts of information.32 Competition authorities have statutory power to compel the provision of information. For example, the EC has extensive information-collecting powers under the Merger Regulation. It can request the supply of information, including from third parties, or obtain it under a formal Commission decision.33 It can conduct inspections at premises and examine books and records. Further, it can conduct interviews to collect information from consenting persons.34

Clarke (n 21), 200. Ibid. 33 Slaughter and May, ‘The EU Merger Regulation: An Overview of the European Merger Control Rules’ (January 2018) 11, https://​www​.slaughterandmay​.com/​media/​64572/​the​-eu​-merger​-regulation​ .pdf. 34 Ibid. 31 32

Mergers  237 In the US, the agencies can issue Civil Investigative Demands in the event that the parties refuse to cooperate fully on a voluntary basis. In Australia, section 155 of the CCA enables the ACCC to require merger parties and others to provide it with information, including orally under oath, ‘if the Commission … has reason to believe that a person is capable of furnishing information, producing documents or giving evidence relating to a matter [that] … is relevant to the making of a decision by the Commission’.35 E Remedies Remedies in this context refer to options available to the merging parties to overcome a finding by a competition authority (or court) that the proposed merger is anticompetitive and therefore contravenes the relevant provision of the competition law. The parties may opt to withdraw a pre-merger notification and reassess the nature of the proposed merger to reduce or avoid the competition issues that have been raised.36 This may provide the parties with (most of) the benefits sought from the merger, while reducing any competitive harm to an acceptable level and possibly increasing welfare through access to benefits such as increased efficiency. Frequently, this involves providing an undertaking to the competition authority or to the court that the parties will take certain actions. This may involve counteracting structural change that would otherwise result from the merger by selling off assets (for example, subsidiaries, brands, particular activities or licences). The success of such a remedy depends on the existence of a willing buyer with ‘the means and incentive to preserve competition in the affected market’.37 Less often, an undertaking may provide a behavioural undertaking in relation to the conduct of the post-merger entity, such as an undertaking to: continue supply to certain customers for a specified time; not enter into exclusive supply arrangements with customers or suppliers; creating firewalls in relation to the passage of information, particularly pricing information; and accepting mandatory licensing. Behavioural undertakings are less favoured because they require ongoing monitoring (which may be particularly problematic in multinational mergers) and because changing market circumstances may be used as an excuse to void the undertaking. Structural undertakings generally more directly address competition concerns. However, as the OECD notes: ‘Where divestiture would be impracticable or disproportionate in order to remedy the adverse effects arising from a merger, behavioural remedies might sometimes be preferable … especially in the case of mergers with vertical elements, and where markets are quickly developing and future developments are difficult to anticipate.’38 CCA, s 155(1), (2)(b)(iii). In Australia, under the informal process, the merger parties may stop discussions without the ACCC reaching a final view about the merger or, following an adverse view, they may opt to offer court-enforceable undertakings, as provided for in section 87B of the CCA. 37 OECD, ‘Remedies in Merger Cases’ (Directorate for Financial and Enterprise Affairs Competition Committee, 30 July 2012) 223, http://​www​.oecd​.org/​daf/​competition/​RemediesinMergerCases2011​.pdf. To illustrate, when Frito-Lay sought to acquire the Australian snack foods business operated by Smiths, the result would have been the second-largest player acquiring the largest player. Frito-Lay undertook to divest sufficient plant and brands to create a new business equivalent to the pre-merger Frito-Lay business. This undertaking was accepted by the ACCC as a basis for allowing the merger but was subject to the ACCC approving the buyer of the new business. 38 Ibid, 20. Examples of acceptance of behavioural undertakings include Comcast Corp/NBC 2011; Ticketmaster/Live Nation (2010); PepsiCo Inc/Pepsi Bottling (2010). For a comparison of undertakings in the US, the EU and China, see Deborah Healey, Zhang Chenying and Jack Coles, ‘Sustaining the 35 36

238  Research handbook on methods and models of competition law In all three jurisdictions, undertakings may be offered by the parties but cannot be imposed by the competition authority. In the US, undertakings are normally offered before a merger is completed, referred to by the US authorities as ‘fix-it-first’. Such undertakings may be rejected if ‘an order will be necessary to enforce and monitor the ongoing obligations’. In the EU, the EC may accept remedies in either Phase I or Phase II of a merger investigation. In Australia, parties may offer an undertaking when requesting an informal clearance or following the ACCC’s completion of such a review. In either case, undertakings are court enforceable. An alternative in Australia for parties that unsuccessfully sought informal clearance may be to withdraw and instead seek an authorization if they believe that there will be sufficient public benefit from the merger to outweigh the public detriment (that is, the substantial lessening of competition).

Box 10.2 THE AUTHORIZATION PROCESS A distinctive feature of Australian competition law is the formal provision for exemption of proposed conduct that would contravene Part IV of the CCA, which prohibits anticompetitive conduct, including mergers.39 This is termed an authorization. The basis for authorization is that the conduct will result in a public benefit which exceeds the resulting public detriment. The term ‘public benefit’ is not defined in the CCA but has been taken to mean ‘anything of value to the community generally’.40 This provides considerable flexibility, accommodating responses to changing circumstances. Although authorization is most frequently sought based on efficiency benefits resulting from the conduct, other benefits claimed have included environmental improvements, avoiding unemployment, enabling industry rationalization, and obtaining improved public health outcomes. Thus, authorization provides an important means of reconciling the competition objectives of the law with other policy objectives. Particularly where a merger results in significant public benefits, it provides an alternative means of avoiding prohibition of the transaction based solely on the competition effects. Between 1974 and 2007, applications for authorization in relation to mergers were assessed by the Commission which made an administrative decision in relation to the claimed net public benefits. The process was controversial for several reasons. One was its very public nature. Applications for authorization were listed on a public register, and comments concerning the effects of the merger were sought from the public. It was argued that this was necessary because conduct that was otherwise prohibited in the public interest may have been permitted. Second, the process was viewed as potentially, and often in fact was,

Status Quo: The Use of Conditions in Chinese Merger Clearance’ (2017) 10 Tsinghua China Law Review 1. 39 The exception is misuse of market power for which an exemption was not available. However, changes to the test in 2017 mean that authorization may now be sought for this conduct. 40 Re 7-Eleven Stores (1994) ATPR 41-357 at 42,677. See also Queensland Co-operative Milling Association Ltd (1976) ATPR 40-012 at 17,242. However, when the mergers test was changed from prohibiting a merger that created or increased dominance to a substantial lessening of competition test (perhaps as a sweetener for businesses), public benefit in the context of authorization of mergers explicitly required consideration of increased exports, import replacement, and indeed anything related to international trade to be considered as a public benefit.

Mergers  239 too slow to be a commercially realistic option. It was also unpredictable. This resulted in part from the provision for de novo review by the Australian Competition Tribunal41 on application by an interested party, and the potential for further appeal on points of law to the Federal Court of Australia and ultimately to the High Court of Australia. To address at least the second of these concerns, from 2007 until November 2017, applications for authorizations of mergers went directly to the Australian Competition Tribunal, thereby eliminating a step in the process. However, the Tribunal was heavily reliant on the Commission for input concerning the merger as it had no staff to undertake this function. The Tribunal was subject to a statutory timeline of three months, with the possibility of a further three-month extension. Interestingly, no application in relation to a merger was made to the Tribunal until Murray Goulburn proposed acquiring Warrnambool Cheese and Butter in 2013 – the matter was discontinued when the target was acquired by Saputo Inc. Three subsequent applications were accepted by the Tribunal.42 Following the Harper Review, authorization applications must again be made to the Commission, not to the Tribunal. A review of the Commission’s decision may be sought from the Tribunal, but this is no longer a de novo review. F Penalties In the US and the EU, penalties may be imposed if the competition authorities are not notified of a non-exempt merger or if conditions offered to overcome the competition issue associated with the merger are not complied with. In the US, for example, failure to notify a merger can result in a statutory penalty of up to US$40,654 per day from the date of completion of the transaction, although the actual penalty imposed tends to depend on the circumstances. In the EU, fines of up to 10 per cent of turnover may be imposed. In addition, penalties may be imposed if misleading information is provided during the course of a merger investigation – for example, in May 2017, the EC fined Facebook €110 million for providing misleading information in relation to its acquisition of WhatsApp.43 In Australia, where there is no mandatory pre-merger notification, a range of penalties may be applied if the court ultimately finds a merger to be anticompetitive. For corporations, the maximum penalty per offence for breaches relating to anticompetitive conduct, including

41 The Tribunal comprises a judge as the presidential member, together with two lay members, one of whom is an economist and the other is a person with business experience. 42 AGL applied to acquire Macquarie Generation in 2014: AGL, ‘AGL Completes Acquisition of Macquarie Generation Assets’ (Media Release, 2 September 2014), https://​www​.agl​.com​.au/​about​ -agl/​media​-centre/​asx​-and​-media​-releases/​2014/​september/​agl​-completes​-acquisition​-of​-macquarie​ -generation​-assets; Sea Swift applied to acquire Toll Marine Logistics in 2016: Sea Swift, ‘Settlement of Sea Swift Purchase of Toll Marine Logistics’ (Media Release, 12 September 2016), https://​www​.seaswift​ .com​.au/​news/​tml​-acquisition​-news/​6457​-settlement​-of​-sea​-swift​-purchase​-of​-toll​-marine​-logistics; and Tabcorp Holdings and Tatts Group Ltd applied for authorization to merge in 2016: Tabcorp, ‘Tabcorp and Tatts to Combine to Create a World-Class, Diversified Gambling Entertainment Group’ (Media Release, 19 October 2016), https://​www​.tabcorp​.com​.au/​news​-media/​media​-releases/​tabcorp​-and​-tatts​ -announce​-intention​-to​-combine​-th. 43 European Commission, ‘Commission Fines Facebook €110 Million for Providing Misleading Information about WhatsApp Takeover’ (Press Release, 18 May 2017), http://​europa​.eu/​rapid/​press​ -release​_IP​-17​-1369​_en​.htm.

240  Research handbook on methods and models of competition law mergers, is the greater of $10,000,000 or, if the court can determine that a ‘reasonably attributable’ benefit was obtained, three times that value; or, if the court cannot determine a benefit, 10 per cent of annual turnover in the preceding 12 months.44 Significantly, mergers are the only conduct contravening the Australian competition law for which a court may order a divestiture. However, there has been only one significant divestiture.45 Other penalties include non-punitive orders (such as a community service order), injunctions, award of damages, and disqualification of directors. G

Post-merger Review

A key aspect of any merger process should be a subsequent review of merger decisions. Given the difficulty of undertaking such a review, until relatively recently this has not occurred very frequently. Determining the effectiveness of merger oversight is far from easy. However, it is important because the periodic ex post analysis of actual effects of enforcement decisions a few years after these are taken is a tool that can help agencies better understand the impact of their interventions and improve future decision-making.46 Increasingly, these studies are being undertaken and they take a variety of forms. The simplest approach for a horizontal merger is to compare prices for products common to both parties pre merger with those of the merged entity. However, price differences may capture the effect of changes such as those in costs and demand, which are not the result of the merger. To avoid this issue, a method known as Difference in Difference (DID) can be used. To control for non-merger influences on price, this uses price changes in other products that are comparable in all respects except for the merger.47 This latter control group ‘captures’ the non-merger effects so that the difference between the two groups is the result of the merger. While relatively simple and cost effective, problems with this approach include: ●● ●● ●● ●●

identifying an appropriate control group; the suitability of the merger parties may influence the outcome of the merger; overlapping or simultaneous events make comparisons difficult; timing: evaluation should only be undertaken once the effects of the merger are sufficiently established – generally taken to be between two and three years after the merger – but, if adaptation is gradual, a longer time may need to be allowed; and ●● it focuses on price, where non-price effects may be significant.48 The second generation of post-merger evaluations addressed ‘certain econometric and causation issues that previously received inadequate attention but had the potential to produce 44 ACCC, ‘Fines and Penalties’, https://​www​.accc​.gov​.au/​business/​business​-rights​-protections/​fines​ -penalties​#restrictive​-trade​-practices. 45 Australian Meat Holdings (n 23). 46 OECD, ‘Reference Guide on Ex-post Evaluation on Competition Agencies’ Enforcement Decisions’ (April 2016) 3, http://​www​.oecd​.org/​daf/​competition/​Ref​-guide​-expost​-evaluation​-2016web​ .pdf. 47 John Kwoka, ‘Second Capacity Building Workshop on Ex-post Evaluation of Competition Agencies’ Activities’ (OECD Directorate for Financial and Enterprise Affairs Competition Committee, Working Party No 2 on Competition and Regulation, 13 June 2016) 6, https://​one​.oecd​.org/​document/​ DAF/​COMP/​WP2(2016)5/​en/​pdf. 48 Ibid.

Mergers  241 misleading results’.49 Among these are simulations which can be used to estimate ‘how prices, demand, and market shares might have changed had the merger gone ahead absent the CA’s [competition authority’s] intervention’,50 or presumably if it was prevented from proceeding: ‘Simulation refers to techniques based on (i) explicit formal modelling of the nature of competition in the market, then (ii) calibrating the model with real world information (sometimes estimated econometrically), before (iii) using it to assess how the equilibrium will change with and without an event/intervention.’51 Both the FTC and the EC have undertaken studies of this type. However, ‘although the simulation approach has the advantage of explicitly using economic theory to identify the counterfactual, it is very sensitive to modelling assumptions, and is not appropriate to be applied to some markets’.52 It is necessary to assess not only the decision to allow or prohibit a merger. Where a merger is allowed to proceed because the parties have offered (and the competition authority has accepted) remedies, the effectiveness of these remedies should also be assessed. The questions to be answered are these: Was the remedy adhered to? What effect did it have on the market? These studies are typically based on surveys and interviews of key market players, industry associations, market experts and other stakeholders … This information is then analysed … Usually, but not always, these studies do not involve econometric based assessments of quantitative data, but are based on analysis of factual information, plus some before and after examination of quantitative data (e.g. on market shares or prices).53

These studies can provide useful insights and have been undertaken in both the US and the EU, but they rely on voluntary supply of information and this may limit their usefulness.54

IV METHODOLOGY [M]erger analysis does not consist of uniform application of a single methodology. Rather, it is a fact-specific process through which the Agencies, guided by their extensive experience, apply a range of analytical tools to the reasonably available … evidence to evaluate competitive concerns in a limited period of time.55 Ibid. Peter Ormosi, ‘Stocktaking on Evaluation’ (OECD Directorate for Financial and Enterprise Affairs Competition Committee, Working Party No 2 on Competition and Regulation, 11 June 2012) 7, http://​www​.oecd​.org/​officialdocuments/​publicdisplaydocumentpdf/​?cote​=​DAF/​COMP/​WP2​%282012​ %295​&​doclanguage​=​en. 51 Ibid, 13–14. For an outline of the process, see European E & M Consultants, ‘Merger Simulation Models’, https://​www​.ee​-mc​.com/​fileadmin/​user​_upload/​ccr​_en/​Merger​_Simulation​_Models​.pdf. 52 Ibid, 14. 53 OECD, ‘Reference Guide on the Ex-Post Evaluation of Competition Agencies’ Decisions’ (April 2016) 38, http://​www​.oecd​.org/​daf/​competition/​Ref​-guide​-expost​-evaluation​-2016web​.pdf. 54 For an example of merger evaluation, see Daniel S Hosken, Luke M Olsen and Loren K Smith, ‘Do Retail Mergers Affect Competition? Evidence from Grocery Retailing’ (Federal Trade Commission 2015), http://​docplayer​.net/​41516834​-Do​-retail​-mergers​-affect​-competition​-evidence​-from​-grocery​ -retailing​-daniel​-s​-hosken​-luke​-m​-olson​-loren​-k​-smith​-september​-2015​.html. This analyses whether the assumption that entry into retailing is relatively easy and so generally mergers are not likely to substantially lessen competition. It finds that prices often rose post-merger in concentrated markets but not in unconcentrated markets. 55 US Horizontal Merger Guidelines 2010 (n 4), 1. 49 50

242  Research handbook on methods and models of competition law As noted above, the US and Australia assess mergers according to whether they are likely to substantially lessen competition, while the EU now applies a test that assesses whether the merger will substantially impede competition. Despite this difference, all three consider the future with the merger and the future without the merger (the ‘with and without’ test) in order to identify the likely effect of the merger on the competitive process. The difficulty with this is that neither outcome is currently known. This is particularly problematic when the market is undergoing change, perhaps due to technology or a shift in demand. Consequently, competition authorities have made it clear that they will not take into account events that are merely speculative. However, what determines when an outcome is sufficiently probable to be taken into account?56 Furthermore, in assessing whether a merger damages the competitive process, it is very important to recognize that the process is dynamic rather than static – existing rivals will react to the change in market conditions: Where unilateral effects occur, other market participants’ responses may vary. In some situations, other market participants may respond in a pro-competitive way and (at least partially) attempt to offset the merged firm’s behaviour. Alternatively, it may be more profitable for other market participants to simply support the merged firm’s conduct …57

A

The Meaning of Substantial Lessening of Competition

The term ‘substantial lessening of competition’ is by no means precise or uncontroversial, as illustrated by the McWane case in the US.58 Courts have been prone to interpret it to mean that the effect of the alleged conduct is ‘not insignificant’, ‘more than de minimus’ or ‘non trivial’. Depending on the circumstances, even a small absolute reduction in competition may be found to be a substantial reduction. In an Australian case that related not to a merger but rather to an allegedly anticompetitive agreement, Lockhart J stated: [T]he word ‘substantial’ is used in a relative sense. The very notion of competition imports relativity. One needs to know something of the businesses carried on in the relevant market and the nature and extent of the market before one can say that any particular lessening of competition is substantial.59

A merger may damage competition, that is, result in a substantial lessening of competition. This occurs when the competitive constraint imposed by rivals is reduced such that firms have less incentive to improve their offer to customers or to innovate to increase efficiency. To

56 Commerce Commission v New Zealand Bus Ltd [2006] HC WN CIV-2006-485-585. In ACCC v Metcash, the Court found that the Commission’s counterfactual was flawed because without evidence of the existence of an alternative buyer, the claim that there was an alternative buyer was speculative. This was upheld on appeal. See Australian Competition and Consumer Commission v Metcash Trading Ltd [2011] FCA 967. 57 ACCC, ‘Merger Guidelines’ (November 2008) 22, https://​www​.accc​.gov​.au/​system/​files/​Merger​ %20guidelines​%20​-​%20Final​.PDF. 58 In re McWane Inc and Star Pipe Productions, Ltd, Federal Trade Commission, Docket No 9351 (30 January 2014), https://​www​.ftc​.gov/​system/​files/​documents/​cases/​140206mcwaneorder​.pdf; ‘Dissenting Statement of Commissioner Joshua D. Wright, In the Matter of McWane, Inc. et al, Docket No 9351’, https://​www​.ftc​.gov/​system/​files/​documents/​public​_statements/​202211/​140206mcwanestatement​.pdf. 59 Radio 2UE Sydney Pty Ltd v Stereo FM Pty Ltd (1982) 42 ALR 557, 562.

Mergers  243 determine whether this is likely, it is necessary to identify what indicators to consider and how they may interrelate, and how an evidentiary base can be established. B

Factors Used to determine the Effects of a Merger

Before analysing the likely competition effects of a proposed merger, it is necessary to identify the relevant market or markets within which competition concerns arise.60 Once this is done, attention turns to the factors used to assess the effects of a merger on competition. These are explained in the merger guidelines published in each of the jurisdictions and are discussed below. 1 Market shares and market concentration If there are numerous firms in a market and the merging parties separately or jointly do not have a relatively large market share, it is unlikely that the merger will substantially lessen competition. Given this, the Herfindahl-Hirschman Index (HHI)61 is used as a ‘clearing’ mechanism or to identify ‘safe harbours’. Generally, thresholds are set by identifying high, medium and low HHI values (the values differ between jurisdictions), then specifying the maximum increase in the HHI for each of these that will avoid further scrutiny. For example, for the US, the values are as shown in Table 10.1. Table 10.1

US HHI values

Market category

HHI range

Safe harbour

Unconcentrated

2,500

Increase 100–200 a concern; increase over 200 presumed a substantial lessening of competition

Source: US Horizontal Merger Guidelines 2010, 19.

However, market share data may be misleading. First, their accuracy depends on the market definition. Second, market share is not a good indicator of market power. A firm with a very large market share may not possess market power because entry barriers are low, while a firm with a very small market share may be able to engage in strategic conduct that determines the market price (this may be possible in some electricity markets). Another example is where the target is an innovator or a maverick such that its removal via a merger may remove the procompetitive effect of differentiation from the market. Consequently, competition authorities on some occasions may bypass the safe harbours.

60 For a discussion concerning the necessity of market definition, see William Blumenthal, ‘Why Bother? On Market Definition under the Merger Guidelines’ (Statement before the FTC/DOJ Merger Enforcement Workshop, 17 February 2004), https://​www​.justice​.gov/​atr/​why​-bother​-market​-definition​ -under​-merger​-guidelines. For a contrary view, see Rhonda L Smith, ‘Market Definition: Going, Going, Gone? Developments in the United States’ (2010) 18 Competition and Consumer Law Journal 119. 61 This is calculated as the square of the market share of each market participant.

244  Research handbook on methods and models of competition law 2 The height of barriers to entry In markets with few players and relatively high market shares, merger parties clearly have an incentive to argue that if the merged entity engaged in anticompetitive activity it would attract entrants and so its conduct would be unprofitable. Even a monopolist would be constrained. However, ‘[t]he prospect of entry into the relevant market will alleviate concerns about adverse competitive effects only if such entry will deter or counteract any competitive effects of concern so the merger will not substantially harm customers’.62 Thus, assuming that few competitors will remain post merger, or that there will be significant unilateral effects from the merger, the height of barriers to entry becomes the key to whether the merger will damage the competitive process or not. Much has been written about the meaning of barriers to entry in the context of competition law.63 For the present purpose, it is sufficient to define a barrier to entry as anything that enables supra-normal profits64 to persist in the long run, causing harm to consumers through higher prices, less choice of product and supplier, and possibly reduced innovation. The nature of barriers to entry varies but they can be categorized as shown in Table 10.2.65 Table 10.2

Barriers to entry

Nature of barrier to entry

Comments

Structural barriers

Structural barriers to entry are stable market features that deter entry. They can be the result of legal requirements or a function of costs, including sunk costs, economies of scale, and absolute cost.

Strategic barriers

Strategic barriers to entry result from strategic behaviour in oligopolistic markets. Conduct, such as entering into long-term contracts with key customers or excessive investment in technology, may adversely alter market structure. However, strategic behaviour may raise barriers to entry even when market structure is unchanged. For example, entry may be deterred by signalling or by reputation for extremely aggressive responses when entry occurs. Strategic conduct that is entry-deterring is not always easy to assess. In some circumstances, conduct that would not normally be regarded as entry-deterring may be assessed as having such an effect – for example, investment in large amounts of excess capacity on the basis that the capacity will be needed in future.

Barrier to expansion

Often, entry occurs on a small scale with the intention of later expansion to become a full service provider. However, sometimes firms find that expansion is not possible. This is particularly a problem in network industries (for example, industries such as electricity, water and telecommunications). Consequently, despite entry, incumbents may not be subject to effective competitive constraint, despite the presence in the market of these firms.

US Horizontal Merger Guidelines 2010 (n 4), 28. See, for example, Joe S Bain, Barriers to New Competition (Harvard University Press 1956); George J Stigler, The Organization of Industry (University of Chicago Press 1968); Richard Schmalensee, ‘Sunk Costs and Antitrust Barriers to Entry’ (2004) 94 American Economic Review 471. 64 Normal profit is the profit required to retain capital in its present use, that is, it is equal to at least what could be earned in its next best use. 65 For a more detailed discussion of barriers to entry, see US Horizontal Merger Guidelines 2010 (n 4), 27–9; European Commission, ‘Guidelines on the Assessment of Horizontal Mergers under the Council Regulation on the Control of Concentrations between Undertakings’ (2004/C 31/03) paras 68–75, http://​eur​-lex​.europa​.eu/​legal​-content/​EN/​ALL/​?uri​=​celex:​52004XC0205(02) (EC Horizontal Merger Guidelines); Gunnar Niels, Helen Jenkins and James Kavanagh, Economics for Competition Lawyers (2nd edn, Oxford University Press 2016) ch 3, section 3.3; RE Caves and ME Porter, 62 63

Mergers  245 While it is true that entry, or the threat of entry, is a constraint on a firm or firms that might otherwise possess substantial market power, it is important to know whether entry will (rather than could) occur; whether ‘look-alike’ entrants are likely to be successful, or whether entrants must differ in some significant way from incumbents if they are to enter successfully; if the incumbent expects entry to constrain it, does it respond by accommodating entry or does it engage in entry-deterring strategies; and finally, is entry relatively routine/passive (unlikely, given the nature of entry and its unsettling impact on incumbents) or are there specific, targeted strategies that an entrant may need to adopt in order to enter successfully?66 Thus, it is not sufficient to alleviate concerns about the effect of the merger on competition by simply pointing to examples of firms coming into the market. 3 Countervailing power Countervailing power exists when one or more buyers can effectively bypass a supplier that possesses substantial market power. This is relevant where post merger there will be few effective competitors left in the market and where barriers to entry are high. A large buyer for whom the relevant product is very significant in terms of its costs, quality and/or profitability may vertically integrate and self-supply the product.67 Alternatively, the buyer may be able to sponsor third-party entry or imports because it is able to contract for supply of a sufficiently large quantity and for long enough to reduce entry risks and thereby overcome barriers to entry. Another possibility is that the buyer may be able to refuse to buy other products of the supplier where those products are supplied in competitive markets. Perfect countervailing power provides a competitive constraint that eliminates economic rents.68 Not all large buyers possess countervailing power, although they may have some bargaining power – ‘[p]owerful buyers are often able to negotiate favourable terms with their suppliers’, but may not be able fully to protect themselves from the exercise of market power.69 In addition, even if powerful customers are able to protect themselves from the exercise of market power by a supplier, this does not assist less powerful customers who may be subject to price discrimination. Therefore, ‘[c]ountervailing buyer power cannot be found to sufficiently off-set potential adverse effects of a merger if it only ensures that a particular segment of customers … is shielded from significantly higher prices, or deteriorated conditions after the merger’.70

‘From Entry Barriers to Mobility Barriers: Conjectural Decisions and Contrived Deterrence to New Competition’ (1977) 91 Quarterly Journal of Economics 241. 66 US Horizontal Merger Guidelines 2010 (n 4), 29. 67 The disadvantage of this is that there are capital costs and the probable loss of economies of scale and hence increased cost. 68 See EC Horizontal Merger Guidelines (n 65), paras 64–7. The US Horizontal Merger Guidelines 2010 (n 4) discuss buyer power in terms of the bargaining power. They discuss the extent to which buyer power negates competition concerns associated with a merger, ibid, 27. 69 US Horizontal Merger Guidelines 2010 (n 4), 27. Bargaining power affects the division of economic rents between parties. 70 EC Horizontal Merger Guidelines (n 65), para 67.

246  Research handbook on methods and models of competition law C

Mergers and Efficiencies

One of the most controversial areas in merger assessment is the treatment of efficiencies claimed to result from the merger. Although not all mergers are claimed to deliver efficiencies, many, if not most, of the controversial mergers are. Efficiency benefits may be especially significant in the case of non-horizontal mergers.71 So long as the market remains competitive, efficiency benefits from a merger are likely to be shared with consumers in the form of lower prices and/or improved products.72 However, even if they are not shared, if there are resource savings from the merger, net welfare increases because these saved resources can be reallocated to other uses. Figure 10.2 is based on the naïve Williamson trade-off model,73 from which he concluded that a loss of allocative efficiency could be compensated for by an increase in production efficiency. Initially, price is P1, output is Q1 and the consumer surplus is ABP1. Now assume that all firms in the market merge, creating a monopoly. This will cause the price to rise to P2, output to decrease to Q2 and consumer surplus decreases – DEFP1 is transferred to the merged firm as monopoly rent and BEF is the deadweight loss. However, the merger facilitates the adoption of new technology requiring less inputs on average per unit of output and so lowers costs to AC’=MC’. Given this, the new profit-maximizing output is Q3 and the price is P3 – the former is greater and the latter less than would have been the case absent the cost reduction. This causes a further increase in monopoly profits but an increase in consumer surplus rel-

Figure 10.2

Williamson trade-off model

71 See, for example, Lynne Pepall, Dan Richards and George Norman, Industrial Organization: Contemporary Theory and Empirical Applications (Wiley 2008) chs 16–18. 72 Alternatively, although the price remains above the competitive level, it will be lower if double marginalization is eliminated than if there is vertical separation in the supply chain. 73 See OECD, ‘The Role of Efficiency Claims in Antitrust Proceedings’ (Best Practice Roundtables on Competition Policy 2012) 17–18, http://​www​.oecd​.org/​daf/​competition/​EfficiencyClaims2012​.pdf.

Mergers  247 ative to the merger without the cost saving. Ignoring the transfer between consumers and producers, there will be a net increase in welfare if the cost saving from the merger exceeds the deadweight loss. Williamson and others concluded that only quite a small decrease in cost is sufficient to result in a net benefit in terms of welfare, and in these cases the merger should be permitted.74 To avoid the loss of potential welfare increases, the ultimate objective of competition policy, it might be expected that if the merger is efficiency enhancing it would be allowed to proceed. However, the merger test is not an efficiencies test, it is a competition test. So, a merger that increases efficiency but substantially lessens competition or creates/ increases dominance may be prohibited. Efficiencies are treated differently in the three jurisdictions under discussion. In the US, provision for an efficiencies defence was included in the 1968 Horizontal Merger Guidelines, but was seen as exceptional.75 However, efficiency gains were seen as a threat to smaller rivals and so courts were not sympathetic to efficiency claims. In the 1984 Guidelines, efficiencies ceased to be treated as a defence for an anticompetitive merger; instead, efficiencies were included in the competition analysis to the extent that they were likely to enhance competition. Dynamic efficiencies are discussed in more detail in the 2010 Guidelines, but they caution that while research and development cost savings from a merger may be substantial, they may not be cognizable efficiencies because they are difficult to verify or may result from anticompetitive reductions in innovative activities.76 Efficiency claims have succeeded, as in the FTC’s investigation of a proposed merger between Genzyme Corp and Novazyme Pharmaceuticals Inc in 2004,77 while in other cases, such as FTC v HJ Heinz Co and Milnot Holding Corporation,78 they failed. Treatment by the court of a proposed merger between Anthem Inc and Cigna Corp in 2017 demonstrates the continuing uncertainty regarding the approach to efficiencies.79 The EU was initially hostile towards claimed efficiencies but has given them more attention in recent years.80 It has been claimed that the ‘EU has traditionally treated efficiencies more as an offence then [sic] a defense’.81 However, in its 2004 Guidelines, the EU stated that it would, in certain circumstances, take into account efficiencies generated by a merger when undertaking a competition analysis. Although efficiency claims have been discussed in a number of mergers, in 2012 the OECD stated that ‘to date the European Commission has cleared no merger solely on the basis of efficiencies’.82 The US and the EC authorities generally are sceptical of efficiency claims, in part because the parties are prone to exaggerate their claims, but also because more than 50 per cent of all

For criticisms of the Williamson model, see ibid, 5–6. US Department of Justice, ‘1968 Merger Guidelines’ (1968) paras 10, 16, https://​www​.justice​ .gov/​archives/​atr/​1968​-merger​-guidelines. 76 US Horizontal Merger Guidelines 2010 (n 4), 30. 77 OECD, ‘The Role of Efficiency Claims in Antitrust Proceedings’ (n 73), 20. 78 Ibid, 21–2. 79 United States of America v Anthem Inc and Cigna Corporation, USCA 17-5024 (2017), https://​ www​.justice​.gov/​atr/​case​-document/​file/​971316/​download. 80 Clarke (n 21), 189; see also OECD, ‘The Role of Efficiency Claims in Antitrust Proceedings’ (n 73), 22. 81 Vlatka Butorac Malnar, ‘The Role of Efficiencies in Merger Control: Comparative EU-USA Perspective’ (2008) 29 Zbornik Pravnog fakulteta Sveučilišta u Rijeci 851, 852, https://​hrcak​.srce​.hr/​file/​ 63801. 82 OECD, ‘The Role of Efficiency Claims in Antitrust Proceedings’ (n 73), 23. 74 75

248  Research handbook on methods and models of competition law mergers decrease shareholder value.83 There are various reasons for this, including difficulties in integrating physically separated plants, the cost and difficulty of shedding staff, and the limits to management capacity. Hence it is relatively unlikely that efficiency claims, for which there is usually little evidence, will facilitate an otherwise anticompetitive merger. Thus, while the competition authorities in both jurisdictions accept that efficiencies may be taken into account in merger analysis, they do not accept the total welfare standard advocated by Williamson.84 Consequently, they require not only that the claimed efficiencies are merger specific and verifiable, but also that they benefit consumers.85 The difficulty in meeting this standard explains the relatively minor role played by efficiencies in merger analysis in these jurisdictions.86 In Australia, although the procompetitive effect of efficiencies can be balanced against any anticompetitive effects in assessing a merger, efficiency benefits may be the basis for an application for authorization, as in the case of AGL Energy Ltd and Macquarie Generation.87 D

Failing Firm Defence

Especially when economic conditions are difficult, claims may be made by merging parties that, absent the merger, the target will fail and consequently, with or without the merger, in the near future the target will no longer be present in the market.88 Competition authorities accept that ‘a merger is not likely to enhance market power if imminent failure … of one of the merging firms would cause the assets of that firm to exit the relevant market’.89 However, such claims may be made in circumstances where the merger is likely to be viewed as anticompetitive. Consequently, while competition authorities accept the possibility of a failing firm defence of a merger,90 they are generally sceptical about such claims.

83 Martin Roll, ‘Reasons Why Most Mergers Destroy Shareholder Values’ (February 2014), https://​ martinroll​.com/​resources/​articles/​strategy/​reasons​-why​-most​-mergers​-destroy​-shareholder​-value/​. For an overview of the performance of mergers, see Dennis W Carlton and Jeffrey M Perloff, Modern Industrial Organization (4th edn, Pearson 2015) 50–3. 84 Jonathan M Orszag and Loren K Smith, ‘Toward a More Complete Treatment of Efficiencies in Merger Analysis: Lessons from Recent Challenges’ (2016) 16(1) Antitrust Source, americanbar​.org/​ groups/​antitrust​_law/​publications/​the​_antitrust​_source/​. They argue that procompetitive efficiency gains from mergers are often inadequately recognized. 85 Clarke (n 21), 190; EC Horizontal Merger Guidelines (n 65), paras 85–8. 86 There is a problem in claiming efficiencies in support of an anticompetitive merger: if a firm gains market power due to the merger, why would it pass through the efficiency savings to its customers? 87 Application for Authorisation of Acquisition of Macquarie Generation by AGL Energy Ltd [2014] ACompT 1, http://​www​.judgments​.fedcourt​.gov​.au/​judgments/​Judgments/​tribunals/​acompt/​2014/​ 2014acompt0001. 88 The failing firm defence is also known as the exiting firm defence or the failing company doctrine. See Oxera, ‘The Failing-Firm Defence: A “Get Out of Jail Free” Card in Mergers?’ (Agenda 2014), https://​www​.oxera​.com/​Latest​-Thinking/​Agenda/​2014/​The​-failing​-firm​-defence​-a​-get​-out​-of​-jail​-free​ .aspx. 89 US Horizontal Merger Guidelines 2010 (n 4), 32. 90 In the US and the EU, this is a defence; in Australia, failing firm arguments are included in the competition analysis. For a discussion of the differences between the jurisdictions, see OECD, ‘The Failing Firm Defence’ (Best Practice Roundtables on Competition Policy 2009) 30–1, https://​www​.oecd​ .org/​daf/​competition/​mergers/​45810821​.pdf.

Mergers  249 The merging parties must provide evidence to satisfy a number of criteria that establish a failing firm defence. They must establish that, absent a takeover, the target company would be forced out of the market because of financial difficulties ‘in the near future’. To establish this, it is not sufficient simply to establish that the firm’s revenue and/or market share is falling or that it is making losses91 – for example, this could be simply because marketing support has been withdrawn in anticipation of a sale. Rather, it must be shown that the failing firm is unable to be ‘saved’ by changes in management, rationalization of staff and/or products, renegotiation of supply and distribution arrangements, and the like. If this requirement is satisfied, it must be shown that, absent the acquisition, the target company’s assets would inevitably exit the market. Finally, it must be shown that the business would not be acquired by another purchaser that would raise fewer competition concerns.92 Even if these conditions are satisfied, the question is still why a buyer would acquire the business as a going concern when, by waiting and allowing it to fail, the assets could be acquired piecemeal at a lower price. The answer is likely to be that there is some advantage in acquiring the business as a going concern – for example, acquiring the customers. If the business fails, typically customers redistribute between remaining firms approximately in proportion to their market shares. However, if the business is acquired as a going concern, most customers remain with the merged entity. Clearly, this has implications for competition where there are economies of scale to be exploited. Unsurprisingly, failing firm defences tend to fail more often than they succeed.93 The position is that the ‘failing firm defense still applies rarely and narrowly and will only succeed in the presence of strong factual evidence’.94

Ibid, 176. To establish this, the target business may refuse to supply documentation to other bidders who are expected to offer less, reflecting expected competitive market conditions. 93 Examples of successful failing firm cases in the EU include Commission Decision of 14 December 1993 Relating to a Proceeding Pursuant to Council Regulation (EEC) No 4064/89 (Case No IV/M.308 – Kali + Salz/MdK/Treuhand) Commission Decision 94/449/EC [1993] OJ L186/38; Commission Decision of 11 July 2001 Declaring a Concentration to Be Compatible with the Common Market and the Functioning of the EEA Agreement (Case COMP/M.2314 – BASF/Eurodiol/Pantochim) Commission Decision 2002/365/EC [2001] OJ L132/45; Commission Decision of 2 September 2013 Declaring a Concentration to be Compatible with the Internal Market and the Functioning of the EEA Agreement (Case COMP/M.6360 – Nynas/Shell/Harburg Refinery) Commission Decision 2014/C 368/04 [2013] OJ C368/5; Commission Decision of 26 January 2011 Declaring a Concentration to Be Incompatible with the Internal Market and the EEA Agreement (Case COMP/M.5830 – Olympic/Aegean Airlines) Commission Decision 2012/C 195/10 [2013] OJ C195/11. In the US, the first failing firm case was International Shoe Co v Federal Trade Commission, 280 US 291 (1930). In Australia, a successful failing firm argument was made in P & M Quality Smallgoods Pty Ltd – Proposed Acquisition of Hans Continental Smallgoods Pty Ltd [2009]; VIP Steel Packaging Pty Ltd – Proposed Acquisition of Certain Steel Drum Assets from National Can Industries Pty Ltd [2012]. 94 Anna Tzanaki, ‘Time to Get Rid of the Failing Firm Defense? Some Thoughts in the Wake of Nynas/Shell and Olympic/Aegean’ (Competition Policy International, 2 January 2014), https://​www​ .competitionpo​licyinternational​.com/​time​-to​-bid​-farewell​-to​-the​-failing​-firm​-defense​-some​-thoughts​-in​ -the​-wake​-of​-nynas​-olympic​-aegean/​. This article is sceptical as to the usefulness of the failing firm defence as it is applied. 91 92

250  Research handbook on methods and models of competition law E

Quantitative versus Qualitative Assessment

Given the factors likely to determine the impact of a merger on competition, the next issue concerns how that effect is to be determined. Today there is an array of potential quantitative methods that can be employed for this purpose, including critical loss analysis, price tests, and diversion ratios,95 through to the use of complex simulation models. The simpler, less data-intensive techniques are frequently used, while simulation modelling is used but less frequently. F

Simulation Modelling

The original claim for merger simulation was that it would allow investigators to avoid much of the standard analysis and instead to carry out a direct analysis of the precise effect of the merger. The more extreme claims were that on the basis of data on market shares, current prices and industry elasticity, the effect of a merger could be predicted to a reasonable degree of accuracy, including allowing for any claimed post-merger marginal cost reductions. The result was that the period from about 2000 to 2005 saw a substantial growth in the use of merger simulations.96 The OECD explains that often the confidence intervals attached to the simulations are very wide and so there is considerable uncertainty attached to the estimates provided by the simulation. This is attributed to ‘difficulties with estimating elasticities precisely and to the difficulty of modelling the exact nature of competition within a market’.97 Further, generally simulations are based on the assumption that the only change in the market is the merger itself; and the effect measured is usually price, so that non-price-related effects, which may be important, are ignored. Finally, data deficiencies often limit the use of merger simulations, especially in small economies such as Australia.98 Despite these issues, simulations have been used in a number of merger cases. In the EU, these include Volvo/Scania 2000,99 Philip Morris/Papastratos 2003,100 Oracle/PeopleSoft 2005,101 BHP

95 See, for example, OECD, ‘Economic Evidence in Merger Analysis’ (Best Practice Roundtables for Competition Policy 2011) 24–6, http://​www​.oecd​.org/​daf/​competition/​EconomicEvid​ enceInMergerAnalysis2011​.pdf. 96 Ibid, 32. 97 Ibid, 33. 98 Merger simulations were employed in the Toll/Patrick case in Australia in 2006: ibid, 32. 99 Commission Decision of 14 March 2000 Declaring a Concentration to be Incompatible with the Common Market and the Functioning of the EEA Agreement (Case No COMP/M. 1672 – Volvo/Scania) Commission Decision 2001/403/EC [2001] OJ L143/74. 100 Commission Decision of 2 October 2003 Declaring a Concentration to be Compatible with the Common Market (Case No COMP/M.3191 – Philip Morris/Papastratos) Commission Decision 2003/C 258/04 [2003] OJ C258/4. 101 Commission Decision of 26 October 2004 Declaring a Concentration to be Compatible with the Common Market and the Functioning of the EEA Agreement (Case No COMP/M.3216 – Oracle/ PeopleSoft) Commission Decision 2005/621/EC [2005] OJ L218/6.

Mergers  251 Billiton/Rio Tinto 2008,102 Kraft Foods/Cadbury 2010,103 and Unilever/Sara Lee Body Care 2010.104 G

Natural Experiments

An alternative technique, and one that is much less data intensive, is the use of natural experiments. Natural experiments ‘relate to past variations in competitive conditions in an industry that can be used to provide indications of the likely effects of a merger’.105 Situations that provide the basis for a natural experiment include the same market but at a different time; a market with the same product dimension but in a different location; and a different market in the same location but one which has features very similar to the one under consideration. For example, ‘if the merging firms compete in some locales but not others, comparisons of prices charged in regions where they do and do not compete may be informative regarding post-merger prices’.106 Although it may be difficult to determine whether the identified effect is attributable to the relevant cause, natural experiments require relatively little data and so have been increasingly used.107 Examples of use in the US include Nestlé and Dreyer’s Grand Ice Cream Inc 2003,108 Whole Foods and Wild Oats 2007,109 Dean Foods and Foremost 2010,110 and Staples and Office Depot 2016,111 while in the EU examples of usage include Blackstone and Acetex 2005,112 INEOS and Kerling 2008,113 and INEOS and Solvay 2014.114 Despite increased willingness on the part of competition authorities to use quantitative analyses of mergers, qualitative assessment of the factors that influence mergers, based on information relevant to the specific merger, remains the most common basis for analysis. In oligopolistic markets, particular attention needs to be paid to how the merger may change the 102 Withdrawal of Notification of a Concentration (Case No COMP/M.4985 – BHP Billiton/Rio Tinto) [2008] OJ C312/16. 103 Commission Decision of 6 January 2010 Declaring a Concentration to be Compatible with the Common Market (Case No COMP/M.5644 – Kraft Foods/Cadbury) [2010] OJ 29/4. 104 Commission Decision of 17 November 2010 Declaring a Concentration to be Compatible with the Internal Market and the EEA Agreement (Case No COMP/M.5658 – Unilever/Sara Lee Body Care) [2010] OJ C108/06 105 OECD, ‘Economic Evidence in Merger Analysis’ (n 95), 308. 106 US Horizontal Merger Guidelines 2010 (n 4), 3, para 2.1.2. 107 OECD, ‘Economic Evidence in Merger Analysis’ (n 95), 230–2. 108 In re Nestlé Holdings, Inc; Dreyer’s Grand Ice Cream Holdings, Inc; and Dreyer’s Grand Ice Cream, Inc, Federal Trade Commission, Docket No C-4082 (12 November 2003). 109 In re Whole Foods Market, Inc, and Wild Oats Markets, Inc, Federal Trade Commission, Docket No 9324 (7 August 2007). 110 Federal Trade Commission v Whole Food Market, Inc and Wild Oats Markets, Inc, Case 09-1020 (23 January 2009). 111 In re Staples, Inc and Office Depot, Inc, Federal Trade Commission, Docket No 9367 (19 May 2016). 112 Commission Decision of 13 July 2005 Declaring a Concentration Compatible with the Common Market and the Functioning of the EEA Agreement (Case No COMP/M.3625 – Blackstone/Acetex) Commission Decision 2005/839/EC [2005] OJ L312/60. 113 Commission Decision of 30 January 2008 Declaring a Concentration to be Compatible with the Common Market and the Functioning of the EEA Agreement (Case No COMP/M.4734 – INEOS/ Kerling) [2007] OJ C174/11. 114 Commission Decision of 8 May 2014 Declaring a Concentration to be Compatible with the Internal Market (Case No COMP/M.6905 – INEOS/Solvay/JV) [2014] OJ C273/11.

252  Research handbook on methods and models of competition law relationship between various parties, especially with customers and suppliers, and how these relationships may be used to raise entry barriers – that is, to change the dynamics of the market. H

Sources of Evidence

To determine whether a proposed merger is likely to be anticompetitive, what sources of evidence would potentially assist and how can evidence be obtained?115 The merging parties clearly have an interest in ensuring that the competition authorities are supplied with information concerning the merger. Documents and data created for normal business purposes are likely to be more useful than those created specifically for submission to the competition authorities. Also, the terms of the sale contract may be useful.116 Useful insights into the businesses of the parties may be obtained from former employees, assuming that they no longer have any ties to the businesses. They can provide industry information and may be able to provide an opinion as to the likely market outcomes resulting from the merger. Customers of the merging parties can provide information about the relevant market and past activity in that market, such as responses to previous price increases. As well, they can provide information as to their purchasing behaviour, and whether they are concerned that prices will increase post merger or that other supply terms will be adversely affected. They can explain how they expect to react to any such price increase, what options they have when responding to such increases, and the attractiveness of alternative products or suppliers. However, if the products of the merged entity are acquired as inputs into products sold in a downstream market that is relatively uncompetitive, the customer may be able to pass through the price increase and so may not be concerned about price increases post merger. Competitors can provide information about the market and how it operates. However, information from this source must be treated with care. Competitors are likely to complain if the merger is expected to increase competitiveness in the market but are unlikely to do so if they expect to benefit from higher prices. Others who may assist in explaining how the market works and/or in providing opinions as to the effect of the merger include suppliers, distributors, industry experts and economists.

V

SOME CONCLUDING THOUGHTS

The prohibition on anticompetitive mergers is a fundamental plank in the competition law platform. Its role is to prevent mergers that change the competitive structure in ways that enable coordinated conduct or that increase unilateral market power. Changes in recent years, such as more open economies and online sales, have introduced new or more significant constraints into many markets and so have diminished the risk that mergers in such markets will be anticompetitive. However, multinational businesses face significant cost and delay in dealing with different laws in different jurisdictions and differences in the processes employed by different competition authorities. In response to these issues, there has been increasing alignment of laws and processes in the three jurisdictions discussed in this chapter, but significant differences remain, in part due to philosophical differences in belief about how businesses behave. See, for example, US Horizontal Merger Guidelines 2010 (n 4), 4–5. See ibid, 4, para 2.2.1.

115 116

11. Competition law in Japan, Malaysia and the Philippines: an overview Mel Marquis1

I INTRODUCTION The last decade has seen the introduction of a competition law for the first time in many developing countries in Asia. Typically, these new laws were based in part on either US or EU competition law, but they often incorporated additional objectives and reflected the existing legal, institutional and cultural structures of those Asian countries. In the context of this process, these developing countries and their newly established competition authorities face problems that may limit the full effectiveness of their nascent competition regimes if left unaddressed, such as lack of independence from government, shortage of funds, and difficulty securing employees with appropriate skills to implement the law. This chapter focuses on competition law in three jurisdictions: two countries with relatively new competition laws, and one with an old established law. The first is Malaysia, which passed its Malaysian Competition Act (MyCA) in 2010. The second is the Philippines, which, despite long-standing competition law provisions, only passed the Philippine Competition Act (PCA) in 2015. Contrasting with these newer laws, Japan has had a competition law for over 70 years. In this chapter, Japan’s Anti-Monopoly Act (AMA) is compared to the equivalent legislation in Malaysia and the Philippines. At a high level, the aim of competition law in all three jurisdictions is to promote and protect competition. Thus, the PCA refers to preventing undesirable economic concentration and penalizing anticompetitive conduct. However, it is accepted that competition law is also a means of encouraging economic development and promoting welfare. The stated objectives of Japan’s AMA include fair and free competition; enhanced levels of employment and national income; and development of the national economy.2 Similarly, both the Malaysian and Philippines Acts refer to promoting economic development.3 Fairness is also a focus of competition law in these countries. In Japan, this is reflected in the prohibition on unfair trade practices, such as the abuse of a superior bargaining position or the unfair inducement of consumers to conclude contracts. In the Philippines, section 2 of the PCA provides, inter alia, that the state must promote free and fair competition.

1 The author is very grateful to the editors, Professors Deborah Healey and Rhonda Smith, for input and corrections that significantly improved the text. He received valuable feedback on a previous draft from Shila Dorai Raj, Cassie Lee, Cid Butuyan, Graciela Base, Iwakazu Takahashi, Toshiaki Takigawa, Akinori Uesugi, Etsuko Kameoka, Shūya Hayashi, Shingo Seryo, and Tadashi Shiraishi. 2 AMA, Art 1. 3 The Philippine Competition Commission (PCC) helped the National Economic and Development Authority (NEDA) to formulate a strategy outlined in Chapter 16 of the Philippine Development Plan 2017–22, approved by NEDA in February 2017.

253

254  Research handbook on methods and models of competition law

II

COMPETITION LAW PROCESS

A

Legal Provisions

1 Anticompetitive agreements While the competition laws of the three jurisdictions prohibit anticompetitive agreements, they differ significantly in their approach. The AMA distinguishes between horizontal and vertical agreements. The MyCA does not distinguish between types of agreements, except for cartel conduct. Section 14 of the PCA establishes three categories of agreements: a per se category; a horizontal ‘object or effect’ category; and another ‘object or effect’ category of agreements that may be justified. In Malaysia, hard-core cartel conduct is deemed to have the object of significantly restricting competition (s 4(2) of the MyCA), and in the Philippines price-fixing and bid-rigging are per se illegal under section 14 of the PCA. In Japan, Article 3 of the AMA, read in conjunction with Article 2(6), prohibits horizontal agreements, including price-fixing arrangements, if they are an ‘unreasonable restraint of trade’. This is interpreted as being a substantial risk to competition due to the creation, maintenance or strengthening of market power. Consequently, cartel conduct in Japan is not technically per se illegal4 and a breach can only be established if market power is created or strengthened.5 Non-cartel agreements and concerted practices are prohibited by section 4(1) of the MyCA if they have the ‘object or effect of significantly preventing, restricting or distorting competition in any market for goods or services’. In the Philippines, horizontal agreements are prohibited if they have the object or effect of substantially lessening competition and they either limit or control production, markets, technical development or investment, or if they divide or share markets.6 In Japan, vertical restraints fall under the rubric of ‘unfair trade practices’.7 Unfair trade practices are prohibited by Articles 2(9) and 19 of the AMA, as supplemented by the Japan Fair Trade Commission’s (JFTC’s) ‘General Designations’ (2009). Market power need not be demonstrated (see above). In all three jurisdictions, conduct may be exempt from competition law scrutiny under certain conditions. For example, in Japan resale price maintenance (RPM) may be justified on efficiency grounds, while in Malaysia individual exemptions can be claimed by virtue of section 5 of the MyCA, which establishes conditions roughly similar to those that apply in Europe under Article 101(3) of the Treaty on the Functioning of the European Union (TFEU). In the Philippines, non-horizontal agreements are ‘not necessarily’ prohibited if they contribute to improving the production or distribution of goods and services, or to promoting technical or economic progress, while allowing consumers a fair share of the resulting benefits.

4 OECD, Competition Law in Asia-Pacific: A Guide to Selected Jurisdictions (2018) 88, http://​www​ .oecd​.org/​daf/​competition/​Competition​-Law​-in​-Asia​-Pacific​-Guide​-2018​.pdf. 5 Supreme Court, judgment of 17 December 2010, Higashi Nihon Denshin Denwa v JFTC, 64(8) Minshu 2067. 6 The factors guiding the PCC’s assessment are described in s 26 of the Act and in rule 7 of the PCC’s Implementing Rules and Regulations (IRRs) of 2016, http://​phcc​.gov​.ph/​implementing​-rules​ -regulations​-philippine​-competition​-act. 7 OECD (n 4), 88.

Competition law in Japan, Malaysia and the Philippines  255 2 Abuse of dominance Both the Malaysian and the Philippine competition laws prohibit abuse of a dominant position by a business that possesses substantial market power.8 Each includes a definition of ‘dominance’ within the Act.9 For the former, an enterprise is dominant if it can adjust prices or output without effective constraint from competitors. In both jurisdictions, the conduct of a dominant firm will only contravene the Competition Act if it has an anticompetitive effect.10 In Japan, unilateral conduct may be treated as an unfair trade practice. The AMA lists conduct regarded as unfair trade practices (Article 2(9)) and this has been supplemented by a number of practices designated by the JFTC.11 Articles 3 and 2(5) of the AMA are lesser-used provisions that prohibit ‘private monopolization’ (shiteki dokusen) where an enterprise ‘excludes or controls’ the activities of other enterprises and thus substantially restrains competition. 3 Merger control Malaysia currently has no comprehensive ex ante merger approval system, although some (voluntary) sector-specific regimes do exist. By contrast, the competition laws of both Japan and the Philippines prohibit mergers that are anticompetitive.12 In Japan, a transaction is prohibited if it could substantially restrain competition in a particular field of trade. In the Philippines, mergers are prohibited if they substantially lessen competition. Both the AMA and the PCA have compulsory notification of mergers over a certain value. In Japan, notification is required if total domestic sales of one party exceed 20 billion yen (US$180 million) and total domestic sales of the other party exceed 5 billion yen (USD$45 million). Notification does not in itself require ‘control’, and acquisitions of minority shares may be subject to mandatory notification.13 In the Philippines, a proposed transaction must be notified if its value exceeds 2 billion pesos (around US$38 million).14 ‘Value’ refers to either the target’s assets in the Philippines or to the gross revenues generated by those assets in or into the Philippines. At least one of the parties must have consolidated sales or assets in the Philippines exceeding 5 billion pesos (US$95 million).15 As noted above, the MyCA does not contain a merger provision. The antitrust prohibitions against abuse of dominance and restrictive agreements could be used in relation to mergers in some scenarios – but, as witnessed in Europe before 1989, such ex post control comes too late,



MyCA, s 2; PCA, s 15. MyCA, s 2; PCA, s 27 and rule 8 of the PCC’s Implementing Rules and Regulations (IRRs). 10 Sections 3.6–3.9 of the MyCC Guidelines on the Chapter 2 Prohibition. 11 See further Etsuko Kameoka, Competition Law and Policy in Japan and the EU (Edward Elgar Publishing 2014) 91–3; Toshiaki Takigawa, ‘Japan’ in Mark Williams (ed), The Political Economy of Competition Law in Asia (Edward Elgar Publishing 2013) 11, 33–7. 12 In Japan, the main provisions are contained in Chapter IV of the AMA. In the Philippines, relevant rules are contained in sections 16–23 of the PCA, in conjunction with rule 4 of the 2016 IRRs, the Rules on Merger Procedure, four Clarificatory Notes, the Merger Review Guidelines, and the Guidelines on the Computation of Merger Notification Thresholds. 13 See, for example, JFTC Media Release (26 May 2015) (acquisition of 20.9 per cent of voting rights, thus exceeding the 20 per cent threshold). 14 Rule 4, s 3 of the IRRs as amended in 2018. 15 On filing, timing and other practical points, see Rachelle Diaz and Edan Cañete, ‘The Philippine Competition Act and the Rules on Merger Control and Antitrust: A Guide for International Firms’ (2018) 9(1) Journal of European Competition Law & Practice 42. 8 9

256  Research handbook on methods and models of competition law with generally irreversible consequences for welfare.16 It is also cumbersome and susceptible to errors. It is acknowledged that a developing country may have an interest in delaying or avoiding the introduction of merger control. Such an approach can support industrial policy goals by enabling mergers that achieve economies of scale,17 and it can mitigate substantial resource and personnel costs in enforcement (though some savings may be offset by potentially higher ex post enforcement costs). Nevertheless, the premise of this strategy may be questioned: if a well-designed merger review system is established, mergers apt to enhance economies of scale will not be blocked if they are procompetitive.18 New discussions among Malaysian officials about the introduction of merger control have been provoked by the Uber-Grab merger (which the Malaysia Competition Commission awkwardly had to examine under its authority to police abuse of dominance), although other factors also may have played a role in these discussions. The Minister for Domestic Trade and Consumer Affairs (MDTCA) has been quoted as saying that Malaysia ‘needs to get ready to analyse and further develop its competition law in order to tackle more high-profile cases to prevent and resolve future market monopoly’.19 It has also been reported that the Malaysian Competition Commission (MyCC) will consult with other agencies as part of a new effort to amend the MyCA to cover mergers and acquisitions, presumably in the form of ex ante review.20 It is unclear what the new measures would be, or what implications they would have for other provisions of the MyCA. What does seem clear is that, when the MyCC gains authority to conduct systematic merger control, it needs to be given greater financial and human resources to succeed (see below). B

Competition Law Enforcement

1 Competition authorities Of particular significance for the successful operation of competition law in a jurisdiction is the independence of the competition authority from government,21 and this varies significantly between the three countries. Key formal indicators in this regard are the way in which members of the commission are appointed, the ease with which commissioners can be removed, and the

16 Wilson Tze Vern, ‘Competition Law in Malaysia: Renaissance and the Road Ahead’ (2013) 2 Malayan Law Journal 23, 35. 17 May Fong Cheong and Yin Harn Lee, ‘Malaysia and Singapore’ in Mark Williams (ed), The Political Economy of Competition Law in Asia (Edward Elgar Publishing 2013) 215, 243. 18 Furthermore, the need to assess mergers helps to develop analytical skills and sectoral knowledge within the enforcement agency. Cassey Lee, ‘Competition Law Enforcement in Malaysia: Some Recent Developments’ (2014) 51 Malaysian Journal of Economic Studies 77. In the absence of authority to conduct merger reviews, the MyCC has used market studies, for example in the pharmaceutical and building material sectors, to enhance its knowledge of certain industries. 19 See Nur Haziqah A Malek, ‘Grab-Uber Merger Probe to Conclude Soon’, The Malaysian Reserve (28 November 2018), https://​themalaysianreserve​.com/​2018/​11/​28/​grab​-uber​-merger​-probe​-to​-conclude​ -soon. David Fruitman kindly referred the author to this news report. 20 ‘Sector Regulators Meet to Discuss Competition Issues’, Bernama (21 December 2018), http://​ bernama​.com/​en/​news​.php​?id​=​1677784. 21 For discussion of the advantages of independent competition law enforcement, see OECD Secretariat, ‘Background Paper, Independence of Competition Authorities – From Designs to Practices’ (DAF/COMP/GF(2016)5, 21 November 2016).

Competition law in Japan, Malaysia and the Philippines  257 Table 11.1

Competition authorities: appointments and tenure

Authority

Composition

Appointed by

Special conditions

Tenure

Japan Fair Trade

Chairman and 4

Prime Minister with

Must have knowledge

5-year term, with

Commission

commissioners

consent of the Diet

and experience in law or possibility of renewal; economics

only removable for cause (Art 31)

Malaysian Competition

Chair and 7–9

Prime Minister on the

4 members represent

3 years, plus one

Commission

commissioners

recommendation of the

the government, 1 of

renewal; Prime Minister

MDTCA Minister

whom must be from the

may remove any

Philippines Competition Chair and 4 Commission

commissioners

President

MDTCA

commissionera

Must have appropriate

Non-renewable 7–year

professional experience, term; secure tenure probity and independence

Note: a This can be done in the interest of the ‘effective and economical performance of the functions of the Commission’ (MyCA, s 11(1)). The Prime Minister is thus not authorized to terminate a commissioner at will, but neither is there a need to justify termination on grounds of, for example, misconduct or incapacity.

financing of the commission. Table 11.1 outlines the process for appointing members of the competition authority in each of the three countries. Chairs of the JFTC have traditionally been former officials of the powerful Ministry of Finance. This strong leadership contributes to the JFTC’s capacity to push its agenda and to overcome vested interests. The JFTC also occupies a powerful bureaucratic status as an external and independent bureau of the Cabinet Office. Similarly, in the Philippines the Commission Chairman has a rank equivalent to cabinet secretary. Also important is the appointment of commissioners who are suitably qualified. The process in the Philippines specifically recognizes this. Appointees in Malaysia appear generally to be qualified: the MyCC as at December 2018 comprised a former judge of the Court of Appeal as Chairman, two government representatives (though the Act calls for four), two academics specializing in law and economics respectively, two solicitors, and a former Public Works Department official. However, the background of some of the commissioners gives cause for concern about perceived or actual independence. As explained below, other provisions of the MyCA establish additional links between the MyCC and the government. Given their role and multiple tasks, it is essential that the commissions are well resourced. The JFTC’s high rank as an external bureau of the Cabinet enables it to compete with ministries for resources and gives it credible authority over industries regulated by the Ministry of Economy, Trade and Industry and other government bodies. As a consequence, the JFTC is a large and well-resourced agency – its budget has increased over the years and for FY 2017 was 11.22 billion yen (about US$102 million). In 2017, it had 832 staffers, including 438 investigators and a number of senior economists. In the Philippines, the resources provided to the Philippine Competition Commission (PCC) – an agency just getting started in a developing country – have generally been sufficient, but signs of stress are beginning to appear. The PCC’s initial budget of 300 million pesos rose to 421 million pesos in 2017 and 457 million pesos (US$8.7 million) in 2018. As of December 2018, 160 employees worked at the PCC, including 46 lawyers and 20 economists. While the agency’s incremental budget increases were necessary and appropriate, PCC salaries are not sufficiently attractive to recruit and retain qualified staff. Having developed expertise with the agency, some employees are lured away to other government offices or to the private sector. The fate of a proposal submitted

258  Research handbook on methods and models of competition law to the Department of Budget and Management for higher salaries to boost recruitment and retention remains to be seen. At the far end of the spectrum, the MyCC appears significantly under-resourced. In 2017, its budget was reported to be 4.2 million ringgit (just over US$1 million), a tiny amount for a US$300 billion economy.22 Government funding, instead of steadily increasing, has trended in the opposite direction: after a start-up grant of 10.5 million ringgit, the annual allocation dipped to 5 million ringgit or less between 2014 and 2017. During the same period, the MyCC incurred a net loss of at least 1.5 million ringgit per year. The sustainability of this pattern is doubtful, and the introduction of merger control would cause additional budgetary strain unless adjustments are made. The MyCC’s human resources likewise appear to pose concern: in 2017, the combined staff and leadership of the MyCC was around 50–60 persons, which is insufficient given the size of the economy and the agency’s responsibilities. Under the AMA, the JFTC has broad enforcement and investigation powers in relation to competition, but its portfolio extends to other issues including fair business dealings and the pass-on of Japan’s consumption tax. The JFTC can initiate legislation and prepare opinions for the Diet to promote the AMA’s purposes (Article 44(2) of the AMA). Like the JFTC, the PCC has a range of powers and functions.23 It is empowered to investigate cases; impose sanctions and apply remedies; issue advisory opinions and guidelines on competition matters; conduct market studies; intervene in administrative and regulatory proceedings; and review economic and administrative regulations. The PCC’s enforcement powers are further specified in provisions contained in Chapter VII (ss 31 et seq) and in the PCC’s 2017 Rules of Procedure.24 The functions of the MyCC also extend beyond investigation and enforcement.25 For example, it advises the MDTCA Minister or any other public or regulatory body on any competition-related matter; alerts the MDTCA Minister to anticompetitive effects in legislation; and makes recommendations to the MDTCA Minister on reforming competition laws. However, in carrying out its functions, the MyCC is not fully autonomous. It is ‘responsible to the Minister’, who may give the agency written directions of a general character if they are consistent with the competition laws, and the Commission must give effect to such directions.26 These provisions, while not unique to Malaysia, are problematic. The ability of the Commission to set its own priorities is restricted by the Minister, who may require it to focus on, for example, pharmaceuticals (which led to a market study of prescription medicines and entry by generic substitutes), tyres, beverages, professional bodies and bid-rigging. While

22 The majority of the MyCC’s receipts for 2017 (around 2.8 million ringgit) consisted of management and development grants from the government. Article 27(2)(b) of Malaysia’s Competition Commission Act 2010 (MCCA) provides that fees and charges imposed by the MyCC are used as partial funding for the agency. While this mechanism seems unwise insofar as it could incentivize the imposition of (higher) fines, or be perceived to do so, in practice the provision has had limited effect. In 2017, fines and penalties amounted to 152,000 ringgit, less than 4 per cent of receipts for that year. Going forward, Malaysia would do well to consider reforming this mechanism so that, as in the Philippines for example, fines are paid into the national treasury (PCA, s 51). 23 PCA, s 12. 24 See http://​phcc​.gov​.ph/​wp​-content/​uploads/​2017/​09/​2017​-PCC​-RULES​-OF​-PROCEDURE​-1​ .pdf. 25 Its functions are enumerated in MCCA, s 16, III and IV. 26 MyCA, s 18(1), 18(2) respectively.

Competition law in Japan, Malaysia and the Philippines  259 these areas may be entirely appropriate, there are few guarantees that ministerial goals will always be free from interests driven by rent-seeking or other extraneous factors. The MyCC is required to liaise with sectoral regulators. A Special Committee on Competition meets periodically to discuss common interests and to find consistency in the application of competition provisions.27 The Committee includes the Communications and Multimedia Commission, the Land Public Transport Commission, the Central Bank, the Energy Commission, the National Water Services Commission, the Securities Commission and the Malaysian Aviation Commission. To some extent, the Committee gives the MyCC an opportunity to coordinate monitoring and action in sectors beyond its formal reach. It is submitted that more autonomy (and, to reiterate, more resources) should be granted to the agency. It would also be advisable to put distance between the MyCC and the MDTCA Ministry, without sacrificing appropriate communication and coordination enabling the MyCC to carry out its important advocacy functions,28 and to reduce the number of commissioners representing the government to zero. In any event, the MyCC should be elevated to a rank equivalent to that of a ministry. Finally, responsibility for oversight of the MyCC should be reassigned to the Parliament. Such steps would fortify the agency and increase its access to resources while enhancing the regime’s credibility. 2 Other non-judicial agencies In Malaysia, other regulators with competition powers are shown in Table 11.2. A memorandum of understanding (MOU) between the MyCC and the Central Bank of Malaysia entered into on 5 June 2014 has clarified areas of consultation and cooperation with regard to competition issues affecting the financial sector. In both Japan and the Philippines, while other economic regulators interact with the competition authority they have no direct enforcement role in competition matters. In the Philippines, the Office for Competition of the Department of Justice (DOJ) may undertake preliminary investigations and – with the aid of sister departments within the DOJ – criminal prosecutions. Cooperation between the DOJ and the PCC is facilitated by an MOU of 2018. In addition, other authorities – such as the Energy Regulatory Commission, the Communications and Multimedia Commission, the Insurance Commission and the Philippine Ports Authority – play Table 11.2

Malaysian regulators with competition powers

Agency

Role

Malaysia Communication and

Enforcement of competition-related provisions under Malaysia’s Communications and

Multimedia Agency

Multimedia Act (Act 588)

Energy Commission

Enforcement of competition-related provision under Malaysia’s Energy Commission Act (Act 610)

Malaysian Aviation Commission

Application of competition rules under the Malaysian Aviation Commission Act 2015 (Act 771), particularly Part VII (Competition)

MyCA, s 39. For example, the MDTCA’s current participation on the MyCC’s Advocacy and Communication Committee (also composed of MyCC members, as well as representatives from a consumer association and the media) is a logical mechanism that permits, inter alia, the timely flow of information to the MyCC about new government initiatives that may affect markets. A grant of greater autonomy to the MyCC would not necessitate the abandonment of this communication channel. 27 28

260  Research handbook on methods and models of competition law a relevant role in technical and economic regulation. In Japan, various ministries have worked closely with the JFTC to develop principles governing application of the AMA in some sectors. In the fields of energy (the Ministry of Economy, Trade and Industry) and telecommunications (the Ministry of Internal Affairs and Communications), this has led to joint guidelines. 3 Judicial enforcement In all three countries, there is provision for appeal of administrative decisions made by the competition authority. In Malaysia, the Competition Appeal Tribunal (CAT) hears appeals against MyCC decisions at first instance.29 The CAT may confirm, vary or set aside the Commission’s decisions.30 It can also cancel or modify financial penalties and adopt other appropriate relief. When reviewing a decision, the CAT sits as a tribunal of at least three judges and rules by majority. Its President is a judge of the High Court. At the end of 2018, the CAT had 11 members; a maximum of 20 may be appointed, each serving for up to six years. A party wishing to challenge a decision of the CAT may lodge an application for judicial review before the High Court, with further possibility of appeals to the Court of Appeal and the Federal Court. The High Court may also adjudicate private claims.31 Any person harmed directly as a result of an infringement of the section 4 or section 10 prohibitions may file a follow-on or stand-alone action before the High Court, again subject to possible further appeals. In Japan, as in the Philippines, administrative challenges to a decision of the JFTC or the PCC go direct to court. In Japan, the court may rule on criminal cases; decide private cases under either the AMA or Japan’s Civil Code (see below); and hear shareholders’ derivative suits under the Companies Act 2005. With regard to administrative appeals, following amendments to the AMA in 2013, the Tokyo District Court has exclusive jurisdiction (under Article 85 of the AMA) to review JFTC acts such as cease-and-desist orders and surcharge orders, subject to subsequent review by the Tokyo High Court and, ultimately, the Supreme Court.32 All such actions are governed by Chapter IX of the AMA (Articles 77 et seq), as well as the Administrative Case Litigation Act (ACLA). The legislator has scrapped the ‘substantial evidence rule’, whereby the High Court had to accept the JFTC’s reasonable findings of fact; appellants are now able to introduce new evidence before the District Court. These reforms might lead to enhanced judicial review, which under administrative law principles has traditionally been limited to an abuse-of-discretion standard.33 In the now-abolished hearing procedure, the JFTC’s self-scrutiny was not limited by that standard and it thus had wider scope to correct its own mistakes, but in practice it rarely reversed itself.34 Closer scrutiny of the JFTC’s decisions might lead to more rigorous investigations in terms of fact-finding, procedure and

29 The CAT is governed by Part V of the MyCA. See the CAT webpage at http://​trp​.kpdnkk​.gov​.my/​ tribunalpersaingan/​index​.php​?lang​=​en. 30 MyCA, ss 44, 51. 31 MyCA, s 64(1). 32 As of March 2017, five challenges to JFTC orders in cartel and bid-rigging cases had been brought before the Tokyo District Court. See JFTC, Annual Report (2016) 90–91 (in Japanese), www​.jftc​.go​.jp/​ soshiki/​nenpou/​index​.files/​h28nennpou​.pdf. 33 See ACLA, Art 30: ‘The court may revoke an original administrative disposition made by an administrative agency at its discretion only in cases where the disposition has been made beyond the bounds of the agency’s discretionary power or through an abuse of such power.’ 34 Kameoka (n 11), 120 (citing the JASRAC case).

Competition law in Japan, Malaysia and the Philippines  261 analysis.35 However, the impact of the reforms remains to be seen. In particular, the courts will have to clarify which aspects of the JFTC’s decisions are discretionary and thus deserving of deference under the ACLA and which aspects are properly subject to greater scrutiny. In the Philippines, appeals against decisions of the PCC are heard by the Court of Appeals, with further appeal to the Supreme Court.36 Where a criminal case is brought by the DOJ, or if a civil claim for direct injury is filed under section 45 of the PCA, section 44 assigns jurisdiction to the relevant Regional Trial Court. The Regional Trial Court also has jurisdiction if a (stand-alone) claim is brought under Article 28 of the Civil Code.

Box 11.1 APPEAL PROCESS IN THE PHILIPPINES: AN EXAMPLE FROM THE TELECOMS SECTOR An early high-profile merger case in the telecoms sector concerned the joint acquisition of VTI by the two major incumbents, PLDT and Globe Telecom. When the parties challenged the PCC’s jurisdiction, the Court of Appeals deemed the acquisition approved because PLDT and Globe had substantially complied with the simple notice requirements under the PCC’s interim rules that applied in the period prior to 18 June 2016, the effective date of the Implementing Rules and Regulations (IRRs). Under the interim rules, the PCC could only check that the acquiring companies had provided the required information; it had no jurisdiction to carry out a substantive review. The PCC had thus abused its discretion by intervening. The outcome of the case was bad for the PCC and for the Philippines, since, in effect, a duopoly succeeded in eliminating a competitive threat. However, future notifiable transactions will be subject to full review under the PCA and the IRRs. Furthermore, a Sino-Filipino telecoms consortium, Mislatel, was granted a licence in November 2018 to enter the market and challenge the duopoly. It is too soon to say how this new entry will affect market conditions. Both Japan and the Philippines have a provision for criminal law sanctions for competition law breaches; Malaysia does not.37 In the Philippines, and for practical purposes in Japan, criminal fines and custodial sanctions can be imposed only in cases involving cartels and bid-rigging.38 Exercise of this power in Japan resides with the District Courts. These courts have been reluctant to send natural persons to jail. Instead, the two dozen-plus fully prosecuted cases (about

35 Mel Marquis and Shingo Seryo, ‘The 2013 Amendments to Japan’s Anti-Monopoly Act’, Competition Policy International – Asia Column (2014) 8, https://​papers​.ssrn​.com/​sol3/​papers​.cfm​ ?abstract​_id​=​2508199. 36 PCA, s 39. 37 In Malaysia, certain misconduct incidental to collusion in procurement (for example, bribery-related conduct) may trigger individual criminal liability under other penal provisions, such as those of the Anti-Corruption Act 1997 or the Penal Code, but liability under such provisions does not depend on the anticompetitive nature of the activity. 38 For example, while Article 89 of the AMA provides for criminal sanctions for breach of the prohibition against private monopolization, this may be seen as a vestige of Japan’s early transplantation of certain provisions of the Sherman Act, which technically made monopolization offences criminal. As in the US (and the rest of the world), the offence of private monopolization is never prosecuted in Japan as a criminal matter.

262  Research handbook on methods and models of competition law one per year) have resulted in suspended sentences. The stigma attached to cartel conduct in Japan still seems insufficient to achieve effective deterrence. With regard to procedure, the JFTC, having investigated alleged cartel conduct, may refer the case as a criminal matter to the Public Prosecutors office (PPO). However, the JFTC cannot require the PPO to prosecute, and the two agencies have not always agreed on their approach to particular cases. The JFTC does not have exclusive authority in relation to bid-rigging as the Ministry of Justice can prosecute bid-rigging under the Criminal Code and on the basis of rules against fraudulent interference with bids. In the Philippines, the PCC may file a criminal complaint with the DOJ, which triggers a preliminary investigation by the Office for Competition that may ultimately lead the DOJ to prosecute the case before the Philippine courts. Private actions for damages are available in all three countries. In Malaysia, any person who suffers loss or damage as a result of anticompetitive conduct can sue for damages. This is not limited to situations where the MyCC has made a finding of infringement, and so stand-alone actions or follow-on actions may be brought. In the Philippines, independent civil action for damages caused by anticompetitive conduct is prohibited until the PCC has terminated its preliminary inquiry into the conduct. Perhaps unsurprisingly, given Japan’s long history with competition law, private claims in competition disputes are far more significant in Japan than in either of the other two countries; however, even in Japan private enforcement does not effectively secure redress for cartel victims.

Box 11.2 PRIVATE ENFORCEMENT IN JAPAN More than 200 private claims have been filed in Japan since the mid-1990s.39 Only opt-in class actions are available.40 While governmental authorities have often claimed damages in bid-rigging cases, cartel victims suffering diffuse harm are seldom compensated.41 Nevertheless, the AMA42 provides a basis for no-fault civil liability (single damages) in follow-on damages cases brought before the Tokyo District Court by a direct or indirect purchaser. It is not necessary to show negligence or intent,43 but causation and damages must be established. If damages are ‘extremely difficult’ to prove, Article 248 of Japan’s procedural code permits judges to estimate the quantum. In practice, such estimates tend to be conservative.44 A claimant may seek a prohibitory injunction, but only in relation to unfair trade practices and only if the claimant demonstrates actual or likely ‘significant damage’ (ichijirushi songai).45 Antimonopoly tort cases are not subject to the above restrictions. In particular, a claimant: (i) may file a case before any District Court; (ii) need not wait for a final JFTC order; and (iii) if seeking an injunction, may ask for a cessation of any violation of the AMA, not

39 Simon Vande Walle and Tadashi Shiraishi, ‘Competition Law in Japan’ in John Duns, Arlen Duke and Brendan Sweeney (eds), Comparative Competition Law (Edward Elgar Publishing 2015) 414, 436. 40 Simon Vande Walle, Private Antitrust Litigation in the European Union and Japan (Maklu 2013) 63–4. 41 Ibid, 143. 42 AMA, Art 25 read with Arts 26 and 85-2. 43 AMA, Art 25(2). 44 Vande Walle (n 40) 71. 45 AMA, Art 24.

Competition law in Japan, Malaysia and the Philippines  263 just unfair trade practices. Four particular provisions in Chapter V of the Civil Code are relevant here. First, under Article 709, any person who has infringed the rights of others by violating the AMA is liable for damages. Second, Article 703 permits actions for unjust enrichment (futō ritoku). Third, Article 719(1) provides for joint and several liability. Fourth, prescription under Article 724 is either three years from the time a plaintiff becomes aware of the damage or 20 years from the time of the wrongful act, whichever comes first. By comparison, Article 26(2) of the AMA gives plaintiffs three years to bring a case once the JFTC’s cease-and-desist or payment order becomes final. 4 Penalties Contravention of competition law generally results in the levying of a fine on the company and/or individuals concerned. Typically, the aim is deterrence of re-offending and offending by others, rather than punishment. Under Japan’s AMA, there are two types of sanctions for contraventions – administrative sanctions and criminal sanctions – although, as noted above, the latter rarely lead to actual prison time. Administrative sanctions consist of cease-and-desist orders and surcharge payment orders (that is, fines). Differentiated surcharges apply according to the type of conduct – with limited possibilities of adjustment.46 For a serious infringement (cartels or ‘control-type’ monopolization), the base rate is 10 per cent of relevant domestic sales made during the infringement, subject to a ceiling of three years.47 For monopolization by exclusion, the rate is 6 per cent. Most unfair trade practices only trigger a surcharge (generally 3 per cent) if there is a repeated infringement. The varying rates are all currently subject to the three-year cap. The rate will be increased where: (i) the enterprise has been the subject of a surcharge within the past 10 years (15 per cent of sales for manufacturers); (ii) the enterprise has played a ‘major role’ in a cartel within the last 10 years (again, 15 per cent of sales for manufacturers); or (iii) the enterprise belongs to both categories just described (20 per cent for manufacturers). Furthermore, the surcharge will be decreased by 20 per cent if each of the following conditions is satisfied: (i) the enterprise stops infringing; (ii) it is neither a recidivist nor a ‘major role’ cartelist; and (iii) it has infringed for less than two years. Significant changes to Japan’s surcharge system are likely in response to the findings of a Study Group in 2016 and 2017. This will provide the JFTC with discretion to reduce a surcharge to reward cooperation. Proposed changes which may be accepted by the Diet include a relaxation of the three-year retroactivity cap; the adoption of a uniform surcharge rate; and the replacement of the concept of ‘sales of relevant goods or services’ with a broader concept of ‘basic amount of sales’, which could lead to augmented surcharges.48 These amendments will make the surcharges more flexible. In Malaysia, the MyCC can issue cease-and-desist orders for anticompetitive agreements and abuse of dominance, and it can also impose pecuniary penalties.49 The maximum penalty

AMA, Arts 7, 20. This is reduced to 4 per cent for SMEs and rates of, respectively, 3 per cent and 2 per cent for retailers and wholesalers, reduced to 1.2 or 1.0 per cent if a retailer or wholesaler is an SME. 48 See further Etsuko Kameoka and Mel Marquis, ‘Competition Law in Japan: An Overview of Developments from 2016 to Mid-2017’ Competition Policy International – Asia Column (November 2017) 1, 4–5, https://​www​.competitionpo​licyinternational​.com. 49 MyCA, s 40(1)(c), read together with MCCA, s 17(2)(b). 46 47

264  Research handbook on methods and models of competition law is 10 per cent of worldwide turnover during the entire period of infringement – potentially as far back as 2012. In the Philippines, the PCC can impose a fine of up to 100 million pesos (around US$2 million) for breaches of the PCA.50 For a repeat offender, the fine must be between 100 million pesos and 250 million pesos (between US$2 million and US$5 million). Fines are trebled where the violation involves basic necessities or prime commodities.51 These maxima seem too low and should be reconsidered by the Philippine legislature. As explained earlier, criminal sanctions may be applied in cartel cases in both Japan and the Philippines. For serious infringements, Japanese courts can hand down prison terms of up to five years (in practice, suspended sentences) and fines of up to 5 million yen (US$45,000) on natural persons.52 For an enterprise, the maximum criminal fine is 500 million yen (US$4.5 million)53 and reductions mitigate combined sanctions when a surcharge has already been levied. If it is found that a corporate representative knew of an unlawful plan but failed to act (or if the representative or executive was aware of a violation but failed to rectify it), that person may incur a criminal fine of up to 5 million yen.54 However, criminal prosecution is only possible if the JFTC files an ‘accusation’ (kokuhatsu).55 Between 2001 and 2015, criminal sanctions were imposed on 68 individuals (no time served) and on 74 enterprises.56 Since the JFTC and PPO may assess cases differently, coordination between the two is essential. As the JFTC states in its Policy on Criminal Accusation and Compulsory Investigation, it will only refer cases that involve either recidivism or vicious conduct that has a widespread impact on the livelihood of the people.57 Bid-rigging cases meet the latter criterion with some regularity. In the Philippines, criminal penalties are applied for price-fixing, bid-rigging, market-sharing and output-restriction cartels.58 Prison terms of two to seven years may be imposed, plus a fine of 50–250 million pesos. It is doubtful that these sanctions make up for the deficiency of low civil fines.

III METHODOLOGY Each of the three competition commissions has issued guidelines explaining various aspects of competition law and their approach to its enforcement. Guidelines issued by the MyCC, for example, cover subjects such as Complaints Procedures, Market Definition, Anticompetitive Agreements, Abuse of a Dominant Position, Financial Penalties and Leniency. The educative PCA, ss 29, 30. See also rule VI of the PCC’s Rules of Procedure (n 24). PCA, s 41. 52 AMA, Arts 89 et seq. 53 AMA, Art 95. 54 AMA, Art 95-2, 95-3. 55 Akira Inoue, Antitrust Enforcement in Japan (Wolters Kluwer 2012) 282–91. 56 Japan (2016), ‘Contribution to the OECD Global Forum’ (DAF/COMP/GF/WD(2016)24, 16 November 2016) 5 fn 3. 57 See https://​www​.jftc​.go​.jp/​en/​policy​_enforcement/​cartels​_bidriggings/​anti​_cartel​_files/​ 2015policy​_on​_criminalaccusation​.pdf. 58 PCA, s 30. In the case of bid-rigging, Rule XXI of the 2016 revised IRRs for the application of the Government Procurement Reform Act 2003 provides an additional basis for criminal liability. Rule XXII allows the government to file civil claims. Rule XXIII adds the possibility of suspension orders for one or two years. Additional sanctions apply to complicit public officials. 50 51

Competition law in Japan, Malaysia and the Philippines  265 role of such guidelines – for the public and the commercial sector, but also for new agency staff – is especially important in the period after a competition law has been passed. In addition, the commissions may undertake market studies and sector surveys. Such initiatives are routine in Japan, for example, where the JFTC has carried out more than 30 surveys in the last 15 years. These instruments are important in building a nuanced understanding of the structure of particular markets and in identifying distortions that may require enforcement and/or advocacy efforts. Achieving in-depth sectoral knowledge through market studies is particularly important in Malaysia, where, as noted above, there is no provision in the MyCA establishing ex ante merger review. As Lee observes, ‘[a]dvocacy-related activities are clearly important to create awareness amongst the various stakeholders such as consumers, NGOs, industry associations and the media’.59 A

Initiation of Investigation

In each of the three countries, the competition commission may undertake an investigation if it suspects that conduct is anticompetitive or in response to a complaint. However, in Malaysia, the commission may be directed by a Minister to undertake an investigation. The commissions in Japan and the Philippines are under no such obligation. In the Philippines, the PCA indicates that the PCC opens investigations upon (inter alia) referrals from regulatory agencies, but the PCC reserves the option of denying ‘due course’ to such a referral in specific circumstances.60 Likewise, in Japan, neither a ministry nor a sectoral regulator can compel the JFTC to undertake an investigation. The JFTC can conduct administrative investigations or, in cases that may lead to criminal sanctions, compulsory investigations. In relation to the former, the Commission can undertake on-site inspections and can seize materials61 and issue appearance orders, interrogate parties and issue orders.62 In compulsory investigations, under a search warrant, the JFTC may enter private premises and seize material (including postal items) as evidence of misconduct.63 In Malaysia, MyCC investigating officers are vested with all the powers of a police officer in relation to a police investigation. The MyCA confers power on the Commission to require the provision of information, to retain documents and to access records. Search and seizure require a warrant unless the time needed to acquire it may adversely affect the investigation or its outcome. Similarly, the PCC has broad investigative powers under the PCA. The PCC’s information gathering is supported by the power to issue a subpoena, and it may apply to the court for an inspection order. The PCC’s powers when conducting inspections (dawn raids) are governed by Supreme Court Resolution A.M. No. 19-08-06-SC of 10 September 2019.

Lee (n 18) 80. PCC Rules of Procedure (n 24), Rule II, Art I, s 2.3. Where a referral is denied due course, the PCC must give notice to the regulator concerned within ten days of denial (as would be the case if the PCC chose not to open an investigation based on a verified complaint). 61 Where this occurs on private premises, parties can refuse to accept investigation. 62 AMA, Art 47. 63 AMA, Arts 101–103. 59 60

266  Research handbook on methods and models of competition law

Box 11.3 THE CAMERON HIGHLANDS FLORICULTURE ASSOCIATION AND PRICE-FIXING The investigative process in Malaysia is illustrated by a matter involving the Cameron Highlands Floriculturist Association (CHFA). Following a media report about an industry association meeting on 4 March 2012, at which members agreed to increase the price of cut flowers by 10 per cent, the MyCC began a formal investigation. The Commission notified the Association of the violation by letter. In response, the Association explained that prices had not increased since 2008. As the conduct was price-fixing, it was per se illegal. The minutes of the Association’s meeting were used as the main evidence for the infringement. In its final decision of 6 December 2012, the MyCC ordered the CHFA to cease and desist from fixing prices for flowers, undertake that its members would refrain from any anticompetitive practices in the relevant market, and publicize these remedial measures in mainstream newspapers.64 B

Time Limits

Investigations may be subject to time limits, either legislative or self-imposed. Various motivations may drive such time limits, such as efficiency or legal certainty objectives, a desire to adjust the speed of enforcement in response to increasingly fast-moving markets, or a combination of rationales. In the Philippines, the legislature decided to formalize such time limits. Specifically, within a 90-day window, and after considering the evidence it uncovers through its fact-finding or preliminary inquiry, the PCC must either issue a resolution closing the case if it finds no contravention, or proceed with a full administrative investigation if it has reasonable grounds to do so.65 Whereas the applicability of time limits to antitrust investigations varies by jurisdiction, in the field of merger assessment they are indispensable since proposed mergers are particularly time sensitive. In this regard, Japan and the Philippines are broadly similar. In Japan, the JFTC must review a notified transaction within 30 days, during which time the parties may not implement the merger. In complex transactions, the JFTC may request ‘Reports’ containing additional information. Once it accepts all Reports, the agency has another 90 days to conduct a Phase II investigation (a ‘secondary review’). Likewise, the Philippines has a ‘Phase I’ of 30 days, during which parties to a notified merger cannot close their deal.66 If necessary, Phase II takes 60 days from the request for more information. If the parties fail to notify, the operation is void and fines apply. C

Assessment in Cartel Cases: Technically, a ‘Rule of Reason’ Persists in Japan

Unsurprisingly, in the Philippines and Malaysia, cartel conduct is deemed to be anticompetitive once an agreement has been established, although the many sectoral exclusions limiting

For discussion of the case, see Lee (n 18) 81–2. PCC Rules of Procedure (n 24), Rule II, Arts I, II. 66 The requirements regarding determination of control for acquisitions, the notification procedure, deadlines and other details are all provided in the IRRs. 64 65

Competition law in Japan, Malaysia and the Philippines  267 the scope of application of the MyCA in Malaysia remain an issue. Japan, however, is formally an outlier compared to these other two countries but also relative to the wider world, since, as noted earlier, the cartel prohibition is contravened only if market power is created or strengthened. In practice, the JFTC has dealt with this anomaly by defining relatively narrow markets for the cartelized product.67 Although its ability to fight cartels has not been seriously compromised, it is submitted that an explicit per se rule for unambiguous cartels would assist enforcement and help to achieve greater coherence in Japan’s AMA. D Leniency In all three jurisdictions, the competition commissions may grant leniency for cartel conduct, although under slightly different conditions.68 In Japan, at the time of writing, leniency is available on a sliding scale for up to five applicants prior to the commencement of an investigation. Japan’s Leniency Program took effect on 4 January 2006 and by March 2017 the JFTC had received more than a thousand applications.69 However, questions remain about the Program’s effectiveness.70 In Malaysia, immunity is available to an applicant that admits involvement in a cartel unknown to the MyCC and who provides information or other cooperation that significantly assists the agency in identifying or investigating the infringement.71 To qualify for leniency in the Philippines, an applicant must have acted promptly to terminate its participation in the illegal activity, it must provide full cooperation, and it must not have coerced another party or functioned as the leader or originator of the misconduct. The details of the PCC’s Leniency Program were published in December 2018.72 Assuming that the Supreme Court approves the PCC’s proposed Rules on Inspection Orders, the new Leniency Program signals that cartel enforcement may ramp up soon. The Program allows individuals to apply independently of organizations, and the PCC may refer an applicant to the DOJ’s Witness Protection programme. The Leniency Program sets rules on matters such as markers, qualifying criteria, the perfecting of conditional leniency, and the extent to which leniency for an organization protects its employees. Notably, only two applicants may qualify for, respectively, immunity and a fine reduction. None of the three countries provides for immunity from private litigation or other protection from private claims, such as removal of joint and several liability.

Takigawa (n 11), 26–7. For Japan, see AMA, Arts 7-2(10) et seq, as supplemented by the JFTC’s Rules on Immunity or Reduction of Surcharges, http://​www​.jftc​.go​.jp/​en/​legislation​_gls/​antimonopoly​_rules​.files/​immunity​ .pdf. For Malaysia, see MyCA, s 41 and the MyCC 2014 Leniency Guidelines, http://​www​.mycc​.gov​ .my/​sites/​default/​files/​handbook/​MyCC​_Guideline​-on​-Leniency​-Regime​.pdf. For the Philippines, see PCA, s 35 and the PCA 2018 Leniency Program, https://​phcc​.gov​.ph/​rules​-leniency​-program. 69 Toshiyuki Nambu, ‘A Successful Story: Leniency and (International) Cartel Enforcement in Japan’ (2014) 5 Journal of European Competition Law & Practice 158 (underlining the positive results of the Leniency Program and identifying several explanatory factors for this outcome). 70 Steven Van Uytsel, ‘Anti-Cartel Enforcement in Japan: Does Leniency Make the Difference?’ in Caron Beaton-Wells and Christopher Tran (eds), Anti-Cartel Enforcement in a Contemporary Age: Leniency Religion (Hart 2015) 81. 71 See further MyCC, Leniency Guidelines (n 68) (describing conditions of eligibility and benefits, the granting of markers, revocation of leniency, etc). 72 Cited above (n 68). 67 68

268  Research handbook on methods and models of competition law E

Agreements and Mergers

For non-cartel conduct, competition authorities must not only establish the anticompetitive effects of the conduct, they must also show that certain legal thresholds have been satisfied. For example, in the case of mergers, an authority must establish that the merger falls outside of any applicable safe harbour provision. The same may be true in the field of restrictive agreements. Thus, in Malaysia, the MyCC Guidelines73 establish a ‘soft’ safe harbour for non-hard-core horizontal agreements where combined market share does not exceed 20 per cent (s 3.4 of the Guidelines). For agreements between non-competitors, the soft safe harbour will apply if all the parties have a market share of less than 25 per cent on all relevant markets. However, the MyCC may intervene if an agreement below those thresholds has significant anticompetitive effects. No safe harbour applies to minimum RPM (s 3.14). In Japan, RPM, as well as non-price and internet-related vertical restraints, for example, are evaluated in line with the JFTC’s Distribution Guidelines of June 2017.74 Under these Guidelines, most vertical restraints other than RPM benefit from a safe harbour if the seller’s market share does not exceed 20 per cent.

Box 11.4 THE UBER-GRAB MERGER IN SOUTHEAST ASIA In March 2018, Grab acquired Uber’s entire Southeast Asia business in exchange for a 27.5 per cent stake in its operations in the region. This merger provides an illustration of the different methodologies adopted by the Philippines and Malaysian competition authorities – Japan was not involved. Prior to merging, the parties were the major suppliers of e-hailing services. At the time of the merger, Uber had already ceased operating in the Philippines owing to the expiry of its licence. The merger was given a green light by the PCC, despite its view that Grab operates as a virtual monopolist.75 In order to obtain approval, the merged entity had to submit to service, quality and price controls to ensure fair treatment of consumers, and it had to commit that it would no longer require drivers to enter into exclusivity agreements.76 A third party will be appointed to monitor adherence to these undertakings. The aim was to replace the competitive constraint previously imposed on Grab by Uber’s presence on the market. In October 2018, the PCC fined the two companies a total of 16 million pesos (US$305,000) for previously failing to preserve key elements of the status quo, thus defying an interim order requiring them to do so.77 Lacking a merger provision, the MyCC in response to the Uber-Grab merger announced that it would monitor the market. This has been complemented by new regulations relating to ride-hailing services introduced in the interest of a level playing field. In addition, the

73 See http://​www​.mycc​.gov​.my/​sites/​default/​files/​handbook/​MYCC​-4​-Guidelines​-Booklet​ -BOOK1​-10​-FA​-copy​_chapter​-1​-prohibition​.pdf. 74 See http://​www​.jftc​.go​.jp/​en/​legislation​_gls/​imonopoly​_guidelines​.files/​170616​.pdf. 75 PCC Media Release (10 August 2018). The PCC acted motu proprio under rule 4, s 1 of the IRRs. 76 For a complete list of requirements, see Jon Russell, ‘Grab-Uber Deal Wins Philippines Approval but “Virtual Monopolist” Concern Remains’, TechCrunch (10 August 2018), https://​techcrunch​.com/​ 2018/​08/​10/​grab​-uber​-deal​-gets​-philippines​-approval. 77 PCC Resolution of 11 October 2018.

Competition law in Japan, Malaysia and the Philippines  269 Malaysian government has restricted the amount of commission to be passed on to the companies at 20 per cent, while the rate for taxi drivers using e-hailing applications will be a maximum of 10 per cent. Unlike most of the other countries affected by the merger, it appears that in Malaysia the response has been that existing competitors have expanded and new suppliers have entered the market. F Dominance Traditionally, monopolization cases in Japan have been uncommon. This is because the JFTC has habitually applied the unfair trade practices framework (see above) to examine conduct that could be characterized as monopolization. As a result, the JFTC benefits from a softer standard of proof: a ‘tendency to impede fair competition’. However, the agency is occasionally more willing to pursue monopolization cases, especially in liberalized sectors. In particular, the JFTC prioritizes a monopolization case if the contested conduct results in a market share exceeding 50 per cent, and if it impacts citizens’ lives. In NTT East,78 the incumbent had squeezed rivals who sought to compete on the market for high-speed internet access – specifically, by charging an access price that exceeded the price paid to the incumbent by its own customers. Resonating with EU case law, NTT East makes clear that margin squeeze is a ‘stand-alone’ infringement; application of the AMA is generally not precluded by sector-specific regulation. In Malaysia, the Commission generally considers a market share exceeding 60 per cent to be indicative of dominance. In the Philippines, section 27 of the PCA and rule 8 of the PCC’s IRRs set out the factors for assessing whether an entity is dominant; generally, a firm is regarded as dominant if it has a share of 50 per cent or more. The PCC has not yet published a guideline for abuse of dominance. Section 10(2) of the MyCC Guidelines provides a non-exhaustive list of abusive conduct, including unfair prices; limitation of production, market access or technological development; and refusal to supply.79 However, it should be recalled that such conduct may in effect be exempt due to the MyCA’s significant sectoral exclusions. Section 3.6 of the MyCC Guidelines states that ‘even if an enterprise is dominant, it should not be stopped from engaging in competitive conduct that benefits consumers even if inefficient competitors are harmed’. Further, the prohibition does not apply where a dominant enterprise acts with a ‘reasonable commercial justification’.

Box 11.5 APPROACH TO DOMINANCE IN MALAYSIA: EXAMPLES In 2013, the MyCC opened a margin squeeze case against Megasteel.80 It then reversed course entirely in 2016. Megasteel was vertically integrated and dominant in hot rolled coil Supreme Court of Japan, 17 December 2010, Heisei 21 (gyō-hi) No. 348, 57(2) Shinketsushū 215. The MyCC’s approach to particular practices is explained in sections 3.11–3.31 of the Guidelines. 80 Case No MyCC/002/2012, Megasteel, MyCC Decision of 15 April 2016. For conduct characterized by the MyCC as discrimination under the MyCA, s 10(2)(d), see Case MyCC (ED) 700-1/1/2015, My E.G. Services, MyCC Decision of 24 June 2016 (confirmed by the CAT in its decision of 28 December 2017 and at the time of writing pending further appeal), http://​www​.mycc​.gov​.my/​sites/​default/​files/​ Section​-40​-Notice​-of​-Finding​-of​-an​-Infringement​-by​-My​-Services​-Berhad​.pdf. Fournier suggests that My E.G. was more accurately a tying case and criticizes the lax approach to market definition and market 78 79

270  Research handbook on methods and models of competition law (HRC), which is an essential input for cold rolled coil (CRC). Melewar, the complainant and Megasteel’s competitor – through its subsidiary Mycron – on the downstream CRC market, claimed that Megasteel had charged a high price for the essential input while often undercutting Mycron’s price for CRC. The MyCC in its final decision highlighted several government-induced distortions in the form of tariffs and bureaucratic barriers, and a general policy of protection of the local champion. According to the MyCC’s brief explanation, while Mycron in 2012 had a downstream share of 47 per cent, Megasteel’s share was only 6 per cent. But the decisive factor was that Megasteel’s price for CRC had in fact exceeded that of Mycron.81 The MyCC did not hesitate to enforce the abuse of dominance provision in a more recent case where the dominant firm held exclusive rights. In 2018, after investigating the conduct of Dagang Net Technologies in the provision (as National Single Window) of trade facilitation services, the MyCC found provisionally that Dagang Net had abused its dominant position by refusing to supply new or additional electronic mailboxes to end-users of its customs-related software that used front-end software sold by suppliers not authorized by Dagang Net, and by imposing exclusivity clauses.82 G

Sanctions, Undertakings and Remedies

Once an investigation is complete, in all three countries the competition commission makes an administrative decision concerning the illegality of the conduct. Then, especially in Malaysia and the Philippines, it must be determined whether the conduct is exempt. For example, in Malaysia, parties to an anticompetitive agreement are not liable if there are significant identifiable technological, efficiency or social benefits from the agreement, not reasonably available without the anticompetitive effects from the agreement, and the detrimental effect is proportional to the benefits provided by the agreement – the onus of proof lies with the parties. If the conduct is not exempt, the commission must decide what sanction to apply.83 In each of the three jurisdictions, the competition commission has a range of instruments available without imposing pecuniary penalties. Under the Comprehensive and Progressive Trans-Pacific Partnership, effective as from 30 December 2018, each contracting party agrees to establish a procedure for accepting commitments in non-cartel cases. Such a procedure will facilitate early resolution of cases where parties subject to an investigation come forward with proposals designed to restore effective competition, which then become binding if accepted by the authority. On this basis, Japan has in fact adopted a commitment procedure.84 In some sense, this move resurrects the old consent decision procedure that applied prior to 2005, power in that case. See Knut Fournier, ‘The MyEG Case’ (2018) 39 European Competition Law Review 366. 81 MyCC, Megasteel (n 80), [47]–[50]. 82 MyCC Media Release (11 July 2018). 83 In Japan and the Philippines, if a cartel case is prosecuted in criminal proceedings, only the courts can impose sanctions. 84 JFTC Media Release (26 September 2018) (in Japanese) (describing this procedure, known as kakuyaku tetsuzuki). An English translation of the Policies Concerning Commitment Procedures is available at https://​www​.jftc​.go​.jp/​en/​legislation​_gls/​antimonopoly​_rules​ _files/​policies​_concerning​_commitment​_procedures​.pdf​?fbclid​=​IwAR1OnlpzPrEJOnT6PbME0Dd​​ _25uw1vQioidM2nnTH7M76pJAtCpPzyQlwpU.

Competition law in Japan, Malaysia and the Philippines  271 except that the new mechanism enables the JFTC to close a case without imposing surcharges. Even prior to the introduction of a formal framework, the JFTC was already prepared to accept commitments to resolve an investigation.85 In Malaysia, the MyCC may likewise accept undertakings, and may subject them to conditions, in order to settle a matter (MyCA, s 43(1)). It may then close the file without a finding of infringement and without imposing a penalty (MyCA, s 43(2)). Once accepted, an undertaking must be ‘available for inspection by the public in a manner determined by the Commission’ (MyCA, s 43(3)). Sections 43(4), 40 and 42 of the MyCA together indicate that the MyCC may enforce any undertaking it has made binding by bringing an action before the High Court. By the end of 2018, the MyCC had resolved six cases by way of undertakings. In the Philippines, an investigation by the PCC can be closed by non-adversarial remedies – a binding ruling, an order to show cause, or a consent order. Section 37 of the PCA allows the PCC to employ certain non-adversarial administrative remedies. For example, an entity in doubt as to whether a contemplated act or agreement is lawful may apply for a binding ruling in its favour and a successful application will not be subject to administrative, civil or criminal action. To obtain a consent order, an investigated entity may propose terms and conditions to resolve the case. These must include, inter alia, the payment of damages to any injured private party. Alternatively, under section 28, the PCC can decide not to apply the provisions of the Act or the IRRs for a limited time if it determines that enforcement is unnecessary to attain the objectives of the Act, and if forbearance is consistent with the public interest and consumer welfare. In some respects, an order under section 28 may resemble an individual or ‘block’ exemption. More details will emerge in 2019, when the PCC issues its anticipated Rules on Forbearance. In relation to pecuniary penalties, at the date of writing, the Japanese system is non-discretionary: once the JFTC has determined that a violation has occurred, it must impose a fine and the rate is predetermined. This contrasts with Malaysia, where Fining Guidelines explain the methodology for determining the amounts of fines.86 In particular, the MyCC considers the seriousness and impact of the infringement, the degree of fault, the role the enterprise played in the infringement, and any recidivism. Typical aggravating and mitigating factors may apply. The sanctions imposed by the MyCC have included fines of 248,000 ringgit (US$60,000) in the Sibu Confection and Bakery price-fixing case; 20 million ringgit (US$4.8 million) in the MAS/AirAsia market-sharing case (that is, 10 million ringgit for each enterprise); and 33 million ringgit (US$8 million) in the Tuition and Day Care price-fixing case. Moreover, in February 2017, the MyCC adopted a proposed decision involving the insurance industry, which would impose a combined fine (on 22 insurance firms and their association, PIAM) of 213 million ringgit (US$51 million). In its 2014 Leniency Guidelines, the MyCC stated that it intends to impose a ‘high financial penalty’ for cartel conduct and so future cartel investigations may lead to substantial fines, given the MyCC’s ability to use the number of years of infringement as a multiplier. Finally, as the Philippines only launched non-merger enforcement in the summer of 2017, the PCC has not yet had the occasion to impose fines to sanction anticompetitive behaviour. This will likely change soon as at the time of writing the PCC is 85 JFTC Media Release (1 June 2017) (undertaking by Amazon Japan to eliminate MFN-type clauses from its contracts). 86 MyCC Fining Guidelines, ss 3.1–3.5, http://​www​.mycc​.gov​.my/​sites/​default/​files/​handbook/​ Guildline​-on​-Financial​-Penalties​.pdf.

272  Research handbook on methods and models of competition law conducting several antitrust investigations. However, the above-described rigid and low limits imposed on fines by the PCA may prove to be an obstacle to effective deterrence, at least with regard to enterprises with substantial financial resources.

IV CONCLUSION This chapter has reviewed competition regimes of varying ‘ages’ and varying stages of economic development. Japan obviously has the oldest regime, with seven decades of experience. Originally influenced by US antitrust law, the Japanese system is generally converging towards the EU model and global standards – especially in relation to investigative procedures, early case resolution and discretionary fining powers. However, it retains Japanese characteristics, in particular its concern about fairness. With regard to procedure, Japan’s commitment to robust due process protections is still evolving. Overall, Japan illustrates how worldwide convergence is a qualified concept that is more meaningful when one speaks of convergence of particular dimensions, standards and criteria, as opposed to convergence as a general or homogenized process. Malaysia and the Philippines fall within the broad category of developing (and ‘tiger cub’) economies. For a developing economy, Malaysia is relatively high performance and has less economic inequality, while the Philippines has been enjoying impressive and sustained growth for several years but still faces numerous challenges typical of the developing world. The experiences of these countries in establishing and building competition law systems differ from that of Japan, and they differ from each other. It is submitted that the Malaysian competition law system has significant shortcomings and requires a substantial legislative overhaul to address (i) problems of sector exclusions, fragmented competences and incoherence; (ii) omission of ex ante, cross-sectoral merger control; and (iii) the insufficient independence (and appearance thereof) of the MyCC from the Malaysian government. The problem of unduly broad sector exclusions will likely be exacerbated as Malaysia liberalizes fields such as energy and telecoms. Reforms under consideration may address the need for merger control but may be undermined by those sector exclusions. Doubts as to the MyCC’s independence stem from, inter alia, the processes for appointing commissioners and the role of the Minister in requiring investigations and effectively setting the Commission’s enforcement priorities. Moreover, the MyCC is under-resourced – additional staff and money are required. With regard to both the breadth of agency competence and buffers against potential political pressure, Malaysia should follow the examples set by Japan, the Philippines and other jurisdictions. Nevertheless, engaging with political actors, participating as an active stakeholder in legislative and regulatory processes, and pursuing other opportunities for advocacy are all vital functions. The agency thus needs to be autonomous from, but also judiciously connected to, government processes – a delicate but not impossible balance. Despite its difficulties, the MyCC has accomplished much. Its willingness to enforce competition law against state-owned enterprises and privileged enterprises with exclusive rights helps to bolster the credibility of the regime, and its sustained advocacy and outreach activities have enhanced the business community’s general awareness of competition law obligations – particularly among SMEs. In the Philippines, the antitrust prohibitions only entered into force in August 2017. It is difficult to gauge the political commitment to credible enforcement and to developing a national competition culture. However, encouraging signs include a commitment to reviewing govern-

Competition law in Japan, Malaysia and the Philippines  273 ment policies and legislation to remove unnecessary anticompetitive measures. The PCA is fairly cohesive and coherent, and provision for the Commission’s independence is at least formally well designed. Under the PCA, the PCC has unambiguous jurisdiction over competition issues in regulated sectors. Early strong leadership of the PCC – with commissioners expert in law, economics and development policy – has positioned the PCC well. Of course, the agency faces challenges. These include staff recruitment and retention, as well as maximum fines that are fixed in the PCA at a level too low to deter deep-pocketed enterprises, including multinationals. The PCC is, however, blessed with a wide range of other enforcement tools to assist its efforts to change the behaviour and culture of commerce.

12. Building an efficient system of protection of competition in Serbia on its path to the EU Dragan Penezic and Zoran Soljaga

I INTRODUCTION Competition policy represents a very important instrument for the development of a modern and competitive market economy. The enforcement of a competition Act is only a part of competition policy, which assists transition and is the foundation of a future competitive environment. In a developing country in transition, such as Serbia, other instruments of competition policy (foreign trade and investment policy, privatization, public procurements, licensing and concessions, as well as liberalization) are also important for the development of an efficient economy. The introduction of competition policy in Serbia faced several challenges. First, state presence in the market was significant and the new policy was not welcomed by some political and business elites. Second, the level of knowledge and awareness of competition law and policy among relevant stakeholders was very low. Third, the regulatory regime was not procompetitive, and many regulations contained restrictions preventing market entry and limiting competition in the market. Protectionism and support for state-owned companies and national champions were characteristics of the economy. The shift towards a market economy and competition policy was primarily initiated by external pressure rather than internal need. In that sense, the EU and the process of EU integration played a key role, starting at the beginning of the new millennium. The Republic of Serbia made major steps towards joining the EU by signing the Stabilisation and Association Agreement with the EU in 2008 and starting the membership negotiation process in 2014. The view of the authors is that, despite the various challenges which the EU is currently facing, the EU integration process is a unique opportunity for Serbia to accelerate reforms and to continue with the alignment of its legal and economic systems to the needs of a competitive economy. Competition policy is one of the main pillars of the EU accession process. The implementation of an efficient competition law regime in a candidate country is a precondition for closing the negotiation and accession to the EU. The candidate country must establish a legal framework in line with the EU and must build an independent institution with adequate administrative capacity. Finally, it must achieve a credible enforcement record. The first modern Law on Protection of Competition was adopted in Serbia in 2005. That law has undergone several amendments to correct the shortcomings of the legal framework and better align with the EU competition law. Further amendment is still necessary to achieve EU standards. A brief history of earlier laws follows.

274

Building an efficient system of protection of competition in Serbia  275

II

INTRODUCING COMPETITION LAW IN THE REPUBLIC OF SERBIA

Competition law was introduced into the Republic of Serbia at the end of the 1990s, during the Federal Republic of Yugoslavia (which, at the time, encompassed Serbia and Montenegro). The first law relating exclusively to the protection of competition was the Federal Antitrust Law of 19961 (the 1996 Law). It prohibited the abuse of dominance and monopolistic agreements under the general principles enshrined in then Articles 81 and 82 of the Treaty of Rome. The 1996 Law established a Federal Anti-Monopoly Commission (AMC) for enforcement purposes. The AMC was not an independent entity but acted as a collegial management body within the Federal Ministry of Economy and Internal Trade. The 1996 Law contained sanctions, including fines for undertakings and their responsible persons, and prison sentences of between six months and five years for responsible persons. Additional penalties included the publication of offences, the confiscation of proceeds, and the imposition of bans on some business activities. The 1996 Law was criticized for many shortcomings, including that it did not regulate concentrations (mergers and acquisitions). The lack of merger control can be explained in two ways: the intention of the legislature not to sanction the activities aimed at obtaining a dominant position, in order to enable state-owned enterprises to grow and become internationally competitive; or simply a failure to acknowledge these types of measures in the EU law that served as a role model. The AMC became operational in January 1999. Initially, it mostly dealt with price control in key sectors. It did not analyse economic justification for the price increases but rather made decisions because prices had been increased above the average rise in prices on the domestic market. A position of dominance was assessed solely on the basis of production share and not on the basis of the sales participation rate, when actually many monopolists faced competition from imports.2 Each increase in retail prices above an average price increase was sanctioned by ordering the reduction in price to the level of the envisaged price growth. An emphasis was put on monitoring the prices of products that were vital for the standard of living of citizens. The competition policy of this period was characterized by a complete lack of supporting regulations, which meant that competition law was regulated only by the 1996 Law itself. Furthermore, the very essence of competition was not dealt with and the primary task was to combat inflation. This was partially due to the lack of experience in competition law. The 1996 Law did not yield any significant results, but it did indicate what might be done in the future and it provided the basis for future competition policy in Serbia.

Official Gazette of the Federal Republic of Yugoslavia, No 29/96. An additional problem was the authenticity of the data based on which the AMC assessed market shares. The AMC used only the data of the Federal Statistical Office, which recorded the data of only the socially owned enterprises and not the private ones, which means that the data were incomplete and also not current because they referred to the preceding year, at best. 1 2

276  Research handbook on methods and models of competition law

III

ADOPTION OF THE FIRST MODERN COMPETITION LAW IN 2005

Modern competition law in Serbia arrived when the Law on the Protection of Competition was passed in 2005 (the 2005 Law).3 The 2005 Law was mainly a result of Serbia’s progress in meeting the requirements of the EU integration process. Serbia needed to do its homework in order to start negotiating the Stabilisation and Association Agreement.4 The 2005 Law was not the result of a high level of awareness in Serbia of the need to have modern competition legislation, or a desire by decision-makers to put the market in order and implement competition and efficiency. Indeed, it is highly unlikely that the decision-makers were at all willing to establish an independent institution and vest significant powers in it. Given that the EU served as an external catalyst of reforms in Serbia at the time, and that the entire legislative activity focused on harmonizing Serbian legislation with the EU acquis, it seemed natural to take EU competition law as a model for shaping the Serbian competition system. This meant transposing the EU model without taking proper account of the specificities of the Serbian legal and economic system, and this approach was criticized by the academic community.5 Thus, the form of the 2005 Law was a result of Serbia’s commitment to EU integration, the fact that the system itself was not prepared for such legislation, and, to a certain extent, the lack of expertise of the drafters. The 2005 Law was a major step forward, but it had many deficiencies. The 2005 Law formulated a clearer competition policy for Serbia. The policy mainly focused on establishing a legal framework that would set out fundamental rules of operation and behaviour for market participants and establish an independent institution to ensure that the law was implemented. The 2005 Law incorporated provisions modelled on Articles 101 and 102 of the Treaty on the Functioning of the European Union (TFEU). It allowed market participants to file any agreement potentially containing restrictive clauses, referred to as a ‘prohibited agreement’, to the Commission for the Protection of Competition (CPC) for exemption. Even though the EU in 2004 abolished individual exemptions, the Serbian lawmakers assessed that these should remain part of the Serbian system for a transitional period to allow the market to familiarize itself with the new competition rules. This was important insofar as the CPC itself used individual exemptions to gain experience and develop its own practice in assessing restrictive agreements. Its practice was a guide for market participants on how they should make their operations compliant. To ensure enforcement and legal certainty, the 2005 Law provided that by-laws should be enacted to clarify the intentions of the lawmaker; however, this legal framework was never completed. The lack of clarity in the 2005 Law was to some extent overcome by numerous interpretations by the CPC, which, due to the government’s inaction, mainly relied on the laws and practices of the EU Member States and the European Commission. The lack of by-laws governing horizontal and vertical agreements placed a heavy burden on the CPC’s

Law on Protection of Competition (Official Gazette of the Republic of Serbia, No 79/05). The law was adopted on 16 September 2005 and the negotiations of the Stabilisation and Association Agreement began on 10 October 2005. 5 Boris Begović and Vladimir Pavić, ‘Jasna i neposredna opasnost: Prikaz novog Zakona o zaštiti konkurencije’ [‘Clear and Present Danger: An Overview of the New Law on Protection of Competition’] (in Serbian) (2009) 57(2) Annals of the Faculty of Law in Belgrade 70. 3 4

Building an efficient system of protection of competition in Serbia  277 weak administrative capacities and left some issues unaddressed, creating uncertainty on some aspects for market participants. The 2005 Law did not provide for the de minimis rule, so the CPC also received requests for exemptions with respect to agreements that did not have a significant impact on Serbian markets. Such agreements could have been compliant even without notifying the CPC, had there been de minimis or block exemption rules in place. Even though the CPC in the first years of its operation made a number of decisions on restrictive agreements, the most important advance was related to the abuse of a dominant position. Despite its limited capacities and experience, the CPC completed a number of complex cases in this area. It entered the spotlight when it found an abuse of the dominant position in the raw milk purchase market6 (Imlek) against one of the largest Serbian companies. This resulted in a greater interest by the public in the CPC, its operation and its competences. Further, a case related to telecommunications7 (SBB-DTH) and another related to the money transfer market8 (EKI-Tenfore, collective dominance) were landmark cases, sending a message that the CPC would not shy away from tackling complex cases and big market players. The legal provision defining a dominant position contained a presumption that an undertaking with a share of more than 40 per cent had a dominant position. The CPC did not determine its decisions based on this presumption, instead presenting evidence to verify whether there actually was a dominant position. Nevertheless, experts suggested that this presumption should be eliminated from the 2005 Law. Concentration (merger) control was first introduced by the 2005 Law. The law obliged merging parties to notify concentrations if the joint turnover was above €10 million in the preceding financial year in the Serbian market, or if there was more than €50 million in revenue generated on the international market and one merging party was registered in Serbia. This meant that the CPC was heavily burdened by many merger filings in the initial years of the application of the 2005 Law. Notifications also involved acquisition of control over undertakings with negligible economic power and insignificant revenue, where the other participant in a concentration (the acquirer of control) alone met one and/or the other alternative condition. The thresholds were fixed at so low a level that the CPC was often criticized by both the business community and the expert public, including the European Commission.9 Most of the criticism focused on the concept of control for concentrations. Filing fees for concentrations were also very high, which created financial and administrative pressure on the undertakings and the business community. One of the most significant successes of the 2005 Law was the establishment of the CPC as an independent institution. The CPC managed to build up and safeguard its independence in a system that was only just learning about the role of independent institutions as the ‘fourth branch of government’. Naturally, the CPC did not become a strong institution instantly. The 2005 Law laid the foundations for the CPC’s future operations and development: it provided 6 CPC v Imlek and Mlekara Subotica, http://​www​.kzk​.gov​.rs/​kzk/​wp​-content/​uploads/​2012/​03/​ Resenje​-Imlek​-25​.01​.2008​.pdf. 7 CPC v Serbia Broadband – Srpske kablovske mreže, http://​www​.kzk​.gov​.rs/​kzk/​wp​-content/​ uploads/​2012/​03/​Resenje​-SBB​-26​.12​.2008​.pdf. 8 CPC v Eki Transfers and Tenfore, http://​www​.kzk​.gov​.rs/​kzk/​wp​-content/​uploads/​2012/​03/​ Resenje​-EKI​-Transfers​-12​.01​.2010​.pdf. 9 European Commission, Commission Working Staff Document: Serbia 2008 Progress Report (5 November 2008,  SEC(2008) 2698 final) 35, https://​ec​.europa​.eu/​neighbourhood​-enlargement/​sites/​near/​ files/​pdf/​press​_corner/​key​-documents/​reports​_nov​_2008/​serbia​_progress​_report​_en​.pdf.

278  Research handbook on methods and models of competition law for the financial independence of the CPC by stipulating that it would be financed using its own funds, generated mainly from concentration fees. Finances from the state budget were provided to the CPC only in the first year of its operation, as was specified in the law. Moreover, the manner in which the CPC’s Council was elected ensured its independence. The members of the Council were proposed by the Association of Lawyers, the Association of Economists, the Bar Chamber, the Chamber of Commerce and the Serbian government. Even though the proposals of individual proponents may have depended on their own particular interests, their variety was a guarantee of the impartiality and independence of the CPC’s operation. The CPC had one important deficiency: it did not have the power to impose fines. This could only be done by a minor-offence court in proceedings that were instituted based on the CPC’s decision. This provision proved to be ineffective: not one breach of competition was sanctioned while the 2005 Law was in effect. The application of the 2005 Law thus resulted in the CPC approving a large number of concentrations; it found several clear cases of abuse of dominant position and a few restrictive agreements.

IV

BUILDING A MORE EFFICIENT INSTITUTION: A NEW LAW IN 2009

The current Law on Protection of Competition, passed in July 2009,10 became effective on 1 November 2009 (the Law). It introduced key tools for the CPC, such as the power to impose fines against undertakings, to conduct dawn raids and to exempt undertakings from paying fines if they submit the appropriate evidence to the CPC. The substantive law provisions are in line with the EU acquis and contribute to the modernization of the protection of competition. Article 1 of the Law provides that it governs the ‘protection of competition on the market of the Republic of Serbia, in order to achieve economic prosperity and well-being of the society, and especially the benefit of the consumers’. The Law contains an effects provision. This means that it applies to ‘all actions and practices performed on the territory of the Republic of Serbia’ equally as it applies to activities ‘performed outside of its territory when they produce or are likely to produce an effect on the territory of the Republic of Serbia’. The Law applies to undertakings within the EU concept: ‘all natural and legal persons who, directly or indirectly, on a permanent or ad hoc basis, conduct business operations in the trade in goods and services, regardless of their legal status, form of ownership, citizenship or state of origin (hereinafter: undertakings)’ (Article 3). It thus applies to both public and private undertakings – that is, public enterprises, companies, entrepreneurs and other undertakings performing activities of public interest – unless this would prevent them from performing activities or tasks assigned to them by public authority.11

Law on Protection of Competition (Official Gazette of the Republic of Serbia, No 51/09). The draft of a new Competition Act has been included in the legislation plan for 2019/20 (2019 Draft). It is more extensive than earlier laws, with more detailed provisions on procedure to align it to the new General Administrative Procedure Act from 2016: Official Gazette of the Republic of Serbia, No 18/2016, https://​www​.paragraf​.rs/​propisi/​zakon​-o​-opstem​-upravnom​-postupku​.html. 10 11

Building an efficient system of protection of competition in Serbia  279 A

Restrictive Agreements

Article 10(1) of the Law defines restrictive agreements as ‘agreements between undertakings whose purpose or effect is a significant restriction, distortion, or prevention of competition in the territory of the Republic of Serbia’. In terms of the types of restrictive practices, Article 10(2) provides that restrictive agreements can include ‘contracts, certain contractual provisions, express or tacit agreements, concerted practices, as well as decisions of association of undertakings’. Both these provisions are in line with EU competition law. Article 10(2) refers to a list of examples of restriction of competition, in similar terms to Article 101(1) of the TFEU. It includes agreements that: ●● directly or indirectly fix purchase or selling prices, or any other trading conditions; ●● limit or control production, markets, technical development, or investment; ●● apply dissimilar conditions to equivalent transactions with other trading parties, thereby placing them at a competitive disadvantage; ●● make the conclusion of contracts subject to acceptance by the other parties of supplementary obligations that, by their nature or according to commercial usage, have no connection with the subject of such contracts; and ●● share markets or sources of supply.12 All restrictive agreements that have not been exempted from prohibition are void and prohibited. The Law differentiates between individual exemptions (Article 12) and exemptions by categories of agreement (Article 13). Both these forms of exemption must meet criteria referred to in Article 11, which essentially echoes Article 101(3) of the TFEU. In order to be exempted, a restrictive agreement must meet two positive and two negative conditions. The positive conditions are, first, that the agreement must contribute to improving the production or distribution of goods or to promoting technical or economic progress; and, second, that it must give consumers a fair share of the resulting benefits. The negative conditions are, first, that the agreement should not impose restrictions which are not indispensable to the attainment of these objectives; and, second, that it should not eliminate competition in the relevant market or in a substantial part of it. An individual exemption can be approved by the CPC for a period not exceeding eight years. The lawmakers decided to retain the individual exemption, a relic from the EU law. Regrettably, the CPC takes a strict position in its Guidelines for applications for individual exemption.13 Its position is that if it opens an investigation to establish whether an agreement is restrictive, the fact that the agreement meets the exemption criteria is of no consequence for the decision on whether to suspend or terminate the procedure. The parties to the agreement can be sanctioned even if the agreement meets all four conditions referred to in Article 11. In the EU system, the parties can use the exemption conditions (Article 101(3)

12 The 2019 Draft allows for self-assessment of restrictive agreements, along with the continuing possibility of notification. Negative clearance, where parties may seek an advisory declaration that the notified conduct is legal, is also proposed for both restrictive agreements and abuse of dominance. Furthermore, the CPC considers introducing settlement proceedings for cartels, but also for other restrictive practices – including the abuse of dominance. 13 The CPC instruction on the submission of a request for individual exemption, http://​www​.kzk​ .gov​.rs/​kzk/​wp​-content/​uploads/​2011/​08/​UPUTSTVO​-Podnosenje​-zahteva​-za​-pojedinacno​-izuzece​ -restriktivnih​-sporazuma​-od​-zabrane​.pdf.

280  Research handbook on methods and models of competition law of the TFEU) as a valid defence before the European Commission. The CPC interpretation of this provision means that a person can be sanctioned for applying an agreement that actually contributes to the efficiency of production or distribution and increases the resulting benefit for the consumers. This would be contrary to the nature and the objective of antitrust policy. It is hoped that the next amendment to the Law will abolish individual exemptions. Agreements of minor importance are excluded from the scope of Article 10 of the Law as they do not have a significant impact on the restriction, distortion or prevention of competition. The Law defines such agreements using fixed market share thresholds: ●● ●● ●● ●●

10 per cent for horizontal agreements; 15 per cent for vertical agreements; 10 per cent for combined horizontal/vertical agreements; and 30 per cent if the agreements concluded between various participants have a similar impact on the market, and if the individual market share of each participant does not exceed a 5 per cent threshold in each individual market where the effects of the agreement are manifested.

The Law contains a list of hardcore provisions that the parties should not include in an agreement. Horizontal hardcore restrictions are price fixing, limitation of production or sales, and division of the sourcing market, while vertical hardcore restrictions include price fixing and the division of markets. B

Abuse of a Dominant Position

Article 15(1) of the Law provides that an undertaking has a dominant position in a relevant market: ●● when competition in the market does not exist or it is insignificant; ●● when an undertaking has considerable competitive advantage over its competitors taking into account the volume of its market share, economic and financial power, access to sourcing and distribution markets, as well as legal and actual obstacles for market entry of other undertakings.14 This definition departed from the definition in previous law, which was based on EU case law.15 In the 2009 Law, the definition of the term ‘dominant position’ was based on the German competition law and the legislator also maintained the legal presumption of dominance for undertakings whose market share is above 40 per cent. These provisions were criticized on the grounds that they are not fully in line with the provision of the EU treaty and the EU practice, particularly the provision on legal presumption of dominance. In order to remedy these shortcomings, the Serbian legislature amended the law in 2013 and the ‘EU definition’ of dominance was re-established, while the legal presumption of dominance was removed. 14 See also the German Gesetz gegen Wettbewerbsbeschränkungen, Article 19, paragraph 2, sentence 1. In addition to those mentioned already, the German law also refers to the following criteria: relation with other Unternehmungen (undertakings); actual or potential competition through Unternehmungen resident both inside and outside of the areas covered in the application of this law; the ability to redirect their offer and demand to other goods and services; and the ability of the other party in the market to turn to other Unternehmungen. 15 Case 85/76 Hoffmann-La Roche (Vitamins) [1979] ECR 461, paras 38–39.

Building an efficient system of protection of competition in Serbia  281 Additionally, the amended definition contains the list of qualitative and quantitative factors that should be used for the purpose of determining the existence of the dominant position. In summary, the Law now provides a general prohibition on the abuse of a dominant position in a market. Article 16 defines abuse of a dominant position and provides a list of actions that can lead to abuse of competition that mirrors the restrictions referred to in Article 102 of the TFEU. C Concentrations The Serbian merger control system is modelled closely on the EU Merger Regulation.16 The Law provides for three types of concentration: ●● mergers and other statutory changes in which a merger of undertakings occurs, within the meaning of the law governing the status of companies; ●● the acquisition by one or more undertakings of direct or indirect control, within the meaning of Article 5(2) of the Law, over another or more undertakings; and ●● the creation of a joint venture by two or more undertakings in order to create a new undertaking or to gain a joint control, within the meaning of Article 5(2) of the Law, over an existing undertaking operating on a lasting basis. Article 19(1) of the Law prohibits concentrations of undertakings if they would ‘significantly restrict, distort or prevent competition in the market of the Republic of Serbia or in any part of it, and especially if such restriction, distortion or prevention is the result of the creation or strengthening of a dominant position’. Even though the legal norm can be interpreted as constituting an application of the dominance test only, the CPC also applies the SIEC (Significant Impediments of Effective Competition) test, thereby following the trend in the EU and in other countries. The Law contains a list of criteria to be taken into account when assessing concentrations. It focuses directly on the structure of the relevant market as indicated by the market concentration of participants, their economic and financial power, and their level of competitiveness. The role of actual and potential competitors, and legal and other barriers to entry are also relevant. Vertical relationships in the particular market could be relevant, especially with respect to alternatives available to suppliers and users. Dynamic factors – especially trends in supply and demand, technical and economic progress, and consumer interests – all need to be considered. The Law exempts four types of concentration from the merger control process: ●● where a bank or other financial institution or insurance company, in the course of its regular business operations, temporarily acquires stocks or shares for further sale, provided that it sells them within one year from the date of acquisition and does not use them to influence the business decisions of the undertaking in terms of its conduct in the market; ●● where an investment fund acquires a stake in an undertaking with the aim of fully preserving the value of its investment and provided that it does not affect the competitive behaviour of the undertaking in the market;

EU Regulation No 139/2004.

16

282  Research handbook on methods and models of competition law ●● where a joint venture aims to coordinate market activities between two or more undertakings that retain their legal independence, such a joint venture being subject to assessment in accordance with the Law (Articles 10 and 11); and ●● where control is exercised over an undertaking in bankruptcy. The Law sets out a notification procedure. Concentrations must be notified to the CPC if the following notification thresholds, as defined under Article 61, are met: (1) the total annual revenue of concentration participants in the international market in the preceding financial year exceeds €100 million, with at least one participant generating revenue in Serbia exceeding €10 million; and (2) the total annual revenue of at least two concentration participants in Serbia in the preceding financial year exceeds €20 million, with at least two participants generating revenue in Serbia exceeding €1 million each.17 The notification must be filed with the CPC within 15 days of the date of the conclusion of the agreement, or after the announcement of a public call (the offer or closure of a public offer), or after acquiring control over an undertaking. The Law provides for a standstill clause, so the concentration participants must delay the implementation of concentration until the CPC decision. The timing of the procedure is strictly limited and is conducted in two stages. Only after the formal confirmation that notification is complete can the one-month statute of limitation for passing the final decision begin to run (stage one). The CPC President can make this decision in summary procedure. Stage two can begin when the CPC President passes his or her decision on ex officio investigation of the concentration. The final decision concerning the notified concentration must be passed within four months; otherwise, the concentration is deemed to be approved. A concentration can be investigated ex officio if: ●● the CPC finds that the combined market share of concentration participants in Serbia is at least 40 per cent; ●● there are reasonable indications that the concentration does not meet the approval conditions; or ●● the conditions have not been met (Article 62(1)). The CPC sends a Statement of Objections to the participants requesting their arguments. The parties can put forward commitments so that the concentration can meet the conditions for the approval. Most of the CPC’s cases to date – around 100 per year – have related to concentrations, with the vast majority of concentrations approved under the summary procedure. An investigation procedure has been used mainly in cases of mergers that lead to the new entity acquiring a high market share. The CPC has always been willing to consider remedies and commitments. It has prohibited concentrations in two cases,18 but both decisions were quashed 17 The 2019 Draft proposes changes to merger filing thresholds, mainly to exclude transactions with little connection to the jurisdiction. 18 Sunoko d.o.o./Hellenic Sugar Industry SA (2012), http://​www​.kzk​.gov​.rs/​kzk/​wp content/ uploads/2012/01/SUNOKO-Helenic.pdf; Primer C d.o.o./C market d.o.o. (2007), http://​www​.kzk​.gov​ .rs/​kzk/​wp​-content/​uploads/​2012/​11/​DELTA​-resenje​-u​-ponovnom​-postupku​-26​.11​.2007​.pdf.

Building an efficient system of protection of competition in Serbia  283 by the court. Even though it has the legal power to do so, the CPC has not yet imposed any sanctions for failure to notify. D

Institutional Organization

The CPC was established in 2006 as an independent and autonomous national competition authority. The CPC President and the four members of the CPC Council are elected by the National Assembly for a five-year term (Articles 23(2), 24(1)). The Law provides that the CPC President and Council members are elected from the ranks of distinguished experts in the field of law and economics, with at least 10 years of relevant professional experience, who have achieved significant recognized successes or practice in the relevant field – particularly in the areas of competition protection and European law. Finally, they should enjoy a reputation for being objective and impartial persons. This system was continued under the Law, which gives the CPC wider powers. The intention was to ensure the independence of this institution by electing its officials in the National Assembly. The candidates are first heard by the parliamentary Committee for the Economy, which votes on whether to approve the proposal. If approved, the proposal is then sent to the plenary for a vote and final election. The purpose of the five-year term is to maintain the composition of the CPC’s Council even after the expiry of the National Assembly’s four-year term, which provides additional safeguards to its operational independence. The CPC is accountable for its operation to the National Assembly, to which it submits its annual work reports. The CPC also enjoys financial independence, which has allowed it to retain and improve its resources. The CPC is financed from its own budget and the finances are secured mostly from concentration fees. The CPC thus does not depend on the state budget for finances, which means that it is not affected by the austerity measures that have been in force in the Serbian public sector with respect to salaries and new recruitment. Nevertheless, this system of financing is not sustainable in the longer run and may affect the development of the institution. An expert study of the Serbian competition protection policy was carried out in 2011 by the United Nations Conference on Trade and Development (UNCTAD).19 One of the recommendations was that the CPC should be provided with predictable, sound and adequate sources of financing. As the report notes, the CPC is the only competition protection authority in Europe that secures its finances mainly from concentration fees. For instance, concentration fees in the EU account for only between one-fifth and one-quarter of revenue. The uncertainty characterizing CPC financing particularly intensified after 2009, when the Law gave the CPC the power to impose fines on undertakings. However, if the Administrative Court quashes a CPC decision, it must pay interest on the amount of fine to be reimbursed. Thus, the CPC is exposed to financial risk if the parties are successful in the Administrative Court. This unique legal provision jeopardizes the CPC’s finances. It had to pay €750,000 to a company that was successful in obtaining the annulment of the CPC’s decision at the court.

19 UNCTAD, Voluntary Peer Review of Competition Policy: Serbia (2011), http://​unctad​.org/​en/​ pages/​PublicationArchive​.aspx​?publicationid​=​934.

284  Research handbook on methods and models of competition law E

Internal Organization

The Law established two CPC bodies, the CPC President and the CPC Council, in an attempt to rectify the procedural deficiencies and delays of the earlier law. Nevertheless, giving more powers to the President raised concerns. The President has the widest possible powers, including the power to decide whether, when and on what basis to start an antitrust investigation; what investigative activities to take in particular proceedings; and whether the CPC staff will conduct a dawn raid. In addition, the President alone has the power to decide on concentrations in the summary procedure and on the initiation of the investigative procedure (second phase merger control). The CPC staff perform technical tasks related to CPC responsibilities and are headed by a Secretary who is appointed by the Council. An authorized official case handler conducts investigations, including obtaining evidence to establish the factual situation; collecting statements of parties and witnesses; obtaining data, documents and objects; and conducting inspections. This also includes preparing and drafting the (final) decision in cooperation with the rapporteur from the CPC Council. The CPC currently has two operational divisions working on all cases, one dealing exclusively with concentration control and the other addressing all other anti-competition (restrictive) practices. There are two advisory divisions, one for economic advice and the other for legal issues. F

Competencies of the CPC

The competences and functions of the CPC are specified in Article 21. They can be broadly divided as follows: (1) the implementation and application of the Law; (2) the supervision of the market using sector analyses; (3) advising the government and state authorities on issues related to the competition protection law and policy; and (4) cooperating at the international level in the competition protection area. The basic role of the CPC is to establish breaches of the Law and to impose measures for instituting effective competition and fines. It has at its disposal all modern investigation instruments, such as the power to: ●● ●● ●● ●●

conduct dawn raids; collect information from all undertakings and state authorities; pass provisional measures; and take statements from witnesses and third parties and impose fines and procedural penalties.

Procedural rules are regulated by the Law, with a subsidiary application of the Law on General Administrative Procedure.20 Given that the Law on General Administrative Procedure goes back to the time of Yugoslavia, and that it was intended for procedures conducted by state authorities, its application should be reduced and the number of procedural provisions in the Law should be increased. Indeed, the CPC is a body that applies a specific procedure that involves investigative activities, communication with parties, the taking of statements, and the conduct of dawn raids. In addition, appeals are not allowed against the CPC’s procedural

20 Law on General Administrative Procedure (Official Gazette of the Federal Republic of Yugoslavia, No 33/97 and No 31/01 and Official Gazette of the Republic of Serbia, No 33/10).

Building an efficient system of protection of competition in Serbia  285 decisions; instead, lawsuits can be instituted at the Administrative Court. Given the high fines and the behavioural and structural measures that can be imposed on undertakings, the procedure run by this authority requires special rules that will guarantee the highest standard of protection for parties and that will enable, at the same time, an efficient procedure. This is not possible to achieve with a law that is intended for a typical bureaucracy. The first dawn raids were conducted in the summer of 2015, despite the fact that the Law had commenced in 2009. There were many reasons for the delay, including the need to train staff and equip them with the tools to conduct successful investigations. The European Commission provided the funds for purchasing IT forensic equipment under a project to support the development of the CPC. In order to coordinate the competences of the relevant authorities, the CPC signed a cooperation agreement with the Ministry of the Interior in relation to implementing inspections and using digital forensic tools.21 Positive developments in the efficiency of the CPC have occurred since 2015, following inspections in a number of cases that were intensively covered in media reports and were well noted by the business community.22 A significant number of CPC activities have also focused on international cooperation. The CPC is active in many regional and international competition forums, such as the International Competition Network, the OECD Regional Centre for Competition,23 and the Sofia Competition Forum. Of particular relevance for the CPC’s development and for Serbian competition protection policy was Serbia’s participation in the Voluntary Peer Review on Competition Law and Policy conducted by UNCTAD in 2011.24 Many of the recommendations provided in the report were accepted and implemented in the 2013 amendments to the Law. In addition, the CPC signed memoranda of understanding with all neighbouring countries, as well as with many other regional and European countries, in order to improve the exchange of experiences and information. These activities are very important for a relatively new competition authority and help to overcome the obstacles encountered by every new policy or institution. G

Judicial Control

The CPC is not the only institution that is competent to implement the Law and its by-laws. The Law provides that an administrative dispute can be instituted before the Administrative Court of the Republic of Serbia against final decisions of the CPC. The 2005 Law provided for the competence of the Supreme Court in appeals. Most CPC decisions at that time were quashed in the Court on the basis of procedural issues with CPC decisions and proceedings. In addition, the court did not have a proper understanding of the specificities of the Law and of proceedings run by the CPC. The 2009 Law vested this competence in the Administrative Court, a first instance court with its decisions appealed to the Supreme Court of Cassation using the extraordinary legal remedy. However, the Administrative Court has not upgraded its capacity to deal with essen21 CPC press release regarding the execution of the Cooperation Agreement between the CPC and the Ministry of the Interior, http://​www​.kzk​.gov​.rs/​sporazum​-o​-saradnji​-ministarstva​-unutrasnjih​-poslova​-i​ -komisije​-za​-zastitu​-konkurencije​-i​-ugovor​-o​-saradnji​-u​-oblasti​-digitalne​-forenzike. 22 CPC v N sport d.o.o. and retailers of sports shoes (2016), http://​www​.kzk​.gov​.rs/​en/​pokrenuti​ -postupci​-protiv​-16​-ucesnika​-na​-trzistima​-veleprodaje​-i​-maloprodaje​-sportske​-odece​-obuce​-i​-opreme; CPC v Frikom (2017), http://​www​.kzk​.gov​.rs/​en/​komisija​-pokrenula​-postupak​-protiv​-i. 23 The OECD Regional Centre for Competition is based in Budapest. 24 UNCTAD (n 19).

286  Research handbook on methods and models of competition law tial competition law issues. Even though the CPC is now far more successful in presenting its cases in court, the court decisions still lack interpretation of law and facts that would shed light on the application of antitrust legislation in Serbia. Given the significance that the competent courts have in the development of competition law in the EU, it is clear that the limited capacity of the Administrative Court to adjudicate cases in the antitrust area has a negative effect on the development of case law and on the development of the CPC itself. The CPC alone is unable to attain a significant level of development as the quality of the operation of these two institutions is related and interdependent. Once the CPC receives essential comments about its work and the way in which it applies the legal rules for the protection of competition in court decisions, the quality of its work will improve. The Administrative Court may completely replace a decision of the CPC with its own judgment, but this has not happened in practice. Only when the Administrative Court improves its work will there be a complete and efficient institutional system for the protection of competition in Serbia.

V

PROMOTION OF COMPETITION CULTURE – A CHALLENGE FOR THE NEW INSTITUTION

Serbia is still a country in transition towards a market economy, so increased awareness of competition issues is a very important goal for all relevant stakeholders. The current Serbian economic policy is oriented to an open-market approach with the aim of EU accession. It calls for a modern, transparent and effective implementation of competition legislation, in line with the EU acquis, representing a basic pillar of the efficient market economy. The best way to do this is a broad public ‘coalition’ for introducing the benefits of a competitive market. Whereas the ‘foreign factor’ – the EU – has had a major impact on introducing competition law in Serbia, the CPC has the role of promoting competition and raising awareness among stakeholders about competition policy. This task involved determining whether the government and public authorities should be the initial targets, since many regulations contained norms which were not in accordance with the competition rules, or whether the business community should first be addressed, since it was not familiar with the basic competition rules. In addition, at the time of introducing the competition law – and it has remained the case until recently – there were no courses that related to competition protection at the Faculty of Economics and Law at the University of Belgrade. The CPC has intervened on a number of occasions, providing its opinion on regulations. There was, however, no system that enabled a lawmaker to address the CPC. Pursuant to the Law, the CPC is responsible for providing opinions on draft laws, and the European Commission has emphasized on several occasions that the CPC should be a priori consulted on the regulations with relevance for regulating market conditions. Progress has been made over the past several years, and the CPC’s provision of opinions has been significantly intensified – in spite of the fact that there is no obligation to submit drafts and proposals to the CPC. The CPC’s opinions are not binding, but they are taken into account. Furthermore, the CPC publishes opinions on its website. This achieves a double impact: it contributes to notifying the wider public of the CPC’s opinions on draft regulations and norms with regard to the application of competition rules, and it also allows the CPC to exert pressure on lawmakers to take its opinions into consideration.

Building an efficient system of protection of competition in Serbia  287 The CPC is becoming a more recognized institution and its work has become increasingly transparent. All CPC decisions are published on its website, including both their operative parts and their statements of reasons. This enables all stakeholders to be informed about the manner of assessment of market activities and of the CPC’s methodology and practice. The CPC has signed several cooperation agreements with national bodies, independent agencies (sector regulators), the National Bank and others to enable all the institutions to learn about competition issues through the exchange of knowledge and experience and to notify the CPC of irregularities within its scope of work. The CPC has close cooperation with the Serbian Chamber of Commerce and with other associations with regard to raising awareness on competition protection in Serbia. Going forward, the CPC must develop a comprehensive programme and implement joint activities aimed at raising awareness on the importance of competition. The programme should aim to inform citizens and consumers about competition protection rules, about the way in which competition infringements are detected, and about the ways of reducing these infringements.

VI

THE EU LAW AND PRACTICE – INTERPRETATIVE INSTRUMENTS OF THE NATIONAL COMPETITION RULES

Serbia signed the Stabilisation and Association Agreement (SAA) with the EU in 2008 and became an associated country to the EU. This Agreement is double-binding, with precise rights and obligations for both parties. Article 73 of the SAA addresses competition in the market and establishes an obligation on the Serbian side to apply EU competition rules, antitrust and state aid, and to create independent enforcement authorities. Hence, the Serbian Competition Authority can use EU rules and practice as interpretative instruments when applying national competition rules. Throughout the SAA institutions, such as the SAA Committee or the Subcommittee on Internal Market and Competition, the EU Commission supervises the work of the Serbian Competition Authority and its practice. Moreover, it provides guidance to the harmonization process and the priorities of the national institution. A

Competition Provisions of the EU Stabilisation and Association Agreement

The SAA and the Law are the foundations of competition in Serbia. The SAA is a model for extending the application of the EU competition law to non-EU Member States. Article 16, paragraph 2 of the Constitution of the Republic of Serbia25 provides: ‘Generally accepted rules of international law and ratified international treaties shall be an integral part of the legal system in the Republic of Serbia and applied directly, and ratified international treaties need to be in accordance with the Constitution.’ In addition, Article 194 of the Constitution lays down that laws and other general acts may not be in non-compliance with ratified international treaties and generally accepted rules of international law. The consequence of this is that the SAA is an integral part of the legal system of the Republic of Serbia and sits above the Law on Protection of Competition and other by-laws in the legal hierarchy of the country. 25 The Constitution of the Republic of Serbia (Official Gazette of the Republic of Serbia, No 98/2006).

288  Research handbook on methods and models of competition law The protection of competition is laid down in Article 73, which reads: 1. The following are incompatible with the proper functioning of this Agreement, insofar as they may affect trade between the Community and Serbia: (a) all Agreements between undertakings, decisions by associations of undertakings and concerted practices between undertakings which have as their object or effect the prevention, restriction or distortion of competition; (b) abuse by one or more undertakings of a dominant position in the territories of the Community or Serbia as a whole or in a substantial part thereof; (c) any State aid which distorts or threatens to distort competition by favouring certain undertakings or certain products. 2. Any practices contrary to this Article shall be assessed on the basis of criteria arising from the application of the competition rules applicable in the Community, in particular from Articles 81, 82, 86 and 87 of the EC Treaty and interpretative instruments adopted by the Community institutions.

When applying competition provisions, the CPC and the courts are obliged, in conformity with the cited SAA provision, to apply the criteria arising from the application of the competition rules and EU interpretative instruments applicable in the EU. However, this obligation is not applicable in all circumstances because the SAA is an agreement that regulates trade between Serbia, on the one hand, and the EU and its Member States, on the other. Therefore, there is only an obligation to act in accordance with the SAA and to apply the SAA to the competition protection rules in that case. If trade is not between Serbia and the EU, the authorities are obliged to act in compliance with the regulations adopted by national authorities. However, this does not assist in providing legal certainty. If authorities, the CPC, and the courts had a different approach to determining the infringement of competition depending on whether or not the situation concerned was regulated by the SAA, a parallel competition system would be established. This would jeopardize the rights and interests of market participants, and this would have a negative impact on the Serbian economy. Therefore, although there is no legal formal obligation on the authorities to apply the EU criteria to a non-SAA situation, the CPC should ensure a uniform application of regulations for the jurisdiction-based enforcement, and should be guided by the EU law and practice at all times. The SAA competition provisions provide the CPC and the courts with the interpretative instruments for national provisions, but they do not lay down an obligation for national regulators to apply EU regulations.26 Authorities are not obliged to apply the EU regulations directly because Serbia is not an EU Member State, nor do the SAA and the Serbian Constitution provide for this possibility. Therefore, in the competition procedures, where there are legal gaps or ambiguities in the interpretation of national legislation, or when some issues are regulated differently in the Serbian law as compared to the EU law, the CPC may use the criteria arising from the application of EU rules (including the decisions made by the European Commission and the EU courts) as auxiliary interpretative instruments. The competent court in Serbia has not commented on the application of the SAA in practice, but the manner of application – which was validated by the Croatian Constitutional Court’s decisions

26 Boris Begović and Vladimir Pavić, ‘Jasna i neposredna opasnost II: čas anatomije’ [‘Clear and Present Danger II: Anatomy Lesson’] (2010) 58(2) Annals of the Faculty of Law in Belgrade 337.

Building an efficient system of protection of competition in Serbia  289 in the Republic of Croatia – has been described.27 The Constitutional Court of the Republic of Croatia found that when assessing the possible disturbance of market competition, the Croatian institutions competent for the protection of market competition are authorized and required to apply the criteria arising from the application of the competition rules of the European Communities and from the interpretative instruments adopted by the Community institutions. Therefore the SAA and the Interim Agreement oblige the Croatian competition institutions to apply not only the Croatian competition law but also to take account of the relevant law of the European Community … the criteria, standards and interpretative instruments of the European Communities are not applied as the primary source of law, but only as an auxiliary instrument of interpretation … When applying such harmonized legislation, government institutions are obliged to apply it in the same way as the European Communities, that is in the sense and spirit of the legislation pursuant to which the harmonization was carried out …

Whereas the SAA does not contain any provisions on concentrations, but only on the infringement of competition law, the question that arises is whether the EU law may be used as an interpretative instrument for assessing concentration. This question has been raised in academic articles,28 and the solution proposed is to interpret this part of the Serbian legal framework ‘in the sense and spirit’ of the EU law, having in mind that, after the adoption of rules that are in compliance with the EU law, they will be applied in the same manner as they are applied in the EU. Finally, the SAA provides the opportunity for Serbian competition law to reflect that of the EU before Serbian formal membership. In these circumstances, the existence of some legal gaps and an incomplete legal framework, accompanied by limited knowledge of the EU competition law, requires Serbia to take significant steps to achieve the goal. The adoption of guidelines, instructions and the definition of clear practice through the decisions of the CPC and competent courts would be a good step forward. B

The EU Commission as the Supervisor of the Serbian Competition Authority

The EU is probably most responsible for the fact that modern competition law was introduced in Serbia in 2005. As noted earlier, a state that wishes to accede to the EU is obliged to accept and apply the competition protection rules applicable in the EU. Having in mind that this obligation is also laid down in the binding agreement concluded between the EU and a candidate country, it is not just a matter of political will, but is also a legally binding norm. Therefore, the European Commission, as the EU body responsible for supervising the implementation of the SAA, ensures that the competition law is adequately applied in Serbia and that an institution is established to ensure the enforcement of the law and to take all the steps necessary to ensure the efficient application of the law. The SAA expressly provides for the establishment of institutions responsible for monitoring the SAA implementation. Article 42 of the SAA provides for the establishment of an Interim Committee for the Implementation of the Interim Agreement that will supervise the SAA application and implementation. Following its establishment, the Interim Committee also

Constitutional Court Decision U-III-1410/2007, t. 6‒7. Ivana Rakić, ‘European Union Merger Control Rules as a Source of Law in the Republic of Serbia’ (Collected Papers of the Law School in Niš, No 67, 2014) 273. 27 28

290  Research handbook on methods and models of competition law established five subcommittees for monitoring the implementation of the SAA in certain areas, including the Subcommittee on Internal Market and Competition. The Interim Committee and subcommittee members are the representatives of the Republic of Serbia and the EU. At these meetings, reports on the activities of the CPC are considered, including reviews of the enforcement record, legislative activities and the improvement of administrative capacities. Current issues may be brought up at these meetings if they refer to the application of competition rules in a situation that affects trade between the EU and Serbia. Besides the above-mentioned supervision mechanism, the EU annually assesses the progress Serbia makes within its European integration process – the so-called Progress Report. The European Commission draws up a report on the progress that Serbia has made in its EU association process. The report contains information on the implementation of reforms that lead to the full achievement of the Copenhagen criteria (the EU accession criteria). The Progress Report contains a special chapter on competition law, which includes state aid. For this purpose, the Serbian CPC submits regular annual reports to the European Commission. These reports contain data on legislative activities (the Law, regulations, guidelines and other acts drawn up by the CPC), data on the CPC’s institutional and administrative capacities (job classification scheme, number of employees, work organization), and data on the cases dealt with during the year concerned. The 2018 Progress Report29 indicated that the Serbian legislative framework is broadly in line with the acquis and that the activities of the CPC have been enhanced, but that its enforcement record must be strengthened – in particular, by imposing higher fines and by intensively promoting its leniency programme. Previous Progress Reports have suggested the enhancement of judicial capacity to solve complex antitrust cases. These activities taken by the European Commission aim at full harmonization of the Serbian legislation and the practice of Serbian competition authorities with the EU rules. In previous years, the Progress Reports have also contained negative remarks regarding the laws, by-laws and practice of the Serbian authorities. These remarks have always been carefully considered in Serbia and have often formed the basis for amendments to the Law. The European Commission and the SAA guarantee the independence of the CPC. Each time its financial or operational independence has been challenged, the European Commission has appeared to defend it. In its reports and through addresses by its representatives to the Serbian authorities, the European Commission has always promoted the significance of independence for a competition authority. Besides implementing the Law, the CPC actively contributes to the EU accession negotiations. Under the government decision establishing the EU Accession Process Coordination Body and negotiating groups, the CPC participates in the activities of four negotiating groups: (1) for Chapter 8, the Negotiating Group on Competition; (2) for Chapter 10, the Negotiating Group on Information Society and the Media; (3) for Chapter 14, the Negotiating Group on Transport; and (4) for Chapter 15, the Negotiating Group on Energy. Within these activities, the CPC prepares documents and other contributions necessary for an analytical review of legislation (screening), which refers to the level of compliance with the EU acquis communautaire, data on the implementation of regulations, institutional and administrative

29 European Commission, Commission Working Staff Document: Serbia 2018 Report (17 April 2018,  SWD(2018) 152 final), https://​ec​.europa​.eu/​neighbourhood​-enlargement/​sites/​near/​files/​20180417​ -serbia​-report​.pdf.

Building an efficient system of protection of competition in Serbia  291 capacities, implementation obstacles, and plans relating to further alignment and application of legislation.

VII

ENFORCEMENT BY THE CPC

The CPC has 13 years of practice. It has built a solid enforcement record, with several abuse of dominance cases, restrictive agreements, decisions against public companies, and almost 1,000 merger control decisions. Just like other newly established competition authorities, the CPC faces the issue of determining priorities. This is a pivotal issue and affects (1) the scope of application of the competition rules in the country; (2) the direction of practice development, and consequently of the professional competence of the authority; and (3) whether the aim is to establish a strong and bold authority or to maintain it at the level of a typical administration. Serbian academics have criticized the current selection of priorities.30 The CPC focuses on the control of concentrations, which is said to create a ‘built-in’ error because low concentration notification thresholds have been set by law. This was particularly so in relation to the 2005 Law. The business community, foreign investors and the academic community believed that these thresholds hindered business and economic development and imposed unwarranted administrative costs. The EU voiced similar criticisms.31 In addition, the low thresholds were accompanied by high concentration fees, which resulted in a negative attitude towards the CPC. The 2009 Law changed the concentration notification thresholds, but the CPC still deals with approximately 100 concentrations annually. The thresholds were, however, designed to target only those concentrations that have significant impact on the market. The amendments introduced by the 2009 Law focused the CPC’s capacity on tackling serious infringements. The Law now empowers the President of the CPC to decide summary proceedings, which accelerates decision-making and provides Council members with an opportunity to focus on competition and concentration infringement that is examined within the investigation procedure (second phase). In addition to a large number of concentrations, the CPC has determined a significant number of cases on abuse of dominance, which have made it widely known to the public. During the first years of its work, the CPC made several decisions relating to the abuse of dominance in areas such as the telecommunications sector, banking services and the dairy industry. The CPC prioritized the investigation of an abuse of dominance in relation to restrictive agreements, particularly the cartels. Perhaps this is a consequence of the fact that Serbian society still considers the monopoly or dominant position a cause of many problems that occur in the market. Setting priorities to some extent depends on the familiarity of market participants and others with competition and the Law. The level of awareness of the protection of competition and its concepts in Serbia is at a low level. This has hindered the work of the CPC because there has

30 Mirko S Vasiljević and Dušan Popović, ‘Competition Law Enforcement in Serbia: Six Years of Staggering Along’ (2012) 13 European Business Organization Law Review 141. 31 European Commission, Commission Working Staff Document: Serbia 2008 Progress Report (5 November 2008,  SEC(2008) 2698 final) 35, https://​ec​.europa​.eu/​neighbourhood​-enlargement/​sites/​near/​ files/​pdf/​press​_corner/​key​-documents/​reports​_nov​_2008/​serbia​_progress​_report​_en​.pdf.

292  Research handbook on methods and models of competition law been a low level of complaints and information that might indicate infringement of the Law. It has been difficult for the CPC to initiate investigations without such information. Furthermore, the leniency programme introduced in 2009 has not so far yielded any significant results. Finally, one of the weak points of the CPC enforcement is its fining policy. Under the Law, the maximum fine for competition infringement is up to 10 per cent of the company’s turnover generated in Serbia in the year preceding the year in which the proceedings were initiated. However, the CPC has imposed significant fines in only a few cases. Two of these were the military uniform case,32 in which the CPC determined a fine of 5.85 per cent for bid rigging, and the Frikom case,33 in which the CPC determined a fine of 4 per cent of the company’s turnover for the abuse of a dominant position in the ice cream market. In other cases, average fines were between 0.5 and 2 per cent. However, a 13-year period has been enough for the CPC to establish the foundations for its operation and to develop appropriate capacities and practice to define more specifically the goals of its fight for a free and competitive market in the Republic of Serbia. Progress has been made in the past few years, in which the CPC has initiated several major investigations of restrictive agreements and cartels. In 2015, the CPC conducted its first dawn raid and it has used this instrument for data gathering on several occasions since. Recent decisions related to bid rigging in public procurement procedures suggest that the CPC is focused on bid-rigging cases. This is desirable and may contribute to the promotion of the CPC and to raising awareness about the protection of competition in Serbia, because the CPC’s activities protect public finances and taxpayers’ money.

VIII CONCLUSION The strategic development of competition law in Serbia should be based on two important factors: (1) institutional development and enforcement of the CPC; and (2) constant improvement of the legal framework. Having in mind the starting position in 2005, when the first law on modern competition rules was passed, progress to date is significant. However, additional capacity building is needed in order to develop one or more credible institutions that will remain stable in political and economic turmoil, which constantly threatens Serbia and the western Balkans region. The independence of the CPC is the key factor for the success of competition law in Serbia. The Serbian Competition Authority has shown that it has some financial and operational independence, which must be preserved and strengthened. Its administrative capacity has significantly increased from the previous period. But it still lacks a credible enforcement record and a more sophisticated interpretation of the competition law. Therefore, it is necessary to continue to reform the efficiency of the CPC’s institutional and administrative organization. Further education of case handlers and the employment of experienced economists and lawyers with deep knowledge of competition law and economics can contribute to increasing the number of cases before the CPC and to improving the quality of the assessment and the establishment of reliable practice. However, it is not easy in Serbia to find experts in com32 CPC v Amm Immovables d.o.o. (2015), http://​www​.kzk​.gov​.rs/​kzk/​wp​-content/​uploads/​2014/​03/​ povreda​-konkurencije​-frikom​.pdf. 33 CPC v Frikom a.d. (2012), http://​www​.kzk​.gov​.rs/​kzk/​wp​-content/​uploads/​2014/​03/​povreda​ -konkurencije​-frikom​.pdf.

Building an efficient system of protection of competition in Serbia  293 petition law and economics because competition policy is relatively new in Serbia and, as indicated earlier, the universities still do not recognize the importance of competition law courses in their regular programmes. Capacity-building programmes financed by the EU and other international organizations are still necessary for the development of the CPC’s work. The Commission now has about 30 case handlers, currently divided into two departments: Merger Control and Antitrust. This is sufficient for the internal division of tasks and to organize departments for specific industries. The grouping of case handlers enables them to acquire experience and knowledge in one specific industry. This, in turn, allows the CPC to deal with cases from that particular industry or sector of the economy in an efficient manner. A review of the CPC’s enforcement establishes that merger cases (notably, extraterritorial mergers) comprise the largest group of cases. In addition, there is an increase in the number of sectoral inquiries, which is a significant step forward for the CPC. Sectoral inquiries allow the CPC to identify sectors where there is a competition issue, either structural or other, and to take measures – such as to initiate investigations or to advise the government and other institutions to take action to improve competition conditions in that market. The CPC has also increased its competition advocacy activities. However, non-enforcement activities cannot replace antitrust investigations and infringement decisions. The low level of fines,34 even for the most serious violations, and the ineffective leniency programme prevent the CPC from discovering and sanctioning cartels in Serbia. In the period 2016–18, the CPC adopted nine competition infringement decisions, of which two relate to the abuse of a dominant position and seven relate to restrictive agreements. So, the CPC must define its annual priorities carefully in order to maximize its administrative potential within its legal competencies. It should clearly define feasible plans and priorities. The anti-cartel battle and severe sanctions for anticompetitive behaviour should be at the top of its priorities. The progress of the CPC’s enforcement and organization, along with further alignment with the EU legal framework, represents a cornerstone of the development of competition law in Serbia. With strategic prioritization of its enforcement, the CPC could became a good example of the successful agency that helps to speed the economic processes in the emerging market economy and to improve the welfare of consumers in Serbia. To achieve this goal, its work should be constantly evaluated, internally and externally, in order to assess whether the enforcement of the competition law reflects the needs of the market and the expectations of the CPC and other relevant stakeholders.

34 CPC v N Sport and sport retailers (2017) (resale price maintenance): the CPC imposed a fine of only 0.85 per cent of annual turnover, http://​www​.kzk​.gov​.rs/​kzk/​wp​-content/​uploads/​2017/​12/​R​-40​-02​ -892017​-312​.pdf; CPC v Inter Turs Plus (2017) (transport company). For the abuse of dominant position (excessive pricing), the fine was 1.3 per cent of annual turnover, http://​www​.kzk​.gov​.rs/​kzk/​wp​-content/​ uploads/​2017/​11/​Resenje​-InterTursPlus1​.pdf.

13. Merger review updates in Latin America Fernando M. Furlan

I INTRODUCTION According to Global Competition Review, Latin America has three competition authorities listed for enforcement rating: Brazil, Chile and Mexico.1 Brazil has four stars (up to five), and its competition enforcement is considered very good. Chile and Mexico both have three-star ratings and their competition enforcement is judged good. Each of the three countries has done a remarkable job in building competition awareness, contemporary legislation, functional systems and solid institutions. Brazil changed its Competition Law in 2011 and adopted a more contemporary and efficient merger review and investigation system. The country shifted from a three-agency structure to a single structure, the Administrative Council for Economic Defense (CADE), but also maintained a government department in charge of advocacy, the Secretariat for Economic Monitoring, within the Ministry of Economy. All competition decisions are made by an administrative tribunal subject to judicial review by a first-level federal judge. Chile has evolved from a voluntary review system to a mandatory one, based on a threshold notification system, and has developed a solid, efficient and specialized competition judicial review structure. Mexico has also made important modifications to its competition system. The Federal Competition Commission is now a solid institution in Mexico’s economy. Before proceeding, it is important to remember that other Latin American jurisdictions have made tremendous advances in building their competition regimes and institutions. According to Peña,2 the development of competition law in Latin America can be divided into three stages: (i)

a ‘preliminary stage, when a few countries in the region had some basic and vague legislation with poor enforcement’; (ii) a ‘second stage, when most Latin American countries either modernized their existing competition laws or introduced modern ones with the guidance of international organizations’; and (iii) a ‘stage of consolidation of the existing regimes … with the introduction of substantial amendments to their laws based on their own experience and on the ideas taken from the best practices agreed at the international forums, with significant debates and strong political support from their governments’.

Global Competition Review, Rating Enforcement (2015). Julian Peña, ‘The Consolidation of Competition Law in Latin America’ (2011) 11 Antitrust Chronicle, https://​www​.competitionpo​licyinternational​.com/​the​-consolidation​-of​-competition​-law​-in​ -latin​-america/​. 1 2

294

Merger review updates in Latin America  295 There has been consolidation of the existing regimes, with the introduction of substantial amendments to laws based on experience and on ideas taken from the best practices agreed at international forums, with significant debates and strong political support from their governments. This chapter briefly analyses the Chilean and Mexican competition laws and policy and then turns to a more detailed analysis of Brazil in the context of these comments.

II CHILE Since 1 June 2017, a new mandatory pre-merger control regime has been in force in Chile, replacing the old voluntary filing system. The pre-merger control system makes it compulsory to notify any projected transaction to the National Economic Prosecutor’s Office (FNE) when certain turnover thresholds are met. The FNE then examines whether the transaction substantially reduces, or has the potential to reduce, competition in the market. It is, therefore, mandatory to notify any transaction in Chile that (i) has an impact in Chile; and (ii) meets the two-pronged thresholds test. That test is based on turnover, not on market share: ●● aggregate turnover of UF 2,500,000 (approximately US$100 million) of the merging parties; and ●● individual turnover of UF 450,000 (approximately US$18 million) of at least two of the merging parties (the relevant turnover corresponds only to sales in Chile). In cases where it is not obligatory to notify a transaction, the companies may voluntarily notify the FNE to apply for clearance, provided the transaction has not been implemented at the time of the notification. In such cases, the same rules as for mandatory notifications apply, though less burdensome information is required. If these transactions are not notified voluntarily, the FNE may still investigate them within one year of their completion. The merger control regime also provides for a simplified merger notification, including cases where the projected operation does not produce vertical or horizontal overlaps. The simplified notification basically means that less information is required when notifying the operation. A post-closing notification requirement exists when a 10 per cent interest or more is acquired from a competitor. Penalties may be applied by the Tribunal for the Defence of Free Competition (TDLC) for infringing a notification obligation, implementing a transaction during the suspension period, not complying with the conditions subject to which the operation was approved, implementing an operation that has been prohibited, and notifying an operation supplying false information for its appraisal. The penalties are the same as for any other anticompetitive conduct. The FNE is responsible for the mandatory pre-merger notification merger control regime, which involves a two-stage investigation process with increased fines for breaches of the rules. The regime also includes: ●● an obligation to suspend a transaction closing prior to clearance; ●● a two-phase review period of 30 business days (for initial review) and 90 business days (where an in-depth review is required); and ●● fines for failure to notify or unauthorized completion of approximately US$15,000 per day.

296  Research handbook on methods and models of competition law The two specialized Chilean competition authorities, the FNE and the TDLC, enforce the competition law. The FNE is the administrative agency responsible for the investigation and prosecution of infringements of competition. It also provides technical support to the TDLC and supervises compliance with decisions of the TDLC. The FNE is an independent institution that functions under the responsibility of the Ministry of Economy and acts in the general interest of the public. In relation to mergers and acquisitions, the FNE can initiate an investigation to assess a transaction’s compatibility with competition. Such an investigation may start ex officio upon a submission by the parties to the operation or upon a third party’s complaint. Depending on the results of the investigation, the FNE may close the case without further action, initiate a procedure before the TDLC, or negotiate a settlement, which is subsequently submitted for approval to the TDLC through the fast-track procedure. The TDLC is an independent specialized tribunal with the status of a court that hears and decides competition cases in the first instance. As such, it is competent to decide on merger cases submitted by the parties themselves, the FNE or third parties in non-contentious procedures, as well as claims filed by the FNE or third parties that are allegedly affected by an executed operation in contentious procedures. The TDLC is also authorized to approve or reject the outcome of negotiations between the merging parties and the FNE in the fast-track procedure. The TDLC comprises five expert members: three lawyers and two economists. Decisions by the TDLC regarding operations of concentrations in non-contentious or contentious procedures are subject to appeal to the Supreme Court. Appeal is open to all parties that were part of, or had intervened in, the merger review process before the TDLC. The Supreme Court analyses both issues of fact and law; more specifically, it analyses whether the decisions made were well motivated and whether the remedies required by the TDLC to overcome the competition risks are proportional and adequate. A Notification The notification of mergers and acquisitions is mandatory and must be submitted before closing the transaction. However, the TDLC may analyse a merger prior to its implementation, as well as post implementation, upon the request of the FNE and/or third parties with sufficient interest. Essentially, there are three procedural scenarios: (i) pre-merger consultation with the TDLC; (ii) pre-merger consultation with the FNE, whereby the negotiated outcome may be submitted for approval by the TDLC in a fast-track procedure; or (iii) implementation of the transaction without consulting any of the authorities, thereby assuming the risk of a non-contentious or contentious procedure (depending on whether or not the operation is materialized) initiated by the FNE or a third party. B

Pre-merger Consultation with the TDLC

As stated previously, parties to a transaction that potentially raises competition concerns may file a voluntary consultation with the TDLC. The FNE and third parties with sufficient interest can also initiate a non-contentious procedure. The TDLC may approve the operation, approve the operation under conditions (remedies), or block the operation.

Merger review updates in Latin America  297 C

Pre-merger Consultation with the FNE (Fast-track Procedure)

Parties to the merger can inform the FNE about an anticipated merger and, where there are potential competition risks according to the FNE, negotiate an extra-judicial agreement, which will normally concern mitigation measures to neutralize any such potential risk. An agreement that reflects a negotiated outcome is subsequently submitted to the TDLC for approval in the so-called fast-track procedure, which may take between one and two months. Such a decision is not open to appeal to the Supreme Court. D

Contentious Procedure Post Merger

A contentious procedure can be initiated post merger by the FNE or any third parties with sufficient interest, provided that the operation was not assessed by the TDLC in a non-contentious procedure. The TDLC can order preventive measures and, by final decision, can order the parties to terminate the operation, modify the operation, and/or impose a fine of up to approximately US$20 million. In general, parties to a merger with effects in Chile do not consult with the TDLC, especially given the length of the procedure. A procedure before the TDLC can take up to eight months or more, with an additional four to six months in the case of an appeal to the Chilean Supreme Court. There is no possibility of negotiating remedies with the TDLC and/or the Supreme Court. The parties can suggest or offer conditions or modifications, but it is the final decision of the courts that determines whether to accept the suggested remedies or whether to impose other remedies. Only in cases where the operation raises serious competition risks that need complex mitigation measures may the companies involved opt for a consultation pre-operation with the TDLC. In other cases, companies generally choose to implement without consultation or – in a case where the operation raises certain competition issues that, relatively speaking, can be easily remedied – to submit the operation to the FNE and negotiate the eventual mitigation measures. An additional option, as a matter of good business practice and to reduce the risk that the FNE may initiate an investigation ex officio merely based on public information, is to approach the FNE informally regarding the transaction, giving the parties legal and economic reasoning as to why the transaction does not raise competition issues and/or the eventual remedies to mitigate any potential risk. Therefore, the decision whether to approach the TDLC or the FNE essentially depends on whether the operation concerned raises competition issues, or whether it is likely that the FNE will start an investigation ex officio or a third party will file a complaint with the FNE and/or initiate a non-contentious or contentious procedure with the TDLC. E

Failing to Notify

According to Article 48 of the Competition Law,3 parties that fail to notify a concentration are subject to a sanction of up to 20 UTA (approximately US$16,500) for each day from the completion of the transaction. Fines are, in principle, made public. The FNE initiated one



3

Decree Law 211 of 1973.

298  Research handbook on methods and models of competition law such procedure regarding the acquisition of a chain of movie theatres, requesting, in addition to divestitures, a sanction of approximately US$1 million on each party. The case was subsequently settled and no fine was imposed.4 F

Types of Transactions

The Competition Act does not define the type of transactions that may be subject to a merger review. The TDLC and the FNE generally refer to ‘operations of concentrations’ or ‘concentrations’, a concept that is interpreted rather broadly. The TDLC defines operations of concentration as acts or agreements through which an independent entity merges or acquires decisive influence in the management of another independent entity, or as a result of which two or more such entities jointly participate in a business or establish a joint venture, thereby reducing the independence between them in a significant and structural manner. The FNE’s Merger Guidelines identify the following as operations of concentration or concentrations: mergers, acquisitions of shares or assets, associations, and, in general, acts or agreements with the objective or effect that two or more economically independent entities form one single entity taking common decisions, or become part of one and the same group of companies. The Guidelines state that the relevant issue is the change of competitive incentives, which is an economic concept. The principles of merger analysis also apply to agreements between competitors that decrease the level of autonomy or modify the form in which competitive decisions are made and, as such, produce effects on the market structure that are similar to a concentration. In general, mergers, acquisitions of controlling interests, and joint ventures are regarded as ‘concentration’. Restructurings or reorganizations within a company or a group of companies are unlikely to qualify as concentrations, although there are no formal precedents on this issue. ‘Control’ is not defined in the Competition Act. In relation to issues of decisive interest or control, the authorities generally rely on the definition of ‘controlling interest’ provided in the Securities Market Law. That law defines the controller of a company as any person or group

4 On 27 June 2012, the FNE filed a complaint before the TDLC to challenge the acquisition by Chile Films SA of the Hoyts cinema chain, which was the second-largest player in the market. Prior to the transaction, Chile Films owned Cinemundo, the third-largest cinema chain in Chile. With the deal, Chile Films has become the leading player in the market. This was a landmark case in Chile’s competition law since it was the first time that the FNE had filed a complaint to challenge a closed merger. In its complaint, the FNE concluded that the relevant product market involved was the exhibition of first-run commercial movies in multiplex movie theatres. Regarding the relevant geographical market, the FNE considered the areas in which the merger produced a significant overlap and – according to international standards – used a radius of 6 kilometres to assess the actual and potential effects of the transaction. The FNE concluded that the merger produced anticompetitive concerns in three zones of the capital Santiago and in the whole city of Valparaiso. The FNE requested that the TDLC impose fines totalling US$2 million against the buyers, due to their dominant position in the market and the fact that they knew and foresaw that the merger would create actual or potential anticompetitive effects. Despite such circumstances, the buyers did not decide to file a consultation before the TDLC in advance of closing. Neither asked the FNE to review the transaction. In addition, the FNE requested that the TDLC issue a divestiture order concerning the three areas (two in Santiago and one in Valparaiso) in which the merger produced the most serious anticompetitive concerns.

Merger review updates in Latin America  299 of persons with a joint action agreement that, directly or through other individuals or legal entities, participates in the ownership and has the power to: ●● ensure the majority of votes in the shareholder meetings, elect the majority of the directors, and appoint the administrator; or ●● decisively influence the management of the company. A ‘decisive influence’ in the management of a company is defined as the ability to control at least 25 per cent of the voting capital of the company. However, if the following exceptions apply, the management will not be considered to hold a decisive influence: ●● another person, or group of persons, controls an equal or larger percentage of the voting capital of the company; ●● the management does not control more than 40 per cent of the voting capital of the company and the controlled percentage is lower than the sum of the holdings of the other partners or shareholders with more than 5 per cent of the equity; or ●● when the Chilean Securities and Insurance Supervisor so determines in consideration of the distribution and dispersion of the ownership of the company. Resolution 667, issued by the FNE, established that a transaction must be notified when the two following conditions are met: ●● the sum of sales in Chile of the agents subject to a concentration operation are equal to or exceed UF 1.8 million (approximately US$70 million) during the last fiscal year (calendar year) in which the notification is verified; and ●● the sales in Chile of at least two of the agents subject to a concentration operation are, separately, equal to or exceed UF 290,000 (approximately US$11 million) during their last fiscal year in which the notification is verified. Foreign-to-foreign transactions may be subject to review where the operation influences any market in Chile. Effects on the Chilean market are less likely if the target has no sales in the jurisdiction. Both the TDLC and the FNE explicitly mention joint ventures as an example of operations of concentration5 that may be subject to merger control review. No distinction is made between types of joint ventures. Although there are very few precedents on joint ventures, it can be assumed that the decisive element is whether a joint venture causes a structural change to the competitive behaviour of previously economically independent companies. The FNE understands that joint ventures should be assessed as concentrations rather than as horizontal agreements where the effects are structural in nature. The TDLC has been adopting a methodology similar to that of the European Commission when assessing ‘full-function’ joint ventures with possible coordinative effects between the parent companies.6



Joint ventures are, in fact, considered to be operations of cooperation, not concentration. See Council Regulation (EC) No 139/2004, https://​www​.legislation​.gov​.uk/​eur/​2004/​139/​pdfs/​eur​ _20040139​_adopted​_en​.pdf. 5 6

300  Research handbook on methods and models of competition law G

Powers to Investigate a Transaction

The TDLC can review any merger post implementation upon the request of the FNE and/or a third party, provided that the transaction was not submitted and cleared in a non-contentious procedure prior to implementation. The statute of limitation for filing a claim post operation is not entirely clear but is likely to be three years from the time at which the transaction was implemented. H

Deadlines for Notification

Once the jurisdictional thresholds are met, the notification is compulsory and must be filed before closing the transaction. A formal binding agreement is not required prior to notification. The TDLC requires the parties to submit a document that describes the transaction and shows that the intention to complete the operation is ‘real’ and ‘serious’. A letter of intent or memorandum of understanding will generally meet this criterion. I

Phases of the Review Process

A non-contentious procedure with the TDLC provides for one phase. After filing, the TDLC invites interested parties to provide information. Any parties with a legitimate interest can join the procedure. All parties to the procedure are given the opportunity to submit additional arguments and information (such as expert reports). The FNE issues a report regarding the compatibility of the operation with the Competition Act. The TDLC subsequently calls for a public hearing. There are no set timelines for these stages after publication of the submission by the TDLC.7 The fast-track procedure is generally shorter. The FNE’s Merger Guidelines provide for a term of 60 days between the submission of the information and the agreement, but the term may be extended by agreement between the parties. The 60-day term will start running once the parties allow the FNE to make the transaction public. The agreement is reviewed in a single hearing within five days of the receipt of the information (which may be extended if the TDLC requests additional information). The TDLC will make its decision – either an approval or rejection of the agreement – within 15 working days of the hearing. Although the timelines are rather strict, the procedure generally takes between one and two months from the submission of the agreement. The fast-track procedure has certain shortcomings. The decision is binding on the parties to the agreement only and, theoretically, third parties can challenge the transaction post implementation before the TDLC. More importantly, third parties can frustrate the procedure by challenging the operation before the agreement is submitted to the TDLC. There is no possibility for pre-notification discussions with the TDLC or its support staff. The parties may engage in discussion with the FNE, either prior to a voluntary submission to the TDLC or to discuss the possibility of the fast-track procedure. The FNE generally welcomes these discussions in the pre-submission stage. 7 According to TDLC statistics, the average duration of non-contentious procedures (which include both merger cases and other competition cases) is around eight months.

Merger review updates in Latin America  301 J

Accelerated Procedure for Review

As indicated above, the Competition Act provides for a fast-track procedure whereby the parties notify to the FNE and, if necessary, negotiate a settlement, which is accordingly submitted to the TDLC for approval. The TDLC is not bound by the recommendation that has been accepted by the party or parties to the operation. K

Substantive Test

The substantive test applied by the TDLC and the FNE is to establish whether the market power of the merged entity will increase and, as a result, the risk of abuse of market power or the risk of coordinated practices will increase. Chilean authorities apply a rather broad definition of market power and often focus on the effects on the market dynamics and market structure, rather than on the actual risk of abuse or collusion post operation or the causal link between the operation and that risk. This test is closer to the test for substantial lessening of competition than to an ‘abuse test’. Indeed, the FNE has informally stated that the substantive test is designed as a risk-effect test, like the substantial lessening of competition test, rather than as a structural test such as the dominance test. The anticompetitive risks analysed by the authorities are unilateral effects and coordinated effects, in addition to potential foreclosure effects resulting from vertical integration. Under the unilateral effects analysis, the authorities assess whether the loss of competition between the parties enables the merged group to exercise market power and increase prices. Coordinated effects are assessed where the market structure is such that companies would consider it possible, economically rational, and therefore preferable to adopt a course of action aimed at selling at increased prices on a sustainable basis. For the TDLC, it is irrelevant whether the possible coordination is explicit or tacit. It is also not necessary that the companies in the coordinating group adopt any of this conduct; it is sufficient that they have the ability and incentive to do so, which puts the competitive market structure at risk. So far, there have been no decisions regarding pure conglomerate mergers, product extension mergers, or operations that may eliminate a potential competitor. It is most likely that the authorities will consider conglomerate effects and the elimination of potential competition if the issues are raised by future cases. In Falabella/D&S,8 a controversial merger case in Chile, the TDLC had to decide on a proposed merger between two of the main Chilean retail companies. The merger combined the main operator of department stores with the largest supermarket chain in Chile. The TDLC adopted a concept of ‘integrated retail’, which considered the market power derived from the integration of complementary activities.9

Resolution No 24 – Case NC 199-07. On 31 January 2008, the TDLC decided not to clear the merger of Falabella and D&S. The Tribunal’s main arguments to justify its decision were that (i) the materialization of the merger would produce an enormous change in the market structure, creating a firm that would be the dominant actor in integrated retail and in practically all its segments – department stores, home improvement stores, supermarkets, real estate, and credit cards – and that would have the capacity to transfer this market power to 8 9

302  Research handbook on methods and models of competition law The FNE provided guidance on its assessment method in its 2012 Horizontal Merger Guidelines, indicating in general terms the fundamentals, principles and methodology used in the case of an investigation into a concentration conducted by the FNE or in the case of a request for an opinion by the TDLC. The method and the factors taken into consideration are comparable to those used by other competition authorities in, for example, the EU and the US. Chilean authorities do, moreover, use international precedents and doctrine for guidance. L

Prohibition of Transactions

In the case of a review in a non-contentious procedure, initiated by the merging parties or by the FNE and/or third parties showing sufficient interest, the TDLC can clear the operation (by declaring that the operation does not violate the Competition Act), clear the operation subject to conditions, or prohibit the operation. In the case of a review in a contentious procedure ex post, the TDLC can decide that the operation qualifies as a violation of the Competition Act and order the unwinding or modification of the transaction. In the fast-track procedure, the TDLC can either reject or approve the extra-judicial agreement signed by the parties or between the party and the FNE. The TDLC must state the facts, law and economic basis upon which its decision is based. The competition risks should be sufficiently real, imminent, severe, serious and likely. The TDLC has prohibited a transaction on two occasions (Falabella/D&S and Quiñenco/ Terpel Chile10) although the Supreme Court overruled the latter decision approving the deal under conditions. M

Negotiation of Remedies

None of the merger review procedures provides for the negotiation of remedies with the TDLC. The parties can propose remedies when voluntarily filing a transaction or during a procedure. The TDLC will consider these proposals but is free to impose any conditions of operation on any remedies that it deems necessary. There is some possibility of negotiating remedies with the FNE. Prior to or during a non-contentious procedure before the TDLC initiated by a voluntary filing, the parties may discuss the issue of remedies with the FNE. The merging party or parties may also settle a contentious case initiated by the FNE, whereby the settlement is likely to be related to com-

other retail areas in which it could choose to enter in the future; (ii) in that scenario, given the size of the Chilean economy and the relevance of the barriers to entry in the different business areas that form the integrated retail market, it is unlikely that an entrant, if there is one, can impose competitive pressure in a reasonable period of time; (iii) if the merger had been approved, there would have been a substantial and lasting decrease in the competition conditions in a market that represents a very relevant part of the consumption decisions of Chileans, with expected harmful effects – in terms of welfare – over prices, quantities and quality of the products involved; (iv) the efficiencies and synergies that the merger parties presented were not well enough proved and, even if they had been, they would not have complied with the minimum requirements to be accepted in a concentration analysis because they could not have compensated for the anticompetitive risks that the operation would have generated, had it been approved; and (v) the Tribunal stated that there are no conditions or mitigation measures that are sufficient and effective enough to compensate or minimize the risks to free competition of the operation. 10 Resolution No 39 – Case NC 399-11.

Merger review updates in Latin America  303 mitments that the party or parties accept to mitigate potential competition concerns. However, as was noted earlier, the recommendations and settlements are not binding on the TDLC. Negotiations on remedies are more effective in the fast-track procedure. The TDLC can either approve the negotiated outcome or reject it. However, it lacks the authority to modify the agreement reached. N

Standard Approach for Divestitures and Other Remedies

There are no guidelines regarding best practices for divestitures or other remedies. Recent decisions by the TDLC and opinions issued by the FNE provide guidance as to the authorities’ position towards conditions and timing and demonstrate that their practice regarding divestitures is moving towards the best practices developed by the European Commission and the US Department of Justice. Completion prior to materialization of a divestiture has been allowed, with transitional measures and conditions imposed, to secure the value and competitiveness of the business to be divested and to avoid coordination during the divestiture period – for example, an administrative hold separate. The TDLC may, in certain cases, find that divestitures are to be carried out by means of auction, but on other occasions the parties may be left free to negotiate with potential buyers. Conditions in respect of the potential buyer relate to issues such as sufficient financial recourse, experience, and independence from the merged group. Until now, the TDLC and the FNE have not conditioned the divestiture on prior approval of the buyer, although in certain cases involving concentrated markets, and where the selected purchaser business was an existing competitor, the TDLC set a condition that the divestiture was to be notified to the court for prior review. Divestiture periods range from six to 18 months, depending on the complexity of the sale, whether the operation will be implemented pending divestiture, and whether the sale is part of a wider global package. These periods may be extended upon a reasonable request from the party or parties concerned. The FNE will monitor and supervise compliance with the measures adopted. In recent cases, the authorities have insisted on the appointment of a monitoring trustee to facilitate the FNE’s task. O

Appeals and Judicial Review

Decisions by the TDLC regarding mergers are subject to appeal before the Supreme Court. Appeal is, in principle, open to all parties that were part of, or intervened in, the merger review process before the TDLC. P

Partial Conclusions

Chile has just approved and regulated its mandatory merger notification system after decades of voluntary submissions. It is a great challenge for Chile’s competition institutions, such as the FNE and the TDLC, as they launch brand new merger investigation procedures. Thresholds and timing are deemed fair for the country’s economic idiosyncrasies.

304  Research handbook on methods and models of competition law

III MEXICO Article 10 of the Mexican Federal Economic Competition Law (Economic Competition Law) provides that the Federal Economic Competition Commission (FECC) is an autonomous entity with its own legal personality and patrimony, independent in its decisions and operation, professional in its performance, and impartial in its actions. Its purpose is to guarantee free market access and economic competition, as well as to prevent, investigate and combat monopolies, monopolistic practices, concentrations and other restrictions to the efficient functioning of the markets. Mexico’s 2014 Competition Law formalizes the implementation of important amendments to the Mexican Constitution that primarily cover telecommunications, media and antitrust. The Economic Competition Law aims to restore the structure of the antitrust regulator and update Mexican antitrust regulations by adopting Organisation for Economic Co-operation and Development (OECD) practices and standards, as well as tracking and satisfying the mandates of the constitutional amendments. The new law (i) creates two regulators with broad new powers; (ii) implements a new regulatory regime for general antitrust matters; (iii) implements a special framework for antitrust matters related to telecommunications and media; and (iv) creates specialized courts to resolve antitrust and concentration matters. The two regulators are: ●● the FECC, which oversees all matters except telecommunications and media; and ●● the Federal Telecommunications Institute (IFT), the successor to the Federal Telecommunications Commission, which oversees all matters related to telecommunications and media. The FECC is vested with broad authority and powers to eliminate barriers to competition, mandate divestiture of assets, and issue regulations and opinions. It has power to investigate and determine the absence of effective competition, the existence of barriers to competition and free access or essential resources (covering inputs and facilities) that may generate anticompetitive effects in a market, and the FECC’s authority to resolve the same, even ex ante, by imposing corrective structural or behavioural interim measures, including the determination of access conditions and prices. Except in those cases where the FECC orders the divestiture or sale of assets, rights or equity interests, the implementation of the FECC’s resolutions is not subject to a judicial stay pending the outcome of litigation. Preventive measures may be imposed but may be suspended if the economic agent provides adequate guarantees. An Investigation Unit within the FECC conducts antitrust (anticompetitive practices and concentrations) inquiries and related analyses. If the unit concludes that there is a case to be pursued, it can open an administrative process in which it will act as the prosecutor, the alleged violators will be defendants, and the FECC will preside over and decide the case.

Merger review updates in Latin America  305 A Concentrations Concentrations that meet the thresholds set forth in Article 86 of the Economic Competition Law must be authorized by the FECC before they are carried out: ●● when the transaction or series of transactions that gives rise to them, regardless of the place of execution, imported into the national territory, directly or indirectly, more than the equivalent of 18 million times the general minimum daily wage for the Federal District; ●● when the transaction or series of transactions that gives rise to them involves the accumulation of 35 per cent or more of the assets or shares of an operator whose annual assets in Mexico or annual sales originating in Mexico are equivalent to 18 million times the minimum daily wage for the Federal District; or ●● when the transaction or series of transactions that gives rise to them involves an accumulation in the national territory of capital assets that is more than the equivalent of 8.4 million times the general minimum daily wage for the Federal District and there is a concentration involving two or more operators with annual sales originating in the country or assets jointly or separately owned in the country equivalent to more than 48 million times the daily general minimum wage in force for the Federal District. Additionally: ●● the parties may not close a transaction until the FECC clears it (closing after the lapse of the ten-day period without the issuance of a stop order, permitted under the existing Economic Competition Law, is no longer allowed); ●● the FECC is permitted to seek information from third parties and government agencies; ●● the parties forming part of the concentration may propose conditions (related to future structure or administration of the concentrated assets to address the adverse effects of the proposed market concentration) in order to secure the approval of the concentration prior to the FECC ruling; and ●● the FECC clearance will be effective for six months, extendable for one additional six-month period. B Penalties Penalties (administrative and criminal) and fines11 have been strengthened, such as the divestiture of assets, the regulation of essential resources, and the barring of individuals from participating as agents, officers or members of the board of directors because of their direct or indirect participation in illegal antitrust or market concentration practices. Administrative fines are up to 10 per cent of a violator’s revenue, in addition to civil and criminal liabilities exposure.

11 Fines may be imposed on individuals and companies acting as authors or facilitators of anticompetitive behaviour, including, among others, unlawful mergers, not notifying a merger meeting of the mandatory thresholds, not complying with merger remedies or commitments, false testimony, and the submission of false information. Penalties may include, among others, professional disqualification; partial or total divestiture of assets, rights, shares or partnership interests; and prohibition on participating in public procurement.

306  Research handbook on methods and models of competition law C

Partial Conclusions

Mexico has undertaken a national effort in recent years to modernize the country’s legislation, bureaucracy and practices with a view to increasing legal certainty and creating a business-friendlier environment. This has resulted in important amendments to the Constitution and in significant changes in competition policy and enforcement. Among the recommendations of the OECD 2020 Mexico Competition Policy Peer Review are the following, which concern the relationship with the IFT and the challenges of the digital economy: ●● competition authorities should continue to cooperate and clarify through the adoption of guidelines how cases are allocated between the IFT and the FECC; ●● the merger control exemption granted to transactions by non-preponderant players in the telecommunications and broadcasting sector should be eliminated; and ●● the IFT and the FECC should consider further developing the economic analysis of mergers and should rely more on new tools, such as statistical and econometric analysis.12

IV BRAZIL Effective competition law enforcement in Brazil is relatively young as the country’s economy opened to free trade and free markets less than three decades ago. Despite its youth, competition enforcement in Brazil has grown soundly, especially in the last decade or so, with new and successful investigative tools such as the leniency programme, authorized dawn raids, and wiretapping. Competition decision-making in Brazil has shown maturity also by sanctioning state-owned giants, long-standing cartels, and huge and aggressive quasi-monopolists with record fines and other penalties.13 The Brazilian competition authority, CADE, has struggled since 2002 to confirm its prerogatives in the banking and financial sectors against a government decision14 of 2001 that deemed the Central Bank of Brazil (BACEN) responsible for scrutinizing conduct and reviewing mergers and acquisitions of banks and other financial institutions. Finally, CADE and BACEN approved Joint Normative Act No 1 at the end of 2018, which provides for approval by BACEN of mergers and acquisitions linked to maintaining the stability of the national financial system, as well as notifications from CADE to BACEN on administrative procedures aimed at investigating anticompetitive conduct involving institutions supervised by BACEN. In 2018 the Senate approved a bill,15 which, at the time of writing, is under scrutiny in the House, to turn the memorandum of understanding between CADE and BACEN into specific legislation, giving more legal certainty to the market.

12 OECD (2020), OECD Peer Reviews of Competition Law and Policy: Mexico, http://​www​.oecd​ .org/​daf/​competition/​oecd​-peer​-reviews​-of​-competition​-law​-and​-policy​-mexico​-2020​.htm. 13 Petrobrás SA case (Administrative Process No 08700.002600/2014-30); Oxygen and Cement Cartel cases (Administrative Processes No 08012.009888/2003-70 and 08012.011142/2006-79 respectively); AB Inbev (Ambev) case (Administrative Process No 08012.010028/2009-74). 14 Parecer AGU GM – 020, 5 April 2001. 15 PLP 499/2018.

Merger review updates in Latin America  307 Competition enforcement in Brazil underwent a deep change between 2011 and 2012 as the merger review system shifted from post-merger to pre-merger review, bringing many challenges to the country’s competition authorities as discussed below. A brief review of the historical background to the creation of competition law and policy in Brazil, as well as the current institutions and specific legislation dealing with conduct and mergers and acquisitions, follows. In addition, this section analyses Brazil’s competition legislation, which has been effective since 29 May 2012, bringing important changes in relation to the unification of institutions, new merger review mechanisms, and new criteria for the submission of transactions and for imposing fines and other sanctions. A

Historical Background

The history of competition law and policy in Brazil dates from 1962, when CADE was created through Law No 4.137 as a consulting government commission with powers to present recommendations to the executive branch, as at that time the state-controlled prices in most sectors and many of the country’s largest enterprises were either state-owned or publicly sanctioned private monopolies. Law No 8.884/94 transformed CADE into an independent administrative tribunal with extensive new powers. Its decisions became final and appealable only to the courts. The promulgation of the new competition law in Brazil followed major economic changes brought by the 1988 Constitution and the ‘Real Plan’, which created a new currency and implemented strict fiscal policies, as well as reopening the economy through the elimination of international trade barriers and launching a broad privatization process. All of these economic adjustments carried intensive and progressive in-flow of foreign capital into the country which, among other effects, culminated in many mergers and acquisitions. The Brazilian Constitution provides an unequivocal basis for competition policy. Article 173, paragraph 4 provides that ‘[t]he law shall repress the abuse of economic power that aims at the dominance of markets, the elimination of competition, and the arbitrary increase of profits’. Article 170 provides that the ‘economic order’ shall be ‘based on the appreciation of the value of human labour and on free enterprise’. It establishes that some principles must be respected, including ‘free competition’, ‘consumer protection’, ‘private property’ and ‘social role of property’.16 B

2011 Brazilian Competition Law

On 30 November 2011, Law No 12.529 was enacted, restructuring the Brazilian Competition Policy System (BCPS). By its provisions, the current CADE is comprised of two main bodies: one decision-making, the Competition Tribunal, and the other providing instructions and recommendations, the Superintendence. These replaced the Secretariat of Economic Law. In addition to these structural changes, the new law redefines the functions of the Secretariat for Economic Monitoring within the Ministry of Economy (SEAE), focusing on the activities of 16 Inter-American Development Bank and Organisation for Economic Co-operation and Development, ‘Competition Law and Policy in Brazil: A Peer Review’ (2005), http://​www​.oecd​.org/​dataoecd/​12/​45/​ 35445196​.pdf.

308  Research handbook on methods and models of competition law ‘competition advocacy’, including within the public sector. The ‘new CADE’ and SEAE form the new BCPS. C

Pre-merger Analysis System

One of the main changes brought by the 2011 law was the introduction of the pre-merger analysis system, in line with most world-leading antitrust jurisdictions. Beforehand, transactions had to be notified to competition authorities up to 15 business days after consummation (binding agreement). The pre-merger system avoids legal uncertainty and transactional costs associated with any decision by CADE to divest in a scenario where acts subsequent to the transaction were already accomplished and implemented. Companies involved in a merger or acquisition seek economic advantage of synergies, making a separation of operationally complex business processes and routines more difficult. Intuitively, it is also more difficult to reverse a merger or acquisition undertaken than to prevent it beforehand – that is, when it is just a set of intentions and understandings. D

Agility, Legal Certainty and Economic Timing

Non-complex merger cases are initially scrutinized by CADE’s General Superintendent and may be approved without consulting the Tribunal, except for the possibility of a call-back of the case by a member of the Tribunal. If there is no call-back, Tribunal involvement will only be triggered when the Superintendent concludes that the operation generates, or may generate, harm to competition and therefore requires some state intervention. In these circumstances, the case will be sent to the Tribunal for a final decision. Another significant change introduced by the 2011 law is related to the role of the Federal Prosecutor’s Office. Under the system, the Office’s opinion on merger and acquisition cases will no longer be mandatory. That, however, does not preclude the general participation of the Office in line with its institutional functions and prerogatives, including the protection of diffuse interests of society. An additional important innovation is the creation of the Department of Economic Studies of CADE, serving both the Superintendent and the Tribunal. This legal innovation responds to the growing sophistication of antitrust economic analysis, often permeated by complex econometric opinions and studies. The 2011 law also strengthens the independence and autonomy of the members of the Tribunal and the Superintendent. The terms of the President and board members of the Tribunal are extended from two to four years, removing the possibility of renewal. The terms of board members (or commissioners) will be unmatched to allow substitutions to occur in order to preserve as much institutional integrity as possible. The Superintendent, in turn, will have a renewable two-year term. E

New Criteria for Notification of Mergers and Acquisitions

A change offering more flexibility in the analysis of mergers and acquisitions was the amendment of the criteria for notification of transactions to reduce the number of transactions submitted to CADE. Prior thresholds led to an excessive number of notifications with many

Merger review updates in Latin America  309 having little competitive impact. Past law provided that operations in which any of the participants show revenues exceeding BRL 400 million (approximately US$115 million) in Brazil, or that result in a concentration equal to or greater than 20 per cent of the relevant market, should be submitted. Under the 2011 legislation, the criterion of market share was removed and ‘additional criteria’ on revenue were introduced whereby the second (or other) parties to the operation must have Brazilian revenue of at least BRL 75 million (approximately US$17 million) and the main party must have revenue of BRL 750 million (approximately US$170 million). That is, currently any operation performed by a large economic agent, albeit in conjunction with one or more other small economic agents, should be compulsorily notified to CADE for review. With the new criteria, operations without any impact on the market are no longer notified, saving public and private resources and time. However, as a safeguard to CADE, within one year from the date of consummation of a transaction it may require the submission of operations that do not fit the revenue criteria described generically by the law.

V CONCLUSIONS Competition policy in Latin America has evolved extraordinarily in the last decades. The three Latin American jurisdictions briefly analysed here have done noteworthy work in building competition consciousness, up-to-date laws, well-designed systems and recognized institutions. It is important to note that other Latin American jurisdictions, even though not discussed here, have also made tremendous advances in building their competition regimes and institutions. Chile has finally adopted a mandatory merger notification and review system17 as its economy is becoming more complex and concentrated. This is a very significant change in the Chilean competition model and is expected to bring more control and supervising powers to this increasingly sophisticated South American economy. Mexico approved a new bill formalizing the implementation of important amendments to the Mexican Constitution aimed at modernizing its economy, bureaucracy and institutions. Those changes primarily cover telecommunications, media and antitrust and are already in effect. The two new regulators, the FECC and the IFT, have broad and new powers. They must also implement a new regulatory regime for general antitrust and for telecommunications and media, as well as create specialized courts to resolve competition matters. That will allow Mexico to deliver important results on sound competition enforcement and advocacy activities.18 Brazil has celebrated eight years of a renewed institutional framework, with a bigger and more powerful CADE and the awaited pre-merger notification system, both through the prom-

17 Ferrada Nehme, ‘Mergers in Latin America: A Chilean, Mexican and Brazilian Approach’ (2018) 1(1) Macau Journal of Brazilian Studies, http://​www​.chambersandpartners​.com/​guide/​practice​-guides/​ location/​272/​8474/​2412​-200. 18 Charles McConnell, ‘Competition Courts Helped Open Mexican Economy, Says Judge’, Latin Lawyer (14 June 2017), https://​latinlawyer​.com/​article/​1142840/​competition​-courts​-helped​-open​ -mexican​-economy​-says​-judge.

310  Research handbook on methods and models of competition law ulgation of Law 12.529/11. With such changes, Brazil was able to eliminate its M&A backlog and allow a larger portion of its taskforce to focus on anticompetitive practices such as bid-rigging.19 Beyond legislative changes, it is crucial that Brazilian competition enforcers and the business community adopt the spirit and fundamentals of this new legislative approach. Examples and experiences from other jurisdictions were very helpful, as was a constructive, open and respectful dialogue between enforcers and business representatives, as well as clear and extensive guidelines. The transition period from the old legislation and practice to the new system was crucial to its success. Challenges include continuous focus on investigation and scrutiny of massive conduct cases.20 Comparing the recent advances of the three jurisdictions, one may conclude that, regarding the specialized judicial review of competition cases, Brazil is still far behind its two Latin American neighbours, even though it has initiated a process of specializing its federal judges in matters such as competition, regulation and trade.21 The Brazilian and Chilean developments in merger review systems deserve praise for they were long wished for, each for their own reasons: Brazil for a pre-merger system instead of a post-merger one, and Chile for a mandatory merger notification system. Finally, Mexico’s example shows how the country is able to adopt a national effort to modernize its institutions and legislate towards a friendlier approach to business and investment.

19 John M Connor and C Gustav Helmers, ‘Statistics on Modern Private International Cartels, 1990–2005’ (American Antitrust Institute, Working Paper No 07-01), http://​papers​.ssrn​.com/​sol3/​papers​ .cfm​?abstract​_id​=​944039​#PaperDownload; Gisela Ferreira Mation, ‘As Ações Civis para Cessação e Reparação de Danos Causados por Condutas Anticoncorrenciais no Brasi’ (III Prêmio SEAE/2008). 20 Lava Jato cartel case; São Paulo City Metropolitan Transport System cartel case; Brasilia region fuel cartel case, among many others. 21 Fernando de M Furlan, Especialização Judicial (Singular 2017).

PART III PARTICULAR ISSUES

14. The interface between intellectual property rights and competition law: implications for public health in sub-Saharan Africa Mor Bakhoum

I INTRODUCTION This chapter deals with the interface between intellectual property rights (IPRs) and competition law. It emphasizes the dual role of competition law vis-à-vis IPRs: first, as an innovation policy instrument and, second, as an access and dissemination tool. The implications of this dual role are drawn in the pharmaceutical sector in a discussion about applying competition law as an innovation, access and dissemination instrument. The interplay among IPRs, competition law and public health from the perspective of sub-Saharan Africa is also discussed. Following Section I, the chapter contains a discussion in Section II of the parallel developments in intellectual property (IP) and competition law since the entry into force of the Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS). This section emphasizes the increased protection of IPRs, in addition to their strategic commercialization. At the same time, there has been a more aggressive application of competition law to IP-related restraints, with the aim of protecting innovation and enhancing access and dissemination. This trend goes beyond the restrictive approach of TRIPS competition-related provisions which focus only on technology transfer. Section III then discusses the dual role of competition law vis-à-vis IPRs, as an innovation policy instrument and as an access and dissemination instrument. In line with this discussion, Section IV deals with the interface between IP, competition law and public health. It emphasizes IP-related restrictions on competition in the pharmaceutical sector that affect both innovation and access. The discussion is informed by the enforcement experience in the EU and the US. With regard to innovation-related restraints, the chapter relies on the finding of the 2009 EU report on the pharmaceutical sector. On the issue of access, reverse payment settlement cases on both sides of the Atlantic are discussed. Section V of the chapter argues that in complement to the ‘built in’ flexibilities within the TRIPS agreement, competition law should be a legal instrument that fosters innovation and access in the pharmaceutical sector. This approach has huge potential for developing countries, which are enacting and enforcing competition law at a very fast rate. Section VI discusses the interface between IP, competition law and public health from the perspective of sub-Saharan Africa. Examples of how some laws have dealt with the interface between IPRs and competition law are given. Some relevant cases in the pharmaceutical sector are also discussed.

312

IPRs and competition law: public health in sub-Saharan Africa  313

II

PARALLEL DEVELOPMENTS IN IP AND COMPETITION LAW

IPRs and competition law have experienced tremendous changes since the adoption of the TRIPS agreement in 1994.1 Competition-related provisions embodied in TRIPS were a concession to developing countries that feared the detrimental consequences of a strong exercise of IPRs.2 They were conceived as an instrument that limits and controls the exercise of IPRs and facilitates the transfer of technology to developing countries. As a consequence, the protection of the local – and, in the case of developing countries, often weaker – licensee against abuses of the IP right holder was regarded in TRIPS as a key objective of competition law. Not surprisingly, the most detailed treatment of competition aspects in TRIPS is devoted to the regulation of anticompetitive practices in licensing agreements.3 Twenty years after the entry into force of TRIPS, parallel and intertwined developments have taken place both on the IP front and on the competition law front. These developments have changed the nature of competition law vis-à-vis IPRs, which have moved from a transfer-of-technology instrument to an innovation policy instrument.4 The main developments in IP and competition law relevant to this interface are discussed in this section. A

The Changing Nature of IPRs: Enhanced Protection and Strategic Commercialization

In the last two decades, IPRs have expanded in both qualitative and quantitative terms. New types of technologies, works and trademarks have been found eligible for protection;5 rights holders have been granted new exclusive prerogatives;6 the terms of protection for certain subject matter have been extended;7 and enforcement mechanisms and remedies have been strengthened and taken on an increasingly punitive character.8 The scope of protection in patent law seems to be extended, with too many patents of dubious quality being granted. This

1 The developments in this section draw from the previous work of the author. See Mor Bakhoum and Beatriz Conde Gallego, ‘TRIPS and Competition Rules: From Transfer of Technology to Innovation Policy’ in Hanns Ullrich, Reto M Hilty, Matthias Lamping and Josef Drexl (eds), TRIPS Plus 20: From Trade Rules to Market Principles (Springer 2016). The author thanks Beatriz Conde Gallego for accepting that part of this chapter is based on their joint paper. 2 Ibid. 3 Ibid. 4 Ibid. 5 See, for example, Directive 96/9/EC of the European Parliament and of the Council of 11 March 1996 on the legal protection of databases [1996] OJ L77, 20–28. 6 In the copyright field, for example, new forms of digital infringement related to the protection of technological protection measures have been created. See, for example, Directive 2001/29/EC of the European Parliament and of the Council of 22 May 2001 on the harmonization of certain aspects of copyright and related rights in the information society [2001] OJ L167, 10–19. 7 See, for example, Directive of the European Parliament and of the Council of 27 September 2011 amending Directive 2006/116/EC on the term of protection of copyright and certain related rights [2011] OJ L265, 1–5; Sonny Bono Copyright Term Extension Act, Pub L No 105-298, 112 Stat 2827 (1998). 8 See, for example, Directive 2004/48/EC of the European Parliament and of the Council of 29 April 2004 on the enforcement of intellectual property rights [2004] OJ L195, 16–25.

314  Research handbook on methods and models of competition law phenomenon results not only from the weakness of patent systems, but also the practice of some companies to patent extensively as part of a commercial strategy.9 IPRs are a core element of that part of the economic system in which productive assets, products and processes are largely intangible rather than physical in nature. The shift towards what some scholars have called ‘intellectual capitalism’10 has had at least the following two consequences. First, as firms have become aware of the growing value of their intangible assets, they have developed new commercial strategies to fully exploit them.11 Both the tendency to accumulate large IPR portfolios in order to use them as bargaining chips in licensing negotiations12 and the tendency to move away from production and distribution and to concentrate on research while deriving income from licensing royalties are examples of this development. Though aggressive commercial IP strategies are not necessarily at odds with legitimate competition, the fact is that IPRs are increasingly used as strategic tools to influence precisely the kind of (dynamic) competition they are expected to promote.13 This threat to innovation is problematic in technology-exporting countries, as well as in technology-importing countries. In this context, exclusion based on IPRs may be more problematic as innovation processes are increasingly cumulative and collaborative in nature. Second, as a response to the increasing value of IPRs, groups representing the interests of IP owners have an incentive to organize and push for expansive IP protection from their national legislators.14 Far from being restricted to the national level, this development towards higher standards of protection has been paralleled at the international level as well. From an international perspective, there is a proliferation of the so-called TRIPS-Plus agreements over recent years. Hence, through bilateral and regional trade agreements, standards of IPR protection of developed countries have been transferred to developing countries without due regard for the socio-economic, political and social context of the latter.15 Parallel and in reaction to the enhanced protection of IPRs through the TRIPS-Plus agreements, there is a growing question at the international level on how best to use the flexibilities embodied within TRIPS. The development in the framework of Doha with the Doha Declaration on IP and public health and subsequent Article 31bis to allow World Trade

The phenomenon of defensive patenting. See Ove Granstrand, Industrial Innovation Economics and Intellectual Property (Svenska Kulturkompaniet 2010) 18ff. 11 See Steven Anderman, ‘The IP and Competition Interface: New Developments’ in Steven Anderman and Ariel Ezrachi (eds), Intellectual Property and Competition Law: New Frontiers (OUP 2011) 6ff. 12 On this issue, see Dietmar Harhoff et al, ‘The Strategic Use of Patents and Its Implications for Enterprise and Competition Policies’ (Tender for No ENTR/05/82, Final Report, 2007), https://​core​.ac​ .uk/​reader/​2782468. 13 See Section III.A. 14 Stressing the significance of public choice theory for explaining expansion in IP law, Josef Drexl, ‘Immaterialgüterrechte zwischen Innovationsförderung durch Monopole und Wettbewerbsbeschränkung’ in Forschungsinstitut für Wirtschaftsverfassung und Wettbewerb e.V. (ed), Innovation und Wettbewerb, Referate des XLII. FIW-Symposion (Heymann 2009) 34; Christina Bohannan and Herbert J Hovenkamp, Creation without Restraint: Promoting Liberty and Rivalry in Innovation (OUP 2012) 133ff (focusing mainly on how US copyright law has been captured by interest groups). 15 For an overview and analysis of the most significant free trade agreements, see Pedro Roffe, ‘Intellectual Property Chapters in Free Trade Agreements: Their Significance and Systemic Implications’ in Josef Drexl, Henning Grosse Ruse-Khan and Souheir Nadde-Phlix (eds), EU Bilateral Trade Agreements and Intellectual Property: For Better or Worse? (Springer 2014) 17ff. 9

10

IPRs and competition law: public health in sub-Saharan Africa  315 Organization members to issue compulsory licences for export are prime examples of this development.16 In this vein, competition law is also an essential flexibility.17 B

The Changing Role of Competition Law: From Technology Transfer to Innovation Policy

On the competition law front, important developments have also taken place since the entry into force of the TRIPS agreement. On the one hand, there is an expansion of the number of jurisdictions that have enacted competition laws. From a substantive point of view, competition law has become an innovation policy instrument. More than 100 jurisdictions have now adopted and are applying competition laws, including many sub-Saharan African countries.18 This development results in some cases from the awareness of the benefits of competition law as a regulatory instrument that polices the behaviours of companies in the market and limits the interventionist role of the state in the market. To mention a few, emerging markets such as South Africa, China, India, Brazil and Mexico are becoming global players in competition law. Their laws sometimes go beyond the focus on efficiency and consumer welfare to incorporate other values into the goals of their competition laws. China’s competition law has raised questions about the treatment of international businesses active in China. The Coca Cola19 merger, which was prohibited by the Ministry of Commerce, raised early suspicions about the roles of industrial and protectionist policies in the approach of the Chinese competition authorities. On the other hand, smaller developing jurisdictions have not always enjoyed a great deal of discretion when enacting their competition laws and defining their competition policies. Hence, the enactment of competition law in developing jurisdictions often resulted from specific requirements in bilateral trade agreements20 or specific requirements of international lending institutions. Competition law in this context is conceived as part of a negotiation package that developing countries must adopt. As a consequence, they have limited discretion to define the goal of their competition laws in accordance with their development goals and their socio-economic and political context. Mostly imported competition norms ignore the development needs and aspirations of receiving countries.21 From an international perspective, the international harmonization of IPRs contrasts with the national (or regional) dimension of competition law. Tentative efforts to discuss and elaborate an international framework

Bakhoum and Conde Gallego (n 1). Ibid. 18 See http://​unctad​.org/​en/​pages/​SGStatementDetails​.aspx​?OriginalVersionID​=​83; https://​www​ .oecd​.org/​daf/​competition/​Challenges​-Competition​-Internat​-Coop​-2014​.pdf. 19 See, for discussion, Hogan & Hartson, ‘Antitrust Update: MOFCOM Discloses Details Concerning Rejection of Coca-Cola/Huiyuan Transaction’ (27 March 2009), https://​s3​.amazonaws​.com/​documents​ .lexology​.com/​2f933ddf​-f4f4​-4c9c​-88f8​-013f6f6fdcf4​.pdf. See, for discussion, Xiaoye Wang and Jessica Su, ‘China’s Anti-Monopoly Law: Agent of Competition or Engine of Industrial Policy?’ in Daniel Zimmer (ed), The Goals of Competition Law (Edward Elgar Publishing 2012). 20 See, for example, the CARIFORUM agreement, discussed below. 21 Mor Bakhoum, ‘A Dual Language in Modern Competition Law: Efficiency Approach versus Development Approach and Implications for Developing Countries’ (2011) 34(3) World Competition 495. 16 17

316  Research handbook on methods and models of competition law for competition law have so far failed.22 Convergence of substantive norms, especially in the framework of the ICN,23 and cooperation between competition authorities offer alternative solutions to the issue of uncoordinated enforcement and inconsistent decisions. Another development that should be highlighted is the increasing significance of the concept of dynamic competition in the competition laws of many developed jurisdictions. Competition in innovation is not a new phenomenon. The idea that it is not price competition, but competition in innovation, that leads to greater economic growth can be traced back to Joseph Schumpeter’s works in the 1940s.24 Yet, competition laws have largely relied on neoclassical models, which mainly take into account the short-term effects of impugned business behaviour on price and output. Only relatively recently have competition law enforcers recognized innovation as a proper parameter of competition.25 Thus, it was only in the mid-1990s that the concept of innovation markets was introduced and discussed as an analytical tool to assess the effects of a merger on innovation.26 This development, which is taking place in developed jurisdictions, contrasts with the situation in developing jurisdictions, which tend to integrate more and more non-economic goals in their competition laws. Hence, it is not uncommon to have goals such as protection of employment, competitiveness of small and medium-sized enterprises, industrial policy and public interest in the competition law of developing countries.27

22 For discussion, see Josef Drexl, ‘International Competition Policy after Cancún: Placing a Singapore Issue on the WTO Development Agenda’ (2004) 27(3) World Competition 419. 23 See 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). 24 Joseph A Schumpeter, Capitalism, Socialism and Democracy (Harper & Brothers 1942; reprinted Routledge 1996) Ch VII. However, while Schumpeter accepted, and even advocated, restrictive practices as a way to promote innovation and growth, it was JM Clark who provided a theoretical framework for research into more dynamic and more realistic theories and definitions of competition. See, for example, JM Clark, ‘Competition: Static Models and Dynamic Aspects’ (1955) 45(2) American Economic Review 450; JM Clark, Competition as a Dynamic Process (Brookings Institution 1961). 25 Josef Drexl, ‘Anticompetitive Stumbling Stones on the Way to a Cleaner World: Protecting Competition in Innovation without a Market’ (2012) 8(3) Journal of Competition Law and Economics 507, 513. 26 Richard J Gilbert and Steven C Sunshine, ‘Incorporating Dynamic Efficiency Concerns in Merger Analysis: The Use of Innovation Markets’ (1995) 63(2) Antitrust Law Journal 569. The concept of innovation markets found its way into the US IP Guidelines of 1995. See US Department of Justice and Federal Trade Commission, ‘Antitrust Guidelines for the Licensing of Intellectual Property’ (6 April 1995) 10, http://​www​.justice​.gov/​atr/​public/​guidelines/​0558​.pdf. In Europe, the Guidelines that accompanied Commission Regulation (EC) No 772/2004 of 27 April 2004 on the application of Article 81(3) of the Treaty to categories of technology transfer agreements also included it. See Guidelines on the application of Article 81(3) of the Treaty to categories of technology transfer agreements [2004] OJ C101, 2–42 para 25. 27 For discussion, see Michal S Gal and Eleanor Fox, ‘Drafting Competition Law for Developing Jurisdictions: Learning from Experience’ in Michal S Gal, Mor Bakhoum, Josef Drexl, Eleanor M Fox and David J Gerber (eds), The Economic Characteristics of Developing Jurisdictions: Their Implications for Competition Law (Edward Elgar Publishing 2015).

IPRs and competition law: public health in sub-Saharan Africa  317

III

THE INTERFACE BETWEEN IP AND COMPETITION LAW: THE ROLE OF COMPETITION LAW VIS-À-VIS IP

The parallel (and interrelated) developments on IP and competition law discussed in the previous part of this chapter are relevant to the discussion of the interface between IP and competition law. From an enforcement perspective, the developments in competition law are a reaction to the new and strategic use of IP in the market that has been described in this chapter. This has resulted in the application of competition law to preserve innovation and also to ensure access and dissemination of IP-embodied products. Before analysing this dual role of competition law vis-à-vis IP, it is necessary to discuss, at the outset, the interface between IP and competition law. A

The Interface Between IP and Competition Law

The interplay between IP and competition law is dynamic. From a dual and opposite approach, with a conception of competition law as an instrument that limits and controls the exercise of IP, the fields have turned out to be complementary, pursuing the same goal of innovation. The backwards-looking approach of the TRIPS competition-related provisions,28 which are designed as a limit to the exercise of IPRs in favour of innovation, are not in line with the modern approach to the relationship between IP and competition law as being complementary.29 The IP system gives incentive for innovation and the development of cultural markets. IP provides a temporary protection in this regard. However, IPRs do not operate in a vacuum, but are rather exercised in a given market whose rules apply fully to the rights concerned. As a market regulatory tool, competition law determines the framework within which competition takes place in a market, thus also in IP-related markets. To function properly, the IP system needs open and competitive markets. By precluding imitation of the protected subject matter, IPRs redirect competition towards the development of substitutable products, a kind of competition that is considered to produce greater social gains than pure price competition. Whether new technologies and products are developed in the first place, whether they may enter the market, and whether they may finally reach consumers will, however, very much depend on the competitive conditions present in the market.30 Thus the important role of competition law in defining the general framework within which IPRs are exercised. The protection granted to the rights holders as such does not constitute the actual reward. IPRs only give rights holders the opportunity to extract a reward from the market31 by commercializing their invention.

28 Which focus on technology transfer. See, for discussion, Hanns Ullrich, ‘Expansionist Intellectual Property Protection and Reductionist Competition Rules: A TRIPS Perspective’ in Keith E Maskus and Jerome H Reichman (eds), International Public Goods and Transfer of Technology under a Globalized Intellectual Property Regime (Cambridge University Press 2005). 29 Bakhoum and Conde Gallego (n 1), 9: see, especially, the reference in fn 33. 30 Ibid, 3. 31 See Hanns Ullrich and Andreas Heinemann, ‘Die Anwendung der Wettbewerbsregel auf die Verwertung von Schutzrechten und sonst geschützten Kenntnissen’ in Ulrich Immenga and Ernst-Joachim Mestmäcker (eds), Wettbewerbsrecht, Band I, EG/Teil 2, Kommentar zum Europäischen Kartellrecht (Beck 2011) 146.

318  Research handbook on methods and models of competition law Hence, the extent to which IPRs are successfully commercialized depends on the market, which in the end decides which products (inventions) are successful.32 Competition law plays a dual and complementary role vis-à-vis IP. On the one hand, competition law preserves innovation against some forms of exercise of IPRs that may hinder innovation. It does so by ensuring that innovative products enter the market freely. Competition law performs this task by addressing conduct blocking or hindering market access of already developed new or improved products.33 On the other hand, competition law has to ensure that IP-embodied products are disseminated according to market rules. In both dimensions, competition law complements the IP system as an innovation policy tool.34 B

Competition Law and Innovation-related Restraints

In line with the objective of IP to foster innovation, competition law aims also to foster innovation. Competition law ensures that innovative products enter the market. In other words, competition law guarantees that competition by substitution, on which the IP system relies, works properly. By doing this, competition law may address conduct blocking or hindering market access of new or improved products.35 A couple of examples of how competition law is applied to IP-related restraints of competition illustrate this point. Licensing agreements are a good instrument for the diffusion of technology and follow-up innovation. Yet, they may be used as legal tools to restrict innovation by incorporating certain clauses that limit the freedom of the licensee or third parties to innovate. An example is non-compete obligations included in licensing agreements that prohibit the licensee from using third-party technologies, which may foreclose the market for new competing technologies. This is very likely when the licensor has substantial market power and many licensees are tied to the technology.36 Another example is contractual clauses that prevent the licensee from using and/or developing its own competing technology, as well as broad grant-back clauses that directly affect the competitiveness of the licensee’s technology and incentives to further invest in its development. Moreover, restrictions imposed on the licensing parties’ ability to conduct independent research and development may eliminate an important source of innovation in the market, especially where few technologies are available.37 While innovation restraints in the licensing context should not be underestimated,38 a greater threat for competition in innovation and, at the same time, a bigger challenge for competition

Bakhoum and Conde Gallego (n 1). Ibid. 34 Ibid, 2. 35 Ibid. 36 Ibid. 37 Ibid. 38 Both US and EU competition rules contain detailed and complex provisions by which the overall positive effects of licensing and the necessity to safeguard the right holder’s incentives to license are balanced against the risk that some licensing clauses result in restricting innovation. See Guidelines on the application of Article 101 of the Treaty on the Functioning of the European Union to technology transfer agreements [2014] OJ C89, 3–50, paras 128ff; US Department of Justice and Federal Trade Commission (n 26) 10ff. 32 33

IPRs and competition law: public health in sub-Saharan Africa  319 law enforcement arises both from the strategic use of IPRs in standardized markets39 and from patenting strategies that have the potential to negatively influence competitors’ research and development (R&D) activities.40 In contrast to the examples above, the competition law treatment of so-called strategic patenting is far less developed. As previously noted, the creation of large patent portfolios is a legitimate commercial strategy that largely reflects the increasing value that companies assign to their IP.41 In certain situations, however, the strategic filing of patent applications may have a negative impact on innovation. In this context, the 2009 Pharmaceutical Sector Inquiry Report of the European Commission identified certain patenting strategies that may interfere with the development of competing medicines by reducing the incentives of other originator companies to continue their own R&D efforts.42 This kind of patenting practice may likewise occur in other fields of technology.43 In all such cases, the motivation of the patent applicant is not so much to protect its own invention as predominantly to block innovative efforts of competitors.44 So far, there is no precedent on the application of the competition rules to blocking patenting strategies.45 In 2007, the European Commission began an investigation against Boehringer Ingelheim after a competing pharmaceutical company had raised concerns that Boehringer’s

39 A detailed examination of issues regarding technology standards and standard essential patents is made by Peter Picht, ‘From Transfer of Technology to Innovation through Access: TRIPS and Standards-Based Markets’ in Hanns Ullrich et al (eds), TRIPS Plus 20: From Trade Rules to Market Principles (Springer 2016). 40 Bakhoum and Conde Gallego (n 1). 41 See Section II.A. 42 European Commission, Pharmaceutical Sector Inquiry: Final Report (2009) 16, 19, http://​ec​ .europa​.eu/​competition/​sectors/​pharmaceuticals/​inquiry/​staff​_working​_paper​_part1​.pdf. Generally, the Sector Inquiry Report is concerned with two different sets of strategic patenting practices: on the one hand, those that aim at extending the exclusivity period to delay market entry of generics and, on the other hand, those whose objective is to block substitutive innovations by competitors. Regarding the latter, blocking patents can be applied either to broaden the applicant’s own field of activity (defensive blocking) or to limit the scope of action of competitors (aggressive blocking). See Hanns Ullrich, ‘Strategic Patenting by the Pharmaceutical Industry: Towards a Concept of Abusive Practices of Protection’ in Josef Drexl and Nari Lee (eds), Pharmaceutical Innovation, Competition and Patent Law: A Trilateral Perspective (Edward Elgar Publishing 2013) 244, 248. 43 For an analysis of motivations for patenting across different technology areas, see Harhoff et al (n 12), 231ff. 44 See Ullrich (n 42), 248. 45 In contrast, the Commission and the two European courts have confirmed the application of Article 102 of the Treaty on the Functioning of the European Union to a strategic use of patent procedures aimed at delaying generic entry in the AstraZeneca case. See Commission Decision of 15 June 2005 relating to a proceeding under Article 82 of the EC Treaty and Article 54 of the EEA Agreement, Case COMP/A.37.507/F3 – AstraZeneca, http://​ec​.europa​.eu/​competition/​antitrust/​cases/​dec​_docs/​ 37507/​37507​_193​_6​.pdf; GC, Case T-321/05 AstraZeneca v Commission [2010] ECR 2010 II-2805; CJ, Case C-457/10 P AstraZeneca v Commission [2012] ECR I-770. On the relevance of the AstraZeneca rulings for the assessment of blocking patents, see Josef Drexl, ‘AstraZeneca and the EU Sector Inquiry: When Do Patent Filings Violate Competition Law?’ in Josef Drexl and Nari Lee (eds), Pharmaceutical Innovation, Competition and Patent Law: A Trilateral Perspective (Edward Elgar Publishing 2013) 312ff.

320  Research handbook on methods and models of competition law patent applications had the potential to block its competing medicines. On the suggestion of the Commission, however, the case was settled between the companies.46 C

IP, Competition Law, and Access- and Dissemination-related Restraints

Access and dissemination of IP-embodied products are part of the innovation rationale behind the IP system. This would be incomplete if IP-embodied products did not reach the final consumer. Yet, different forms of anticompetitive practices may harm the efficient functioning of markets where IP-protected products are commercialized and may thereby compromise the access of final consumers to IP-protected products. Hence, in addition to safeguarding the openness of markets for innovation, competition law also plays a role in facilitating access and dissemination of IP-protected products. In contrast to its application to innovation restraints, competition law in these cases follows a more classical approach, focusing mainly on price competition and, to a certain extent, also ensuring that consumers may have access to a great variety of products. Conduct restricting access to IP-protected goods may take different forms. IP-related markets can, in principle, be affected by the same kind of anticompetitive practices as any other market. Hence, IP-related markets are not immune to cartels.47 Horizontal price-fixing agreements between IP right holders, for example, reduce output and increase the prices of IP-protected products. Because of their pernicious effects on competition, there is consensus on their prohibition by competition law. Beyond these more traditional forms of anticompetitive practices, a particular type of agreement among competitors in the pharmaceutical industry has drawn the attention of competition law enforcers in recent years. Basically, it concerns situations where a brand-name pharmaceutical company, as patent holder, and a generic producer agree to settle either a patent infringement suit or a dispute concerning the validity of the patent under terms that require, first, the generic manufacturer not to produce and/or to distribute the patented product until the expiration of the patent, and second, the patent holder to ‘compensate’ the generic company for staying out of the market. This practice is commonly known as a ‘pay-for-delay’ agreement or, since it often involves a payment from the patentee to the alleged infringer, a ‘reverse payment’ settlement agreement. Both in the US and in Europe, the competition agencies have perceived such arrangements as an attempt to allocate markets and preserve monopolistic conditions48 and have condemned them as clear

46 See European Commission, ‘Antitrust: Commission Welcomes Improved Market Entry for Lung Disease Treatments’ (Media Release, 6 July 2011), http://​europa​.eu/​rapid/​press​-release​_IP​-11​-842​_en​ .htm. 47 A great variety of examples of horizontal agreements in copyright-related markets is provided in a study mandated by the World Intellectual Property Organization: Max Planck Institute for Intellectual Property and Competition Law, ‘Copyright, Competition and Development’ (2013) 83–101. 48 Announcing the Commission’s decision on the Servier case, then Competition Commissioner Joaquín Almunia stated: ‘Servier had a strategy to systematically buy out any competitive threats to make sure that they stayed out of the market. Such behaviour is clearly anti-competitive and abusive. Competitors cannot agree to share markets or market rents instead of competing, even when these agreements are in the form of patent settlements. Such practices directly harm patients, national health systems and taxpayers.’ See European Commission, ‘Antitrust: Commission Fines Servier and Five Generic Companies for Curbing Entry of Cheaper Versions of Cardiovascular Medicine’ (Media Release, 9 July 2014), http://​europa​.eu/​rapid/​press​-release​_IP​-14​-799​_en​.htm.

IPRs and competition law: public health in sub-Saharan Africa  321 violations of competition law.49 In the US, the Supreme Court has already had the opportunity to pronounce on the legal assessment of this kind of patent settlement.50 In addition, price and sales restrictions in licensing agreements may have a detrimental effect on the distribution of IP-protected goods. However, due to the specificities of the distribution of copyright-protected works, it is in the copyright-related markets that anticompetitive practices engaged in by distributors are most relevant.51

IV

IP, COMPETITION LAW AND PUBLIC HEALTH

The discussion in this chapter of the IP-related restrictions of competition that may negatively affect both innovation and access is very relevant to public health. In the pharmaceutical sector, also, restrictions of competition may affect both innovation and access. The topic of IP, competition law and public health has attracted much attention recently. This was the focus of the UN Conference on Trade and Development (UNCTAD) Intergovernmental Experts (IGE) meeting in 2015.52 On both sides of the Atlantic, there is increased scrutiny by competition authorities of behaviours in the pharmaceutical sector that negatively impact the timely entry of generics into the market or the reduction of innovative drugs for the benefit of consumers.53 Anticompetitive behaviours in the pharmaceutical sector may affect both innovation and access. In the EU, since the publication of the Sector Inquiry Report on the pharmaceutical sector in 2009, these have come under increased scrutiny.54 Hence anticompetitive practices in the pharmaceutical sector may take different forms and correspond to the different catego49 In Europe, see European Commission, ‘Antitrust: Commission Fines Lundbeck and Other Pharma Companies for Delaying Market Entry of Generic Medicines’ (Media Release, 19 June 2013), http://​europa​.eu/​rapid/​press​-release​_IP​-13​-563​_en​.htm​?locale​=​en; European Commission, ‘Antitrust: Commission Fines Johnson & Johnson and Novartis €16 Million for Delaying Market Entry of Generic Pain-Killer Fentanyl’ (Media Release, 10 December 2013), http://​europa​.eu/​rapid/​press​-release​ _IP​-13​-1233​_en​.htm; European Commission, ‘Antitrust: Commission Fines Servier and Five Generic Companies for Curbing Entry of Cheaper Versions of Cardiovascular Medicine’ (Media Release, 9 July 2014), http://​europa​.eu/​rapid/​press​-release​_IP​-14​-799​_en​.htm. At the time of writing this chapter, no public version of these decisions was yet available. For an overview of the Federal Trade Commission’s practice, see Timothy A Cook, ‘Pharmaceutical Patent Litigation Settlements: Balancing Patent & Antitrust Policy through Institutional Choice’ (2001) 17(2) Michigan Telecommunications and Technology Law Review 417, 437ff (commenting particularly on In re Schering-Plough Corp 136 FTC 956 (2003); Federal Trade Commission v Watson Pharmaceuticals, Inc, 611 F Supp 2d 1081 (CD Cal 2009); and Federal Trade Commission v Cephalon, Inc, 551 F Supp 2d 21 (DDC 2008)). 50 See Federal Trade Commission v Actavis, 570 US 133 (2013), https://​www​.supremecourt​.gov/​ opinions/​12pdf/​12​-416​_m5n0​.pdf. 51 See Max Planck Institute for Intellectual Property and Competition Law (n 47). 52 See UNCTAD, ‘The Role of Competition in the Pharmaceutical Sector and Its Benefits for Consumers’ (TD/RBP/CONF.8/3, 27 April 2015), http://​unctad​.org/​meetings/​en/​SessionalDocuments/​ tdrbpconf8d3​_en​.pdf. 53 See the different cases discussed in this chapter: Case T-472/13, Lundbeck v Commission, box no 1; Case C-142/18 J&J v Commission (‘Fentanyl’); Federal Trade Commission v Actavis, 570 US 133 S Ct 2233 (2013), https://​www​.supremecourt​.gov/​opinions/​12pdf/​12​-416​_m5n0​.pdf; Case T-691/14 Servier v Commission EU: T: 2018: 922. 54 See the contribution by the European Commission to the UNCTAD Roundtable on the Role of Competition in the Pharmaceutical Sector and Its Benefits for Consumers (unpublished, 2015, on file with the author).

322  Research handbook on methods and models of competition law ries of behaviours addressed by antitrust law: horizontal and vertical agreements, unilateral conduct, and mergers which may negatively affect innovation and access in the pharmaceutical sector. Such practices may also be classified taking into account both the ambit of IP law and competition law. Category one concerns practices that are within the ambit of the patent system (Italian Pfizer, AstraZenecca) and involve misuse of the patent and regulatory system that may impact innovation and access. Category two concerns behaviours that are outside of the boundaries of the patent system (such as ever-greening), while category 3 comprises practices half-way between the other two categories (pay-for-delay cases, Federal Trade Commission v Actavis, Lundbeck).55 In line with the discussion in this chapter of the effect of IP-related behaviours on innovation and access, there will be subsequent elaboration on IP-related restrictions of competition in the pharmaceutical sector that affect innovation and IP-related restrictions that affect access and dissemination.56 Of course, some behaviours may affect both innovation and access and dissemination. Some cases will be discussed as illustration. A

IP-related Restrictions and Innovation in the Pharmaceutical Sector

The development of practices that have the potential to affect innovation discussed in this chapter are also very relevant in the pharmaceutical sector. Restrictive clauses in licensing agreements that limit the possibilities of the licensee to develop its own technology or to use a competing technology are also relevant in the pharmaceutical sector. However, strategic filings in the pharmaceutical sector and their potential to hinder innovation and the development of new medicines have attracted a lot of attention. The 2009 Sector Inquiry Report of the pharmaceutical industry evidenced strategic patenting practices in the pharmaceutical industry that have the potential to limit innovation and the availability of new medicines. As seen in the report, such behaviours may at the same time affect access and dissemination. Patenting strategies may be both dilatory (referred to as dilatory patenting57) and innovation restrictive. Both practices have detrimental effects on consumer welfare from static (price) and dynamic (innovation) perspectives. Anticompetitive behaviours that slow down the application of new medicines affect dynamic efficiency by extending the duration of the monopoly of the patent holder.58 Strategic patenting is amongst strategies whose effect may be the limitation of competing innovation. Such practices are not exclusive to the pharmaceutical sector. They are part of the general practices used by patent holders in order not only to broaden their own field of R&D, but to avoid being impaired in their innovation activities by potential competitors. They may also be used as a strategy to limit the scope of R&D and the potential development of new products by competing firms.59 Such practices are referred to as blocking patents, which may have a defensive (defensive blocking patent) or aggressive (aggressive blocking patent) 55 Such distinction was made by Sven Gallasch, ‘Pharmaceutical Antitrust: Use but with Caution’ (Keynote Speech at the UNCTAD Roundtable on the Role of Competition in the Pharmaceutical Sector and Its Benefits for Consumers, 8 July 2015), http://​unctad​.org/​meetings/​en/​Presentation/​CCPB​ _7RC2015​_PRES​_RTPharma​_Gallasch​_en​.pdf. 56 The same distinction was made by the EU Commission in its Sector Inquiry Report. 57 See Ullrich (n 42), 244. 58 Ibid, 242. 59 Ibid, 248–9.

IPRs and competition law: public health in sub-Saharan Africa  323 nature.60 Whereas defensive blocking patents may constitute a legitimate strategy aiming at ensuring that the R&D efforts of the company are not hindered by competitive technologies, aggressive blocking patents aim only at limiting innovation by preventing the development of new and competing technologies. The pharmaceutical sector also experiences patenting strategies which have the potential to prevent the appearance of new drugs in the market. Hence, defensive patenting strategies may have the legitimate aim of creating ‘a protected space for the free development of a new active substance by obtaining not only broad primary patents, but also by filing secondary patents more or less immediately’.61 When such strategies are of an aggressive nature and aim simply at blocking competing innovation, however, they may raise antitrust scrutiny. Such strategies take place also in the pharmaceutical sector, as evidenced by the Sector Inquiry Report. Patent filing strategies raise the general question of the use of the regulatory framework that might be considered abusive and raise antitrust scrutiny. This question raises the issue, from an institutional point of view, of the interdependency between the patent system and competition law. Should competition law intervene as early as the moment of patent filing for practices that might affect dynamic and static (price) competition? If intervention is conceivable in order to protect innovation and guarantee access, the more challenging question is under what conditions competition authorities should intervene.62 The case law is developing slowly on the issue without giving clear answers to the issue of patenting strategies. What the case law confirms, however, is that the filing of a patent may constitute an abuse of dominance in the sense of Article 102 of the Treaty on the Functioning of the European Union, although the applicant may be entitled to a patent under patent law.63 In the same vein, the Pfizer decision of the Italian Administrative Supreme Court confirms that a misuse of the regulatory system constitutes an abuse of a dominant position, which forms part of the broader concept of abuse of right.64 Mergers and acquisitions may also adversely affect access to pharmaceutical innovation. The business model of the pharmaceutical industry, in addition to the traditional investment Ibid. Ibid. 62 On this issue, see Drexl (n 45). 63 Ibid, 312. 64 The Administrative Supreme Court confirmed the Italian competition authority’s decision against the multinational pharmaceutical group Pfizer as it considered several types of conduct of Pfizer punishable. Pfizer’s complex strategy consisted of several types of conduct, including (1) the filing of divisional patent applications and the subsequent request for a Supplementary Protection Certificate, in order to extend the Italian patent protection many years after filing the main patent application; (2) patent-related court litigation hindering market entry of generic companies (shame litigation); (3) actions aimed at preventing the Italian Medicines Agency from granting geneticists marketing authorizations; and (4) application for a further patent protection extension through paediatric experimentation. In its reasoning, the court found that all of these complex conducts, although individually and abstractly legitimate, could correctly be defined as abuse of rights and specifically anticompetitive. According to the court, the doctrine of the abuse of rights includes the existence of a right; the possibility to effectively use such a right in different ways; the exercise of the right in a reprehensible manner, although formally legitimate; and the resulting unjustifiable disproportion between the benefit to the right’s owner and the harm caused to the counterparty. UNCTAD (n 52), sourced from C D’Amore, ‘The Administrative Supreme Court Confirms the ICA’s [Italian Competition Authority’s] Decision to Condemn Pfizer for Abuse of Dominant Position Aimed at Delaying the Market Entry of Generic Pharmaceutical Companies’ (2014) 1 Italian Antitrust Review 77. 60 61

324  Research handbook on methods and models of competition law in R&D, is moving towards buying promising innovator or generic companies with a strong potential for coming up with new chemicals and drugs. This approach is justified by the structural changes that are taking place in the pharmaceutical sector. Many companies are facing ‘soon-to-expire patents’, which has a considerable impact on their sales.65 In addition, the generic companies in emerging markets such as Brazil, China and India are growing, creating a highly global competitive environment in the pharmaceutical sector.66 These changes lead to a shift from an R&D and investment model to mergers and acquisitions, where transnational pharmaceutical corporations buy generic companies with a promising pipeline of new drugs.67 Although such a business model is legitimate, it may negatively impact the emergence of new medicine and may also restrict access to pharmaceuticals. A pharmaceutical company would not launch a new more efficient drug in competition with its own medicine. This would be detrimental to consumers, who continue to pay monopoly prices with a reduction of innovative products. B

IP-related Restrictions and Access and Dissemination in the Pharmaceutical Sector

Known and conventional behaviours such as horizontal, vertical and unilateral conduct may affect access and dissemination when they take place in the pharmaceutical sector. Practices such as cartels, bid rigging, and boycotts are conventional behaviours that aim to fix prices and earn monopoly profits.68 Unilateral conduct by the pharmaceutical industry, such as abusing a dominant position, may also affect access to and dissemination of pharmaceuticals. Refusal to license or charging high royalty fees for licence seekers who want to enter the market is also a strategy used by brand companies in order to block the entry of generics in the market. As pointed out, charging a high licensing fee has the same effect as a refusal to license.69 In addition to those traditional approaches of blocking generic competition, new more sophisticated strategies are used in order to prevent generic competition and thereby affect access and dissemination. The Sector Inquiry Report of the European Commission identified a number of these practices. Among such practices are patent filing strategies, which consist of filing numerous patent applications for the same medicine. This practice, also known as ‘patent cluster’ or ‘patent thickets’,70 may not only affect innovation as already seen, but also access and dissemination. These common practices have the sole objective of delaying or blocking the entry of generic medicines.71 A high number of patents or pending patent applications lead to uncertainties that limit the ability of generic companies to enter the market.72 Filing a ‘divisional patent’ application is also identified as a strategy used to create legal uncertainties and to delay the entry of affordable generics in the market.73 Divisional patents UNCTAD (n 52), 7. Ibid. 67 Ibid. 68 Ibid, 10. 69 Ibid, 6. 70 See EU Commission, ‘Executive Summary of the Pharmaceutical Sector Inquiry Report’ (2009) 10, http://​ec​.europa​.eu/​competition/​sectors/​pharmaceuticals/​inquiry/​communication​_en​.pdf. 71 Ibid, 10. 72 Ibid. See also UNCTAD (n 52), 6. 73 See EU Commission (n 70), 11. 65 66

IPRs and competition law: public health in sub-Saharan Africa  325 do not extend the scope of the original patent application or the protection period. But they extend the examination period and de facto delay the entry of generics. Those practices may be legitimate from a perspective of patent law as they are in principle allowed by the regulatory framework. They have, however, the goal and the consequence of unduly delaying the entry of generics, extending the monopoly pricing of the originator company. As pointed out, ‘the denser the patent cluster or divisional patent, the more difficult it will be for a generic company to bring its generic version to market’.74 The mere existence of those clusters of secondary patents creates uncertainties and impedes the timely entry of generics. Additional strategies documented in the EU Sector Inquiry Report that may also effect access to and dissemination of pharmaceuticals include litigation strategies, oppositions and appeals, and practices aiming at extending the life cycle of the medicine, also known as evergreening.75 With regard to evergreening, the Novartis case decided by the Supreme Court of India is an interesting example.76 In this case, the Indian Supreme Court concluded that for known molecules that have benefited from patent protection elsewhere, only an enhanced efficacy may justify granting a patent. Among practices that have attracted attention recently and have led to close scrutiny from the antitrust authorities are the so-called reverse payment settlement agreements. Reverse payment settlements, known also as pay-for-delay, are horizontal behaviours by pharmaceutical industries to delay or prevent the entry of generics into the market. Patent settlements are legitimate and sometimes legal practice. In the pharmaceutical sector, generic companies may challenge the validity of a patent or a protection granted. Originators, on the other hand, may claim that a generic company is infringing its patent. Settlement of such dispute may be beneficial since it reduces the costs of litigation and allows early generic entry. However, when the sole purpose of such settlements is to avoid competition from generics and to maintain high prices for pharmaceuticals, they trigger antitrust scrutiny. Hence, if settlements between an originator and a generic company aim only at preventing competition by generics in exchange for transfer of value, they raise significant concerns from an antitrust perspective. A number of investigations conducted in the EU and the US confirm this strategy of pharmaceutical companies to enter into settlement agreements with the sole objective of preventing or delaying competition from generics and maintaining artificially high prices for

UNCTAD (n 52), 7; Ullrich (n 42), 245. See EU Commission (n 70), 11ff. 76 Novartis Ags v Union of India (2013) 6 SCC 1 is a landmark decision by the Indian Supreme Court on the issue of whether Novartis could patent the cancer drug Gleevec (imatinib mesylate) in India. The Supreme Court upheld the Indian patent office’s rejection of the patent application in part because of section 3(d) of India’s Patents Act, which requires demonstration of increased therapeutic efficacy for a medicine to deserve a patent. The patent was rejected on the grounds that imatinib mesylate is merely the salt form (mesylate) of an older medicine, imatinib. Novartis was trying to argue that ‘efficacy’ should be interpreted differently and that increased bioavailability of the salt form of imatinib meant increased efficacy, entitling it to a patent on imatinib mesylate. However, it was clarified by the Supreme Court that efficacy must be interpreted as ‘therapeutic effect in healing a disease’ and, even if the bioavailability of the drug was improved, it did not demonstrate enhanced efficacy, and that therefore Glivec was not patentable. This case set an important precedent to facilitate access to essential medicines as there was a significant price gap between the patented version of Glivec (US$5,000) and its generic copy (US$200). See Ravinder Gabble and Jillian Clare Kohler, ‘To Patent or Not to Patent? The Case of Novartis’ Cancer Drug Glivec in India’ (6 January 2014), http://​www​.ncbi​.nlm​.nih​.gov/​pmc/​articles/​ PMC3884017/​#B9. See also Médecins sans Frontières, ‘Q&A on Patents in India and the Novartis Case’ (20 December 2006), http://​www​.msf​.org/​en/​article/​qa​-patents​-india​-and​-novartis​-case. 74 75

326  Research handbook on methods and models of competition law medicines. In July 2013, the Commission fined Lundbeck and several producers of generic medicines for delaying generic market entry of Citalopram.77 In December 2013, the Commission fined Novartis and J&J,78 which concluded an agreement whose aim was to delay the market entry of a cheaper generic version of Fentanyl, a painkiller. This was a straightforward pay-for-delay case, as it did not involve any patent dispute or litigation. In the US, the Actavis decision of the Supreme Court79 sets the legal standard for the appreciation of pay-for-delay cases. After a discrepancy in decisions from lower courts, the Supreme Court concluded that the rule of reason should be applied to reverse payment settlements.

V

TRIPS, COMPETITION LAW AND PUBLIC HEALTH: GOING BEYOND TRIPS FLEXIBILITIES

The issue of the interface between IPR and public health has attracted much attention since the adoption of the TRIPS agreement. In advancing access to pharmaceuticals, many efforts have been directed at operationalizing the ‘built-in’ flexibilities within the IP system, such as the use of compulsory licensing. Yet, the flexibilities within the IP system have turned out to be ineffective in addressing public health issues. The development in the framework of Doha with the Doha Declaration on Intellectual Property and Public Health and the subsequent Article 31bis, which react to the difficulty in using the compulsory licensing mechanism for countries without production capacities, evidence the limits in using the TRIPS-related flexibilities, especially in developing countries. Competition law is another flexibility recognized by the TRIPS agreement as a potential instrument to foster access. Articles 8, 31(k) and 40 of TRIPS, which are usually referred to as the TRIPS competition-related provisions, are largely conceived as legal tools to monitor the exercise of IPRs based on competition considerations.80 TRIPS is the first international instrument that recognizes the need to control IPR-related restrictions of competition. The inclusion of competition provisions was very much influenced by the context in which the agreement was concluded. Competition-related provisions in the TRIPS agreement were conceived as a result of a bargain between developed and developing countries. Hence, developing countries that fear a monopolistic exercise of IPRs managed to have provisions on competition law included in order to mitigate potential abuses in the exercise of IPRs. Competition law in TRIPS is usually associated with technology transfer. The protection of the local – and, in the case of developing countries, often weaker – licensee against abuses of the IP right holder was regarded as a key objective of competition law. Not surprisingly, the most detailed treatment 77 European Commission, ‘Antitrust: Commission Fines Lundbeck and Other Pharma Companies for Delaying Market Entry of Generic Medicines’ (Media Release, 19 June 2013), http://​europa​.eu/​rapid/​ press​-release​_IP​-13​-563​_en​.htm​?locale​=​en. Information of General Court upon time of completion not available (July 2015). 78 See European Commission, ‘Antitrust: Commission Fines Johnson & Johnson and Novartis €16 Million for Delaying Market Entry of Generic Pain-Killer Fentanyl’ (Media Release, 10 December 2013), http://​europa​.eu/​rapid/​press​-release​_IP​-13​-1233​_en​.htm. 79 Federal Trade Commission v Actavis, 570 US 133 (2013), https://​www​.supremecourt​.gov/​ opinions/​12pdf/​12​-416​_m5n0​.pdf. 80 Bakhoum and Conde Gallego (n 1).

IPRs and competition law: public health in sub-Saharan Africa  327 of competition law aspects in TRIPS is devoted to the regulation of anticompetitive practices in licensing agreements.81 In light of the development in the exercise of IPRs and the subsequent reaction of the application of competition law, it would be too restrictive to confine the TRIPS competition-related provision to a simple technology transfer instrument. Hence, as argued elsewhere, TRIPS competition-related provisions can be read as an innovation policy instrument. Competition law, in compliance with TRIPS, may be applied in order to favour both access and innovation.82 However, how effective competition law might be in favouring access and innovation depends on each country. Hence, TRIPS does not create a binding international framework that obliges signatory members to apply competition law to IP-related restrictions of competition: it is optional. TRIPS competition-related provisions give leeway to signatory members to define their own policies when it comes to applying their competition law to IP-related restrictions. Therefore, the effectiveness of competition law in addressing abuses of IPRs depends on the enforcement institutions of each country. This situation creates an imbalance from an international perspective. On the one hand, there is a harmonization, from the top, of the protection of IPRs. On the other hand, the use of competition law is ‘deregulated’ and left to the choice of each member to define its own policy. This flexibility has, however, allowed the flexible use of competition law in order to address different forms of exercise of IPRs. With regard to the specific issue of health in general, and access to medicine in particular, there is a noticeable increased reliance on competition law as an instrument in favour of access to medicine. The different cases discussed in this chapter showcase the importance of using competition in order to address public health issues. Hence, competition law can be used as an instrument that addresses behaviours that hinder innovation as much as behaviours that restrict access and dissemination, as discussed here. It is worth mentioning that using competition law as an instrument that fosters access to medicine should not be conceived as an alternative to the safeguards and flexibilities recognized by TRIPS. Rather, it should be viewed as ‘complementary and providing added value’.83 Advancing public health by means of competition law has proven to be very effective in developed countries. Developing jurisdictions are still developing their competition law framework. However, using competition law in order to foster public health has huge potential.

VI

THE SITUATION IN SUB-SAHARAN AFRICA

IP and competition law are not stand-alone disciplines with different goals and different legal and economic rationales. The theory of complementarity between IP and competition law requires a legislative approach that takes into account the internal incentive structure of the IP system, but also – and more importantly – the need to put safeguards in place that protect the competitive structure of the markets where IPRs are commercialized. Keeping IP-related markets competitive is a task assigned in part to competition law. This section sheds some light

Ibid. Ibid. 83 See Jonathan Berger, ‘Advancing Public Health by Other Means: Using Competition Policy’ in Pedro Roffe, Geoff Tansey and David Vivas-Eugui (eds), Negotiating Heath: Intellectual Property and Access to Medicine (Earthscan 2006) 182. 81 82

328  Research handbook on methods and models of competition law on whether and how the theory of complementarity is dealt with in the context of sub-Saharan Africa by looking at how IP-related restrictions of competition are dealt with in some competition legislation. The relevance of the approach to public health will be discussed, with some cases as examples. A

IPR(s) and Competition Law from a Sub-Saharan Africa Perspective

Competition law and policy are relatively new in most African countries.84 According to a needs assessment conducted by the African Competition Forum,85 no fewer than 24 countries have competition laws. Although some countries have a long tradition of having a kind of competition law, which in fact was much closer to price control regulations, an approach of competition law guaranteeing the protection of free competition by the prohibition of cartels and abuses of dominance is relatively new. It was only in the 1990s, following the wave of liberalization and privatization, that most African countries began to draft and implement competition laws. Even Nigeria, which is the leading economic giant in West Africa, did not have a competition law or a competition authority until 2019. When it comes to enforcement, limited results have been achieved, partly due to the lack of institutional, human and financial capacities of the newly created competition authorities. The treatment of IP-related restrictions of competition is rather rudimentary. Some laws do not even refer to the exercise of IP as a potential form of anticompetitive behaviour.86 However, this does not mean that competition authorities do not have the legal tools to address IP-related restrictions of competition. Anticompetitive licensing agreements or abuses of dominant position involving IPRs may be dealt with by a competition authority. The issue is more complex, and any possibility of enforcement lost, if a competition law expressly exempts IP from its application or gives the option for an exemption. In Gambia, for instance, the Competition Act expressly exempts IP-related issues from its application. To the 84 This section draws from the author’s previous research. See, in particular, Mor Bakhoum, ‘Balancing “Incentive to Innovate” and “Protection of Competition”: An African Perspective on Intellectual Property Rights and Competition Law’ in Ruth L Okediji and Margo A Bagley (eds), Patent Law in Global Perspective (OUP 2014). 85 African Competition Forum, Full Needs Assessment Report (18 March 2011). 86 In Zambia, for instance, section (3)(a) of the Competition and Consumer Protection Act No 24 of 2010 asserts that the Act shall not apply to ‘an agreement or conduct insofar as it relates to intellectual property rights including the protection, licensing or assignment of rights under, or existing by virtue of, a law relating to copyright, design rights, patents or trade marks’. In Seychelles, with the exception of section 7(4)(c) of the Fair Competition Act, which exempts abuses of dominant power resulting from the exercise of an IPRs, there is not a comprehensive legal framework addressing the interface of IP and competition law. There is no case law yet on the issue. In the Gambia, IP-related matters are expressly exempted from the application of the Competition Act. See Schedule 1(5) of the Competition Act. In contrast, in Kenya, the Competition Act does apply to IP-related restrictions unless an exemption is granted by the Commission according to section 28(1) of the Act. Also, in Mauritius, the legislation is applicable to IP-related restrictions of competition. Extended guidelines deal with the specific issue of the application of the Act to IP-related restrictions: Competition Commission of Mauritius, ‘Guidelines: General Provisions’ (CCM 7, 2009). In Malawi, the Competition Act does not apply to licensing agreements involving an IPR. See section 3(d) of the Competition Act. In Namibia, the Competition Act does apply to IP-related restrictions unless they are exempted by the competition authority. See section 30 of the Competition Act. In all those jurisdictions, there is no case yet on the application of the competition law to IP-related restrictions.

IPRs and competition law: public health in sub-Saharan Africa  329 contrary, Mauritius has comprehensive provisions in its Competition Act dealing with IP. In South Africa, IP-related restrictions of competition are expressly dealt with in the Competition Act. However, the Act provides a possibility to apply for an exemption from the application of the Act.87 Exempting IPRs from the application of competition law is a rather problematic approach, especially for developing countries. This is in contradiction to the very rationale of the international IP system and the interests of developing countries to use competition law as a balancing tool and thereby to facilitate technology transfer. Fighting IP-related abuses of competition aims at facilitating technology transfer. Absence, or weak enforcement, of competition law in developing countries against abuses of IPRs hinders technology transfer and subsequent potential follow-up innovation. In this regard, one could argue that the legislative and enforcement approach to competition law in sub-Saharan Africa does not yet fulfil its role of fostering innovation. This is also true with regard to the treatment of competition law in the free trade agreements, as exemplified by the EU-Caribbean Forum (CARIFORUM) agreement. No fully fledged agreement encompassing rules on trade, IP and competition law has been concluded yet between the EU and African countries. The interim agreements88 signed by some states bilaterally with the EU do not concern IP or competition law. In the ongoing Economic Partnership Agreements (EPA) negotiations between the EU and African countries, IP and competition law are part of the so-called rendez-vous clause, which means that the parties agree to discuss these clauses in the future. Despite the fact that there is no agreement yet signed between the EU and the different regional groups with which the EU is in discussions, one can already envision the potential approach that the EU will be taking when it comes to competition matters. In this regard, the approach taken by the EU in the agreement it has signed with the CARIFORUM provides useful insights on how competition law will be dealt with.89 The EU-CARIFORUM agreement requires CARIFORUM states to introduce competition laws and create national competition authorities.90 With regard to the substantive approach, the principle of non-discrimination – which does not take into account the need and competitive constraints of local companies – should be applied. The approach taken seems to constitute only an additional market-access tool for EU companies. Do the EPA’s competition rules foster innovation and protect free competition? It is very doubtful. There is a differentiated treatment of IPRs and competition law in the EPAs, as evidenced by the EU-CARIFORUM See n 86. See, for example, the interim agreement signed by the EU with Cameroon on 15 January 2009: Council Decision of 20 November 2008 on the signature and provisional application of the interim agreement with a view to an Economic Partnership Agreement between the European Community and its Member States, of the one part, and the Central Africa Party, of the other part [2009] OJ L57, 1. 89 See Mor Bakhoum, ‘Commerce International, Politique de Concurrence et Accords de Partenariat Economique: Réflexions sur les Enjeux et Perspectives d’un Triptyque’ Agence Africaine pour le Commerce et le Développement’ (10 April 2014), http://​2acd​.org. See also Josef Drexl, ‘Perspectives Européennes sur la Politique de Concurrence dans l’Espace OHADA’ (2011) 3 Revue Internationale de Droit Economique 298 (giving a critical approach to the European policy approach in the free trade agreements); see also South Centre, ‘Fact Sheet No 8: Competition Policy in Economic Partnership Agreements (CARIFORUM Text)’ (SC/AN/TDP/EPA/15, 2008), http://​www​.southcentre​.org/​index2​ .php​?option​=​com​_docman​&​task​=​doc​_view​&​gid​=​707​&​Itemid​=​68. 90 See Articles 127–9 of the Agreement. For a detailed analysis of the obligations, see South Centre (n 89). 87 88

330  Research handbook on methods and models of competition law agreement. Strong emphasis is put on IP protection, whereas competition law remains marginal. Competition law is not approached as a complement of IP that would enhance innovation by protecting free competition. Competition law provisions are only viewed as an additional market-access tool.91 Granting additional market access in complement to stronger IP protection is only in the interests of the IP-exporting countries. Despite this stronger IP protection and more openness of the markets, no meaningful mechanisms to check the potential abuses of IPRs have been put into place. In the EU-CARIFORUM agreement, IP is not expressly mentioned in the competition law provisions, as one would have expected given the complementary of the two fields. This should not, in principle, prevent countries from using their competition law against abuses of IPRs. However, this possibility is only optional since ‘the competition law rules of the EPA contain no express obligation to fight anticompetitive behavior relating to IPRs … and do not provide any guidance on how to apply competition law in IP-related cases’.92 One may argue in the context of the free trade agreements that raising the IP protection standards provides an incentive to right holders to innovate by preventing unlawful copying. However, there is also a risk that a high level of protection, or a monopoly power derived from an IPR, may lead to market foreclosure. As observed, ‘these standards are not necessarily procompetitive in the sense that IP exclusivity will trigger dynamic competition; they may well foreclose markets for what may be more innovative competitors in favour of incumbent right-holders’.93 B

The Relevance for Public Health: Experiences in IP, Competition Law and Public Health

This section94 discusses the developments in sub-Saharan Africa with regard to the issue of access to pharmaceuticals, with an emphasis on how competition authorities have dealt with the issue so far. Although an unprecedented development has taken place over the past years in competition law in sub-Saharan Africa, the creation and effective functioning of competition authorities is still lagging behind.95 Efforts have been made at the national as well as at the regional level to enact and enforce sound competition laws. The Common Market for Eastern and Southern Africa (COMESA), the West African Economic and Monetary Union (WAEMU), the Southern African Development Community (SADC) and possibly the Economic Community of West African States (ECOWAS) are regional integration groups

See Bakhoum (n 89). See also South Centre (n 89). See Josef Drexl, ‘Competition Law and Transfer of Technology in the EC-CARIFORUM EPA’ (unpublished) 17. 93 Josef Drexl, ‘Intellectual Property and Competition: Sketching a Competition-Oriented Reform of TRIPS’ in A Bakardjieva (ed), Festskrift till Marianne Levin (Norstedt Juridik 2008) 269. 94 This section draws extensively from Mor Bakhoum, ‘Intellectual Property, Competition Law and Access to Pharmaceuticals: The Relevance of a “Market Approach” to the Exercise of Intellectual Property Rights’ in WIPO Academy and the WTO Intellectual Property Division (eds), WIPO-WTO Colloquium Papers: Research Papers from the WTO-WIPO Colloquium for Teachers of Intellectual Property Law (2013), https://​www​.wto​.org/​english/​tratop​_e/​trips​_e/​wipo​_wto​_colloquium​_2013​_e​.pdf. 95 According to a needs assessment conducted by the African Competition Forum (n 85), no fewer than 24 countries have competition laws. 91 92

IPRs and competition law: public health in sub-Saharan Africa  331 that deal with competition matters.96 At the national level, South Africa is by far the most advanced country, with sound competition institutions and enforcement authorities. Other countries, such as Mauritius, Zambia and Seychelles, are catching up and are developing their institutions. When it comes to the interface between IP and competition law, some competition laws directly address the issue, whereas others exempt IP from competition law application. Only the South African Competition Commission dealt with a case related to the issue of IP, competition law and access to medicine. The case was eventually settled. Another merger case, which was eventually authorized with conditions, is also relevant. Finally, the newly functioning COMESA Competition Commission has recently cleared a merger that involved two pharmaceutical companies. Those cases will be discussed subsequently. Reference is made to them to demonstrate the relevance of competition law as a public policy instrument for access to pharmaceuticals. A complaint was lodged before the South African Competition Commission against GlaxoSmithKline South Africa (Pty) Ltd (GSK) and Boehringer Ingelheim (Pty) (BI),97 initially for high pricing, but then the Commission extended the investigation to include an alleged violation of section 8(b) and (c) of the Competition Act, which deals with the essential facilities doctrine and exclusionary conduct respectively.98 The case was eventually settled. In particular, GSK and BI were accused of anticompetitive conduct involving the following: ●● GSK abused its dominant position in the market for antiretroviral drugs (ARVs) by charging excessive prices; ●● making the product inaccessible to the general public; ●● refusing to supply a competitor access to an essential facility; ●● a dramatic difference in the price of ARVs sold in South Africa and generic alternatives sold outside South Africa; ●● the existence of patents prevented sale of generic substitutes in South Africa; and ●● patent protection did not require a firm to charge high prices. The Competition Commission concluded its investigation with a finding that GSK and BI abused their dominant position by charging excessive prices, refusing to grant access to essential facilities to a competitor, and engaging in exclusionary conduct. The matter did not come before the Competition Tribunal, as GSK and BI accepted a settlement, which resulted in a drastic reduction in the prices of the pharmaceuticals in South Africa. As part of the settlement, GSK and BI agreed to: ●● grant licences to generic manufacturers; ●● permit licensees to export the relevant ARV medicines to sub-Saharan African countries; 96 Generally, on competition law and policy in regional integration, see Josef Drexl, Mor Bakhoum, Eleanor M Fox, Michal S Gal and David J Gerber (eds), Competition Policy and Regional Integration in Developing Countries (Edward Elgar Publishing 2012). 97 The case was settled. For a discussion of the background, see Berger (n 83) 197–201. For the Commission’s comments on the case, see South African Competition Commission, ‘GSK and BI Issue Anti-Retroviral Licences’ (2004) 15 Competition News 1, http://​www​.compcom​.co​.za/​wp​-content/​ uploads/​2014/​09/​March​-04​-Newsletter​.pdf. 98 For a discussion of the case, see Mfundo Ngobese and Liberty Mncube, ‘Competition Policy in South Africa’s Pharmaceutical Sector: Balancing Competition and Innovation’ (unpublished, 2011).

332  Research handbook on methods and models of competition law ●● where the licensee did not have manufacturing capability in South Africa, permit the importation of the ARV medicines for distribution in South Africa only, provided all the regulatory approvals were obtained; ●● permit licensees to combine the relevant ARVs with other ARV medicines; and ●● not require royalties in excess of 5 per cent of the net sales of the relevant ARVs. Two aspects are worth highlighting in this case. First, the competition law offences of which GSK and BI are accused would have been difficult to tackle using only the IP flexibilities, such as compulsory licensing. Charging high prices, refusing to grant access to essential facilities, or engaging in exclusionary conduct would be difficult to use as grounds for compulsory licensing under the TRIPS agreement. The second interesting aspect of this case is the conditions of the settlements and the commitments accepted by GSK. The different commitments mirror the developments in the framework of Doha with regard to pharmaceuticals, with the introduction of the mechanism of licensing for export for countries without sufficient manufacturing capacities. In Doha, in addition to the declaration on IP and public health, a new mechanism allowing countries without sufficient manufacturing capacities to issue compulsory licences for imports was introduced. Although in theory the mechanism would enhance access to pharmaceuticals, in practice it proved difficult to render operational, as the only instance in which it was tested illustrates.99 It is interesting to note in the GSK case in South Africa that the Doha mechanism set up for countries without manufacturing capacities, which allow countries to issue compulsory licences for exportation, was achieved through competition law. Hence, in its commitments, GSK agreed to permit licensees to export the relevant ARV medicines to sub-Saharan African countries. In addition, GSK agreed, where the licensee did not have manufacturing capacity in South Africa, to permit the importation of the ARV medicine for distribution in South Africa only, provided the regulatory approval was obtained. Those commitments, which constitute the essence of Article 31bis of the TRIPS agreement, were obtained not through importing mechanisms, which turned out to be difficult to use, but by using competition law. Moreover, a price cap of 5 per cent of the net sales of the relevant ARVs allows the control of the prices charged by GSK to licensees. The terms of the commitments go beyond what has been agreed upon in the framework of Doha. In addition, enforcing the Doha measures involves a heavy administrative burden, whereas the Competition Commission can easily monitor that GSK actually respects its commitment. This case displays the efficiency gains of using competition law in addition to IP flexibilities. Competition law control can be more effective and easier to enforce than IP stricto sensu flexibilities. The Aspen/GSK100 merger case dealt with by the South African Competition Commission is another example of the relevance of competition law intervention in order to keep the market open and competitive. Aspen was a large generic pharmaceutical company that wanted to acquire the pharmaceutical component of GSK. During the merger, GSK announced its intention to license ARV to Aspen. The Competition Commission raised a concern about whether

So far, only Rwanda has used the system. For comments on the merger, see South African Competition Commission, ‘Aspen and GSK Merger’ (2010) 34 Competition News 13, http://​www​.compcom​.co​.za/​wp​-content/​uploads/​2014/​09/​ FINAL​-for​-Web​.pdf. 99

100

IPRs and competition law: public health in sub-Saharan Africa  333 or not GSK would allow access of the ARV to other competing firms on the same conditions it had granted Aspen. In order to achieve a more competitive price, the Competition Commission finally approved the merger on the condition that GSK granted licences to other competing firms on a non-exclusionary basis. The condition to grant licences to other competing firms allows price competition in the market that will eventually reduce the prices of pharmaceuticals. Price competition from an access point of view is very relevant for the consumer.101 However, one must bear in mind that R&D is very costly in the pharmaceutical industry. Mergers between competing firms can constitute a way to fund R&D. Therefore, when analysing the actual or potential effects of a merger in the pharmaceutical industry, one must balance the need to allow access to pharmaceuticals with the need to ensure that future innovation will not be hindered. The Competition Commission of the Common Market for Eastern and Southern Africa (COMESA) approved unconditionally a merger between two pharmaceutical companies: Cipla India and CiplaMedpro South Africa Ltd.102 Cipla India is a generic pharmaceutical manufacturing company that does business in various therapy areas. Cipla does not have manufacturing plants in the COMESA market. Cipla India supplies the Common Market primarily through distributors. As for CiplaMedpro, it manufactures and distributes various pharmaceuticals products and provides health care solutions as well. After defining the relevant market as the supply of generic pharmaceutical products in the Common Market, the COMESA Competition Commission determined that (1) ‘the same market concentration would remain post merger as the parties did not compete in the common market before the merger’;103 and (2) ‘import competition was very rife in this market as most of the drugs sold in this market were imported. This would therefore give competitive discipline to the merging parties and restrain them from behaving in an anticompetitive manner.’104 The Competition Commission added that ‘the transaction would not result in the removal of any competitor from the relevant market as generally the parties were not competing pre-merger’.105 Despite the absence of competition between the two firms and the openness of the relevant market to competition, the Commission reveals the existence of structural and regulatory barriers. Those relate to the cost of establishing a distribution network and the various registration processes that the pharmaceutical companies need to undertake before they have the authorization to supply in the Common Market. Regulatory barriers are common in the pharmaceutical industry business. The Competition Commission concluded that the acquisition of CiplaMedpro by Cipla India was not likely to substantially prevent or lessen competition and it will not be contrary to public interest in accordance to with Article 26 (1) and 26 (3) of the Regulations respectively. Further, the assessment of the merger revealed that it was

Bakhoum (n 84). COMESA, ‘Statement of Merger 2 of 2013’, https://​ www​ .comesacompetition​ .org/​ ?page​ _id​ =​ 2020. 103 ‘COMESA Publishes Explanation of First Two Merger Approvals & Receives 4th Deal Filing’, African Antitrust & Competition Law News & Analysis (7 August 2013), http://​africanantitrust​.com/​ 2013/​08/​07/​comesa​-publishes​-explanation​-of​-first​-two​-merger​-approvals​-receives​-4th​-deal​-filing/​ comment​-page​-1/​. 104 Ibid. 105 Ibid. 101 102

334  Research handbook on methods and models of competition law compatible with Article 55 of the COMESA Treaty in that it did not negate the objectives of free and liberalized trade.106

The merger did not raise a competition-related issue that would have been detrimental to access to pharmaceuticals in the Common Market. Competition in importation was not affected and the merging firms were not competitors in the relevant market pre merger. The analysis would certainly have been different if the merging companies were competing in the relevant market and held a dominant position in the distribution. This would have raised competition concerns. The COMESA Commission hints at the issue of public interest, which is one criterion put forward by the COMESA Regulations when analysing a merger. This aspect goes beyond the scope of this chapter, but it would be interesting to see how access to pharmaceuticals relates to the concept of public interest as defined in the COMESA Regulations.

VII CONCLUSION Applying competition law to IP-related restrictions of competition is in line with the policy objective of IPRs to foster innovation and access. A number of practices discussed in this chapter showcase how IP-related restrictions of competition may hinder not only access, but also innovation. The interface between IPRs and competition law is relevant for the issue of access to pharmaceuticals. IP-related restrictions of competition in the pharmaceutical industry may negatively affect access to pharmaceuticals. Such restrictions of competition showcase the limits of the ‘built-in’ flexibilities within the IP system in dealing with practices that may hinder innovation and access to pharmaceuticals. Practices in the pharmaceutical sector that violate the principles of free market and competition can only be dealt with by using competition law. Competition law is therefore a relevant regulatory instrument that complements the IP system. It can not only foster innovation, but also guarantee access in the pharmaceutical sector. The TRIPS agreement already recognizes the relevance of competition in dealing with abuses of IPRs. Conceived in the TRIPS agreement as a flexibility that may foster technology transfer, competition law application to IPRs has evolved to become a suitable tool to foster innovation and access in the pharmaceutical sector. Sub-Saharan Africa countries are steadily enacting competition laws and creating competition authorities. This is, of course, a positive development. Yet, the case law in the application of competition law to IP-related restrictions of competition is very scarce. Some countries even exempt IPRs from competition law application. This approach, which puts an emphasis on IP protection and downgrades the necessary competition control of IP-related restrictions of competition, is of course questionable from a policy perspective. Lack of competition control in the pharmaceutical sector means that final consumers face the risk of paying for medicines at cartelized prices. In the process of building their competition law frameworks, sub-Saharan Africa policy-makers and competition authorities should put an emphasis on dealing with the restrictions of competition involving IPRs. This is important not only for innovation and technology transfer, but also for access to pharmaceuticals. A sound approach to the issue of access to pharmaceuticals requires a combination of complementary regulatory tools. In addition to using the flexibilities within the IP system, competition law should play a central role. Ibid.

106

IPRs and competition law: public health in sub-Saharan Africa  335 Although limited in numbers, the case law in South Africa set an interesting precedent that sub-Saharan African countries should follow by enforcing their competition laws with regard to restrictions of competition in the pharmaceutical industry.

15. Pay for delay in perspective: the impact of adversarial and inquisitorial legalism on pharmaceutical antitrust enforcement Sven Gallasch

I INTRODUCTION Pay-for-delay settlements are patent settlements in the pharmaceutical sector where the brand company as patent holder makes a significant payment or value transfer to the generic entrant allegedly infringing the patent to stay out of the market. These kinds of agreements have attracted considerable antitrust scrutiny and are one of the enforcement priorities in the US and Europe. The US Federal Trade Commission (FTC) has investigated numerous pay-for-delay settlements over the last decades, supported private actions through amicus curiae briefs, and undertaken advocacy efforts by conducting studies into the interplay between intellectual property and antitrust, and sector-specific studies. The pinnacle of these efforts was the agency’s success in the US Supreme Court in FTC v Actavis1 on 17 June 2013. In contrast, the enforcement efforts in Europe seem to be more limited or at least more targeted. Following the pharmaceutical sector inquiry in 2008–09, the European Commission initiated formal proceedings against pay-for-delay settlements in a small number of cases.2 Lundbeck was the first case in which the European Commission reached a final decision.3

FTC v Actavis, Inc, 133 S Ct 2223 (2013). European Commission, ‘Antitrust: Commission Opens Formal Proceedings against Les Laboratoires Servier and a Number of Generic Pharmaceutical Companies’ (European Commission Press Release, 8 July 2009), http://​europa​.eu/​rapid/​press​-release​_MEMO​-09​-322​_en​.htm​?locale​=​en; European Commission, ‘Antitrust: Commission Opens Formal Proceedings against Pharmaceutical Company Lundbeck’ (European Commission Press Release, 7 January 2010), http://​europa​.eu/​rapid/​ press​-release​_IP​-10​-8​_en​.htm; European Commission, ‘Antitrust: Commission Opens Proceedings against Johnson & Johnson and Novartis’ (European Commission Press Release, 21 October 2011), http://​europa​.eu/​rapid/​press​-release​_IP​-11​-1228​_en​.htm. 3 Lundbeck (Case AT.39226), Commission Decision of 19 June 2013 [2013] OJ C 3803 final. Since then, further pay-for-delay decisions have reached the General Court. As in Lundbeck, the Court in Servier (Case T-691/14, Servier v Commission, EU:​T:​2018:​922) upheld the Commission’s general finding that pay-for-delay settlements can constitute a restriction by object (para 418), expressly referring to its Lundbeck judgment (para 412). Article 101(1) of the Treaty on the Functioning of the European Union (TFEU) is contravened as long as (1) the parties are at least potential competitors; (2) the generic undertaking committed itself in the agreement to limit, for the duration of the agreement, its independent efforts to enter one or more EU markets with a generic product; and (3) the agreement is related to a transfer of value from the brand company as a significant inducement that substantially reduced the incentives of the generic undertaking to independently pursue its efforts to enter one or more EU markets with the generic product (para 409). The Court in Servier also had to address whether licensing agreements constitute side-deals facilitating pay-for-delay settlements and a separate infringement of Article 102 of the TFEU relating to a potential abuse of dominant position. Although both issues are interesting, 1 2

336

Pay for delay in perspective  337 Interestingly, the Lundbeck decision was published just two days after the US decision in FTC v Actavis and came to a conclusion very similar to that of the US Supreme Court. Furthermore, the then-Director-General of the Directorate-General for Competition, Alexander Italianer, noted at a symposium in the US: Incidentally, to those of you who are familiar with the Supreme Court’s Actavis opinion, the factors taken into consideration by the Commission will sound familiar. Indeed, the Supreme Court looked at the same factors, in particular the size of the payment including as compared to the expected profits of the generic producer, and the lack of any other convincing justification.4

Meanwhile, the General Court rejected Lundbeck’s appeal of the European Commission’s findings and upheld the decision in its entirety.5 The General Court’s judgment contains a number of direct references to the judgment in FTC v Actavis.6 Observing the course of events in Europe and their proximity to the events in the US raises a number of questions. Is the European Commission the more efficient enforcement agency, or is EU pharmaceutical antitrust enforcement just an imitation game of US efforts? Neither of these assumptions is correct and especially the latter is far too cynical. As is so often the case, the answer lies somewhere in the middle. However, it is probably safe to assume that the European Commission was able to benefit from the enforcement efforts and expertise of its US counterpart. After all, it is to be expected that enforcement agencies do share information and insights regarding their understanding of a sector, or a type of anticompetitive conduct, at least on a general level. In addition to this expertise, pay-for-delay settlements have also attracted significant attention from academic economists and legal scholars in the US. One can only assume that this kind of expertise has at least partially informed the European Commission’s investigations in the pharmaceutical sector, which could be described as an ‘agency learning effect’. Nonetheless, this does not fully explain how the European Commission managed to arrive at the same conclusion as the FTC in relation to enforcement for pay-for-delay settlements with a single investigation, as opposed to the numerous investigations and lawsuits undertaken by the FTC in a way that can only be described as an uphill battle. The key to answering these questions lies in the respective legal systems in which the FTC and the European Commission operate. This chapter investigates the impact that the adversarial and inquisitorial systems have on public competition law enforcement, particularly in relation to pay-for-delay settlements in the pharmaceutical sector. In this analysis, the judicial perception of distinct features of pay-for-delay settlements, such as the interplay between intellectual property and antitrust, as well as the propensity for settlements, plays a crucial role. Before discussing the US and EU perspectives on the antitrust enforcement for pay-for-delay settlement, the chapter sets the scene. It provides a brief explanation of pay-for-delay settlements, setting out the incentives for parties to conclude such settlements. It also discusses the US Supreme Court judgment in FTC v Actavis and the General Court’s

only the former is of limited relevance to the argument put forward in this chapter and is thus briefly addressed at a later stage. 4 Alexander Italianer, ‘Competitor Agreements under EU Competition Law’ (40th Annual Conference on International Antitrust Law and Policy, Fordham Competition Law Institute, New York, 2013), http://​ec​.europa​.eu/​competition/​speeches/​index​_2013​.html. 5 Case T-472/13, Lundbeck v Commission ECLI:​EU:​T:​2016:​449. 6 Ibid, paras 353, 491, 492, 512, 513.

338  Research handbook on methods and models of competition law judgment in Lundbeck, before investigating the impact of the respective legal traditions on public antitrust enforcement in the US and Europe. It concludes that the agency learning effect does not increase the procedural efficiency of European antitrust enforcement. However, it can add to the robustness of the development of novel theories of harm, reducing the inherent prosecutorial and hindsight bias of the European system.

II

PAY FOR DELAY – THE ISSUE EXPLAINED

Pay-for-delay settlements7 are patent infringement settlements in the pharmaceutical sector between a brand company that is the patent holder and a generic company that wants to enter the market prior to patent expiry and would therefore possibly be infringing the brand company’s patents. Brand companies generally argue that they are entitled to exclude competitors from the market until the patent has expired and that the generic entry constitutes a patent infringement, which entitles the brand company to defend its patents by means of patent infringement litigation. In these situations, settlements are generally encouraged as they end litigation early, reducing the cost and length of litigation while also creating legal certainty.8 One would generally assume that patent settlements are determined by the parties’ likelihood of success in court and should therefore be based on the perceived strength of the underlying patent. However, pay-for-delay settlements are different and should not be regarded as ordinary settlements. Their exclusionary power is derived not from the strength of the underlying patent but rather from a significant value transfer from the brand company to the generic challenger. In pay-for-delay settlements, the patent holder pays the allegedly infringing generic entrant a substantial amount of money in return for the generic company’s acknowledgement of patent validity and the assurance that the generic company will stay out of the market until the date stipulated in the settlement. However, if the assumption of patent validity is correct, such a payment should not be necessary. The validity of the patent as such should ensure that the generic company stays out of the market. In normal circumstances, a payment from the brand company to the generic entrant should only occur if the patent is presumed to be invalid or not infringed; here, the payment would be of compensatory nature for the generic entrant.9 In pay-for-delay settlements, these two scenarios are combined. The generic challenger accepts the validity of the patent, and agrees not to challenge the patent and to stay out of the market, yet nonetheless it receives a significant payment that usually considerably exceeds the litiga-

7 For the purpose of this chapter, the brief explanation offered in this section is very much based on the common understanding of the anticompetitive potential of pay-for-delay settlements from an antitrust point of view. For a more detailed discussion of pay-for-delay settlements, including countervailing arguments in favour of them, see Sven Gallasch, ‘Debunking the Pay for Delay Myth: Pay for Delay Settlements Are No Ordinary Patent Settlements’ (2016) 15 Cambridge Law Journal 89; Sven Gallasch, ‘Activating Actavis in Europe – the Proposal of a “Structured Effects Based” Analysis for Pay for Delay Settlements’ (2016) 36 Legal Studies 683. 8 See Section IV.A.2. 9 See Herbert Hovenkamp, ‘Antitrust and Patent Law Analysis of Pharmaceutical Reverse Payment Settlements’ (15 January 2011), https://​ssrn​.com/​abstract​=​1741162. 

Pay for delay in perspective  339 tion costs.10 The payment goes in the wrong direction and is not based on the strength of the underlying patent, casting doubt over the validity of the patent. Instead, the payment is linked to the reduced threat of competition for the brand company and the value of lost profits for the generic company. Proponents of the legality of these settlements have argued in the past that the risk of entry reflects the probabilistic nature of patents and that large payments by the brand company to the generic challenger can be justified.11 It has been suggested that even a small risk of patent invalidity (for example, 10 per cent) might lead a risk-averse brand company to pay the generic challenger 10 per cent of the expected remaining profits, which can amount to a substantial payment in the pharmaceutical sector, depending on the drug in question and the remaining length of patent protection.12 Alternatively, a payment reflecting the probabilistic nature of a patent and the parties’ own estimation of success is not required to settle such a dispute. It would also be possible to split the remaining period of exclusivity according to the parties’ estimation, thereby allowing the generic company to enter prior to patent expiry.13 However, both scenarios would render the relevant patent unchallengeable – an outcome that is not intended by patent policy. Carl Shapiro rightly argues that a patent should not confer a ‘right to exclude’ as such but rather a ‘right to try to exclude’.14 Yet, turning a patent into an unchallengeable right by means of a pay-for-delay settlement is exactly what happens in the US pharmaceutical sector under the Hatch Waxman Act and this is also possible under the regulatory regime of the European pharmaceutical sector. In the US, the Hatch Waxman Act15 was introduced in 1984 in order to foster innovation in the pharmaceutical sector and increase generic competition by means of an expedited marketing authorization process by the US Food and Drug Administration (FDA).16 Following its enactment, generic companies could rely on an abbreviated new drug application (ANDA) on the clinical data and the relevant pharmaceutical test results of the brand drug, as long as the generic drug was bioequivalent to the brand drug.17 This route made it a lot quicker and far less expensive for the generic company to enter the market. Furthermore, the generic companies were also incentivized to enter the market early by challenging the brand companies’ patents. Such a patent challenge is essentially achieved by applying for marketing authorization with the FDA. As part of this application process, the generic applicant must file a so-called Paragraph IV notification. That document informs the brand company about the intended entry of the generic company, stating that the generic company perceives that the underlying brand

10 Carl Shapiro, ‘Antitrust Limits to Patent Settlements’ (2003) 34 RAND Journal of Economics 391, 407. 11 Robert D Willig and John P Bigelow, ‘Antitrust Policy Toward Agreements That Settle Patent Litigation’ (2004) 49 Antitrust Bulletin 655, 656. 12 Kevin D McDonald, ‘Patent Settlements and Payments That Flow the Wrong Way: The Early History of a Bad Idea’ (2002) 15 Antitrust Healthcare Chronicle 1, 10. 13 Willig and Bigelow (n 11), 658. 14 Shapiro (n 10), 395. 15 Drug Price Competition and Patent Term Restoration Act of 1984, Pub L No 98-417, 98 Stat 1585 (1984). 16 For a detailed discussion of the Hatch Waxman Act, including the legislative intent behind it, see Mike A Carrier, ‘Unsettling Drug Patent Settlements: A Framework for Presumptive Illegality’ (2009) 108 Michigan Law Review 37, 41. 17 21 USC § 355(j)(8)(B).

340  Research handbook on methods and models of competition law drug patents filed with the FDA are either invalid or not infringed.18 The filing of an ANDA with a Paragraph IV notification constitutes an act of patent infringement.19 This notification then allows the brand company to file a lawsuit for patent infringement, triggering a 30-month stay of the FDA generic approval procedure – time that is intended to allow the parties to resolve their patent dispute. In return for risking this kind of patent infringement litigation, the generic challenger receives a period of 180 days of generic exclusivity once the dispute is resolved and it is allowed to enter the market. Until the end of the generic exclusivity period, the FDA is not allowed to accept any further generic ANDA applications. It is exactly this kind of ‘regulatory bottleneck’ that the parties to a pay-for-delay settlement exploit.20 Instead of litigating the patent dispute as intended by the Hatch Waxman Act, the parties settle the dispute and thereby stipulate the date of generic entry, which triggers the period of generic exclusivity. By settling with the first-filing generic challenger, the brand company can therefore foreclose the market and shield its patents from further challenges until the end of the period of generic exclusivity.21 The reason the parties enter into such a pay-for-delay settlement is simple. It creates a win–win situation in which the parties’ interests are aligned. On the one hand, the payment or value transfer from the brand company to the generic entrant generally amounts to at least the expected generic profits following entry. The generic company therefore no longer has an interest in entering the market. On the other hand, the brand company is willing to offer this large payment as the value transfer is still less than the expected loss of brand profits following generic entry, which can be explained by the significant price differential between brand drug prices and generic drug prices.22 In contrast to the Hatch Waxman Act, the relevant European regulation expressly rejects such a patent linkage,23 thereby not creating a US-style regulatory bottleneck. Regulatory authorities that grant generic marketing authorizations in Europe do not take any economic factors such as patents into consideration. The authorization process is not stayed due to patent disputes. This leads to situations where subsequent generic entrants are able to receive marketing authorizations despite a pay-for-delay settlement with the first-filing generic company. Although market foreclosure and the shielding of patents from generic challenges is therefore more difficult to achieve than in the US, it is by no means impossible. Pay-for-delay settlements do take place in Europe and it has been argued elsewhere that the anticompetitive

21 USC § 355(j)(2)(A)(vii)(IV). 35 USC § 271(e)(2)(A). 20 Scott Hemphill, ‘Paying for Delay: Pharmaceutical Patent Settlements as a Regulatory Design Problem’ (2006) 81 New York University Law Review 101, 130. 21 See Carrier (n 16), 51, citing Representative Waxman, who explained that such agreements were an ‘unfortunate, unintended consequence’ of the Act that ‘turned the … legislation on [its] head’. 22 Michael Kades, ‘Whistling Past the Graveyard: The Problem with Per Se Legality Treatment of Pay-for-Delay Settlements’ (2009) 5 Competition Policy International 143, 148. 23 ‘In the interests of public health, authorisation decisions under the centralised procedure should be taken on the basis of the objective scientific criteria of quality, safety and efficacy of the medicinal product concerned, to the exclusion of economic and other considerations’: Regulation (EC) No 726/2004 of the European Parliament and of the Council laying down Community procedures for the authorisation and supervision of medicinal products for human and veterinary use and establishing a European Medicines Agency (2004), recital 13. 18 19

Pay for delay in perspective  341 potential of pay-for-delay settlements in Europe is dependent on the competitive environment and the market structure in the given relevant market in the EU pharmaceutical sector.24 What pay-for-delay settlements on both sides of the Atlantic have in common are the economic incentives on which they are based, as well as the fact that the exclusionary power of the settlement is not based on the underlying patent itself but rather on the payment from the brand company to the generic entrant. They should therefore be regarded not as a patent settlement but rather as a disguised market-sharing agreement between potential competitors, or even a cartel by contract.

III

PAY FOR DELAY – THE ISSUE SOLVED

On both sides of the Atlantic, pay-for-delay decisions by the FTC and the European Commission have reached the highest court (as in the case of the US with the US Supreme Court decision in FTC v Actavis) or have been upheld on appeal in their entirety (as in the case of Europe with the General Court’s judgment in Lundbeck). In FTC v Actavis, the Court recognized the general anticompetitive potential of pay-for-delay settlements, as well as potential legitimate reasons why pharmaceutical companies might enter into a settlement that appears to be anticompetitive. In its majority opinion, written by Justice Breyer, the Supreme Court opted for a rule of reason approach. However, probably the most striking part of the judgment from a general policy perspective was the Court’s finding that pay-for-delay settlements are not generally immune from antitrust scrutiny, stating that ‘it would be incongruous to determine antitrust legality by measuring the settlement’s anticompetitive effects solely against patent law policy rather than by measuring them against procompetitive antitrust policies as well’.25 The Court then turned to the legal analysis of the actual pay-for-delay settlement, considering the level of evidence the plaintiff would have to provide in order to satisfy the burden of proof, as well as the peculiarities of pay-for-delay settlements that exploit the regulatory framework.26 The judgment acknowledged that the exclusionary power of a pay-for-delay settlement can be based on the value transfer from the brand company to the generic entrant as long as it is a large and otherwise unexplained payment, which can ‘provide a workable surrogate for a patent’s weakness, all without forcing a court to conduct a detailed exploration of the validity of the patent itself’.27 According to the Court, the size of the payment is also a good indicator of market power. The payment is geared towards the protection of the large price–cost margins that only a brand company holding market power can realize by charging supra-competitive prices.28 Similarly, the size of the payment can help to infer the anticompetitive harm as the settlement based on the payment reduces the extent or likelihood of competition and even a small reduction of likely competition is sufficient. The Court accepted that a brand company might argue that even a very small risk of patent invalidity justifies a large payment. However, the

Gallasch, ‘Activating Actavis in Europe’ (n 7). FTC v Actavis (n 1), 2230–31. 26 Ibid, 2234–8. 27 Ibid, 2236. 28 Ibid. 24 25

342  Research handbook on methods and models of competition law Court continued by stating that ‘be that as it may, the payment (if otherwise unexplained) likely seeks to prevent the risk of competition. And, as we have said, that consequence constitutes the relevant anticompetitive harm.’29 It is the combination of the size of the otherwise unexplained payment and the prevention of risk of competition that makes the validity of the patent less relevant. The Court therefore requires the plaintiff to satisfy the rule of reason analysis but allows this analysis to be facilitated by ‘Actavis inference’, which ensures that it is less protracted.30 In Lundbeck, the General Court upheld the European Commission’s finding that the pay-for-delay settlement in question constituted a ‘restriction by object’, which led to an infringement of Article 101(1) of the Treaty on the Functioning of the European Union (TFEU) without having to show the actual anticompetitive effects of the agreement. However, one has to be careful not to generalize the judgment entirely. The Court expressly stated that it only had determined the potential competition law infringement of this specific agreement and was not required to lay down a generally applicable test.31 Nonetheless, a number of general principles can be derived from the judgment. First, the Court reiterated the European Commission’s finding that not every patent settlement including a payment or value transfer – even reverse payments – will lead to a competition law infringement.32 Having said that, the Court nevertheless agreed with the European Commission that agreements [that] transform the uncertainty in relation to the outcome of such litigation into the certainty that the generics would not enter the market, which may also constitute a restriction on competition by object when such limits do not result from an assessment, by the parties, of the merits of the exclusive right at issue, but rather from the size of the reverse payment which, in such a case, overshadows that assessment and induces the generic undertaking not to pursue its independent efforts to enter the market.33

From this core statement, repeated in the judgment on numerous occasions, one can derive two principles. First, the brand company is not entitled to ‘substitute their own assessment of the validity of their patents and the infringing nature of the generic undertaking’s products for that of an independent judge while paying the generic undertaking to comply with that assessment’.34 Second, the size of the payment is a relevant factor. If it is linked to averting litigation cost, it might be permissible. When it is linked to the expected generic profits under competition, it is a strong indicator of anticompetitive conduct. The latter was the case in Lundbeck. The European Commission was able to show that the size of the payment was linked to the estimated profits that the generic companies could expect upon entry – a finding

Ibid. See also Aaron Edlin, C Scott Hemphill, Herbert J Hovenkamp and Carl Shapiro, ‘Activating Actavis’ (2013) 38 Antitrust Health Care Chronicle 16; Aaron Edlin, C Scott Hemphill, Herbert J Hovenkamp and Carl Shapiro, ‘Actavis and Error Costs: A Reply to Critics’ (2014) Antitrust Source 1; Aaron Edlin, C Scott Hemphill, Herbert J Hovenkamp and Carl Shapiro, ‘The Actavis Inference: Theory and Practice’ (2015) 67 Rutgers University Law Review 585; Gallasch, ‘Activating Actavis in Europe’ (n 7). 31 Case T-472/13, Lundbeck v Commission (n 5), para 415. 32 Ibid, paras 334, 402. 33 Ibid, para 336. 34 Ibid, para 390. 29 30

Pay for delay in perspective  343 that was not disputed by the parties.35 This finding is not only in line with the above-outlined economic rationale of pay-for-delay settlements but is also made with direct reference to the relevant sections of the US Supreme Court judgment in FTC v Actavis.36 The Court also found that patent validity cannot be equated to a presumption of illegality. It stated that the European Commission was entitled to find that the generic company’s version of the brand drug did not infringe Lundbeck’s process patent due to a substantial amount of internal evidence.37 Effectively, this finding links back to the point made earlier that the brand company cannot substitute its own assessment for that of an independent judge. Essentially, the General Court’s judgment in Lundbeck is based on a very similar rationale to the US judgment in FTC v Actavis. From a case-specific point of view, the European Commission was able to put forward a very strong narrative against the parties, arguing that the settlement was a disguised cartel by contract. This narrative was further supported by incriminating internal statements, including those that showed the applicants’ intent ‘to use “a large pile of [USD]” to exclude generics from the market’.38

IV

PAY FOR DELAY – BETWEEN ADVERSARIAL AND INQUISITORIAL LEGALISM

Looking at these two judgments in isolation, one could get the impression that the European approach to pay-for-delay settlements is heavily influenced by the US jurisprudence in FTC v Actavis, suggesting a significant imitation effect. After all, the European Commission and the General Court came to a very similar conclusion to the US Supreme Court. Further, this is against the backdrop that Lundbeck is the very first pay-for-delay decision that was also upheld on appeal, as compared to the multitude of enforcement efforts by the FTC over the last decades. However, this initial impression changes when the antitrust enforcement efforts in both jurisdictions are viewed against the background of their underlying legal traditions. The US adversarial system and the EU inquisitorial system each have a profound impact on the antitrust enforcement strategies employed by the respective agencies, as well as on the outcomes of their enforcement efforts. These differences become even more apparent in pharmaceutical antitrust enforcement and, in particular, in relation to pay-for-delay settlements, as these agreements combine two elements that are heavily influenced and shaped by legal traditions: first, the judicial preference for settlement; and second, the interface between intellectual property and antitrust. The chapter therefore investigates the impact of the US adversarial system on the FTC’s antitrust enforcement efforts through the example of pay-for-delay settlements, before it turns to the same discussion for the European side.

37 38 35 36

Ibid, para 362. Ibid, para 353. Ibid, para 121. Ibid, para 368.

344  Research handbook on methods and models of competition law A

Pay-for-delay Antitrust Enforcement from the US Perspective

1 The FTC within the adversarial system The US federal antitrust laws are enforced by two separate authorities, the Antitrust Division of the US Department of Justice (Antitrust Division) and the FTC. Both authorities have overlapping agency jurisdiction by design. Historically, ‘Congress has assigned them the identical mission and for the identical purpose’.39 Nonetheless, a few notable and important differences exist between the two authorities. The Antitrust Division has sole jurisdiction over all criminal investigations in antitrust violations. In civil investigations, the jurisdictions of the two authorities technically overlap directly. However, the cases are generally divided based on expertise. Each authority has developed expertise in particular industries due to prior cases. Generally, cases are divided up along these lines. For example, the FTC has extensive expertise in pharmaceuticals and health care, whereas the Antitrust Division has considerable expertise in the airline sector and computer software.40 Both agencies are administrative authorities conducting internal adjudicatory decisions, which means that they have to litigate all matters in front of the ordinary courts. However, whereas the Antitrust Division enforces the Sherman Act, the FTC is vested with the power to enforce the Federal Trade Commission Act (FTC Act). This is a notable and important difference, as section 5 of the FTC Act provides the FTC with the means to prosecute ‘unfair methods of competition in or affecting commerce, and unfair or deceptive acts or practices in or affecting commerce’.41 The administrative litigation42 of section 5 allows the FTC to scrutinize at least the same kind of conduct that is within the scope of section 2 of the Sherman Act.43 The FTC can therefore rely on the standards and case law of the Antitrust Division under the Sherman Act but technically enforces the FTC Act. The supposed major advantage of this public enforcement strategy is the avoidance of spillover effects caused by private antitrust enforcement, which has long been the dominant driver of antitrust enforcement in the US.44 39 Daniel A Crane, The Institutional Structure of Antitrust Enforcement (Oxford University Press 2011) 28. 40 Ibid, 39. 41 15 US Code § 45(a)(1). Section 5 of the FTC Act was designed to have a broader scope than the Sherman Act and the Clayton Act, enabling the FTC also to prosecute types of conduct that are not, or not yet, covered by these two Acts, filling the gaps in antitrust enforcement. 42 In order to enforce section 5 of the FTC Act, the authority issues a complaint, which is then litigated before an independent Administrative Law Judge (ALJ) at the FTC. The procedure combines similar features of trial and appellate litigation, such as oral hearings and the gathering of evidence by means of discovery. The FTC is represented by ‘complaint counsel’ from the FTC’s Bureau of Competition. Once the ALJ has issued its initial decision, the decision can then be appealed to the Commission itself, which conducts a de novo review. The Commission, represented by its five commissioners, then issues a final decision. This final decision can be appealed to a federal circuit court of appeals and ultimately to the US Supreme Court by means of writ of certiorari. See DB Hoffman and MS Royall, ‘Administrative Litigation at the FTC: Past, Present, and Future’ (2003) 71 Antitrust Law Journal 319. 43 FTC v Cement Inst, 333 US 683, 694 (1948). 44 Historically, private enforcement has been viewed as the more effective means of enforcement, as opposed to public enforcement. This has led to the concept of the ‘private attorney general’. In order to spur private enforcement, US antitrust laws provide incentives such as the award of treble damages and one-way fee shifting. See David J Gerber, ‘Private Enforcement of Competition Law: A Comparative Perspective’ in TMJ Möllers and A Heinemann (eds), The Enforcement of Competition Law in Europe (Cambridge University Press 2007).

Pay for delay in perspective  345 However, the discussion below shows that this theoretical advantage does not have the desired effect in reality. In fact, antitrust authorities and private plaintiffs litigate and enforce the same laws (in this case, the Sherman Act directly, or indirectly through section 5 of the FTC Act), use the same court system, and apply the same procedural rules. This means that regular courts play an important role in this system and subsequently in the development of the law.45 The problem with this linkage of private and public antitrust enforcement is that despite the link, the impetus for litigation is rather different. Private litigation rarely has public policy in mind. It could therefore be argued that US courts often react to private initiatives and try to adjudicate the private individual case, rather than having public policy consideration in the forefront of their minds. This is described by Stephen Calkins as an ‘equilibrating tendency’, where the courts tend to ensure that only viable private antitrust claims are successful in order to counterbalance potential over-deterrence, causing the adjustment of substantive and procedural standards in antitrust litigation.46 This has led to a situation where the FTC is increasingly faced with constricted liability norms that have been created to curb zealous private litigation.47 As mentioned above, this situation is further exacerbated in relation to pay-for-delay settlements as the courts have to determine antitrust liability not only in general, but also in relation to the exercise and exploitation of patents and the amicable resolution of patent disputes by means of settlement. 2 The perception of settlements Settlements are strongly favoured as a means of dispute resolution in civil matters in the US legal system.48 The reasons for this propensity towards settlements are deeply rooted in the adversarial nature of the US legal system. The system is very much ‘lawyer-heavy’ rather than ‘judge-heavy’.49 Effectively, ‘American lawyers do more because American judges do less’.50 In the adversarial system, opposing counsel interrogate and cross-examine witnesses, select expert witnesses, identify the relevant case law, and provide evidence that has often been obtained through pre-trial discovery. The judge therefore has a rather passive role in the adversarial system,51 leading to a relatively low number of judges as compared to the number

Ibid, 440. Stephen Calkins, ‘Summary Judgment, Motion to Dismiss, and Other Examples of Equilibrating Tendencies in the Antitrust System’ (1985–86) 74 Georgetown Law Journal 1065. For an example describing these tendencies in relation to the US Walker Process Doctrine, see Sven Gallasch, ‘AstraZeneca v the Walker Process – A Real EU–US Divergence or Just an Attempt to Compare Apples to Oranges?’ (2011) 7 European Competition Journal 505. 47 Crane (n 39), 63. 48   Aro Corp v Allied Witan Co, 531 F 2d 1368, 1372 (6th Cir 1976): ‘The efficiency-enhancing objectives of a patent settlement are clear, and “[p]ublic policy strongly favors settlement of disputes without litigation”.’ See also Schlegal Mfg Co v USM Corp, 525 F 2d 775, 783 (6th Cir 1975): ‘The importance of encouraging settlement of patent-infringement litigation … cannot be overstated.’ 49 John Langbein, ‘The German Advantage in Civil Procedure’ (1985) University of Chicago Law Review 823, 846. 50 Robert A Kagan, Adversarial Legalism: The American Way of Law (Harvard University Press 2001) 105. 51 Stephen M Bundy, ‘The Policy in Favor of Settlement in an Adversary System’ (1992) 44 Hastings Law Journal 1, 8. 45 46

346  Research handbook on methods and models of competition law of lawyers in a nonetheless very litigious society.52 In fact, the caseload for the US Courts of Appeals has increased significantly from 47,000 in 1992 to its peak of 70,300 in 2006 and 58,951 in 2017.53 During this entire period, the number of authorized judges has remained constant.54 It is therefore not surprising to find judicial statements such as in Ehrheart v Verizon Wireless, where the Court stated that ‘[s]ettlement agreements are to be encouraged because they promote the amicable resolution of disputes and lighten the increasing load of litigation faced by the federal courts’.55 This increased emphasis on settlements over adjudication is also reflected in the revisions to the Federal Rules of Civil Procedure. For example, the 1983 amendment to rule 16 added the ‘facilitating of settlement’ to the key objectives for pre-trial conferences.56 Other consequences of this kind of ‘lawyer-heavy’ arrangement are the high direct and indirect costs of litigation. The parties generally bear most of the costs of litigation as the costs of discovery, evidence production and legal counsel are not covered by the government. The more extensive the pre-trial discovery is, the longer the preparation of the lawsuit takes; and the more complex the litigation is, the more likely it is that expert witnesses will be required.57 These factors, in addition to the fact that lawyers are often paid on an hourly basis, add to the costs of litigation. A settlement might therefore be the only cost-effective way to end litigation. It might even be the only way that certain litigants can afford the undertaking.58 Indirect cost factors that need to be considered are the delay caused by a judiciary that is under strain due to increased caseloads and the unpredictability of the outcome of litigation. Both factors can be mitigated by settlements. The delays for litigants are extensive if they insist on a trial. For example, in 2014, the average time from filing to trial in US District Courts ranged from 21.8 months to 39.6 months.59 Having to wait for up to three years to get a case heard not only significantly adds to the litigation costs, but also increases the legal uncertainty for the parties. According to Kagan, another factor that needs to be considered in relation to legal uncertainty in the US is the inherent legal unpredictability of the US adversarial system. The US judiciary has to some degree a political character, with judges often receiving little formal 52 Samuel R Gross and Kent D Syverud, ‘Don’t Try: Civil Jury Verdicts in a System Geared to Settlement’ (1997) 44 UCLA Law Review 3. 53 United States Courts, ‘Federal Judicial Caseload Statistics’, http://​www​.uscourts​.gov/​statistics​ -reports/​analysis​-reports/​federal​-judicial​-caseload​-statistics. 54 Samuel P Jordan, ‘Early Panel Announcement, Settlement, and Adjudication’ (2007) Brigham Young University Law Review 55. 55 Ehrheart v Verizon Wireless, 609 F 3d 590, 595 (3d Cir 2010). 56 Fed R Civ P 16(a)(5); for further detail, see John H Langbein, ‘The Disappearance of Civil Trial in the United States’ (2012) 122 Yale Law Journal 522. 57 United States v Glens Falls Newspapers, Inc, 160 F 3d 853, 856–7 (2d Cir 1998): ‘Where a case is complex and expensive, and resolution of the case will benefit the public, the public has a strong interest in settlement. The trial court must protect the public interest, as well as the interests of the parties, by encouraging the most fair and efficient resolution.’ 58 Albert W Alschuler, ‘The Vanishing Civil Jury’ (1990) University of Chicago Legal Forum 1, 5: ‘[T]he civil trial is on its deathbed, or close to it, because our trial system has become unworkable. The American trial has been bludgeoned by lengthy delays, high attorneys’ fees, discovery wars, satellite hearings, judicial settlement conferences and the world’s most extensive collection of cumbersome procedures. Few litigants can afford the cost of either the pre-trial journey or the trial itself.’ 59 United States Courts, ‘U.S. District Courts – Judicial Business 2015’ (2015) Table T-3, http://​ www​.uscourts​.gov/​statistics​-reports/​us​-district​-courts​-judicial​-business​-2015.

Pay for delay in perspective  347 training as judges, having been practising lawyers, prosecutors or political activists in their previous careers.60 Although he concedes that this approach might make the US system more pragmatic and more willing to adapt to changing circumstances, it also entails a higher level of legal inconsistency and unpredictability.61 Furthermore, the fact that the US legal system allows for jury trial in civil litigation adds to the unpredictability of the outcome. Civil juries have been described as an engine for inconsistency as they consist of a random selection of lay citizens who do not need to explain or justify their decision.62 All these factors incentivize the parties to settle disputes. Having said that, the settlement should not be too different to the outcome of the fully litigated trial. Because of the extensive pre-trial discovery, the parties can share all the relevant evidence with each other and therefore predict their chances of success fairly accurately. Settlement negotiations occur on the basis of the same evidence as the trial and are likely to come to a similar outcome.63 Once the parties’ expectations about the probable outcome are aligned, it is in fact most sensible to settle, as any delay is only likely to increase their respective costs.64 Apart from cost-efficiency and the creation of legal certainty, or rather the avoidance of judicial unpredictability, settlements are also encouraged from an institutional perspective in order to keep the complex and ‘lawyer-heavy’ system of US-style adversarial legalism afloat. 3

Intellectual property and antitrust – an interface in flux [T]he essence of a patent grant is the right to exclude others from profiting by the patented invention; thus, a patent is an exception to the general rule against monopolies and to the right of access to a free and open market.65

This quote of the US Supreme Court in 1945 seems to suggest that intellectual property law and antitrust law are at odds with each other and, further, that a patent can be construed as a legal exemption from antitrust liability. Nowadays, it is widely accepted that the two policies are not necessarily at odds with each other, nor in conflict. Both policies try to promote economic welfare and innovation, albeit by different means.66 However, whenever intellectual Kagan (n 50), 112. Ibid. 62 George L Priest, ‘Justifying the Civil Jury’ in Robert E Litan (ed), Verdict: Assessing the Civil Jury System (Brookings Institution 1993) 130. 63 See Robert D Cooter and Daniel L Rubinfeld, ‘An Economic Model of Legal Discovery’ (1994) 23 Journal of Legal Studies 435, 445. 64 It is accepted that this outcome is only possible under a number of strong assumptions, including the equality of adequate legal representation for all parties and the rationality of all parties. However, for the purpose of this chapter, this basic underlying rationale shall suffice. For a more nuanced discussion, see Bundy (n 51). 65 Precision Instrument Manufacturing Co v Automotive Maintenance Machinery Co, 324 US 806, 816 (1945). 66 Mark A Lemley, Herbert Hovenkamp, Mark D Janis and Christopher R Leslie, IP and Antitrust: An Analysis of Antitrust Principles Applied to Intellectual Property Law (2nd edn, Aspen Publishers 2010) 1–13. This notion is also reflected in the most recent Antitrust Guidelines for the Licensing of IP jointly issued by the DOJ and the FTC in 2017. See US Department of Justice and Federal Trade Commission, Antitrust Guidelines for the Licensing of Intellectual Property (2017) 2, https://​www​ .justice​.gov/​atr/​IPguidelines/​download. 60 61

348  Research handbook on methods and models of competition law property policy, especially patent policy, and antitrust policy intersect, the same question arises over and over again: which of the two policies should prevail? The answer to this question has been in flux over the last decades. At times, the scope of intellectual property protection expanded, which led to the recession of antitrust protection and vice versa. Often a pendulum is used as an analogy to describe the situation. In the 1930s–40s and again in the 1970s, the intellectual property and antitrust landscape could be characterized as ‘antitrust heavy’. In the early years, patent misuse cases that involved the tying of unpatented and patented products amounted to prima facie antitrust violations.67 In the 1970s, the Antitrust Division of the Department of Justice issued its infamous ‘nine no-nos’ of patent licensing.68 By the 1950s, the pendulum had already swung back once with the passing of the Patent Act 1952, which contained a number of exemptions to patent misuse doctrine.69 However, the far more decisive swing towards the expansion of patent protection took place with the creation of the Court of Appeals for the Federal Circuit (Federal Circuit) in 1981, the specialized appeals court for patent disputes. In a number of high-profile cases, the Federal Circuit expanded not only the scope of patent protection, but also – and more importantly – its jurisdiction to antitrust law, essentially holding that patent law ‘trumps’ antitrust law. In Nobelpharma,70 the Federal Circuit effectively extended its jurisdiction beyond the issues of patent law to encompass the antitrust laws. Before Nobelpharma, the Court applied its ‘own law’ only to matters that involved its exclusive patent law jurisdiction and relied on the regional circuit law for all other matters, including antitrust law.71 However, with this judgment, the Federal Circuit reversed its position on this issue and expressly changed the precedent by stating: [A]n antitrust claim premised on stripping a patentee of its immunity from the antitrust laws is typically raised as a counterclaim by a defendant in a patent infringement suit. Because most cases involving these issues will therefore be appealed to this court, we conclude that we should decide these issues as a matter of Federal Circuit law, rather than rely on various regional precedents. We arrive at this conclusion because we are in the best position to create a uniform body of federal law on this subject and thereby avoid the ‘danger of confusion [that] might be enhanced if this court were to embark on an effort to interpret the laws’ of the regional circuits.72

Henceforth, the Federal Circuit seems to approach the patent/antitrust intersection firmly from the patent side. It has been suggested that the core antitrust principles have become blurred by

67 Hovenkamp et al (n 66), 1–17, referring to cases such as Mercoid Corp v Mid-Continent Inv Co (Mercoid I), 320 US 661 (1943) and Carbice Corp v American Patents Dev Corp, 283 US 27 (1931). 68 See Antitrust Division, Statement on Patent Licensing (21 September 1972) in 4 Trade Reg Rep ¶13126. As noted by Hovenkamp in ‘IP and Antitrust Policy: A Brief Historical Overview’ (University of Iowa Legal Studies Research Paper 05-31, 2005) fn 4, the list included: (1) licences requiring the purchase of unpatented products or preventing the licensee from dealing in products outside the scope of the patent; (2) grant-back agreements requiring the licensee to assign back to the licensor any improvement patents developed by the licensee; (3) restrictions on the resale of the patented product; (4) preventing the licensee from dealing in products outside the scope of the patent; (5) agreements by licensors not to grant further licences to others; (6) mandatory package licences; (7) royalty provisions not reasonably related to sales; (8) restrictions on the licensee’s use of a product manufactured by a patented process; and (9) resale price maintenance of licensed products. 69 Hovenkamp et al (n 66), 1–18. 70 Nobelpharma AB v Implant Innovations, Inc, 141 F 3d 1059 (CA Fed 1998). 71 Hovenkamp et al (n 66), 5–36. 72 Nobelpharma AB v Implant Innovations, Inc (n 70), 1068 (citations omitted).

Pay for delay in perspective  349 the Court, as it views antitrust through the patent lens.73 In Xerox,74 the Court indeed adopted a very narrow view by means of a very broad and sweeping rule that essentially provided a patent owner with an antitrust immunity for any conduct that was considered to be enforcing the patent right, while at the same time the enforcement of patent rights was almost equated with the exercise or exploitation of the intellectual property right.75 One could argue that this narrow view is a direct result of the case-specific issues that the courts had to deal with and should therefore not be generalized. Unilateral refusal to supply cases and especially compulsory licensing of intellectual property cases are, after all, the most intrusive requests that can be sought under the antitrust laws. However, when considering the horizontal agreements in which the scope of the patent test is considered, the Court’s approach is not too far off the very narrow view expressed by the Court in Xerox. Although the Supreme Court has not provided the defendant in cases concerning horizontal patent settlements with a blanket antitrust immunity akin to Xerox, it accepted that such settlements can only violate the antitrust laws under the most limited set of circumstances – that is, when the settlement is based on sham litigation76 or the litigation involves a fraudulently acquired patent.77 Ultimately, this development not only made the Federal Circuit a ‘major arbiter of antitrust law’, but also laid out the difficult task ahead of the FTC. The pendulum had firmly swung towards the expansion of intellectual property right protection, almost dwarfing the impact antitrust law could have on patent rights. 4 The long road to Actavis These considerations must be kept in mind when examining the FTC’s enforcement efforts in relation to pay-for-delay settlements. The agency has enforced section 5 of the FTC Act

73 James B Gambrell, ‘The Evolving Interplay of Patent Rights and Antitrust Restraints in the Federal Circuit’ (2001) 9 Texas Intellectual Property Law Journal 137, 141–3, 148. Gambrell goes even further, suggesting that the Federal Circuit’s expansion of its jurisdiction is ultra vires as it goes beyond the jurisdiction that Congress had envisaged for the Federal Circuit. 74 In re Independent Service Organizations Antitrust Litigation, 203 F 3d 1322 (Fed Cir 2000). 75 Peter M Boyle, Penelope M Lister and JC Everett, ‘Antitrust Law at the Federal Circuit: Red Light or Green Light at the IP–Antitrust Intersection?’ (2002) 69 Antitrust Law Journal 739, 757. See In re Independent Service Organizations Antitrust Litigation (n 74), 1327, 1328: We have held that ‘if a [patent infringement] suit is not objectively baseless, an antitrust defendant’s subjective motivation is immaterial’. We see no more reason to inquire into the subjective motivation of Xerox in refusing to sell or license its patented works than we found in evaluating the subjective motivation of a patentee in bringing suit to enforce that same right. In the absence of any indication of illegal tying, fraud in the Patent and Trademark Office, or sham litigation, the patent holder may enforce the statutory right to exclude others from making, using, or selling the claimed invention free from liability under the antitrust laws. We therefore will not inquire into his subjective motivation for exerting his statutory rights, even though his refusal to sell or license his patented invention may have an anticompetitive effect, so long as that anticompetitive effect is not illegally extended beyond the statutory patent grant. It is the infringement defendant and not the patentee that bears the burden to show that one of these exceptional situations exists and, in the absence of such proof, we will not inquire into the patentee’s motivations for asserting his statutory right to exclude. (Citations omitted, emphasis added.) 76 Professional Real Estate Investors, Inc v Columbia Pictures Industries, Inc, 508 US 49, 60–61 (1993). 77 Walker Process Equipment, Inc v Food Machinery & Chemical Corp, 382 US 172, 177 (1965).

350  Research handbook on methods and models of competition law on an administrative level by means of complaint and in front of the courts on appeal, as well as filing numerous amicus curiae briefs in support of plaintiffs in private antitrust actions. Discussing all these different strands of the FTC’s enforcement agenda chronologically will highlight the difficulties that the FTC faced from 2001 until 2013, when the US Supreme Court handed down its judgment in FTC v Actavis. (a) Schering-Plough as the starting point Although the FTC’s administrative complaint against Schering-Plough (Schering) was not the first complaint in relation to pay-for-delay settlements,78 it can be described as the starting point. Schering was the first pharmaceutical company to appeal an FTC pay-for-delay decision to the courts, making it the first fully litigated pay-for-delay decision, resulting in a hard-fought battle between the FTC and the pharmaceutical industry that lasted almost a decade. The FTC’s administrative complaint, which was brought on 2 April 2001, focused on settlement agreements between Schering and two generic pharmaceutical firms in relation to a potassium supplement, branded K-Dur 20.79 According to the complaint, the generic company Upsher-Smith intended to enter the market with a generic version of K-Dur 20 in 1995. However, it was sued for patent infringement by Schering, the proprietor of K-Dur 20, which was patent protected until 2006. As part of the settlement, Upsher-Smith agreed not to enter the market with a generic version of the drug until 2001. In return, it received a payment of $60 million. In 1998, in a second settlement on very similar terms, another potential generic entrant, ESI Lederle, agreed not to enter the market with a generic version until 2004, and between January 2004 and June 2006 only with a single generic product. ESI Lederle received $15 million in return. In both cases, the generic companies provided Schering with licences for the generic product, which were deemed to be of little value in relation to the respective payments. The FTC charged the three companies with a violation of section 5 of the FTC Act, alleging that the agreements constituted unreasonable restraints on trade and that the companies had conspired to monopolize the market for potassium supplements.80 According to the FTC’s reasoning, the delay in entry was caused by the respective payments that induced the generic companies not to enter the market. Absent the payments, the delay would not have occurred or at least would have been for a shorter period. In essence, Schering exploited the regulatory peculiarities of the Hatch Waxman Act and effectively sheltered its brand drug K-Dur 20 from any generic challenges until 2004. Following the rejection of the FTC’s complaint by the administrative law judge, this initial decision was appealed to the full Commission. In its final decision, the FTC reversed the administrative law judge’s initial decision, essentially finding the pay-for-delay settlements between Schering and the two generic companies, Upsher and ESI, to be anticompetitive, 78 Scott A Backus, ‘Reversing Course on Reverse Payment Settlements in the Pharmaceutical Industry: Has Schering-Plough Created the Blueprint for Defensible Antitrust Violations Comment?’ (2007) 60 Oklahoma Law Review 393. 79 Federal Trade Commission, ‘FTC Charges Schering-Plough over Allegedly Anticompetitive Agreements with Two Other Drug Manufacturers’ (Federal Trade Commission Press Releases, 2 April 2001), https://​www​.ftc​.gov/​news​-events/​press​-releases/​2001/​04/​ftc​-charges​-schering​-plough​-over​ -allegedly​-anticompetitive. 80 Federal Trade Commission, Administrative complaint in the matter of Schering-Plough Corporation, et al, Docket 9297 (2001).

Pay for delay in perspective  351 advocating what can be described as a presumptive illegality test.81 The opinion, written by Commissioner Leary, set out in detail the agency’s reasoning with regard to the main issues: the antitrust scrutiny of settlements, the need to address the merits of the underlying patent, and the size of the ‘reverse’ payment. The opinion acknowledges that patent settlements are generally not illegal, even though they delay generic entry for some time before the underlying patent expires.82 Settlements not only reduce litigation costs, but also create legal certainty for the litigating parties. Such legal certainty is especially welcome in cases involving patents as the litigation outcome is inherently uncertain. The settlement therefore results in a compromise between the parties, reflecting the differing litigation expectations. However, the key concern with pay-for-delay settlements is that the generic delay is the reflection not of a compromise between the parties based on the litigation costs and likely litigation outcomes, but rather of the payment from the brand company to the generic company that includes consideration for the value of reduced competition. The FTC effectively relied on the above-mentioned economic rationale of the win–win situation for both parties to the settlement, the same rationale underlying the US Supreme Court’s decision in Actavis. Arguing that the payment effectively altered the generic entry date to which the parties otherwise would have agreed, the FTC found that ‘absent the proof of other offsetting consideration, it is logical to conclude that the quid pro quo for the payment was an agreement by the generic to defer entry beyond the date that represents an otherwise reasonable litigation compromise’.83 By holding that pay-for-delay settlements raise a red flag warranting antitrust scrutiny and should be distinguished from most other patent settlements, the FTC expressly dismissed the need to address the merits of the underlying patent dispute. Instead, the FTC focused on the likely effect of the payment on the actual entry date. Although patents are presumed to be valid by statute,84 this presumption does not necessarily translate into a right to exclude.85 Patents are rather probabilistic in nature and should thus only confer the right to try to exclude.86 Furthermore, patent validity is not necessarily linked to patent infringement. A patent can be valid but nonetheless not infringed by the generic entrant. It would thus be wrong to refer simply to the presumption of validity in an inquiry into patent infringement. The FTC also argued that the agency should have a different standard of proof as it is not attempting to assess damages but rather is determining whether the agreement constitutes an unreasonable restraint to trade.87 The agency thus argued that the differences in the burden of proof should also result in different remedies sought, effectively pleading that the remedies 81 This opinion marks a change in the FTC’s approach to pay-for-delay settlements, having previously advocated a per se approach. Federal Trade Commission, In the Matter of Schering-Plough Corporation, et al (Final opinion), Docket No 9297 [2003], 29: ‘We are not now prepared to say that all such payments should be viewed as per se illegal or “inherently suspect”. We believe that this particular case warrants a more extensive analysis of competitive effects, without foreclosing the possibility that a more truncated process would be appropriate in some future case.’ 82 Ibid, 25. 83 Ibid, 26. 84 35 USC § 282. 85 In the Matter of Schering-Plough Corporation, et al, Docket No 9297 (n 80), 30. 86 See Mark A Lemley and Carl Shapiro, ‘Probabilistic Patents’ (2005) 19 Journal of Economic Perspectives 75; Carl Shapiro, ‘Antitrust Limits and Patent Settlements’ (2003) 34 RAND Journal of Economics 391, 395. 87 In the Matter of Schering-Plough Corporation, et al, Docket No 9297 (n 80), 31.

352  Research handbook on methods and models of competition law sought by the FTC do not have to be ‘equilibrated’ in light of private enforcement efforts and the associated potential over-compensation through the award of treble damages. (b) The FTC’s battle in the courts Against the backdrop of these arguments, one has to wonder why it took over a decade to reach the current state of play. The FTC’s reasoning is very similar to the underlying rationale of the US Supreme Court’s decision in Actavis. Yet, the FTC had to defend its reasoning on appeal in front of the Court of Appeal for the Eleventh Circuit, and the circuit courts in general were a lot more divided and a lot less accepting of the agency’s reasoning. While the FTC was writing its final decision in Schering, the District Courts already had diverging views on how, and if at all, pay-for-delay settlements should be within the remit of the antitrust laws. On one side, the direct purchasers of the brand company Hoechst Marion Roussel, Inc were successful with their lawsuit in In re Cardizem Antitrust Litigation, claiming that the pay-for-delay settlement between Hoechst and Andrx Pharmaceuticals violated federal and state antitrust laws.88 On appeal, the Sixth Circuit held that the pay-for-delay settlement constituted a per se violation of antitrust law, as the agreement was a ‘horizontal market allocation agreement’,89 which ‘at its core, [had the aim] to eliminate competition … throughout the entire United States’.90 On the other side, the agency faced District Court decisions rejecting claims of antitrust violations in Tamoxifen and Valley Drug. The Eastern District Court of New York dismissed the claim for antitrust violation in Tamoxifen, where the parties entered into a pay-for-delay settlement, and the generic company agreed not to enter the market with its own generic drug until the patent had expired.91 The District Court held that the plaintiff failed to prove an antitrust injury as the higher prices for Tamoxifen resulted from the existence of the relevant patent and the associated lawful patent monopoly, not from an anticompetitive settlement agreement.92 The FTC acknowledged the District Court’s Tamoxifen judgment in its final decision in Schering, yet argued that the judgment would not contradict its decision nor have an impact on its finding. Crucially, the relevant patent in Tamoxifen was successfully defended by the brand company three times. The settlement therefore did not cause consumer harm. Awarding antitrust damages in this private action would have been unreasonable. According to the FTC, the aim of the administrative action in Schering was not the assessment of damages. Rather, the FTC asserted that its aim was the determination of legal liability with regard to an agreement that might have constituted an unreasonable restraint in trade,93 essentially noting that its administrative action would not require the courts to follow up and ‘equilibrate’ the outcome. The Eleventh Circuit in Valley Drug also rejected per se liability for pay-for-delay settlements. The Court considered the impact of a patent that was found to be invalid in separate proceedings after the pay-for-delay settlement had been concluded. The Court found that a per se liability leading to treble damages would be unreasonable if the claim was merely based on a patent that had subsequently been found to be invalid. This finding was supported by rea-

90 91 92 93 88 89

In re Cardizem CD Antitrust Litigation, 105 F Supp 2d 682 (ED Mich 2000). In re Cardizem CD Antitrust Litigation, 332 F 3d 896, 900 (6th Cir 2003). Ibid, 905. In re Tamoxifen Citrate Antitrust Litigation, 277 F Supp 2d 121 (ED New York 2003). Ibid, 137. In the Matter of Schering-Plough Corporation, et al, Docket No 9297 (n 80), 31.

Pay for delay in perspective  353 soning that ‘patent litigation is too complex and the results too uncertain for parties to forecast whether enforcing the exclusionary right through settlement will expose them to treble damages’.94 The fact that the Court referred to the ‘mere’ invalidity of the patent suggests that it requires aggravating factors, such as sham litigation, putting a strong emphasis on the patent’s presumption of validity. At the time, the FTC did not oppose the District Court judgment. Instead, it used the judgment to highlight the benefits of refraining from an inquiry into the merits of a patent. According to the FTC, such an inquiry would be unreliable and unnecessary. Whereas patent litigation is characterized by true adversary proceedings, a settlement between the parties leads to the alignment of interest.95 However, the parties’ appeal of the FTC’s final decision in Schering proved to be a turning point for the agency’s efforts to combat pay-for-delay settlements. It did not come as a surprise that the parties chose the Eleventh Circuit for their appeal.96 The Court’s rejection of the per se approach and the strong emphasis of the presumption of validity made the Eleventh Circuit the circuit of choice to appeal the FTC’s final decision in Schering. It is fair to say that ‘the court did not disappoint the manufacturers … and blasted the FTC’s rule of reason analysis and finding of antitrust violation’.97 The Court saw itself bound by its decision in Valley Drug and reversed the FTC’s final decision in Schering as, in its view, an antitrust analysis was ill-suited to determine anticompetitive effects for the challenged conduct.98 The reason for the reversal was rather simple, according to the Court: ‘one of the parties owned a patent’.99 The Court went on to state that ‘by their nature, patents create an environment of exclusion, and consequently, cripple competition’, which meant that the anticompetitive effect is already present.100 In the light of these findings, the Court felt that the FTC had made its decision before considering any contrary conclusion.101 The Court therefore opted for the ‘scope of the patent test’, which was in line with Valley Drug, holding that antitrust liability would only be possible if the anticompetitive effects arise from conduct that exceeds the exclusionary scope of the patent in question. By relying so heavily on Valley Drug, the Court rejected the FTC’s reasoning that administrative actions in front of the FTC should have a different standard compared to private actions based on the Sherman Act. The Court also put a very strong emphasis on presumption of validity for patents, seemingly conflating patent validity with patent infringement. Even if there is the presumption of validity, infringement has to be proven. However, the over-emphasis by the Court on the presumption of validity essentially allowed parties to cover acts of infringement by a pay-for-delay settlement, which rather resembles a naked market-sharing agreement.102

Ibid, 33. Ibid, 34. 96 The FTC Act allows corporations to file an appeal against FTC section 5 decisions within any circuit in which the conduct in question is used, or in which the corporations reside or carry on business (15 USC §45(c)). 97 Backus (n 78), 399. 98 Schering-Plough Corp v FTC, 402 F 3d 1056, 1066 (11th Cir 2005). 99 Ibid, 1064. 100 Ibid, 1066. 101 Ibid, 1065. 102 Phillip E Areeda and Herbert Hovenkamp, Antitrust Law: An Analysis of Antitrust Principles and Their Application (2nd edn, Aspen Law & Business 2013) [2046c]. 94 95

354  Research handbook on methods and models of competition law The downward spiral for the FTC’s enforcement efforts in front of the courts did not end with the Eleventh Circuit decision in Schering. In Tamoxifen, the Second Circuit relied extensively on the opinions of the Eleventh Circuit in Valley Drug and Schering and expanded on the Eleventh Circuit’s reasoning.103 In the decision concerning a private antitrust action against a pay-for-delay settlement between Zeneca, Inc and Barr Pharmaceutical, Inc, the Court held that pay-for-delay settlements are per se legal and do not lead to an antitrust violation as long as the underlying patent has not been obtained by fraud or a suit for its enforcement is not objectively baseless.104 The per se legality of settlements in patent disputes was regarded by the Court as paramount, ‘even if it leads in some cases to the survival of monopolies created by what would otherwise be fatally weak patents’.105 Just a few years later, the Court of Appeal for the Federal Circuit weighed in on the issue. In Cipro, the Court of Appeal affirmed the District Court’s finding that a pay-for-delay settlement between Bayer and Barr Pharmaceuticals did not violate antitrust laws.106 It sided with the Eleventh and Second Circuits in finding that the ‘scope of the patent’ test is the correct line of inquiry and that patent settlements are per se legal unless they are obtained by fraud or sham litigation.107 Furthermore, the Court of Appeal expressly rejected the FTC’s argument that a rule of reason analysis should take into consideration an assessment of the strength of the underlying patent.108 In fact, the Court of Appeal held that patent validity should only become an issue when determining whether the patent was obtained by fraud or enforced through sham litigation. This led to a situation where patent law essentially trumped antitrust law and where patents in a pay-for-delay settlement effectively carried a non-rebuttable presumption of validity and infringement.109 (c) The creation of a circuit split Until the Federal Circuit’s decision in Cipro, the FTC had been unsuccessful in appealing for judicial review by the US Supreme Court. The agency had filed a petition for a writ of certiorari twice in the past: once in the case of Schering and again in Cipro. In both cases, the petition was denied. Knowing that the chances of judicial review by the US Supreme Court would increase drastically if two circuit courts fundamentally disagreed on the same legal issue, the FTC focused on creating a circuit split. The next opportunity that presented itself was the FTC’s complaint in FTC v Watson Pharmaceuticals, Inc et al, concerning the pay-for-delay settlement in relation to Solvay Pharmaceutical’s brand drug AndroGel. The In re Tamoxifen Citrate Antitrust Litigation, 466 F 3d 187 (2nd Cir 2005). Ibid, 213. 105 Ibid, 212. 106 In re Ciprofloxacin Hydrochloride Antitrust Litigation, 544 F 3d 1323 (Fed Cir 2008), cert denied, 129 S Ct 2828 (2009). 107 Ibid, 1335. 108 Federal Trade Commission, Brief as amicus curiae in In re Ciprofloxacin Hydrochloride Antitrust Litigation [2008] 15–18, https://​www​.ftc​.gov/​sites/​default/​files/​documents/​amicus​_briefs/​re​ -ciprofloxacin​-hydrochloride​-antitrust​-litigation/​ciprobrief​.pdf. 109 In re Ciprofloxacin Hydrochloride Antitrust Litigation (n 106), 1336: We conclude that in cases such as this, wherein all anticompetitive effects of the settlement agreement are within the exclusionary power of the patent, the outcome is the same whether the court begins its analysis under antitrust law by applying a rule of reason approach to evaluate the anti-competitive effects, or under patent law by analyzing the right to exclude afforded by the patent. 103 104

Pay for delay in perspective  355 violation of section 5 of the FTC Act was based on the reasoning that parties were aware of the fact that Solvay would find it difficult to prevent generic entry based on its patent as the generic companies did not question the validity of the patent itself. Instead, the generic companies ‘amassed substantial evidence that their generic products did not infringe the formulation patent and that the patent was invalid and/or unenforceable’.110 Solvay was therefore not likely to prevail in court and the pay-for-delay settlement ensured that Solvay’s monopoly was extended. Following the dismissal of the complaint by the District Court, the decision was appealed to the Eleventh Circuit. Based on past practice of the Court in relation to pay-for-delay settlements, it was not surprising that the Court affirmed the District Court’s decision.111 The Court expressly declined the FTC’s invitation ‘to adopt a rule that an exclusion payment is unlawful if, viewing the situation objectively as of the time of the settlement, it is more likely than not that the patent would not have blocked generic entry earlier than the agreed-upon entry date’.112 Instead, the Court focused on ‘the potential exclusionary effect of the patent, not the likely exclusionary effect … as one side or the other almost always has a better chance of prevailing, but a chance is only a chance, not a certainty’.113 However, in 2012 the FTC’s losing streak came to an end with the Third Circuit’s decision in K-Dur.114 Interestingly, the Third Circuit had to deal with the very same pay-for-delay settlement that was found by the Eleventh Circuit to be immune from antitrust scrutiny in Schering. The private action filed by direct purchasers of K-Dur challenged the settlement as a Sherman Act violation and this was appealed to the Third Circuit following the grant of summary judgment for the defendants by the District of New Jersey. Following the review of the decisions of the Second, Eleventh and Federal Circuits, the Court expressly rejected the ‘scope of the patent test’ that had been endorsed by the other circuits. Instead, it opted for a ‘quick look’ approach that deemed pay-for-delay settlements to be prima facie unlawful. First, the Court took issue with the ‘scope of the patent test’s almost un-rebuttable presumption of patent validity [for which] it could identify no significant support’.115 A number of FTC studies were cited to show the predominant weaknesses of pharmaceutical patents and allude to the fact that the majority of settlements concluded in the pharmaceutical sector did not include a reverse payment. This suggested that the Court’s approach would not be over-interventionist or dissuade the parties to settle.116 The Court also agreed with the FTC that there is no need to consider the merits of the underlying patent, directly citing the FTC’s final decision in Schering that ‘[a]bsent proof of other offsetting consideration, it is logical to conclude that the quid pro quo for the payment was an agreement by the generic to defer entry beyond the date that represents an otherwise reasonable litigation compromise’.117 Finally, the Court also dismissed the argument in favour of patent settlements by stating that ‘the judicial preference for settlement, while generally laudable, should not displace countervailing public policy objec-

112 113 114 115 116 117 110 111

Federal Trade Commission v Watson Pharmaceuticals, Inc [2009] 21. Federal Trade Commission v Watson Pharmaceuticals, Inc, 677 F 3d 1298 (11th Cir 2012). Ibid, 1312. Ibid, 1313. In re K-Dur Antitrust Litigation, 686 F 3d 197 (3d Cir 2012). Ibid, 214. Ibid, 215, 218. Ibid, 218, citing In re Schering-Plough Corp (Final Order), 136 FTC 956, 988 (2003).

356  Research handbook on methods and models of competition law tives or, in this case, Congress’s determination … that litigated patent challenges are necessary to protect consumers from unjustified monopolies by name brand drug manufacturers’.118 This strong rejection of the scope of the patent test by the Third Circuit allowed the FTC to file a petition for writ of certiorari in the US Supreme Court appealing for judicial review of the issue. After all, the Third Circuit did not just slightly disagree with the case law set by the other circuits; it expressly disagreed with the Eleventh Circuit on the legal assessment of the very same pay-for-delay settlement. It therefore did not come as a surprise when the US Supreme Court granted the appeal in FTC v Watson Pharmaceuticals in order to clarify the law of the land discussed above.119 (d) FTC v Actavis Being granted the petition for writ of certiorari in the case against Solay Pharmaceuticals was, however, only the first hurdle that the FTC had to overcome. Just like the circuit courts, the Supreme Court was divided. Chief Justice Roberts wrote a strongly worded dissent in favour of the proponents of the scope of the patent test, which was joined by Justice Scalia and Justice Thomas. The dissent starts by stating: ‘Solvay Pharmaceuticals holds a patent. … A patent carves out an exception to the applicability of antitrust laws. The correct approach should therefore be to ask whether the settlement gives Solvay monopoly power beyond what the patent already gave it.’120 This opening passage set the tone for the FTC’s task ahead. The agency had to persuade the Supreme Court that, despite the fact that pay-for-delay agreements constitute patent settlements, the Court should place less reliance on the two key presumptions discussed above: the presumption of patent validity, and that settlements should be encouraged and therefore honoured. This is exactly what the majority opinion did. Justice Breyer did not question that anticompetitive effects can arise from the exclusionary nature of a patent but held that a simple reliance on the presumed validity of a patent is tenable in relation to pay-for-delay settlements. In short, the presumption of patent validity should not automatically be equated with antitrust immunity. The majority opinion acknowledged the FTC’s theory of harm and recognized the anticompetitive potential posed by the parties’ aim to prevent the risk of competition. It thus held that ‘it would be incongruous to determine antitrust legality by measuring the settlement’s anticompetitive effects solely against patent law policy, rather than by measuring

Ibid, 217. This case turned into FTC v Actavis. 120 FTC v Actavis (n 1), 2238. See also Justice Roberts’s opinion on the strong presumption that patent litigation should be barred from antitrust scrutiny but for only the narrowest of exceptions (ibid, 2242): The majority is therefore right to suggest that these ‘precedents make clear that patent-related settlement agreements can sometimes violate the antitrust laws’. The key word is sometimes. And those sometimes are spelled out in our precedents. Those cases have made very clear that patent settlements – and for that matter, any agreements relating to patents – are subject to antitrust scrutiny if they confer benefits beyond the scope of the patent. This makes sense. A patent exempts its holder from the antitrust laws only insofar as the holder operates within the scope of the patent. When the holder steps outside the scope of the patent, he can no longer use the patent as his defense. The majority points to no case where a patent settlement was subject to antitrust scrutiny merely because the validity of the patent was uncertain. Not one. It is remarkable, and surely worth something, that in the 123 years since the Sherman Act was passed, we have never let antitrust law cross that Rubicon. 118 119

Pay for delay in perspective  357 them against procompetitive antitrust policies as well’.121 Therefore, the Court did not find that pay-for-delay settlements are anticompetitive. It found that in its view such agreements can be anticompetitive and that the FTC should have been allowed to proceed with its lawsuit having to persuade the circuit court that the agreement in question should be found anticompetitive according to the rule of reason analysis set out in this judgment. 5 Comment What is striking about this discussion is the fact that the FTC’s reasoning in Schering has come full circle and ultimately prevailed in the Supreme Court. The very same line of argument that triggered the long losing streak for the FTC in front of the Second, Eleventh and Federal Circuits is the same line of argument that fundamentally led to the Supreme Court’s review of pay-for-delay settlements. The case law shows how the agency’s enforcement efforts took place at a time when patent protection was expanding in the courts while antitrust scrutiny and liability were receding. During those years, the courts elevated patent law over antitrust law by over-emphasizing validity and the associated strength of patents. With every loss in front of the Eleventh Circuit, the task for the FTC became even more of an uphill battle. The agency not only had to convince courts of its theory of harm, but also had to fight an increasing number of precedents that were deemed persuasive by the lower courts. The scope of the patent test received even more traction with the Cipro decision by the Federal Circuit, the specialist patent court for the US established in 1982. An added difficulty was that the courts were not willing to accept the agency’s argument that an administrative action by the FTC should follow a different standard compared with private actions alleging a violation of the Sherman Act. At the same time, one should not be too quick to say that the agency’s enforcement efforts were wasted. Over the years, the FTC’s theory of harm and arguments against pay-for-delay settlements evolved. They were further substantiated by detailed empirical reports and studies, as well as further advocacy measures. For example, an FTC empirical study found that settlement agreements with compensation from the brand company to the generic led to a delay in generic entry that was 17 months longer than in agreements without such payments. The same study also estimated that pay-for-delay settlements cost American consumers $3.5 billion per year.122 One should not forget that pay-for-delay settlements were initially deemed to be per se violations – an outcome that today would be considered overly interventionist. The K-Dur decision, in particular, and also the US Supreme Court decision in FTC v Actavis were to a degree guided and influenced by the evolution of the theory of harm and the studies and amicus briefs submitted by the FTC over time. One could therefore argue that this evolution was necessary to achieve the optimal outcome from a policy perspective. B

Pay-for-delay Antitrust Enforcement from the EU Perspective

1 The European Commission within the inquisitorial system In contrast to the US, European competition law is enforced by the European Commission’s Directorate-General for Competition (European Commission) as a single authority. The decision-making process is administrative in nature, similar to that of the FTC. However, the Ibid, 2231. FTC, ‘Pay-for-Delay: How Drug Company Pay-Offs Cost Consumers Billions’ (January 2010).

121 122

358  Research handbook on methods and models of competition law European adjudication process differs substantially from the adversarial system in the US. Whereas US antitrust enforcement is predominantly driven by private enforcement, EU competition law enforcement relies on public enforcement. It is not suggested that private enforcement of EU competition law is not encouraged, but the relationship between public and private enforcement can be described as rather hierarchical and complementary.123 With deterrence being the main driver for public enforcement in Europe, private enforcement is encouraged in order to achieve individual compensation for harm caused by anticompetitive conduct. Private enforcement is therefore vital for follow-on actions on a national level but does not play a key role in competition law enforcement on a supranational level. Unlike the FTC, the European Commission is not faced with constricted liability norms at a supranational level that are required in order to keep zealous private enforcement at bay. In addition, the administrative decision-making process of the European Commission is based on a quasi-inquisitorial system rather than the US adversarial system. This means that the decision-making body gathers all the necessary information and evidence required to make an informed decision. The European Commission is often represented by an impartial yet active judge who gathers the evidence from both parties, takes testimonies from witnesses, and finally renders a decision based on the facts.124 In competition law proceedings at a supranational level, however, it is often difficult to identify the plaintiffs in the investigation. Or, they might be identifiable but unorganized. For example, final consumers are victims of a cartel. In these situations, it becomes necessary for the agency to step in and ‘fill this vacancy’, essentially also becoming the prosecutor. Hence the public enforcement process in the EU has been described as being based on an inquisitorial system with a prosecutorial bias.125 Effectively, the European Commission can be regarded as prosecutor, judge and jury in one. The investigation is run by the European Commission, which makes the final decision on the competition law infringement126 and also sets the fine. This system has attracted substantial criticism from a human rights perspective, especially from a fair trial point of view.127 Keeping the human rights issues to one side,128 the prosecutorial bias also raises another concern: the European Commission’s unbalanced evidence gathering. The officials on the case team might focus too heavily on one side of the argument, failing to consider all evidence in front of them or even the counterfactual that would find that the procompetitive effects outweigh the anticompetitive potential. The fact that the European Commission is not free from such bias is highlighted by its decisions in Woodpulp 123 Alexander Italianer, ‘Public and Private Enforcement of Competition Law’ (5th International Competition Conference, Brussels, 17 February 2012) 3, http://​ec​.europa​.eu/​competition/​speeches/​text/​ sp2012​_02​_en​.pdf. 124 See generally Hein Kotz, ‘Civil Justice Systems in Europe and the United States’ (2003) 13 Duke Journal of Comparative & International Law 61. 125 DJ Neven, ‘Competition Economics and Antitrust in Europe’ (2006) 21 Economic Policy 742, 763. 126 Although the final decision is technically adopted by the college of commissioners and therefore a body separate to the Directorate-General for Competition, the college is likely having to reply on guidance by the Director-General for Competition, especially in economically complex cases. 127 See, for example, Donald Slater, Sébastien Thomas and Denis Waelbroeck, ‘Competition Law Proceedings before the European Commission and the Right to a Fair Trial: No Need for Reform?’ (College of Europe Research Papers in Law 5/2008, 2008). 128 It is not suggested that these concerns are unimportant or not worth considering. However, their discussion goes beyond the scope of this chapter.

Pay for delay in perspective  359 and Airtours. In these cases, the EU Courts quashed the European Commission’s decisions, expressly criticizing the agency’s failure to fully consider the counterfactual of arguments and evidence presented during the investigation. Yet, at the same time, these judgments by the EU Courts also represent the remedy for this shortcoming of the inquisitorial system with prosecutorial bias. The judicial review by the General Court and the Court of Justice ensures that biases in the investigatory stage or in the decision-making stage are mitigated, or at least that no adverse consequences flow from such decisions. Interestingly, the second pay-for-delay judgment by the General Court in Servier highlights the effectiveness of judicial review on a European level. The Court has largely upheld the Commission’s finding of a restriction by object that relies on the rationale in Lundbeck in relation to all but one settlement agreement in this case. However, with regard to the agreement between Servier and Krka, the Court found that the Commission had failed to sufficiently prove that a licensing agreement between the parties constituted a side-deal facilitating a pay-for-delay settlement.129 According to the Court, it is the Commission that has the burden of proving that the purpose of the licensing agreement is to conceal a pay-for-delay settlement rather than a licensing agreement concluded under normal market conditions.130 From a procedural efficiency perspective, the inquisitorial system, even with its prosecutorial bias, could be regarded as preferable to the adversarial system as it is less time intensive, resource intensive and costly than the alternative.131 The lack of private enforcement at the supranational level avoids blurring the lines between potentially different standards of proof. The European Commission and the EU Courts can focus on the question of legal liability without having to consider the legitimacy of the claim brought by the private party or, indeed, having to quantify private damages that would result from such a claim. 2 The perception of settlements The European reliance on this inquisitorial system in combination with judicial review has a profound effect on the perception of settlements between private parties.132 The above-mentioned factors in the adversarial system, which lead to the prioritization of settlements in the US such as pre-trial discovery, are not prevalent in Europe. The European Commission therefore does not have to overcome a strong bias in favour of settlements when it is investigating pay-for-delay settlements. Historically, private settlements and even those concerning patent infringement disputes did not have a special status under EU competition law. They were considered to be agreements like any other and were therefore also subject to competition law scrutiny. The European Court of Justice (ECJ) in fact expressly stated in a case concerning a patent settlement that included a licensing clause, as well as a non-challenge clause, that Article 101(1) ‘makes no distinction between agreements whose purpose is to put an end to litigation and those concluded with

Servier (n 3), paras 983–5. Ibid, para 949. 131 Kotz (n 124), 69, 70. 132 The perception of settlements discussed in this section should not be confused with the perception of settlements in the sense of commitment decisions by the Directorate-General for Competition following Article 9 of Regulation 1/2003, or indeed the settlement procedure of cartel investigations by the European Commission. 129 130

360  Research handbook on methods and models of competition law other aims in mind’.133 The Court furthermore expressed the view that non-challenge clauses would not be within the subject matter protected by a patent as they cannot be interpreted ‘as also affording protection against actions brought in order to challenge the patent’s validity, in view of the fact that it is in the public interest to eliminate any obstacle to economic activity, which may arise where a patent was granted in error’.134 This strict approach to settlements including non-challenge clauses has arguably been relaxed with the Technology Transfer Block Exemption Regulation (TTBER) in 2004,135 which was accompanied by relevant guidelines issued by the European Commission.136 According to the guidelines from 2004: non-challenge clauses [in the context of a settlement and non-assertion agreement] are generally considered to fall outside Article 81(1). It is inherent in such agreements that the parties agree not to challenge ex post the intellectual property rights covered by the agreement. Indeed, the very purpose of the agreement is to settle existing disputes and/or to avoid future disputes.137

This statement seems to suggest that the European Commission at least indirectly accepts the beneficial nature of settlements. However, it has to be kept in mind that the guidelines refer to settlements within the context of licensing agreements, not to pure settlements that simply bring litigation to an early resolution.138 In its guidelines on technology transfer agreements, the European Commission endorses settlements in general, acknowledging their potential welfare-enhancing benefits, including cost-efficiency and the creation of legal certainty, yet at the same time qualifying this finding by stating that the endorsement of settlements cannot supersede the general public interest to remove invalid patents as unmerited barriers to innovation and economic activity, relying expressly on the ECJ’s decision in Windsurfing.139 In addition, these guidelines limit the acceptability of patent settlement even further by expressly excluding pay-for-delay settlements, finding that they often ‘do not involve the transfer of technology rights, but are based on a value transfer from one party in return for a limitation on the entry and/or expansion on the market of the other party and may be caught by Article 101(1)’.140 The EU Courts and the European Commission have therefore taken a rather restrictive stance towards settlements. Whereas in the US settlements have been regarded very favourably and have only recently received increasing scrutiny from the authorities, Europe has had a rather sceptical view of settlements from a competition law perspective. This view has been slightly relaxed over the years – however, only in relation to technology transfer agreements.

Case 65/86, Bayer AG v Maschinenfabrik Hennecke GmbH & Heinz Süllhöfer [1988] ECR 5249,

133

15.

Case 193/83, Windsurfing International, Inc v Commission [1986] ECR 611, 92. Commission Regulation (EC) No 772/2004 of 27 April 2004 on the application of Article 81(3) of the Treaty to categories of technology transfer agreements. 136 European Commission, ‘Guidelines on the application of Article 81 of the EC Treaty to technology transfer agreements’ [2004] OJ C 101/2. 137 Ibid, para 209. 138 S Lawrence, ‘The Competition Law Treatment of No-Challenge Clauses in Licence Agreements, An Unfortunate Revolution?’ (2014) 9 Journal of Intellectual Property Law & Practice 802, 809. 139 European Commission, ‘Guidelines on the application of Article 101 of the Treaty on the Functioning of the European Union to technology transfer agreements’ [2014] OJ C 89/3, para 235. 140 Ibid, para 238. 134 135

Pay for delay in perspective  361 3 Intellectual property and antitrust In the EU, the interplay between intellectual property and antitrust is in flux just as it is in the US. However, the starting point of this development in Europe is somewhat different. Whereas in the US the development centres on the level of antitrust immunity that should be afforded to the patent holder, in the EU the focus is on the exceptions to the general application of competition law to intellectual property rights. So, akin to the already discussed differences in relation to the perception of settlements, it can also be argued that this development has also started at opposite ends of the spectrum in the US and Europe. Historically, the EU Courts have made a distinction between the ‘existence’ of an intellectual property right, which cannot be subject to competition law, and the exercise of the intellectual property right, which can generally be subject to competition law scrutiny.141 The ‘exercise’ of the intellectual property right is further linked to the ‘subject matter’ and the ‘essential function’ of the intellectual property right in question. For instance, the subject matter for patents has been described as ‘the exclusive right to use an invention with a view to manufacturing industrial products and putting them into circulation for the first time … as well as the right to oppose infringements’.142 Licensing agreements are, for example, generally regarded as an exercise of an intellectual property right and can therefore be subject to competition law scrutiny. Whereas open exclusive licences are not regarded as contrary to Article 101(1) of the TFEU, closed licences that amount to absolute territorial restrictions are deemed to infringe Article 101(1) and are unlikely to be justified under Article 101(3).143 Also, non-territorial licences can be subject to competition law as shown by Windsurfing.144 Here, the Court found that putting an obligation on the licensee to sell a patented product only in connection with a licensor-approved unpatented product infringed competition law; as such an obligation was not indispensable to the exploitation of the patent.145 The unilateral exercise of intellectual property rights can also lead to an abuse of a dominant position in ‘exceptional circumstances’. In Microsoft, Microsoft was found to have abused its dominant position by refusing to license its intellectual property rights to competitors. Microsoft tried to justify its refusal to license by relying on the exclusionary nature of its intellectual property right and the need to protect its incentive to innovate. This argument was rejected as it is not self-evident that an intellectual property right constitutes an objective justification for the refusal to license. Otherwise, intellectual property rights could never be subject

141 Joined Cases C-56/64 and 58/64, Consten and Grundig v Commission [1966] ECR 299. The distinction is historically based on the wording of the EU Treaty, which affords member states the possibility to justify quantitative restrictions to free trade between member states by means of intellectual property rights protection (Article 36 of the TFEU) and Article 345 of the TFEU, which provide that member states’ systems of property law should be protected. Also, EU competition law is generally within the remit of the European Commission, whereas the creation of system for intellectual property rights protection is a competence of the individual member states. 142 Case C-15/74, Centrafarm BV and Adriaan de Peijper v Sterling Drug, Inc ECLI:​EU:​C:​1974:​114. 143 Case C- 258/78, LC Nungesser KG and Kurt Eisele v Commission [1982] ECR 2015. 144 Case 193/83, Windsurfing International, Inc v Commission (n 134). 145 Ibid, para 57.

362  Research handbook on methods and models of competition law to competition law scrutiny.146 However, the Court did consider that the argument could justify such conduct if it was substantiated and not just of vague and theoretical nature.147 Although this brief discussion of the interplay between intellectual property and antitrust in the EU is far from exhaustive,148 it highlights an important difference to the situation in the US. The exercise of intellectual property rights is not generally regarded as immune from antitrust scrutiny, with few exceptions. Rather, the exercise of intellectual property rights is subject to antitrust scrutiny but can be justified if the conduct in question protects an essential function of the intellectual property right in question. 4 Comment The European perception of settlements, as well as the general European approach to the interplay of intellectual property and antitrust, can be traced throughout the General Court’s Lundbeck decision. It is true that the European Commission and the General Court emphasized the same indicators as the US Supreme Court and the FTC in order to establish the anticompetitive nature of pay-for-delay settlements. Yet, the premise of the analysis is fundamentally different. Whereas in the US these types of agreement were generally regarded as procompetitive that may become anticompetitive in cases of large payments, in Europe these agreements were regarded as an anticompetitive market-sharing agreement that the parties were not able to justify by means of intellectual property rights protection. Indeed, the European Commission established that the pay-for-delay settlements constitute a restriction by object, by reference to the ECJ’s judgment in BIDS, which concerned a market-sharing agreement in exchange for payments between competitors in the Irish beef industry.149 In Lundbeck, the General Court found that the European Commission was entitled to rely on the case law in BIDS as the conduct in question was comparable to the market exclusion agreement in the Irish beef case.150 Furthermore, it was held that the market-sharing agreement could be justified if it did not entail a large payment that played a decisive role in the market exclusion of the generic company.151 Lundbeck pleaded with the Court that the BIDS case should not be applicable as the market-sharing agreement did not take place in relation to intellectual property that concerned the protection of patents. This plea was rejected by the Court, which reiterated that the European Commission took account of the relevant patent, but came to the conclusion that the patent, even if presumed valid, was not able to prevent all competition of citalopram.152 Classifying the pay-for-delay settlement as a market-sharing agreement, one of the most serious restrictions of competition, makes it more feasible for the European Commission to Case T-201/04, Microsoft v Commission [2007] ECR II-3601, 690. Ibid, 698. 148 For a detailed discussion of the interplay between competition and intellectual property, see Ioannis Lianos, ‘Competition Law and Intellectual Property (IP) Rights, Analysis, Cases and Materials’ in Ioannis Lianos, Valentine Korah and Paolo Siciliani (eds), Competition Law: Analysis, Cases and Materials (4th edn, Hart Publishing 2017). 149 Case C-209/07, Competition Authority v Beef Industry Development Society Ltd and Barry Brothers (Carrigmore) Meats Ltd (2009) 4 CMLR 6. The case concerned agreements between the members of the Beef Industry Development Society in Ireland, which consisted of the ten largest meat producers in the country. The agreement envisaged that some of the competitors would leave the market. In return for their market exit, those who stayed in the market would compensate the leaving competitors. 150 Case T-472/13, Lundbeck v Commission (n 5), para 435. 151 Ibid, para 431. 152 Ibid, para 435. 146 147

Pay for delay in perspective  363 satisfy the narrow requirements for a restriction by object that were recently proclaimed in the ECJ’s judgment in Cartes Bancaires. In that case, it was held that restrictions by object should be defined narrowly and should only be found if experience shows that the agreement is likely to cause anticompetitive harm.153

V CONCLUSION A number of conclusions can be drawn from a comparison of the pay-for-delay enforcement efforts in the US and Europe. It is very likely that the European Commission was influenced by the US experience and the enforcement efforts of the FTC regarding the underlying rationale of pay-for-delay settlements. However, the development and establishment of a theory of harm is specific to each jurisdiction due to the case law that the respective antitrust authorities need to consider. This case law does not exist in a vacuum. It reflects the legal traditions of its jurisdiction and needs to be viewed in light of the jurisdiction’s legal system. In the US, the FTC was faced with an uphill battle to persuade the courts that a patent settlement with a reverse payment should be considered anticompetitive. As a matter of public policy, settlements are generally favoured by litigating parties and the judiciary in the US, particularly in the patent context. This is a direct consequence of the adversarial legal system. The fact that the antitrust authorities share the same courts and procedural rules as private litigants also has a direct impact on the courts’ perception of the intellectual property and antitrust interface. Legal standards that were developed to curb zealous private antitrust enforcement are also applied by the courts in the sphere of public enforcement, which has led to constricted liability norms. With every pay-for-delay decision against the agency, this hill became even steeper. In contrast to the FTC, the European Commission has had an easier task to develop a theory of harm for pay-for-delay settlements. It did not have to overcome and argue against case law that heavily favoured settlements, due to its inquisitorial system. The EU case law is also more interventionist in the intellectual property and antitrust interplay. This is due to the fact that the EU Courts do not have to equilibrate the legal standard in light of private damages actions that could lead to over-deterrence. From the outset, the European Commission almost had a ‘clean slate’ for the development of a theory of harm for pay-for-delay settlements. However, one should not leap to the conclusion that EU antitrust enforcement is more efficient than its US counterpart simply because it took the European Commission one upheld decision to reach a very similar enforcement level compared to the numerous efforts that the FTC had to undertake in order to reach the US Supreme Court. The US system might be less cost efficient, but it leads to increased judicial scrutiny, which in return forces the FTC to reflect on and adapt its theory of harm. The European Commission does not have to constantly defend its theories of harm in front of the courts. The scrutiny is initially rather internalized in discussions between the case team, the chief economist team and the legal service. Although anecdotally the development of a theory of harm is significantly shaped by this internal scrutiny, it might nonetheless be prone to the above-mentioned prosecutorial and hindsight bias.

153 Case C‑67/13, P Groupement des cartes bancaires (CB) v European Commission [2014] EU:​C:​ 2014:​2204.

364  Research handbook on methods and models of competition law This leaves us with two systems, both with their own advantages and drawbacks. From a procedural efficiency point of view, the European system should be favoured. However, the increased judicial scrutiny of antitrust proceedings in the US is likely to contribute to the robustness of the development of novel theories of harm. From a European perspective, informal cooperation between the antitrust authorities should therefore be highly beneficial – especially with regard to anticompetitive conduct that has first been detected and investigated in the US, such as pay-for-delay settlements. Ultimately, this should lead to a reduction of prosecutorial and hindsight biases in these cases.

16. The Australian approach to third party infrastructure access under Part IIIA of the Competition and Consumer Act 2010 Alice Muhlebach

I INTRODUCTION When a firm controls infrastructure that its competitor must use in order to participate in a related market, conduct that restricts or denies access to the infrastructure may negatively affect competition in that market. Competition law typically analyses such conduct as a form of abuse of dominance or monopolization, and may require that a competitor be allowed to use (or ‘access’) the infrastructure, departing from the general proposition that competition law does not typically interfere with a firm’s rights by requiring it to deal with a particular person.1 Targeted legislation or regulation may separately impose obligations to provide access. Both approaches can involve controversy as they compel a firm to deal with its competitor and so intrude on the firm’s freedom of contract and private property rights. Australian law incorporates these approaches, as well as a third approach: a legislative regime which provides for ‘declaration’ of an infrastructure service if certain statutory criteria are satisfied (the Regime). The criteria can be satisfied regardless of whether the firm controlling the infrastructure (the provider) has acted anticompetitively in contravention of Australian competition law. The effect of declaration is to create a legal right for any person to negotiate to use the infrastructure service, with recourse to binding arbitration by the Australian Competition and Consumer Commission (ACCC). In the 22 years since its introduction, the Regime has been invoked 28 times; only seven applications have (after the exhaustion of any appeals) resulted in declaration. The Regime has often been invoked in relation to privately owned or operated infrastructure that plays a central role in the Australian economy, such as at major domestic and international airports, or that is used in major Australian industries, such as iron ore and coal mining. In many instances, the regulatory and court processes have been protracted and costly and have not resulted in declaration. Nonetheless, a recent review of Australian competition law (referred to as the ‘Harper Review’2) recommended that the Regime be retained, albeit with some amendment. This chapter begins by briefly describing the treatment of ‘essential facilities’ under US, European and Australian competition law. It then describes the introduction of the Regime, the practical experience regarding its application, and key legal issues arising under it. 1 See, for example, United States v Colgate and Co, 250 US 300, 307 (1919); Case C-7/97, Bronner v Mediaprint [1998] ECR I-7791, [1999] 4 CMLR 112, per Opinion of AG Jacobs (Bronner Opinion); Queensland Wire Industries Pty Ltd v Broken Hill Proprietary Co Ltd (1987) 17 FCR 211, 221. 2 Commonwealth of Australia, Competition Policy Review: Final Report (March 2015) (Competition Policy Review). The report was prepared by Professor Ian Harper, Peter Anderson, Su McCluskey and Michael O’Bryan QC.

365

366  Research handbook on methods and models of competition law

II

COMPETITION AND ANTITRUST APPROACHES TO ‘ESSENTIAL FACILITIES’

A

The Underlying Competition Problem

When a firm controls critical infrastructure and competes in an activity that depends on its use, it may have a profit-maximizing incentive to restrict or deny access to its downstream competitors. If the infrastructure is truly critical to the ability of other firms to compete, that restriction or denial can extend the firm’s market power into, and so harm competition in, the dependent market. For example, a firm that controls a railway track and also competes to supply haulage services may have an incentive to refuse to allow competing haulage businesses to use the railway. If those competitors do not have any alternative to using the railway, the refusal may detrimentally affect competition in the haulage market. B

Approaches to the Competition Problem – United States and Europe

1 United States In the US, the ‘essential facilities’ doctrine was developed under the prohibitions on combinations in restraint of trade (section 1) and monopolization (section 2) under the Sherman Act of 1890.3 In Terminal Railroad Association,4 the owners of railway terminal infrastructure, who used the infrastructure for their own train operations, only allowed competitors to use the infrastructure on materially disadvantageous terms. The Supreme Court held that this constituted collective monopolization and ordered that the owners provide access to owners and non-owners on equivalent terms.5 Later cases developed the approach to essential facilities in other contexts, including the supply of news reports to newspapers,6 electricity transmission infrastructure,7 and a football stadium.8 In one of the most significant decisions on essential facilities, the Seventh Circuit identified that antitrust liability for refusal by one firm to deal with its rival could arise where: (1) (2) (3) (4)

the refusing firm is a monopolist which controls access to an essential facility; the facility cannot be reasonably duplicated by the competitor; the monopolist denies access to the competitor; and it is feasible to grant access.9

Sherman Act, 15 USC (1994). United States v Terminal Railroad Association of St Louis, 224 US 383 (1912). 5 Ibid, 411–12. 6 Associated Press v United States, 326 US 1 (1945). 7 Otter Tail Power Co v United States, 410 US 366 (1973). 8 Hecht v Pro-Football Inc, 570 F 2d 982 (DC Cir, 1977); cert denied 436 US 956 (1978). 9 MCI Communications Corporation v American Telephone and Telegraph Company, 708 F 2d 1081 (7th Cir, 1983); see Re Fortescue Metals Group Ltd (2010) 271 ALR 256, 338 (Pilbara Tribunal Decision). 3 4

The Australian approach to third party infrastructure access  367 Later cases have suggested a possible fifth element: that the monopolist does not have a valid business justification for refusing to deal with its rival (although this is arguably inherent in the four limbs stated above).10 The ‘essential facilities’ doctrine is controversial and has many critics.11 There is considerable doubt about its status, following the Supreme Court’s observations in 2004 that it ‘never recognized such a doctrine’.12 There are also targeted regulatory frameworks in the US, outside antitrust law.13 These laws can oblige those who own or control particular infrastructure, or infrastructure in a particular industry, to make the use of that infrastructure available to others. 2 Europe In Europe, cases involving access to ‘essential facilities’ have been dealt with as a form of refusal to deal that is capable of constituting abuse of a dominant position under Article 102 of the Treaty on the Functioning of the European Union.14 The European Court of Justice’s opinion in Oscar Bronner15 is one of the seminal opinions on the European treatment of essential facilities. Bronner published a newspaper and alleged that its competitor’s refusal to allow Bronner to use its national home delivery system was an abuse of a dominant position. The Court stated that ‘even if’ existing law on refusal to license intellectual property applied in that case,16 it would be necessary for Bronner to establish that: (1) its competitor’s refusal to provide access to its system: (a) would be likely to eliminate all competition by Bronner in the daily newspaper market; (b) was not capable of being objectively justified; and (2) the use of the delivery service was indispensable to Bronner’s business, in that there was no actual or potential substitute for that service.17 The Court considered that these elements were not established since other, albeit less advantageous, distribution methods were available to Bronner18 and there were no technical, legal or economic obstacles that would make it impossible ‘or even unreasonably difficult’ for another publisher to establish a competing delivery service.19 It observed that the ‘indispensability’ element would not be satisfied simply because a small newspaper company might find it

10 Spencer Weber Waller, ‘Areeda, Epithets and Essential Facilities’ (2008) 2 Wisconsin Law Review 359, 362–3. 11 See Phillip Areeda, ‘Essential Facilities: An Epithet in Need of Limiting Principles’ (1990) 58 Anti-Trust Law Journal 841, cited in Pilbara Tribunal Decision (n 9), 336. 12 Verizon Communications, Inc v Law Offices of Curtis V Trinko, LLP, 540 US 398, [25] (2004). 13 For example, such obligations existed in relation to the infrastructure at issue in Verizon Communications, Inc v Law Offices of Curtis V Trinko, LLP, 540 US 398, [25] (2004). 14 Consolidated Version of the Treaty on the European Union [2008] OJ C 115/13, Art 102 (formerly Art 82). 15 Bronner Opinion (n 1). 16 Ibid, [39], [40]. 17 Ibid, [41]. 18 Ibid, [43]. 19 Ibid, [44].

368  Research handbook on methods and models of competition law uneconomic to establish a competing delivery service, and that indispensability should be assessed from the perspective of a business with comparable scale to the incumbent.20 The European position has developed over time21 and it now seems that the European Commission will approach essential facilities cases similarly to other refusals to supply: by considering indispensability (or ‘objective necessity’)22 and elimination of downstream competition (or ‘effective competition’),23 as well as efficiency-based justifications for the impugned conduct.24 It may also consider whether a refusal is likely to lead to consumer harm.25 As in the US, this jurisprudence exists alongside targeted regulatory frameworks, which can oblige the owner or controller of particular infrastructure, or infrastructure in a particular industry, to make the use of that infrastructure available to others.26 3 Australia The Australian equivalent of the US and European prohibitions on unilateral conduct is the prohibition on ‘misuse of market power’ in section 46 of the Competition and Consumer Act 2010 (Cth) (CCA).27 The Full Court of the Federal Court has declined to adopt, under section 46, an ‘essential facilities’ rule based on the US and European cases.28 The High Court decided an appeal from that decision without reference to any essential facilities doctrine.29 In a later case, the High Court held that an electricity generation, distribution, transmission and retail business had contravened section 46 by refusing to allow a competing electricity generator to access its distribution and transmission infrastructure, again without reference to any essential facilities doctrine.30 As in the US and Europe, this Australian case law exists alongside dedicated legislative regimes for third party access: several well-established national regimes regulate telecommunications, natural gas pipeline and electricity infrastructure, and some rail, port, water and Ibid, [45]. Including in decisions by the European Commission, before and after the Bronner Opinion, which have addressed port, airport, railway, electricity transmission grid, pipeline and other infrastructure. See Port of Rodby, Commission Decision 94/119/EC [1994] OJ L 55/52; Sealink/B&I Line, Commission Decision 94/19/EC [1992] OJ L 15/8; and the cases cited in Jonathan Faull, Ali Nikpay and Deirdre Taylor (eds), Faull and Nikpay – The EU Law of Competition (7th edn, Oxford University Press 2014) 465. 22 Communication from the Commission – Guidance on the Commission’s enforcement priorities in applying Article 82 of the EC Treaty to abusive exclusionary conduct by dominant undertakings [2009] OJ C 45/02, [83]. 23 Ibid, [81]. 24 Ibid, [89], [90]. 25 Ibid, [81]. 26 For example, the existence of such regimes in the UK was noted prior to the introduction of the Australian Regime. See Independent Committee of Inquiry, Commonwealth of Australia, National Competition Policy: Report by the Independent Committee of Inquiry (1993) (Committee Report), 245–7. 27 Previously the Trade Practices Act 1974 (Cth) (TPA). 28 Queensland Wire Industries Pty Ltd v Broken Hill Proprietary Co Ltd (1987) 17 FCR 211, 221. 29 See Queensland Wire Industries Pty Ltd v Broken Hill Proprietary Co Ltd (1989) 167 CLR 177. 30 NT Power Generation Pty Ltd v Power and Water Authority (2004) 219 CLR 90. See also Melway Publishing Pty Ltd v Robert Hicks Pty Ltd (2001) 205 CLR 1. 20 21

The Australian approach to third party infrastructure access  369 other infrastructure is regulated under state or territory regimes. ‘Undertakings’ under Part IIIA of the CCA (discussed below) are also used to regulate certain railway and port terminal infrastructure. However, Australian law also provides a third means of imposing obligations to allow third parties to access infrastructure. Under the Regime established by Part IIIA of the CCA, the provider (the owner or operator) of any infrastructure that meets the relevant criteria may become subject to obligations to negotiate the terms of, and potentially provide, access to third parties. The Regime is discussed in detail below.

III

BACKGROUND TO THE REGIME

A

The Committee of Inquiry

In 1992, an independent committee (the Hilmer Committee) inquired into the development of a national competition policy31 and made wide-ranging recommendations, including on access to essential facilities. It identified that ‘essential facilities’ are facilities that exhibit natural monopoly characteristics (such that they cannot be duplicated economically) and occupy a strategic position in an industry.32 It considered that where a firm controls an essential facility and competes in an activity that relies on use of the facility, the firm may have an incentive to refuse or limit access to the facility by its competitors, to the detriment of competition.33 The Committee considered that competition policy should enable access to such facilities, while ensuring that the facility owner was not unduly disadvantaged.34 It concluded that ‘something more’ than section 46 was required35 and that legislative access rights should be created and should be available without requiring the access seeker to prove any anticompetitive purpose on the part of the provider.36 The Committee recommended that a national legislative regime be introduced that could potentially apply to publicly or privately owned facilities in any industry, but that would create access rights ‘sparingly’, focusing on strategically significant sectors.37 The Committee recommended that the regime should allow a person to request that a designated Minister ‘declare’ a right of access to a particular facility;38 the declaration, if made, would create a legal right to access the facility, with recourse to regulated arbitration of disputes.39 The

31 The Committee is often referred to as the ‘Hilmer Committee’ after its chair. For the Committee’s terms of reference, see Committee Report (n 26), 361–3. 32 Ibid, 240. 33 Ibid, 241. 34 Ibid, 241–2. 35 Ibid, 186. 36 Ibid, 248. 37 Ibid, 260. The Committee considered that alternative approaches, such as relying on state-based or industry-based regimes to create these rights, would be less desirable than the regime it proposed as they would not achieve national consistency across sectors. However, in practice, state- and industry-based regimes have been widely applied alongside the Regime. 38 Ibid, 267–9. 39 Ibid, 256.

370  Research handbook on methods and models of competition law Committee recommended that unless the facility owner agreed to the declaration, the Minister should only be able to declare a right of access to a facility if: (i)

the declaration had been recommended by ‘an independent and expert body’ based on its investigations and public inquiries;40 and (ii) the Minister was satisfied: (a) that access to the facility was ‘essential to permit effective competition in a downstream or upstream activity’ (and not merely convenient);41 (b) that declaration was in the public interest, having regard to the significance of the relevant industry to the national economy, and the expected impact of effective competition in that industry on national competitiveness.42 The Committee noted that ‘products, production processes or most other commercial facilities’ would not satisfy this criterion;43 and (c) that the facility owner’s legitimate interests would be protected by a fair and reasonable access fee and other terms, including recognition of the owner’s requirements to use the facility.44 The Committee recognized that the regime would interfere with the ‘important and fundamental principle’ that businesses should be free to choose whether and on what terms to deal with another party, which was ‘not to be disturbed lightly’.45 However, it considered that the regime could have ‘a pivotal role’ as competition was introduced to areas previously reserved to public monopolies.46 It emphasized that the regime should ‘carefully limit’ the circumstances in which access was required, and protect facility owners’ interests, in order to avoid undermining investment incentives.47 The Committee proposed that section 46 should not apply in circumstances where an access right had been created under this regime.48 B

Intergovernmental Agreement

In April 1995 the Council of Australian Governments (COAG) agreed that the Commonwealth would put forward legislation to establish a regime for ‘third party access to services provided by means of significant infrastructure facilities (other than facilities which are products, production processes or intellectual property)’, where certain criteria were satisfied in relation to the facility or service.49 Those criteria echoed but did not precisely reflect the Committee’s proposal. Ibid, 252–3. Ibid, 251. 42 Ibid. 43 Ibid. 44 Ibid, 252. 45 Ibid, 242. 46 Ibid. 47 Ibid, 248, 251. 48 Ibid, 260. 49 Clause 6(1) of the Competition Principles Agreement, between the Commonwealth of Australia, and the States of New South Wales, Victoria, Queensland, Western Australia, South Australia and Tasmania, and the Australian Capital Territory and the Northern Territory, 11 April 1995. 40 41

The Australian approach to third party infrastructure access  371 C

Establishment of the Regime

1 Enacting legislation In 1995, the Commonwealth Parliament enacted legislation (1995 Act) which introduced the Regime50 by inserting a ‘Part IIIA’ into the Trade Practices Act 1974 (Cth).51 The Regime echoed, but did not precisely implement, the Committee’s52 and COAG’s proposals. It did not (and still does not) describe the facilities or services it governs as ‘essential’.53 2 Purpose In 1995, the purpose of the Regime was described as follows: ‘access to certain facilities with natural monopoly characteristics, such as electricity grids or gas pipelines, is needed to encourage competition in related markets, such as electricity generation or gas production’.54 In 2006, the Regime was amended to identify that its objects are to: (a) promote the economically efficient operation of, use of and investment in the infrastructure by which services are provided, thereby promoting effective competition in upstream and downstream markets; and (b) provide a framework and guiding principles to encourage a consistent approach to access regulation in each industry.55

3 The declaration process Under the Regime, a party could apply for ‘declaration’ of a ‘service’ provided by a ‘facility’. Declaration created an enforceable right for any person (not solely the declaration applicant) to negotiate access to the service. The declaration application was to be made to the National Competition Council (the Council), which would make a non-binding recommendation to the ‘designated Minister’.56 The Minister would decide whether to declare the service and, if so, for what period.57 The Council and the Minister could not recommend or make a declaration unless satisfied of certain statutory criteria (discussed below).58 The Minister could declare a service regardless of whether the Council had recommended declaration, but could not specify pricing principles governing access. The Regime allowed a declaration to be revoked if the relevant criteria ceased to be satisfied, and for the Minister’s decisions on declaration and revocation to be reviewed by the Australian Competition Tribunal (Tribunal).59

Competition Policy Reform Act 1995 (Cth) (1995 Act). Ibid, s 59. 52 Sydney International Airport [2000] ACompT 1, [10]. 53 Rail Access Corporation v New South Wales Minerals Council Ltd (1998) 87 FCR 517, 519. 54 Commonwealth of Australia, Parliamentary Debates, Senate, 29 March 1995, Competition Policy Reform Bill 1995 (No 88), 2438 (the Hon Rosemary Anne Crowley). 55 TPA, s 44AA, introduced by the Trade Practices Amendment (National Access Regime) Act 2006 (Cth), Sch 1, Pt 1, item 4. 56 Ibid, s 44F, as inserted by the 1995 Act, s 59. 57 Ibid, s 44H(8), as inserted by the 1995 Act, s 59. 58 Ibid, ss 44G(2), 44H(4), as inserted by the 1995 Act, s 59; now contained in s 44CA. 59 Ibid, ss 44K, 44L, as inserted by the 1995 Act, s 59. 50 51

372  Research handbook on methods and models of competition law Under the Regime, disputes arising during access negotiations could be referred to arbitration by the ACCC.60 The ACCC could determine access terms, require that access be provided, and require extension of a facility.61 However, it could not interfere in certain ways with pre-existing use of, or rights to use, the declared service. It was also required to consider certain mandatory matters (such as the provider’s legitimate business interests and investment in the facility).62 Importantly, the ACCC could decide that the provider was not required to provide access: declaration was not a guarantee that the access seeker would obtain use of the service.63 Arbitration determinations could be reviewed by the Tribunal and enforced by the Federal Court.64 Contracts governing access to declared services could be registered and enforced as if they were an arbitration determination.65 The Regime has been amended several times, but the essential features described above are unchanged.66 4 The availability of declaration (a) The definition of ‘services’ The Regime provided for declaration of ‘services’. A ‘service’ was defined as ‘a service provided by means of a facility’, including ‘the use of an infrastructure facility such as a road or railway line’, ‘handling or transporting things such as goods or people’ and ‘a communications service or similar service’. The definition excluded ‘the supply of goods’, and the use of intellectual property or a production process, except to the extent that those exclusions were ‘an integral but subsidiary part of the service’.67 This definition echoes the Committee’s Report and has not been amended. A ‘provider’ was defined in section 44B to mean the ‘owner or operator of the facility that is used (or is to be used) to provide the service’. This definition has not been amended. The terms ‘access’ and ‘facility’ have not been defined. (b) Services that could not be declared A service could not be declared if access to it was governed by an undertaking accepted by the ACCC, or a regime that had been certified as ‘effective’.68 This position is unchanged. Following subsequent amendments, a service also cannot be declared if it is the subject of an 60 Ibid, ss 44S, as inserted by the 1995 Act, s 59. For a more detailed discussion of the arbitration arrangements, see Geoff Peterson, Morelle Bull and Catherine Dermody, Access Regulation in Australia (Law Book Co 2016). 61 TPA, s 44V, as inserted by the 1995 Act, s 59. 62 Ibid, s 44X, as inserted by the 1995 Act, s 59. 63 Ibid, s 44V(3), as inserted by the 1995 Act, s 59. 64 Ibid, ss 44ZP, 44ZZD, as inserted by the 1995 Act, s 59. 65 Ibid, ss 44ZW, 44ZY, as inserted by the 1995 Act, s 59. 66 The key (and recent) change has been to clarify that the ACCC can, on arbitration, order the provider to expand their facility, as well as to extend it geographically: s 44V(2A), introduced by the Competition and Consumer Amendment (Competition Policy Review) Bill 2017 (Cth) (2017 Bill), Sch 12, item 38; see also Explanatory Memorandum, Competition and Consumer Amendment (Competition Policy Review) Bill 2017 (Cth) (2017 EM), 109. 67 TPA, s 44B, as inserted by the 1995 Act, s 59. 68 Ibid, ss 44G(1), 44G(2)(e), 44H(3), 44H(4)(e), as inserted by the 1995 Act, s 59; the equivalent is now in s 44F(1) of the CCA. See Peterson et al (n 60), 150, 149; Verity Quinn and Rosannah Healy, ‘Part

The Australian approach to third party infrastructure access  373 ‘ineligibility decision’ (effectively, a decision that the declaration criteria are not satisfied),69 or is provided by certain government-owned facilities built following a competitive tender process and subject to separate access obligations.70 (c) The declaration criteria The Regime initially provided that the Council or the Minister could only recommend or make a declaration if they were satisfied of all of the following ‘declaration criteria’ (if not, they had no legal power to recommend or make a declaration): (a) that access (or increased access) to the service would promote competition in at least one market (whether or not in Australia), other than the market for the service; (b) that it would be uneconomical for anyone to develop another facility to provide the service; (c) that the facility is of national significance, having regard to: (i) the size of the facility; or (ii) the importance of the facility to constitutional trade or commerce; or (iii) the importance of the facility to the national economy; (d) that access to the service can be provided without undue risk to human health or safety; (e) that access to the service is not already the subject of an effective access regime; (f) that access (or increased access) to the service would not be contrary to the public interest.71

When applying these criteria, the Council and the Minister were required to consider whether it would be economical for anyone to develop another facility that could provide part of the service72 (and, following subsequent amendments, to have regard to the objects of the Regime).73 The Council could recommend against a declaration if it considered that certain applications were not made in good faith.74

IV

THE EXPERIENCE IN PRACTICE

A

Declaration Applications

Since the Regime was introduced, 28 declaration applications have been made (concerning both publicly and privately owned infrastructure),75 of which 14 concerned railway tracks

IIIA – The Current State of Play’ (2013) 21(4) Australian Journal of Competition and Consumer Law 244 261–6 for detailed discussion of the experience with these avenues. 69 CCA, ss 44G(7), 44H(6C), inserted by the Trade Practices Amendment (Infrastructure Access) Act 2010 (Cth), Sch 2, items 5, 6; the equivalent is now in s 44F(1) of the CCA. 70 TPA, ss 44G(1A), 44H(3A), introduced by the Trade Practices Amendment (National Access Regime) Act 2006 (Cth), Sch 1, Pt 1, items 15, 22; the equivalent is now in s 44F(1) of the CCA. 71 Ibid, ss 44G(2), 44H(4), inserted by the 1995 Act, s 59; now contained in s 44CA of the CCA. 72 Ibid, ss 44F(4), 44H(2), inserted by the 1995 Act, s 59; the relevant provisions were recently repealed. 73 Ibid, ss 44F(2)(b), 44H(1A), inserted by the Trade Practices Amendment (National Access Regime) Act 2006 (Cth), Sch 1, Pt 1, items 12, 20. 74 Ibid, s 44F(3), inserted by the 1995 Act, s 59. 75 See ‘Past applications’ (National Competition Council 2017), http://​ncc​.gov​.au/​applications​-past/​ past​_applications.

374  Research handbook on methods and models of competition law (for transporting iron ore, coal and freight)76 and seven concerned airport infrastructure.77 76 Application No D0007: Application by Carpentaria Transport Pty Ltd for declaration of specified rail freight services provided by Queensland Rail (24 December 1996) (Carpentaria Application); Application No D0008: Specialized Container Transport, Application for declaration under section 44F of the Trade Practices Act for NSW portion of the Sydney–Perth railway (4 February 1997), http://​ncc​ .gov​.au/​images/​uploads/​DERaSCTAp​-001​.pdf (SCT Feb 1997 Railway Application); Application No D0009: New South Wales Minerals Council, Application to National Competition Council by New South Wales Minerals Council Ltd for a declaration recommendation in respect of the Hunter railway line service (2 April 1997), http://​ncc​.gov​.au/​images/​uploads/​DERaMCAp​-001​.pdf (Minerals Council Railway Application); Application No D0010: Specialized Container Transport, Application for declaration under section 44F of the Trade Practices Act for Westrail freight support services (31 July 1997), http://​ncc​.gov​.au/​images/​uploads/​DERaSwaAp​-001​.pdf (SCT July 1997 Railway Application); Application No D0011: Robe River Iron Associates, Declaration application under section 44F(1) of the Trade Practices Act 1974 by Robe River Iron Associates relating to the rail track service provided by Hamersley Iron Pty Ltd’s rail infrastructure facility (September 1998), http://​ncc​.gov​.au/​images/​ uploads/​DERaHaAp​-001​.pdf (Robe Hamersley Railway Application); Application No D0013: Freight Victoria Ltd (trading as Freight Australia), Application for declaration under section 4(1) Trade Practices Act 1974 for Victorian Intrastate Network (1 May 2001), http://​ncc​.gov​.au/​images/​uploads/​DERaFrAp​ -001​.pdf (Freight Australia Railway Application); Application No D0014: Portman Iron Ore Ltd, Rail Access Declaration Application (9 August 2001), http://​ncc​.gov​.au/​images/​uploads/​DERaKEAp​-001​ .pdf (Portman Railway Application); Application No D0015: AuIron Energy Ltd, Application under Part IIIA of the Trade Practices Act 1974 for declaration of the service provided by the Wirrida-Tarcoola Railway Line (12 September 2001), http://​ncc​.gov​.au/​images/​uploads/​DERaAuAp​-001​.pdf (AuIron Railway Application); Application No D0018: Fortescue Metals Group Ltd, Application under Part IIIA of the Trade Practices Act 1974 for declaration of the service provided by BHP Billiton Iron Ore Pty Ltd (11 June 2004), http://​ncc​.gov​.au/​images/​uploads/​DeRaFoAp​-001​.pdf (FMG Newman Railway Application); Application No D0020: Tasmanian Department of Infrastructure, Energy and Resources – Rail Unit, concerning declaration of a service provided by the Tasmanian Railway Network (May 2007) (Tasmanian Railway Application); Application No D0021: The Pilbara Infrastructure Pty Ltd, Application under Part IIIA of the Trade Practices Act 1974 for declaration of the service provided by the Mount Goldsworthy joint venture participants (16 November 2007), http://​ncc​.gov​.au/​images/​uploads/​ DeRaGwAp​-001​.pdf (FMG Goldsworthy Railway Application); Application No D0022: The Pilbara Infrastructure Pty Ltd, Application under Part IIIA of the Trade Practices Act 1974 for declaration of the service provided by Hamersley Iron Pty Ltd’s Pilbara rail network (16 November 2007), http://​ ncc​.gov​.au/​images/​uploads/​DERaHiAp​-001​.pdf (FMG Hamersley Railway Application); Application No D0023: The Pilbara Infrastructure Pty Ltd, Application under Part IIIA of the Trade Practices Act 1974 for declaration of the service provided by Robe River Joint Ventures’ Pilbara railway (18 January 2008), http://​ncc​.gov​.au/​images/​uploads/​DeRaRrAp​-005​.pdf (FMG Robe River Railway Application); Application No D0025: Pacific National Pty Ltd, Application under Part IIIA of the Trade Practices Act 1974 for a declaration recommendation for the services provided by Queensland Rail’s Queensland Coal Rail Network (18 May 2010), http://​ncc​.gov​.au/​images/​uploads/​DERaQRAp​-002​.pdf (PN Railway Application). 77 Four applications were the subject of a single compendious application made by Australian Cargo Terminal Operators Pty Ltd for the declaration of particular services at Sydney International Airport and Melbourne International Airport (November 1996), http://​ ncc​ .gov​ .au/​ images/​ uploads/​ DEAiAtAp​-001​.pdf. The NCC considered and recorded this compendious application as four separate applications: Application No D0003 (regarding certain freight loading and unloading services at Sydney International Airport), Application No D0004 (regarding certain freight loading and unloading services at Melbourne International Airport); Application No D0005 (regarding the Ansett airline’s ramp and cargo terminal services at Melbourne International Airport and Sydney International Airport); and Application No D0006 (regarding the Qantas airline’s ramp and cargo terminal services at Melbourne International Airport and Sydney International Airport). In this chapter, these four applications are referred to as the ‘Four ACTO Applications’. The other three applications dealing with airport infra-

The Australian approach to third party infrastructure access  375 The others concerned a ‘payroll deduction service’,78 a high-pressure gas distribution system, electricity transmission and distribution infrastructure, sewage and sewerage infrastructure, water storage and transport infrastructure, a tram network for transporting sugar cane, and port channel infrastructure.79

structure were Application No D0016: Virgin Blue Airlines Pty Ltd, Applications under Part IIIA of the Trade Practices Act 1974 Requesting recommendation that services provided by Sydney Airports Corporation Ltd be declared (1 October 2001), http://​ncc​.gov​.au/​images/​uploads/​DeAiVi​-001​.pdf (Virgin Blue Application); Application No D0026: The Board of Airline Representatives of Australia Inc, Application for declaration, Part IIIA of the Competition and Consumer Act 2010 Jet Fuel Supply Infrastructure to Sydney Airport: Service No 2: provided by the Caltex Pipeline Facility (26 September 2011), http://​ncc​.gov​.au/​images/​uploads/​DEJFBAAp​-004​.pdf; Application, Part IIIA of the Competition and Consumer Act 2010 Jet Fuel Supply Infrastructure to Sydney Airport: Service No 1: provided by the Sydney JUHI Facility (26 September 2011), http://​ncc​.gov​.au/​images/​uploads/​DEJHBAAp​-004​ .pdf (although two separate documents were submitted, the NCC’s register records a single application) (BARA Application); Application No D0027: Tiger Airways Australia Pty Ltd, Application under Part IIIA of the Competition and Consumer Act 2010 (Cth) requesting a recommendation that the Domestic Terminal Service provided by Sydney Airport Corporation Ltd be declared (3 July 2014), http://​ncc​.gov​ .au/​images/​uploads/​DEAiTAAp​-001​.pdf (Tiger Application). 78 The Australian Union of Students (AUS) sought declaration of the ‘Austudy Payroll Deduction Service’, by which it sought to require the Commonwealth government to implement a system whereby students who received government assistance would be required to identify, as a precondition to receiving their government assistance, whether they wanted to join a student union. Relatedly, it sought to require that the Commonwealth government provide information on the AUS to students, and information to the AUS on students who wished to join the AUS; it also sought to require the Commonwealth to advance students sufficient government assistance payments to allow them to pay their union fees. The service was not declared. See Application No D0001: Application concerning the ‘Austudy Payroll Deduction Service’ (24 April 1996), http://​ncc​.gov​.au/​images/​uploads/​deausap​-001​.pdf (Austudy Application); Re Australian Union of Students [1997] ACompT 1 (28 July 1997). 79 See Austudy Application; Application No D0002: Futuris Corporation, Application for recommendation that a service be declared, Perth high-pressure gas distribution system (30 August 1996), http://​ncc​.gov​.au/​images/​uploads/​DEGaWAAp​-001​.pdf (Futuris Application); Application No D0012: Normandy Power Pty Ltd, NP Kalgoorlie Pty Ltd and Normandy Golden Grove Operations Pty Ltd, Application for declaration of the Western Power Corporation electrical transmission and distribution systems in Western Australia (2 January 2001), http://​ncc​.gov​.au/​images/​uploads/​DEElWaAp​-001​ .pdf (Normandy Power Application); Application No D0017: Services Sydney Pty Ltd, Applications under Part IIIA of the Trade Practices Act 1974 requesting recommendation that Sewage Transmission and Interconnection Services provided by Sydney Water Corporation Ltd be declared (1 March 2004), http://​ncc​.gov​.au/​images/​uploads/​DEWaSSAp​-001​.pdf (Services Sydney Application); Application No D0019: Lakes R Us Pty Ltd, Application by Lakes R Us P/L for declaration of access to certain facilities of Snowy Hydro Ltd (29 September 2004), http://​ncc​.gov​.au/​images/​uploads/​DEWALRAp​-001​.pdf (Lakes R Us Application); Application No D0024: North Queensland Bio-Energy Corporation Ltd, Application under Part IIIA of the Trade Practices Act 1974 for declaration of the service provided by CSR’s Herbert River Narrow Gauge Tram Network (13 March 2010), http://​ncc​.gov​.au/​images/​uploads/​ DERaHRAp​-001​.pdf (Queensland Bio-Energy Corporation Tram Network Application); Application No D0028, Application for a declaration recommendation in relation to the Port of Newcastle (May 2015), http://​ncc​.gov​.au/​images/​uploads/​DEPONAp​-001​.pdf (Glencore Port Application).

376  Research handbook on methods and models of competition law Nine applications were made by or on behalf of miners,80 four by a single airport freight handler,81 four by rail haulage operators,82 three by or on behalf of airlines,83 two by railway operators,84 and two by businesses in the water industry.85 A student union, a brick manufacturer, an electricity generator and a sugar cane processor each made one application.86 Declaration has been sought by parties that had been denied or otherwise did not have use of the relevant service,87 and also by parties that already had use of the relevant service (in which cases, the chief effect of declaration would have been to enable recourse to ACCC arbitration). For example, Virgin Blue Airlines Pty Ltd (Virgin Blue) was using the airport service it succeeded in having declared in 2002,88 and ships carrying coal mined by Glencore Coal Pty Ltd (Glencore) already had use of the channel service Glencore sought to have declared at the Port of Newcastle in 2015.89 As this summary indicates, declaration applications have often been made in relation to industries and infrastructure that play a central role in the Australian economy. The applications concerning airports related to the Sydney and Melbourne Airports, which are critical airports for the transport of passengers and freight to, from and within Australia. The applications concerning dedicated iron ore railways concerned four of the five major railways used to transport iron ore (one of Australia’s largest export commodities) from mine to port in Australia, each of which was privately developed, and controlled by a single user. Some of those railways have been subject to multiple declaration applications. The application regarding the Port of Newcastle concerned critical channel infrastructure at the world’s largest coal export port90 (coal also being a major Australian export), which is a long-standing multi-user facility. However, as outlined below, very few declaration applications have actually resulted in a declaration being made. B

Outcome of Applications

Only seven declaration applications have resulted in a final outcome of declaration, following exhaustion of all reviews and appeals. Three concerned airport infrastructure at major air-

80 Minerals Council Railway Application; Robe Hamersley Railway Application; Portman Railway Application; AuIron Railway Application; FMG Newman Railway Application; FMG Goldsworthy Railway Application; FMG Hamersley Railway Application; FMG Robe River Railway Application; Glencore Port Application. 81 Australian Cargo Terminal Operators Pty Ltd, which made the Four ACTO Applications. 82 Carpentaria Application; SCT Feb 1997 Railway Application; SCT July 1997 Railway Application; PN Railway Application. 83 Virgin Blue Application; BARA Application; Tiger Application. 84 Freight Australia Railway Application; Tasmanian Railway Application. 85 Services Sydney Application; Lakes R Us Application. 86 Austudy Application; Futuris Application; Normandy Power Application; Queensland Bio-Energy Corporation Tram Network Application. 87 For example (non-exhaustively), the Austudy Application and the four applications by Fortescue Metals Group Ltd or its subsidiary The Pilbara Infrastructure Pty Ltd. 88 Re Virgin Blue Airlines Pty Ltd (2005) 195 FLR 242. 89 Application by Glencore Coal Pty Ltd [2016] ACompT 6, 57. 90 See Port of Newcastle, ‘About the Port’, http://​ www​ .portofnewcastle​ .com​ .au/​ Company​ -Information/​About​-the​-Port​.aspx.

The Australian approach to third party infrastructure access  377 ports;91 two concerned railways (Tasmania’s general freight railway and the Goldsworthy iron ore railway);92 one concerned facilities at a port used to export coal;93 and one (subsequently revoked) concerned a sewage transmission and sewer connection service.94 Of the 21 ultimately unsuccessful applications, eight were withdrawn before any decision was made regarding declaration.95 Of the others, nine concerned railways96 and the others concerned a ‘payroll deduction service’, water storage and transport infrastructure, a sugar cane tram network and airport fuel supply infrastructure.97 The Regime, therefore, has, in the majority of cases, been invoked in circumstances where ultimately no final declaration was made. This is particularly significant, given the protracted and costly nature of the declaration process, as described below. C

The Nature of the Declaration Process

The declaration process is inherently complex: an access seeker must seek a recommendation from the Council, the Minister then decides whether to make a declaration; if the declaration is made, the access seeker may then negotiate with the provider and may refer any disputes

91 See two of the Four ACTO Applications, being Application No D0003 (regarding certain freight loading and unloading services at Sydney International Airport) and Application No D0004 (regarding certain freight loading and unloading services at Melbourne International Airport), declarations dated 30 June 1997, http://​ncc​.gov​.au/​images/​uploads/​DEAiAtMD​-001​.pdf. See also the service sought to be declared in the Virgin Blue Application: Sydney Airport Corporation Ltd v Australian Competition Tribunal (2006) 155 FCR 124 (Sydney Airport FC Decision). 92 Tasmanian Railway Application and FMG Goldsworthy Railway Application. 93 Glencore Port Application. In that case, the Minister did not make the declaration sought, but the Tribunal declared the service on review of that decision (Application by Glencore Coal Pty Ltd (No 2) (2016) 309 FLR 358 (Glencore Tribunal Decision) and Application by Glencore Coal Pty Ltd [2016] ACompT 6), the Full Court of the Federal Court upheld the Tribunal’s decision (Port of Newcastle Operations Pty Ltd v Australian Competition Tribunal [2017] FCAFC 124), and the High Court refused special leave to appeal from that decision (Port of Newcastle Operations Pty Ltd v The Australian Competition Tribunal [2018] HCATrans 55 (23 March 2018)). 94 Services Sydney Application. See also Re Services Sydney Pty Ltd (2005) 227 ALR 140 and revocation decision dated 1 October 2009, http://​ncc​.gov​.au/​images/​uploads/​DeWaSSRev​-002​.pdf. 95 Futuris Application, two of the Four ACTO Applications (Application number D0005 (regarding the Ansett airline’s ramp and cargo terminal services at Melbourne International Airport and Sydney International Airport) and Application number D0006 (regarding the Qantas airline’s ramp and cargo terminal services at Melbourne International Airport and Sydney International Airport)); Robe Hamersley Railway Application; Normandy Power Application; Portman Railway Application; PN Railway Application; Tiger Application. In one of those eight cases, the Robe Hamersley Railway Application, the withdrawal followed a decision by the Federal Court that the service sought to be declared involved the ‘use of a production process’ and so was not a ‘service’ susceptible to declaration: Hamersley Iron Pty Ltd v National Competition Council (1999) 164 ALR 203 (Hamersley Decision). 96 These were the services sought to be declared in each of the following applications: Carpentaria Application; SCT Feb 1997 Railway Application; Minerals Council Railway Application; SCT July 1997 Railway Application; Freight Australia Railway Application; AuIron Railway Application; FMG Newman Railway Application; FMG Hamersley Railway Application; FMG Robe River Railway Application. Details and source documents regarding the decisions not to declare these services are available on the website of the National Competition Council, http://​www​.ncc​.gov​.au. 97 Austudy Application; Lakes R Us Application; Queensland Bio-Energy Corporation Tram Network Application; BARA Application.

378  Research handbook on methods and models of competition law to arbitration. The process is generally further extended because parties frequently exercise related review or appeal rights. Of the 28 declaration applications to date, only 12 (including eight applications that were withdrawn) have not been the subject of a review application to the Tribunal.98 Subsequent legal proceedings were commenced in each of the other 16 cases, although in six of those cases applications to the Tribunal were withdrawn before any decision was made.99 Ten cases have proceeded to decision by the Tribunal,100 and in six of them, the Tribunal’s decision was subject to judicial review by the Federal Court (with further appeals from the Federal Court in some instances).101 Only two disputes have been referred to ACCC arbitration.102 While the use of review and appeal avenues is appropriate, it does increase the time required to reach a final outcome. To take two examples: the four most recent applications for declaration of iron ore railways took between approximately two-and-a-half and six years to reach a final outcome; the most recent application, concerning the Port of Newcastle, did not reach a final outcome until March 2018, being less than two months short of three years after the application was made, but even that outcome is now the subject of challenge, as the National Competition Council (NCC) is considering a submission supporting the revocation of that declaration.103 Further, participating in the declaration process is not costless. Rather, interested parties typically provide detailed information and submissions to the Council, which is subject to further consideration by the Minister, and the Tribunal (if the Minister’s decision is reviewed). Unsurprisingly, given the consequences of declaration, the merits of a declaration application, and the relevant legal issues, are often strongly contested by relevant parties. BHP Billiton, the operator of two of the four iron ore railways that have more recently been the 98 The eight withdrawn applications are in n 95. The others that have not been the subject of subsequent legal proceedings are one of the Four ACTO Applications (Application number D0004, regarding certain freight loading and unloading services at Melbourne International Airport); Tasmanian Railway Application; Queensland Bio-Energy Corporation Tram Network Application; BARA Application. 99 The six cases were Carpentaria Application; SCT Feb 1997 Railway Application; Minerals Council Railway Application (although this application was the subject of a Federal Court hearing on a jurisdictional issue: see Rail Access Corporation v NSW Minerals Council Ltd (1998) 87 FCR 517); SCT July 1997 Railway Application; Freight Australia Railway Application; Lakes R Us Application. 100 They were the Austudy Application; one of the Four ACTO Applications (Application number D0003 (regarding certain freight loading and unloading services at Sydney International Airport)); AuIron Railway Application; Virgin Blue Application; Services Sydney Application; FMG Newman Railway Application; FMG Goldsworthy Railway Application; FMG Hamersley Railway Application; FMG Robe River Railway Application; Glencore Port Application. 101 They were Virgin Blue Application; FMG Newman Railway Application; FMG Goldsworthy Railway Application; FMG Hamersley Railway Application; FMG Robe River Railway Application; Glencore Port Application. 102 See ACCC, ‘Access Dispute between Services Sydney Pty Ltd and Sydney Water Corporation, Arbitration Report’ (19 July 2007), https://​www​.accc​.gov​.au/​system/​files/​Fn​%2030​%20​-​%20ACCC​ %20Sydney​%20Water​%20Arbitration​.pdf; ACCC, ‘ACCC Notified of an Access Dispute over Charges at the Port of Newcastle’ (Media Release, 16 November 2016), https://​www​.accc​.gov​.au/​media​-release/​ accc​-notified​-of​-an​-access​-dispute​-over​-charges​-at​-the​-port​-of​-newcastle. 103 Port of Newcastle Operations Pty Ltd v The Australian Competition Tribunal [2018] HCATrans 55 (23 March 2018). For the NCC’s progress regarding the application for revocation, see NCC, ‘Consideration of Possible Recommendation to Revoke Declaration of Service at the Port of Newcastle (Council’s Final Views)’, http://​ncc​.gov​.au/​application/​consideration​-of​-possible​-recommendation​-to​ -revoke​-declaration​-of​-service​-a/​2.

The Australian approach to third party infrastructure access  379 subject of declaration proceedings, has publicly identified that the overall cost of all applications and associated court proceedings regarding those four declaration applications was ‘in the order of hundreds of millions of dollars’.104 This raises the question as to whether the Regime, having established a complex, and sometimes protracted and expensive, declaration process, has in fact achieved benefits that justify the associated costs. D

What Is the Utility of the Regime?

The fact that relatively few declaration applications, and even fewer declarations, have been made may result from the existence and extent of other statutory avenues for securing infrastructure access. It may be that where access is clearly required in order to introduce competition into areas formerly dominated by public monopoly, or otherwise to solve a competition problem, issues have been addressed by alternative statutory avenues. This may explain the Regime’s relatively limited effect, but does it justify its continued existence? This issue was considered by a panel that undertook a comprehensive review of Australian competition law and policy. The panel recognized the existence of many other avenues for access, the relatively scant use made of the Regime, and (in what appeared to be a reference to the declaration applications concerning the iron ore railways – see the discussion of the public interest criterion below) that ‘imposing an access regime upon privately developed single-user infrastructure is more likely to produce inefficiency than efficiency, impeding the competitiveness of Australian industry’.105 However, it nonetheless recommended that the Regime remain in place as a ‘back stop’ for other avenues, and as a possible means of ‘future access regulation of airport and port infrastructure’. Therefore, the Regime appears set to remain an enduring feature of Australian competition law. In order to predict how the Regime may operate in future, it is useful to consider the nature of the legal issues that have arisen to date, and how the Regime’s application has been modified over time.

V

LEGAL ISSUES UNDER THE REGIME

A

Jurisdictional Issues

The Regime provides for declaration of ‘services’. The key controversy about the meaning of ‘service’ has concerned the exclusion of the ‘use of a production process’ from that definition. That exclusion, contained in limb (f), has not been amended since its introduction. This controversy arose in two cases in which iron ore miners resisted declaration of services provided by their privately developed and privately owned railway tracks. The tracks were used solely to support the relevant miner’s business. The miners argued that the railways were part of a production process comprising their integrated mine, rail and port operations, such that use of the track involved the ‘use of a production process’, meaning that that use was not a ‘service’ and could not be declared. 104 BHP Billiton, Submission to the Competition Policy Review Panel (June 2014) 26, 27, http://​ competitionpolicyreview​.gov​.au/​files/​2014/​06/​BHP​_updated​.pdf. 105 Competition Policy Review (n 2), 431.

380  Research handbook on methods and models of competition law In the first case, Kenny J of the Federal Court held that ‘production process’ means ‘a series of operations by which a marketable commodity is created or manufactured’.106 She considered that the ‘highly integrated’ system for producing iron ore was a production process. She observed that the system took ore that did not meet export grades, and used train sequencing, blending and stockpiling activities to produce, at the port, a ‘marketable commodity’: ore that met the grade, composition and quantity requirements for export.107 On this basis, she concluded that the use of part of the railway involved ‘use’ of a production process, because the railway was an integral and essential element of the process used to produce marketable iron ore.108 In the second case,109 a different iron ore miner made a similar argument, based on similar facts, and the High Court accepted that the miner’s integrated system was a ‘production process’ for producing marketable iron ore.110 However, the High Court held that the exception in limb (f) nonetheless did not apply: the service sought to be declared (the use of part of the railway track) involved use of part of the production process, but not use of the production process (that is, the integrated system) itself.111 B

The Declaration Criteria

Contrary to early decisions, the High Court has held that the declaration criteria are an exhaustive statement of the matters to be considered in relation to a declaration application: if all the criteria are satisfied, the Minister must declare the service112 and has no residual discretion to decide otherwise. The declaration criteria regarding promotion of competition (criterion (a)), assessment of whether it is uneconomic to develop another facility (criterion (b)) and the public interest (criterion (f)) have been the most controversial. The others, concerning national significance (criterion (c)), health and safety (criterion (d)) and effective access regimes (criterion (e)), have been materially less controversial, and are not addressed further. 1 Criterion (a): promotion of competition Criterion (a) has been amended since its introduction, and now reads (emphasis added) ‘that access (or increased access) to the service, on reasonable terms and conditions, as a result of a declaration of the service would promote a material increase in competition in at least one market (whether or not in Australia), other than the market for the service’. The italicized words were introduced in 2006, to address concerns that criterion (a) could otherwise be sat-

Hamersley Decision, 213. Ibid, 216. 108 Ibid, 219. 109 BHP Billiton Iron Ore Pty Ltd v National Competition Council (2008) 236 CLR 145. 110 Ibid, 160. 111 Ibid, 160–1. 112 Pilbara Infrastructure Pty Ltd v Australian Competition Tribunal (2012) 246 CLR 379, 423–4 (French CJ, Gummow, Hayne, Crennan, Kiefel and Bell JJ) (Pilbara HC Decision). 106 107

The Australian approach to third party infrastructure access  381 isfied where access113 would promote only a marginal or trivial increase in competition;114 the underlined words were introduced in 2017, to address further concerns that criterion (a) set too low a threshold for declaration. The underlined words have not been applied by any court and were not analysed in the cases addressed below. 2 Dependent markets Criterion (a) requires identification of one or more ‘dependent’ markets in which competition would be promoted by access to the relevant service. For example, where a party sought declaration of the use of a railway track, the dependent market might be a market for rail haulage, or for sale of the hauled goods. (a) Promotion of competition Criterion (a) is applied by undertaking a forward-looking analysis: a comparison of the future with and without access (or increased access).115 To ‘promote’ competition is to enhance ‘the conditions or environment for improving competition’.116 Criterion (a) does not require that access would achieve increased competition, or introduce or ‘unlock’ competition where none previously existed. So, for example, it might be satisfied if access would remove a material barrier to entry: such removal could, of itself, be said to enhance the conditions or environment for improving competition, and it would not be necessary to demonstrate that removing that barrier would in fact result in increased competition. Importantly, criterion (a) does not require that access had previously been restricted or denied: the Regime is not ‘a remedy for a wrong’, but ‘a public instrument for the more efficient working of essential facilities’.117 (b) ‘Access’ (i) Meaning of ‘access’  Early decisions interpreted ‘access’ as a right, ability or opportunity to use the service as brought about by declaration,118 and ‘increased access’ as encompassing use of the service on different terms and conditions than otherwise available, with the opportunity for arbitration following declaration.119 However, the Full Court of the Federal Court rejected that approach when it considered Virgin Blue’s application for declaration of a service at Sydney Airport.120 The Full Court considered that ‘access’ means a ‘right

113 This chapter generally uses the shorthand ‘access’ rather than referring repeatedly to ‘access (or increased access)’ and ‘promote competition’ instead of ‘promote a material increase in competition’. 114 Glencore Tribunal Decision, [85], citing the Explanatory Memorandum to the Trade Practices Amendment (National Access Regime) Bill 2005 (Cth), 4 at item 16 (p 21). 115 Pilbara Tribunal Decision (n 9), 425, citing Sydney Airport FC Decision. 116 See Pilbara Tribunal Decision (n 9), 427, and the cases cited there. 117 Sydney Airport FC Decision, 146. 118 See Sydney International Airport [2000] ACompT 1 [111]; Re Virgin Blue Airlines Pty Ltd (2005) 195 FLR 242, 270–1. 119 Re Virgin Blue Airlines Pty Ltd (2005) 195 FLR 242, 271. 120 Sydney Airport FC Decision, 146–7.

382  Research handbook on methods and models of competition law or ability’ to use the service,121 and that ‘increased access’ means ‘an increased or enhanced ability’ to purchase the use of essential facilities on fair and reasonable terms.122 (ii) Significance of voluntary access  When ‘access’ was interpreted to mean ‘access brought about by declaration’, criterion (a) was applied in a way that considered the incremental effect of declaration on competition. If the provider did not already provide use of the service, criterion (a) tested whether declaration would promote competition relative to that situation. If the provider did already provide use of the service voluntarily, criterion (a) tested whether declaration (‘increased access’) would promote competition relative to that situation (that is, having regard to whatever competition was facilitated by voluntary use) without declaration. This approach focused the analysis on the incremental effect of declaration. If a service was available for use by all who sought it, on fair and reasonable terms, the provider and users of a service could reasonably expect that declaration would not promote competition relative to that prevailing situation, and hence that criterion (a) would not be satisfied and declaration could not occur. Conversely, if a service was not generally available, or was available on discriminatory terms, a user or potential user might seek declaration to obtain improved terms or conditions through ACCC arbitration. However, when the Full Court adopted the broader interpretation of ‘access’ noted above, it stated that criterion (a) does not require consideration of ‘what is factually the current position’123 and that it was ‘an unnecessary constriction’ to read criterion (a) as requiring ‘a detailed factual enquiry heavily dominated by the past and the present’.124 The Full Court observed that there was ‘little doubt’ criterion (a) was satisfied in the case it was considering, since the relevant service was provided by a natural monopoly (an airport) over which the provider exerted monopoly power, and access to the airport was essential in order to compete, and for effective competition, in the domestic air passenger market.125 This analysis did not give weight to the fact that airlines, including Virgin Blue, already had use of the service, or consider whether the making of a declaration would promote competition relative to that which already existed having regard to Virgin Blue’s use of the relevant service. Rather, it appeared to compare the ‘with access’ scenario to a hypothetical scenario in which airlines could not use the airport, and so could not compete in the passenger market.126 The Full Court’s approach to criterion (a) is significant, since it suggested that services could be declared if access would promote competition relative to a hypothetical ‘no access’ scenario, even if declaration would not promote competition above prevailing levels of competition made possible by the voluntary provision of access. This issue arose in the Glencore Port Application, regarding access to channel and associated services at the Port of Newcastle. In that case, the right to operate the Port had been granted to a private operator about a year before the declaration application.127 The relevant service was already available to, and used by, ships transporting ore for Glencore and other

Ibid, 147. Ibid, 147–8. 123 Ibid, 147. 124 Ibid. 125 Ibid, 148. 126 Pilbara HC Decision, 423–4. 127 Glencore Tribunal Decision, [2], [12]–[13]. 121 122

The Australian approach to third party infrastructure access  383 miners,128 but the operator had recently increased prices by approximately 40 to 60 per cent for some vessel types.129 The interpretation of criterion (a) was a critical issue before the Tribunal, and Glencore put forward two alternative contentions. First, it said that ‘access’ means a right of use, and that since Port users did not (and would not) have any such right without declaration, they did not have ‘access’, even though they could already use the service.130 Glencore submitted that on this interpretation, and based on the Sydney Airport FC Decision, criterion (a) was satisfied: access was essential to compete in the dependent markets (including the market for coal production and export); the Port operator had monopoly power regarding the service; and there were no commercially viable alternatives for a person seeking to compete in the dependent markets.131 Second, Glencore said that, alternatively, if Port users’ ability to use the service meant that they did have ‘access’, the relevant question was whether ‘increased access’, via the opportunity to negotiate with recourse to arbitration, would promote competition.132 Glencore submitted that coal producers are price takers, could not pass on increased costs (for example, port charges),133 and were facing fragile market conditions; it said further that the Port’s charges were a substantial component of their margins.134 It submitted that declaration would create the potential for arbitration, and hence for regulated prices, and that this would create greater certainty for producers and others in dependent markets.135 It said that absent declaration, uncertainty about pricing would raise barriers to entry, blunt incentives to invest in and expand mines and associated infrastructure, and impair coal producers’ ability to compete effectively in the coal export market and other dependent markets.136 Therefore, it said that ‘increased access’ would promote competition because it would create greater certainty for market participants.137 The Tribunal found that criterion (a) was satisfied based on Glencore’s first argument.138 It considered itself bound by the Sydney Airport FC Decision139 and noted that if it was not so bound, it would not have found criterion (a) to be satisfied.140 The Full Court of the Federal Court upheld the Tribunal’s decision on this point,141 and the High Court refused special leave to appeal from the Full Court’s decision.142

Ibid, [57]. Ibid, [16]. 130 Ibid, [35], [57]. 131 Ibid, [35]. The coal market was one of several dependent markets Glencore identified: [37]–[39]. 132 Ibid, [35], [122]. 133 Ibid, [131]. 134 Ibid, [132]. 135 Ibid, [35], [131]–[134], [138]. 136 Ibid, [134]. 137 Ibid, [138]. 138 Ibid, [121]. 139 Ibid, [79], [83], [92], [102]. 140 Ibid, [157]. 141 Port of Newcastle Operations Pty Ltd v Australian Competition Tribunal [2017] FCAFC 124, [146]. 142 Port of Newcastle Operations Pty Ltd v The Australian Competition Tribunal [2018] HCATrans 55 (23 March 2018). 128 129

384  Research handbook on methods and models of competition law Criterion (a) was recently amended to introduce the underlined words noted above.143 This amendment is intended to focus criterion (a) on the incremental effect of declaration on competition (contrary to the Sydney Airport FC Decision).144 3 Declaration criterion (b): uneconomical for anyone to develop another facility Criterion (b) has been a highly controversial criterion. At the heart of that controversy is a simple question: should criterion (b) test whether it would be profitable to develop another facility or look at broader questions about whether such development would be efficient, or beneficial, from the perspective of society as a whole? (a) A social cost test? Early decisions interpreted ‘uneconomical’ in criterion (b) to mean uneconomical ‘in terms of the associated costs and benefits of development for society as a whole’.145 ‘Anyone’ was interpreted as excluding the provider: the fact that someone could develop another facility would generally demonstrate that the facility was not a bottleneck (that is, was not so ‘essential’ that denial of access could impede competition), but the provider’s ability to do so was not thought to have the same significance. The concern was that the provider might not share either facility, and therefore their ability to develop another facility would not indicate whether the facility was a ‘bottleneck’.146 The Tribunal applied this test by comparing the costs of accommodating ‘a likely range of reasonably foreseeable demand’ for the service on the existing facility with the costs of that demand being accommodated by two or more facilities.147 The test took into account the possibility that the existing facility might be expanded to accommodate third party demand.148 (b) A natural monopoly test In the Pilbara Tribunal Decision, the Tribunal rejected the ‘social cost’ test in favour of a ‘natural monopoly’ test when considering declaration applications regarding four privately owned and developed iron ore railways. Each railway was used solely to transport iron ore from its owner’s mines to port and did not carry third parties’ ore. The applicant149 was a new entrant miner; it had developed its own railway but sought declaration to benefit itself and other recent entrants. The parties identified three possible interpretations of ‘uneconomical’, which the Tribunal summarized: that: (1) it would not be profitable for anyone to develop the facility (the ‘privately profitable’ test); (2) the total net costs (including social costs) exceed the total net benefits (including social benefits)

CCA, s 44CA(1)(a), introduced by the 2017 Bill, Sch 12, Pt 1, item 2. 2017 EM, 97. 145 See Sydney International Airport [2000] ACompT 1, [204]. 146 Ibid, [201]–[205]. 147 Duke Eastern Gas Pipeline Pty Ltd (2001) 162 FLR 1, [137]. 148 National Competition Council, ‘Final recommendation – Goldsworthy Railway, Application for declaration of a service provided by the Goldsworthy Railway under section 44F(1) of the Trade Practices Act 1974’ (29 August 2008) 74–5. 149 The applications were made by Fortescue Metals Group Ltd (in one case) and its wholly owned subsidiary, The Pilbara Infrastructure Pty Ltd (in three cases). 143 144

The Australian approach to third party infrastructure access  385 of developing another facility (the ‘net social benefit’ test); or (3) a single facility can meet market demand at less total cost than two or more facilities (a ‘natural monopoly’ test).150

This passage reflects the key tension about the role of Part IIIA. The ‘privately profitable’ test reflected the view that the Regime is not intended to apply if another facility can profitably be built because in that situation there is no ‘bottleneck’. That is, an access seeker could compete in the dependent market by using another (profitable) facility, without obtaining access to the facility sought to be declared. That prospect may incentivize the provider to provide access voluntarily but, regardless, the access seeker has a route to market (such that competition is not impeded) via the alternative facility.151 The natural monopoly and net social benefit tests reflected the view that the Regime is concerned both with removing bottlenecks and enhancing efficiency. On this view, declaration should be available if it would be more efficient to share than duplicate a facility, even if duplication would be profitable.152 These views also reveal differences about whether competition problems should be solved by reliance on market forces (such as the constraint arising from the fact that a facility can profitably be duplicated), or by regulatory intervention (that is, a decision maker’s assessment of whether sharing a facility would be more efficient than developing another facility).153 The Tribunal considered that a ‘natural monopoly’ test was most appropriate, given the emphasis on natural monopolies in the background materials to the introduction of the Regime,154 and its view that the Regime focused on effective competition (the subject matter of criterion (a)) and social efficiency (the subject matter of criterion (b)).155 The Tribunal rejected the privately profitable test because it was inconsistent with the Regime’s focus on ‘effective’ competition, and efficiency: it considered that it assessed the bare question whether someone could compete, but not whether they could compete effectively,156 or whether duplicating the facility would be efficient.157 The Tribunal also questioned the assumption, underlying the private test, that firms would behave in an economically rational manner, and noted that rational firms might deny giving access to a potential competitor, even if sharing would be socially optimal: ‘it is far from clear that market forces achieve a better result than regulation as a general rule’.158 The Tribunal also considered that a privately profitable test would create significant, and undesirable, overlap between criteria (a) and (b).159 The Tribunal rejected the ‘net social benefit’ test because it preferred the natural monopoly test for the reasons given above, and because it would cause criterion (b) to substantially overlap with, and perhaps usurp, criterion (f).160

152 153 154 155 156 157 158 159 160 150 151

Pilbara Tribunal Decision (n 9), 386. Ibid, 387, addressing evidence given by Professor Willig in that case. Ibid, 386. Ibid, 387–8, addressing evidence given by Professors Willig, Ordover and Kalt in that case. Ibid, 388–9. Ibid, 389–90. Ibid, 386–7. Ibid, 387. Ibid, 388. Ibid, 389. Ibid, 391.

386  Research handbook on methods and models of competition law The Tribunal applied this natural monopoly test by asking: ‘Can each line provide society’s reasonably foreseeable demand for the below rail service at a lower total cost than if provided by two or more lines?’161 Applying this test, the Tribunal found that it would be uneconomic to develop another railway to provide the service provided by three of the four railways at issue. (c) Adoption of the privately profitable test The Full Court of the Federal Court162 and the High Court163 rejected the Tribunal’s approach on appeal, and instead adopted the privately profitable test. The High Court majority held that criterion (b) requires ‘the decision maker to be satisfied that there is not anyone for whom it would be profitable to develop another facility’,164 using ‘profitable’ in the sense of ‘could reasonably expect to obtain a sufficient return’ on capital employed.165 The majority held that ‘anyone’ included ‘existing and possible future market participants’ and saw no reason to interpret it as excluding the owner of the pre-existing facility.166 The majority considered that ‘textual considerations’, particularly the use of the word ‘anyone’ in criterion (b) and the phrase ‘economically feasible’ in relevant background materials, pointed away from a natural monopoly test and towards a privately profitable test.167 The majority observed that ‘due weight’ must be given to the national and economic objectives of Part IIIA of the CCA, the Regime’s objects, and the intention that the Regime would ‘contribute to national economic efficiency’.168 The majority also observed that the natural monopoly test was undesirably focused on existing conditions (for example, existing costs), whereas it could be difficult or impossible to predict future demand and changes in technology, and where the declaration decision ‘must hold good’ for the whole declaration period.169 It observed that a natural monopoly may not be sustainable if it faces a credible threat of competitive entry into a market170 and that the existence of that threat would contribute to efficiency by inducing the monopolist to produce and price its service efficiently.171 The majority noted that while the privately profitable test could allow duplication of a natural monopoly, avoiding such duplication was not the only inefficiency the Regime sought to guard against.172 Further, it observed that a key flaw of the natural monopoly test was that it would not allow declaration if a facility was not a natural monopoly, even if duplication would be unprofitable. In that situation, only the privately

Ibid, 394–5. Pilbara Infrastructure Pty Ltd v Australian Competition Tribunal (2011) 193 FCR 57, 98 (Pilbara FC Decision). 163 Pilbara HC Decision. 164 Ibid, 411. 165 Ibid, 420. 166 Ibid, 421 disagreeing with the Full Court on this point: Pilbara FC Decision, 94. 167 Ibid. 168 Ibid, 418–19. 169 Ibid, 419. 170 Ibid. 171 Ibid. 172 Ibid. 161 162

The Australian approach to third party infrastructure access  387 profitable test would allow declaration, and hence the promotion of competition in the face of such a bottleneck.173 The High Court also observed that the privately profitable test appropriately predicted whether anyone was likely to develop another facility,174 whereas other tests considered an abstract hypothesis, which did not consider likely market behaviour175 or those who might engage in it.176 It noted that the privately profitable test was the converse of a question which ‘lies at the heart of’ every decision to invest in infrastructure: whether it would be economically feasible to develop another facility.177 As a result of the High Court’s decision, criterion (b) was interpreted to limit the scope for regulatory intervention where market forces might be expected to rectify potential competition problems (that is, distortion of competition in dependent markets). Criterion (b) was recently amended to replace the private profitability test with an express natural monopoly test. This means that criterion (b) now tests whether the relevant facility ‘could meet the total foreseeable demand in the market’ over ‘the period for which the service would be declared’, at ‘the least cost compared to any two or more facilities’.178 The new criterion provides some scope to consider whether a facility could, as a result of an expansion to the facility, accommodate increased demand at least cost (etc). ‘Costs’ will include ‘all costs associated with having multiple users of the facility (including such costs that would be incurred if the service is declared)’.179 This amendment reflects the policy view that the objects of the Regime are best served by a natural monopoly assessment, rather than a privately profitable assessment. 4 Declaration criterion (f): public interest Criterion (f) requires the Council/Minister to be satisfied ‘that access (or increased access) to the service would not be contrary to the public interest’. A ‘very wide’ range of matters may be considered under criterion (f), and its application is particularly suited to the holder of a political office.180 Criterion (f) has recently been amended. (a) Previous application Criterion (f) raises the same issues as criterion (a) concerning the word ‘access’.181 Questions also arise about whether the public interest analysis can encompass matters addressed under other criteria,182 or in an arbitration. The types of factors that have been considered when applying criterion (f) include the effect of access on competition, the direct and indirect costs of regulation arising from declaration

Ibid, 420. Ibid, 413. 175 Ibid, 412–13. 176 Ibid, 413, 418. 177 Ibid, 421. 178 CCA, s 44CA(1)(b), introduced by the 2017 Bill, Sch 12, Pt 1, item 2. 179 CCA, s 44CA(2), introduced by the 2017 Bill, Sch 12, Pt 1, item 2. 180 Pilbara HC Decision, 401. 181 That is, whether ‘access’ refers to ‘declaration’, or to some broader concept; earlier decisions appeared to take the former approach (see Re Virgin Blue Airlines Pty Ltd (2005) 195 FLR 242, 352 onwards), later decisions the latter approach (see Pilbara Tribunal Decision (n 9), 445–6). 182 Glencore Tribunal Decision, [170]. 173 174

388  Research handbook on methods and models of competition law (including distortion of efficient investment or production decisions),183 the effect of declaring a service that is already subject to other forms of regulation,184 and operational consequences of declaration (such as the risk of increased congestion or accidents as a result of access).185 The application of criterion (f) in the Pilbara Tribunal Decision illustrates the nature of the public interest issues that can arise. In that case, the Tribunal applied criterion (f) by undertaking an extensive cost–benefit analysis.186 Having done so, it concluded that the two key benefits from access were those associated with ‘unlocking’ iron ore deposits that might not otherwise be developed (or otherwise expediting development of, or increasing production from, iron ore mines), and the capital savings from avoiding the need to build new railways.187 It considered that the two key costs were the costs of discouraging development of new railways (including foregoing associated technological advancements and cost reductions), and the costs associated with delays to providers introducing new technologies on, or expanding, their own railways.188 The Tribunal noted evidence that a three-month average delay to an expansion of the railway system operated by one of the incumbent miners could result in lost export revenue estimated to be in the order of AU$10 billion.189 Ultimately, the Tribunal concluded that criterion (f) was only satisfied (that is, the benefits of access would exceed the costs) for two of the four railways.190 Although the High Court later held that the Tribunal had exceeded its jurisdiction, it observed that the types of matters the Tribunal had considered could nonetheless be relevant to the Minister’s application of criterion (f).191 This criterion was recently amended to refer to ‘access (or increased access) on reasonable terms and conditions, as a result of a declaration’, and test whether access ‘would promote’ (rather than ‘not be contrary to’) the public interest.192 These changes were intended to require the decision maker to be satisfied that access through declaration (rather than the broader concept of ‘access’) is likely to generate overall gains to the community193 and to impose a higher public interest hurdle to declaration.194 The amendments also require a decision maker applying this criterion to consider the effect of declaration on investment (in infrastructure services, and markets that depend on the relevant service), and the administrative and compliance costs the provider would incur if the service was declared.195

Re Virgin Blue Airlines Pty Ltd (2005) 195 FLR 242, 352–3. Glencore Tribunal Decision, [163]–[167]. 185 Sydney International Airport [2000] ACompT 1, [219]–[220]. 186 Pilbara Tribunal Decision (n 9), 446. 187 Ibid, 467–8. 188 Ibid, 468. 189 Ibid, 472. 190 Ibid, 469. 191 Pilbara HC Decision, 422–3. 192 CCA, s 44CA(1)(d), introduced by the 2017 Bill, Sch 12, Pt 1, item 2 (emphasis added). 193 2017 EM, 101. 194 Productivity Commission, Inquiry into the National Access Regime (Report No 66, 2013) 178–9, http://​www​.pc​.gov​.au/​inquiries/​completed/​access​-regime/​report/​access​-regime​.pdf; Commonwealth of Australia, ‘Australian Government Response to the Productivity Commission and Competition Policy Review Recommendations on the National Access Regime’ (24 November 2015) 6, http://​www​.treasury​ .gov​.au/​PublicationsAndMedia/​Publications/​2015/​National​-Access​-Regime​-Response. 195 CCA, s 44CA(3), inserted by the 2017 Bill, Sch 12, Pt 1, item 2. 183 184

The Australian approach to third party infrastructure access  389

VI CONCLUSION The Regime has now existed for more than 20 years. Over that time, it has been used as a means of seeking access to a wide range of infrastructure, including infrastructure that plays a central role in the Australian economy. The associated processes have, in some cases, been particularly protracted and costly. However, only a small number of services have actually been declared. Nonetheless, a recent review of Australian competition law concluded that the Regime should be retained as a ‘back stop’ to other statutory forms of infrastructure regulation. Following this finding, it appears that the Regime is set to remain an enduring, if unusual, feature of the Australian competition law landscape and that there will be continuing uncertainty regarding its application following the introduction of recent amendments to the declaration criteria.

PART IV ENFORCEMENT

17. The EU method of antitrust enforcement Andreas Stephan

I INTRODUCTION The EU’s administrative system of competition law has emerged as the world’s most common antitrust enforcement model. Its defining characteristics are that it is a purely civil enforcement regime that is non-adversarial. In contrast to jurisdictions where a criminal procedure is employed or where competition cases are heard by a tribunal, Europe’s competition regulator – the European Commission – investigates, finds guilt, and imposes an appropriate penalty without the need for judicial consent. Indeed, there is no truly independent adjudication of an enforcement case unless a party subject to an infringement decision decides to launch an appeal to the Court of Justice of the European Union (CJEU). Despite the administrative design of EU competition law, the penalties imposed – corporate fines of up to 10 per cent of annual worldwide turnover – are unmistakably punitive in their scale and function. Indeed, EU antitrust enforcement is considered criminal in nature by the European Court of Human Rights.1 The primary aim of the EU regime is therefore to punish and deter infringements, with a clear focus on hard-core horizontal cartel behaviour. In contrast to a system that simply seeks to regulate how undertakings (businesses) interact with each other, European antitrust officials describe the problem of cartel harm as akin to ‘cancers on the open market’.2 Although behavioural and regulatory tools can be employed against less serious competition concerns,3 a deterrence-based approach is considered necessary against cartels, because they engage in behaviour that is clandestine and that therefore cannot be easily identified or regulated.4 Although the EU is often held up as a model for antitrust, it is a regime that is still in its relative infancy in terms of active enforcement. Before the late 1990s, EU competition rules were enforced infrequently and were considered to have limited deterrent effect.5 The infrequent fines were low and calculated with such a lack of transparent methodology that

Case C-272/09P, Stenuit v France [1992] ECC 401; Engel v Netherlands (1979–80) 1 EHRR 647. Mario Monti, ‘Fighting Cartels Why and How?’ (Speech to the 3rd Nordic Competition Policy Conference, Stockholm, 11–12 September 2000). 3 For example, Commitment Decisions under Article 9, Council Regulation (EC) No 1/2003 of 16 December 2002 on the implementation of the rules on competition laid down in Articles [101] and [102] of the Treaty [2003] OJ L1/1. 4 In addition, ex post damages cannot completely redress the damage caused to the economy. Antitrust fines in the EU are paid into the general EU budget and Member States’ contributions are reduced accordingly in the following year. Private parties can rely on an infringement decision by the European Commission to sue for damages. See Directive 2014/104/EU of the European Parliament and of the Council of 26 November 2014 on certain rules governing actions for damages under national law for infringements of the competition law provisions of member states and of the European Union Text with EEA relevance [2014] OJ L 349, 1–19. 5 For an excellent discussion of how cartel regulation developed over this period, see Christopher Harding and Julian Joshua, Regulating Cartels in Europe (Oxford University Press 2003) chs IV, V. 1 2

391

392  Research handbook on methods and models of competition law one commentator compared them to a ‘lottery’.6 Over this period, European Commission enforcement and the jurisprudence of the CJEU refined and clarified the way in which Article 101 of the Treaty on the Functioning of the European Union (TFEU), formerly Article 81 of the Treaty Establishing the European Community, was to be enforced. In 1996, the European Commission adopted its first leniency programme for cartels, revising it in 2002 and 2006 and adding a settlement procedure in 2008. In 1998, it adopted guidelines on the method of calculating fines, which were revised in 2006. Yet, the most distinctive feature of European antitrust enforcement is the relationship between EU antitrust rules and those of its Member States. For many years EU competition laws were applied loosely because some Member States of the EU had not developed equivalent domestic antitrust regimes or achieved a sufficient level of harmonization. The relationship today is governed by Regulation 1/2003,7 which gives the European Commission the exclusive power to investigate an infringement of Article 101 of the TFEU, where it initiates proceedings against arrangements that affect trade between Member States.8 The Regulation also ensures consistency in the scope of Article 101 and its domestic equivalents, and created the European Competition Network (ECN) to facilitate information exchange and case allocation, and to further ensure consistency between European authorities and courts. Yet during this time of apparent convergence, national competition authorities have developed their own enforcement tools and approaches. For example, most now employ additional civil and criminal sanctions against individuals, whereas the Commission still only imposes corporate fines. The purpose of this chapter is to critically assess the methodology of the EU antitrust enforcement regime against its stated aim of punishing and deterring hard-core horizontal cartels. It will do this by focusing on four key areas in the sections that follow. Section II explores the substantive cartel prohibition contained in Article 101 of the TFEU. This provides legal certainty to deterrence by identifying categories of behaviour that are treated as anticompetitive ‘by object’, meaning that they are illegal regardless of effects or of whether they were implemented. Yet the boundaries of ‘object’ behaviour are unclear. As the other elements of the prohibition are drawn loosely, over time, the scope of ‘object’ has widened to the point that it may risk capturing behaviour that is benign or that should be subject to an effects analysis. Section III looks at the method of calculating fines. The European Commission adopted guidelines to enhance the transparency of cartel fines but does not wish to make them predictable. Studies have suggested that EU fines may fall well short of the illegal profits earned by cartels, raising this question: what kind of deterrence is the EU aiming to achieve? Section IV considers the use of leniency and settlement. These tools were introduced to improve the efficiency of detection and punishment but they significantly reduce the levels of fines imposed and may simply provide opportunities for infringing firms to act strategically. Section V examines the use of sanctions on the national level as a complement to enforcement by the Commission. The analysis in Sections III and IV points to the need for individual sanctions in addition to corporate fines. As these only occur on the national level, the chapter identifies the obstacles that exist to their use in EU-wide cases. Perversely, they are less likely to be used in the more

6 I Van Bael, ‘Fining à la Carte: The Lottery of EU Competition Law’ (1995) 16(4) European Competition Law Review 237. 7 Council Regulation (EC) No 1/2003 of 16 December 2002. 8 Guidelines on the effect on trade concept contained in Articles [101] and [102] of the Treaty [2004] OJ C101/81, para 13.

The EU method of antitrust enforcement  393 serious and harmful multi-jurisdictional instances of cartel behaviour and are largely limited to smaller local infringements. The chapter concludes that while EU antitrust enforcement is deterrence-enhancing in some respects, there are some significant shortcomings which need to be addressed if the regime is to be truly deterrent in its effect.

II

SUBSTANTIVE RULE: THE ‘OBJECT’ OF RESTRICTING COMPETITION

Under Article 101 of the TFEU, bilateral arrangements that have as their ‘object’ the restriction of competition are deemed to be so serious that they are automatically unlawful, regardless of whether they have any anticompetitive effects or, indeed, whether they were even implemented.9 The assumption is therefore that a restriction of competition is a ‘necessary consequence of the agreement’.10 Where an arrangement does not have the object of restricting competition, the authority must evaluate whether there are anticompetitive ‘effects’. This takes account of the actual conditions in the market, its structure and the economic context in which the undertakings operate. It involves establishing the counterfactual (the world had the arrangement not been entered into) to demonstrate an anticompetitive effect.11 The question of whether an arrangement has the ‘object’ or ‘effect’ of restricting competition for the purposes of Article 101 of the TFEU is of heightened importance because of the wide meaning afforded to the other key elements in the provision. Article 101(1) states: The following shall be prohibited as incompatible with the internal market: all agreements between undertakings, decisions by associations of undertakings and concerted practices which may affect trade between Member States and which have as their object or effect the prevention, restriction or distortion of competition within the internal market. [Emphasis added.]

For Article 101 to apply there is no requirement of an agreement. A ‘concerted practice’ is enough and this can exist where there is any coordination between competitors that ‘knowingly substitutes practical cooperation between them for the risks of competition’.12 Indeed, where there is any direct communication between competitors about pricing, output and selling decisions, there is a presumption that a concerted practice will follow.13 This presumption can be rebutted where there is a plausible justification for the behaviour.14 However, there is no requirement of a ‘meeting of minds’, or any reciprocal behaviour to signal a commitment

9 One might compare this to inchoate offences in criminal law. See David Bailey, ‘Restrictions of Competition by Object under Article 101 TFEU’ (2012) 49 Common Market Law Review 559, 563–4. 10 Opinion of AG Trstenjak in Case C-209/07, Competition Authority v Beef Industry Development Society Ltd [2008] ECR I-9291, para 46. 11 Case T-374/94, European Night Services [1999] ECR II-3141, 136. 12 Joined Cases 40-48/73, 50/73, 54–56/73, 111/73, 113–114/73, Coöperatieve Vereniging ‘Suiker Unie’ UA v Commission [1976] 1 CMLR 295, citing Case 48/69, ICI v Commission (1972) ECR 619, paras 64–68. 13 Case C-49/92, Commission v Anic Partecipazioni SpA [1999] ECR I-4125, para 121; Case C-199/92, Hüls AG v Commission [1999] ECR I-4287, para 162. 14 Joined Cases C-89/85, C-104/85, C-114/85, C-116/85, C-117/85 and C-125/85 to C-129/85, A Ahlström Osakeyhtiö v Commission [1993] ECR I-1307, 71.

394  Research handbook on methods and models of competition law to act on the information,15 or even the need for repeated exchanges between the competitors. One instance is enough.16 In addition, there is no bright line between an agreement and a concerted practice. Neither the Commission nor the courts typically specify whether the relevant conduct satisfies one or the other. Article 101 of the TFEU is deliberately wide in this respect, to ensure that parties cannot easily circumvent the law by relying on collusive strategies that fall short of an explicit agreement. In principle, concerted practice can capture a whole variety of arrangements so long as they do not constitute independent decision-making. In practice, there are serious limits to the Commission’s ability to counter every possible justification for indirect communication or forms of parallel behaviour.17 The meaning of ‘undertaking’ is also very broad. It essentially ‘encompasses every entity engaged in an economic activity regardless of the legal status of the entity and the way in which it is financed’.18 The legal status of the entity and whether it is profit-making are irrelevant. Indeed, even someone who acts in a secretarial capacity or as a consultant to an agreement (and therefore does not actually produce any of the goods or services subject to the agreement) is deemed to engage in economic activity and therefore constitutes an undertaking.19 EU competition law takes a functional approach to defining an undertaking, meaning that an entity can be an undertaking when carrying out one activity but not when carrying out another.20 The only activities an entity can engage in that do not make them an undertaking are those connected with the exercise of the powers of a public authority and instances of solidarity. The former concerns activities that are in the public interest, which usually are engaged for the benefit of all and are not paid for at the point of use.21 The latter relates to ‘the inherently uncommercial act of involuntary subsidization of one social group by another’.22 This might include pension, social security and health care schemes, where different members make varying contributions but everyone is entitled to the same benefit. In addition to these broad elements of the Article 101 prohibition, the importance of object and effect is also heightened by the challenges of successfully relying on the efficiency defence 15 However, conduct cannot be unilateral, meaning that its aims must not be achievable ‘without the express or implied participation of another undertaking’: Case T-41/96, Bayer AG v Commission [2000] ECR II-3383, 71. Discussed in Okeoghene Odudu, The Boundaries of EC Competition Law: The Scope of Article 81 (Oxford University Press 2006) 67. 16 Case C-8/08, T-Mobile Netherlands BV v Raad van bestuur van de Nederlandse Mededingingsautoriteit [2009] ECR I-4529, para 59. 17 Joined Cases C-89/85, C-104/85, C-114/85, C-116/85, C-117/85 and C-125/85 to C-129/85, A Ahlström Osakeyhtiö v Commission [1993] ECR I-1307, 71. 18 Case 41/90, Höfner and Elser v Macroton GmbH [1991] ECR I-1979. 19 Case T-27/10, AC-Treuhand AG v European Commission, Judgment of 6 February 2014. 20 See Case C-49/07, Motosykletistiki Omospondia Ellados NPID v Greece [2009] 5 CMLR 11. ‘The classification as an activity falling within the exercise of public powers or as an economic entity must be carried out separately for each activity to be exercised by a given entity’: para 7. See also Case C-264/01, AOK Bundesverband v Ichtyol Gesellschaft Cordes para 2004] 4 CMLR 22, paras AG25, AG45, 58; Case T-155/04, SELEX Sistemi Integrati SpA v Commission [2007] 4 CMLR 1096, para 90. 21 Case C-309/99, Wouters v Algemene Raad van de Nederlandse Orde van Advocaten [2002] ECR I-577; Case C-205/03, P Federación Española de Empresas de Tecnología Sanitaria (FENIN) v Commission [2006] ECR I-6295, paras 25–26. 22 Case C-70/95, Sodemare v Regione Lombardia [1997] ECR I-3395, para 29; Cases C-159–160/91 Poucet et Pistre v Assurances Generales de France [1993] ECR I-637; Case C-244/94, Fédération Française des Sociétés d’Assurance and others v Ministère de l’Agriculture et de la Pêche [1995] ECR I-4013.

The EU method of antitrust enforcement  395 contained in Article 101(3). This states that the provisions of Article 101(1) may be declared inapplicable where the arrangement ‘contributes to improving the production or distribution of goods or to promoting technical or economic progress, while allowing consumers a fair share of the resulting benefit’ and does not impose restrictions that are dispensable or ‘afford such undertakings the possibility of eliminating competition in respect of a substantial part of the products’. Prior to May 2004, this efficiency defence was subject to a notification regime whereby the European Commission would grant individual exemptions. Since the introduction of Regulation 1/2003 (also known as the Modernization Regulation), the system has moved to one of exceptions, self-assessed by the parties subject to the arrangement. However, Article 101(3) sets the bar very high. The burden falls on the parties to satisfy its criteria and any efficiency gains must be substantial.23 While this means that Article 101(3) can in principle be invoked to save any infringement of Article 101, unlike the US ‘rule of reason’ approach, there is no balancing of procompetitive and anticompetitive effects before a prima facie infringement has been established.24 This makes relying on Article 101(3) risky and unpredictable for businesses. They are more likely to adapt their arrangements to avoid liability altogether. Some legal certainty is provided by safe harbours created through block exemptions. If the criteria of a block exemption are met, then the arrangement automatically benefits from an Article 101(3) exception. These block exemptions seek to strike a careful balance between the need for clarity and the often mixed or inconsistent effects of certain types of agreements. In relation to conduct that is usually beneficial to markets and consumers, block exemptions employ market share thresholds to limit the scope for protecting conduct that is likely to be anticompetitive. One of the widest of these relates to vertical agreements, where, except for a small number of blacklisted clauses such as minimum resale price maintenance, a vertical agreement will fall outside the scope of Article 101 if the market shares of either parties do not exceed 30 per cent.25 Horizontal cooperation agreements benefit from a block exemption on ‘research and development’ and ‘specialization’ agreements. These include joint production, specialization and subcontracting; joint purchasing; joint commercialization; licensing agreements; standardization agreements; and the exchange of some categories of information.26 The relevant market shares for these are lower – the joint market share limits vary between 15 per cent and 25 per cent, depending on the type of agreement. Beyond the circumstances where an arrangement benefits from a block exemption or can satisfy Article 101(3), the EU prohibition 23 Commission Notice, Guidelines on the application of Article 81(3) of the Treaty [2004] OJ C101/08B; Brenda Sufrin, ‘The Evolution of Article 81(3) of the EC Treaty’ (2006) 51 Antitrust Bulletin 915. 24 For some time, there was a suggestion that a rule of reasoned approach existed in the Commission’s analysis of whether an arrangement had an anticompetitive effect, but this was rejected in Case T-112/99, Métropole Télévision SA v Commission (2001) ECR II-2459. 25 Commission Notice, Guidelines on vertical restraints [2010] OJ C130/1; Commission Regulation (EC) No 330/2010 on the application of Article 101(3) of the Treaty to categories of vertical agreements and concerted practices [2010] OJ L102/1. 26 Commission Notice, Guidelines on the applicability of Article 101 of the Treaty on the Functioning of the European Union to horizontal co-operation agreements [2011] OJ C11; Commission Regulation (EC) No 1217/2010 of 14 December 2010 on the application of Article 101(3) of the Treaty on the functioning of the European Union to categories of research and development agreements [2010] OJ L335; Commission Regulation (EC) No 1218/2010 of 14 December 2010 on the application of Article 101(3) of the Treaty on the functioning of the European Union to categories of specialisation agreements [2010] OJ L335.

396  Research handbook on methods and models of competition law can in principle capture a vast spectrum of conduct between competitors, from deliberate hard-core ‘object’ agreements to very informal arrangements that may inadvertently have the ‘effect’ of restricting competition. ‘Object’ and ‘effect’ are alternative conditions. Conduct falls into either one or the other. The operation of Article 101 was given expression in European Night Services: In assessing an agreement under [Article 101] account should be taken of the actual conditions in which it functions, in particular the economic context in which the undertakings operate, the products and services covered by the agreement and the actual structure of the market concerned unless it is an agreement containing obvious restrictions of competition such as price-fixing, market-sharing or the control of outlets. In the latter case, such restrictions may be weighed against their claimed pro-competitive effects only in the context of [Article 101(3)] of the treaty, with a view to granting an exemption …27

According to the Commission’s horizontal guidelines, object is said to apply to agreements ‘presumed to have negative market effects’.28 The Beef Industry case clarified that ‘object’ is a legal presumption that competition is restricted ‘where the necessary consequence of the agreement was the restriction of competition’.29 Odudu interprets this as meaning that a ‘restriction of competition is the inevitable consequence of the agreement’.30 However, it was accepted that ‘account must be taken of not only its content, but also its legal and economic context … only of the elements of the legal and economic context which could cast doubt on the existence of a restriction of competition’.31 As there is no definition of ‘object’ beyond statements of this nature, it is hard to determine its exact limits. Indeed, European competition authorities frequently refer to an arrangement as having the ‘object and/or effect’ of restricting competition, but very rarely undertake any actual effects analysis. One consequence of this uncertainty is that the scope of ‘object’ has widened over time. In particular, in T-Mobile, it was held that ‘in order for a concerted practice to be regarded as having an anticompetitive object, it is sufficient that it has the potential to have a negative impact on competition. In other words it must simply be capable in an individual case’ (emphasis added).32 A similar approach was taken in Allianz Hungaria Biztosito,33 leading to criticisms that object had become ‘overly-expansive and, indeed, rather worryingly vague and broad’.34

Case T-374/94 [1999] ECR II-3141, 136. Commission Guidelines on the applicability of Article [101] of the EC Treaty to horizontal cooperation agreements [2001] OJ C3/2. See also Guidelines on the application of Article [101(3)] of the Treaty [2004] OJ C101/97, para 21. 29 Case C-209/07, Competition Authority v Beef Industry Development Society Ltd and Barry Brothers (Carrigmore) Meats Ltd (BIDS), EU:​C:​2008:​643, para 17 and AG Opinion (n 10) para 44. 30 Okeoghene Odudu, ‘Restrictions of Competition by Object – What’s the Beef?’ (2009) 9(1) Competition Law Journal 11. 31 BIDS (n 29), AG Opinion (n 10) para 50. 32 Case C-8/08, T-Mobile Netherlands BV v Raad van bestuur van de Nederlandse Mededingingsautoriteit [2009] ECR I-4529, para 31. 33 Case C-32/11, Allianz Hungaria Biztosito Zrt and Others v Gazdasagi Versenyhivatal, Judgment of 14.03.2013, ECLI:​EU:​C:​2013:​160. 34 Alison Jones and Brenda Sufrin, EU Competition Law (5th edn, Oxford University Press 2014) 213. 27 28

The EU method of antitrust enforcement  397 It is thought that this expansive creep of ‘object’ was halted by the CJEU in Groupement des cartes bancaires, a case that was set aside because it was found to blur the distinction between ‘object’ and ‘effect’.35 The Court said that the Commission’s case failed to show that the coordination revealed in itself ‘a sufficient degree of harm to competition’.36 However, while the Court made it clear that ‘object’ should be applied sparingly, it failed to set out any real criteria that might prevent future over-reach of ‘object’. Confusingly, it also appeared to draw on earlier cases such as Allianz, which blurred the object/effect distinction. The problem is further compounded by the fact that the CJEU has been reluctant to require any kind of initial or quick-look assessment as to the likely anticompetitive effects of a given arrangement. In GlaxoSmithKline Services Unlimited v Commission, it held that ‘for a finding that an agreement has an anticompetitive object, it is not necessary that final consumers be deprived of the advantages of effective competition in terms of supply or price’.37 As acknowledged by AG Wahl’s opinion in Groupement des cartes bancaires, the difficult question that arises is how exactly one determines whether a novel arrangement has the ‘object’ of restricting competition without any analysis of effects.38 From a deterrence perspective, the expansive nature of Article 101 of the TFEU makes some sense. The wide definitions of ‘undertaking’ and ‘concerted practice’ and the lack of precisely defined limits to the meaning of ‘object’ all make it very difficult for firms to circumvent competition rules by adapting their anticompetitive arrangements. In this respect, ‘object’ is deterrence-enhancing because it punishes unsuccessful attempts at colluding, regardless of their stage of implementation, success or failure. The limited scope for any kind of effects analysis, in conjunction with block exemptions and the availability of Article 101(3), also provides a greater degree of legal certainty than would otherwise be the case.39 Even just an abridged effects test for the applicability of ‘object’ risks firms engaging in cartel-like conduct where they wrongly calculate that such conduct is benign. The drawback with this approach is that the boundaries of ‘object’ are fluid when it comes to novel types of behaviour. The danger is that ‘object’ will capture some arrangements that are likely to be entirely benign, such as cover pricing.40 This is especially so where the arrangement does not benefit from a block exemption and cannot meet the criteria of Article 101(3). As Jones points out, in contrast to the US, Europe has been less willing to shift its treatment of specific categories of bilateral

35 Case C-67/13 P, Groupemont des cartes bancaires (CB) v European Commission, Judgment of 11 September 2014, ECLI:​EU:​C:​2014:​2204. 36 Ibid, paras 56–57. 37 Case C-501/06, GlaxoSmithKline Services Unlimited v Commission [2010] 4 CMLR 2, para 63. 38 Opinion of AG Wahl in Case C-67/13, Groupement des cartes bancaires v European Commission, Opinion delivered on 27 March 2014, ECLI:​EU:​C:​2014:​1958, paras 44–52. 39 See Bailey (n 9) 587. 40 This is where a bidder seeks a credible losing bid from a competitor because they do not wish to win the contract but are required to submit a bid to remain on the procurer’s tender list. Under EU competition law, this is clearly an object arrangement because it involves exchanging information with a competitor about future pricing intentions. However, seeking a cover price is unlikely to be harmful unless there are only two bidders involved. In fact, it is less harmful to competition than the alternative (making an announcement to the world that you do not want to contract), as under cover pricing there will be other bidders who are under the impression that the firm seeking the cover price still wants to win the contract. See Andreas Stephan and Morten Hviid, ‘Cover Pricing and the Overreach of “Object” Liability under Article 101 TFEU’ (2015) 38(4) World Competition 507.

398  Research handbook on methods and models of competition law conduct on the basis of accepted or developing economic principles.41 Article 101 is an area of competition law that has been left behind by the EU’s drive towards greater effects treatment in antitrust. This contrasts with the gradual move of the US from per se to rule of reason in its treatment of many arrangements, reflecting a more cautious approach.42 With the substantive cartel prohibition defined in such wide terms, the focus now turns to the tools of enforcement.

III

THE CALCULATION OF FINES

The stated purpose of fines imposed for breaches of EU competition law is the punishment of the present infringement and the deterrence of future cartels.43 As corporate fines are the only available sanction under EU law, these must be substantial enough to achieve these purposes. They also must be of sufficient magnitude to ensure that the leniency programme (discussed in Section IV) is effective at inducing firms into self-reporting an infringement. Before the European Commission first published guidelines on their calculation, antitrust fines in Europe were criticized for being unpredictable.44 Competition law was one of the few areas where practitioners struggled to advise clients on the likely penalty for given conduct, in part because enforcement cases were fairly infrequent before the introduction of a leniency programme in 1996. To meet this criticism, the European Commission published fining guidelines in 1998 and revised these in 2006.45 The basic amount is calculated with reference to the value of sales according to the gravity and duration of the infringement. This can be up to 30 per cent of the value of sales, but for most cartel arrangements it tends to be around 20 per cent. Gravity takes account of the infringement’s nature, geographic scope and whether it was implemented. Duration involves multiplying the level of value of sales calculated for gravity by the number of years of participation in the agreement. Irrespective of duration, the Commission can also increase the basic amount by between 15 per cent and 20 per cent ‘in order to deter undertakings from even entering into’ cartel arrangements. In addition to the basic amount, the 2006 guidelines allow for up to a 100 per cent increase for recidivism (para 28) and an additional ‘specific increase for deterrence’ (paras 30–31), where there is a ‘need to ensure fines have a sufficiently deterrent effect’. The method for calculating fines in Europe is summarized in Figure 17.1. Despite clearly being focused on deterrence, the basic amount is in fact a series of proxies. Where an activity is treated as per se illegal, no real effects analysis is undertaken – even in calculating the penalty. Instead, the gravity expressed as a proportion of value of relevant sales

41 Alison Jones, ‘Left Behind by Modernisation? Restrictions by Object under Article 101(1)’ (2010) 649 European Competition Journal 657. 42 Most notably in relation to minimum resale price maintenance in Leegin Creative Leather Products Inc v PSKS Inc, 551 US 877 (2007). In White Motor Co v United States, 372 US 261 (1963), the US Supreme Court said it was unwilling to treat arrangements as per se illegal when it knew ‘too little of the actual impact of that restriction’. 43 European Commission, ‘Fines for Breaking EU Competition Law’ (Factsheet, November 2011), http://​ec​.europa​.eu/​competition/​cartels/​overview/​factsheet​_fines​_en​.pdf. 44 See Van Bael (n 6). 45 Guidelines on the method of setting fines pursuant to Article 23(2)(a) of Regulation No 1/2003 [2006] OJ C210/2; Guidelines on the method of setting fines imposed pursuant to Article 15 (2) of Regulation No 17 and Article 65 (5) of the ECSC Treaty [1998] OJ C9/3.

The EU method of antitrust enforcement  399

Figure 17.1

EU 2006 guidelines for the calculation of antitrust fines

and any adjustments for duration provide a rough proxy for cartel harm. It does not involve any real assessment of the extent to which the illegal activity was successfully implemented or had any actual effect on prices or output. As every other aspect of the fining guidelines is an adjustment of the basic amount, it may be suggested that the guidelines simply provide an illusion of scientific rigour.46 In addition, no precise guidance is provided as to how the Commission calculates adjustments for aggravating or mitigating circumstances, or for ‘specific deterrence’ and ‘ability to pay’, which are all identified in the guidelines. Ability to pay is especially controversial, as bankruptcy discounts lack any transparency and may be driven by policy considerations that have little to do with competition – for example, where there is a danger of harm to employment within the EU.47 The absence of any sort of effects evaluation also means that there can be some overly harsh outcomes for firms not fully committed to an anticompetitive arrangement. For example, Hviid and Stephan show that the EU’s approach under the 1998 Guidelines failed to reward firms that had agreed to take part in anticompetitive behaviour but in fact never fully implemented the arrangement, so as to benefit from undercutting the cartel.48 This was despite the fact that the guidelines recognized that fines should be 46 See, for example, James Killick, ‘The 2006 Fining Guidelines: Two Steps Forward but One Step Back?’ (2006) 5(9) Competition Law Insight 5. 47 See Andreas Stephan, ‘The Bankruptcy Wildcard in Cartel Cases’ (2006) Journal of Business Law 511; Andreas Stephan, ‘Price Fixing During a Recession: Implications of an Economic Downturn for Cartels and Enforcement’ (2012) 35(3) World Competition 511. 48 Morten Hviid and Andreas Stephan, ‘The Graphite Electrodes Cartel: Fines That Deter?’ in Bruce Lyons (ed), Cases in European Competition Policy: The Economic Analysis (Cambridge University

400  Research handbook on methods and models of competition law reduced for non-implementation. Recognizing and rewarding cheating or non-implementation with a lower penalty would be deterrence-enhancing. The question that arises, therefore, is the extent to which EU antitrust fines can be said to have a deterrent effect. It is clear from the cartels that have been uncovered in Europe that, while some infringements are inadvertent or committed out of ignorance, the vast majority of hard-core cartel arrangements are deliberate and clandestine. Cartels go to great lengths to conceal their behaviour. This includes staggered price announcements or bids to give the impression of genuine competition; communicating through private email accounts and unregistered mobile phones; using encrypted messages; avoiding the use of documents at meetings or destroying them immediately afterwards; and paying for expenses associated with cartel meetings with cash, or hiding them among other expenses.49 While the calculation of fines in Europe does not involve any kind of economics-based effects methodology, their primary purpose is influenced by the economic theory of optimal deterrence most famously expressed in the work of Gary Becker50 and inspired by the writings of Jeremy Bentham and others.51 At the abstract level, the theory is simple: as anticompetitive behaviour constitutes economic wrongdoing, it is primarily motivated by profit. If, therefore, the combination of sanction and likelihood of detection is substantial enough to make that behaviour unprofitable, firms will desist from engaging in it. Since the introduction of the 1998 Guidelines, the levels of antitrust fines imposed by the European Commission have increased significantly. Fines were at a very low level prior to 1998. They averaged €22 million per undertaking in the period 1998–2005, but increased to around €50 million per undertaking in the period 2006–15. They include instances of very significant fines imposed on individual undertakings, such as €896 million against Saint-Cobain (Car Glass Cartel, 2008) and €725 million against Deutsche Bank (Interest Rate Derivatives, 2013). These final fines would be even higher were it not for the significant discounts provided under the leniency programme and more recently the settlement notice. Indeed, although fines can be a maximum of 10 per cent of worldwide annual turnover,52 between January 2006 and Press 2009), referring to the harsh treatment of Carbide Graphite Group Inc in Case 38.359, Electrical and mechanical carbon and graphite products – Commission Decision of 3 December 2003 [2004] OJ L125. 49 Andreas Stephan, ‘See No Evil: Cartels and the Limits of Antitrust Compliance Programs’ (2010) 31(8) Company Lawyer 231, citing COMP/38.889, Gas Insulated Switchgear [2008] OJ C266/1, 10; COMP/E-1/36.490, Graphite Electrodes [2002] OJ L100/1, 59; COMP/E-1/37.027, Zinc Phosphate [2003] OJ L153/1, 253 as examples. 50 Gary S Becker, ‘Crime and Punishment: An Economic Approach’ (1968) 16(2) Journal of Political Economy 169; see discussion in Wouter PJ Wils, ‘The Commission’s New Method for Calculating Fines in Antitrust Cases’ (1998) 23(3) European Law Review 252. 51 Jeremy Bentham, The Theory of Legislation (Routledge 1931). It is important to note that other theoretical approaches to deterrence exist, but most accept the basic principles of optimal deterrence theory. These include sociological literature such as Michael Wenzel, ‘The Social Side of Sanctions: Personal and Social Norms as Moderators of Deterrence’ (2004) 28(5) Law and Human Behaviour 547; Dan M Kahan, ‘Social Influence, Social Meaning, and Deterrence’ (1997) 83 Virginia Law Review 349; Dennis Chong, ‘Values versus Interests in the Explanation of Social Conflict’ (1996) 144 University of Pennsylvania Law Review 2079. 52 A limit introduced in Regulation 17/62 for the protection of undertakings, now contained in Article 23(2), Regulation 1/2003. See Joined Cases 100–103/80, Musique Diffusion Francaise SA, C Melchers & Co, Pioneer Electronic (Europe) NV and Pioneer High Fidelity (GB) Ltd v EC Commission [1983] 3 CMLR 221, 265.

The EU method of antitrust enforcement  401 December 2015 they averaged around 2.2 per cent of annual worldwide turnover.53 Intuitively, these levels appear to be low if one considers that empirical estimates of average overcharges achieved by cartels range from 16 to 60 per cent.54 In addition, this does not allow for the fact that cartels average around seven years in duration and that the rate of detection may be as low as 13 per cent of cartels in existence.55 Some of these empirical estimates suggest that EU cartel fines are unlikely to outweigh the potential profits earned by cartels – assuming that those arrangements were successfully implemented. Indeed, the level of fines needed to truly ensure that anticompetitive conduct was ‘unprofitable’ would likely bankrupt many firms.56 The Commission is silent about the sort of deterrence it aims to achieve through fines. In commenting on the purpose of the guidelines, Commission officials and publications have spoken of transparency, not predictability. Indeed, one Commissioner for Competition noted: ‘I cannot see how allowing potential infringers to calculate the likely cost/benefit ratio of a cartel in advance will somehow contribute to a sustained policy of deterrence and zero tolerance.’57 In reality, despite the lack of scientific rigour, the enforcement practice of the European Commission has created some certainty and predictability about the level of fines imposed in cartel cases. Fines in abuse of dominance cases under Article 102 remain far less predictable, in part due to the fact that they are infrequent but also because dominance covers a wider variety of practices.58 So then, why is there resistance to making fines more predictable through more detailed guidance? One explanation might be that there is value in some unpre-

53 This figure includes immunity applicants whose fine was zero, but who nevertheless were involved in the cartel and were subject to an infringement decision. The data is taken from statistics published on the website of the European Commission, http://​ec​.europa​.eu/​competition/​cartels/​statistics/​statistics​.pdf. 54 John M Connor and Robert H Lande, ‘Cartel Overcharges and Optimal Cartel Fines’ in SW Waller (ed), Issues in Competition Law and Policy (Vol 3, AMA Section of Antitrust Law 2008) ch 88; John M Connor, ‘Price Fixing Overcharges (rev 3rd edn, unpublished manuscript, February 2014), http://​ papers​.ssrn​.com/​sol3/​papers​.cfm​?abstract​_id​=​2400780; Florian Smuda, ‘Cartel Overcharges and the Deterrent Effect of EU Competition Law’ (ZEW Discussion Paper No 12-050, July 2012), https://​papers​ .ssrn​.com/​sol3/​papers​.cfm​?abstract​_id​=​2118566; Yuliya Bolotova, ‘Cartel Overcharges: An Empirical Analysis’ (unpublished manuscript, 2006), http://​papers​.ssrn​.com/​sol3/​papers​.cfm​?abstract​_id​=​931211; Yuliya Bolotova, John M Connor and Douglas J Miller, ‘Factors Influencing the Magnitude of Cartel Overcharges: An Empirical Analysis of the US Market’ (2009) 5(2) Journal of Competition Law and Economics 361; M Boyer and R Kotchoni, ‘How Much Do Cartels Overcharge?’ (2015) 47(2) Review of Industrial Organization 119; OECD, Hard Core Cartels (2000), http://​www​.oecd​.org/​competition/​ cartels/​2752129​.pdf. 55 There are a number of studies that concur on estimates in this region: Peter G Bryant and Edwin Eckard, ‘Price Fixing: The Probability of Getting Caught’ (1991) 73(3) Review of Economics and Statistics 429; Emmanuel Combe and Constance Monnier, ‘Fines against Hard Core Cartels in Europe: The Myth of Over Enforcement’ (2011) 56 Antitrust Bulletin 235; Peter L Ormosi, ‘A Tip of the Iceberg? The Probability of Catching Cartels’ (2014) 29(4) Journal of Applied Econometrics 549. This is not an exhaustive list. 56 See, in particular, Wouter PJ Wils, ‘Optimal Antitrust Fines: Theory and Practice’ (2006) 29(2) World Competition 183. 57 Neelie Kroes, ‘The First Hundred Days’ (Speech delivered to the International Forum on Competition Law, Brussels, April 2005). 58 This is illustrated by the Commission’s investigations of Google, which have resulted in fines far greater than those imposed on individual undertakings for cartel conduct. See European Commission, ‘Antitrust: Commission Fines Google €4.34 Billion for Illegal Practices Regarding Android Mobile Devices to Strengthen Dominance of Google’s Search Engine’ (Media Release, 18 July 2018); European

402  Research handbook on methods and models of competition law dictability, given that truly deterrent fines (under optimal deterrence theory) are in practice impossible to impose. Another possible explanation is that the Commission takes a broad approach to deterrence, relying on carrot as well as stick. Yet the European Commission does not reward good compliance when calculating fines and only engages in limited advocacy activity.59 Its position has for some years been that ‘it is not appropriate to take the existence of a compliance program into account as an attenuating circumstance for a cartel infringement, whether committed before or after the introduction of such a program’.60 Indeed, it was only in 2011 that they clarified that the existence of a compliance programme would not be taken as an aggravating factor when calculating fines.61 The rationale for this conservative approach is that any infringement of the law constitutes failed compliance, and that rewarding ‘unsuccessful’ compliance programmes would be detrimental to enforcement and deterrence. Stiff sanctions, it is argued, should be incentive enough for a firm to keep its house in order.62 There are also concerns that sham compliance programmes could simply be used as a way of reducing the size of the fine after detection. By contrast, the UK’s Competition and Markets Authority (as well as its predecessor, the Office of Fair Trading) provides extensive compliance materials for businesses and is willing to provide a discount of up to 10 per cent where appropriate compliance efforts are taken – especially where these are after the infringement has been discovered.63 So, on the one hand, there is a strong possibility that EU cartel fines do not exceed cartel profits (at least in cases of successfully implemented cartels), yet EU cartel enforcement relies almost entirely on the use of corporate fines to punish and deter and to motivate compliance. While there are strong arguments for and against providing a discount for compliance programmes, the extent to which EU enforcement is deterrence-enhancing may be subject to some debate. Apart from the possible shortfall in levels of fines, there is the issue of punishing those responsible. Where the cartelization of a market is largely down to corporate decision-making, long-standing cultures of anticompetitive behaviour or poor compliance, then it may be appropriate that only the corporation be punished. However, many cartels are organized by small groups of employees who make efforts to hide their activities from their employer, as well as from customers and authorities. Those individuals may very well have left the firm by

Commission, ‘Antitrust Commission Fines Google €1.49 Billion for Abusive Practices in Online Advertising’ (Media Release, 20 March 2019). 59 See Stephan ‘See No Evil: Cartels and the Limits of Antitrust Compliance Programs’ (n 49). Also contrast the European Commission with the UK’s Competition and Markets Authority (CMA), which does far more engaging with businesses to promote compliance and provides a suite of materials to assist them: CMA, ‘Competition Law Guidance’, https://​www​.gov​.uk/​government/​collections/​competition​ -and​-consumer​-law​-compliance​-guidance​-for​-businesses. 60 COMP/E-23/38.359, Electrical and mechanical carbon and graphite products [2004] OJ L125/45, 3131. 61 European Commission, Compliance Matters: What Companies Can Do Better to Respect EU Competition Rules (November 2011) 21. 62 For arguments against awarding discounts for compliance, see a paper written by the European Commission’s Hearing Officer: Wouter PJ Wils, ‘Antitrust Compliance Programmes and Optimal Antitrust Enforcement’ (2013) 1(1) Journal of Antitrust Enforcement 52. For opposing arguments, see Damien Geradin, ‘Antitrust Compliance Programmes and Optimal Antitrust Enforcement: A Reply to Wouter P.J. Wils’ (2013) 1(2) Journal of Antitrust Enforcement 325. 63 See the online resources available on the CMA’s ‘Competition Law Compliance: Guidance for Businesses’ webpage (n 59).

The EU method of antitrust enforcement  403 the time any corporate fine is imposed. Many national competition authorities in Europe have responded to this by imposing sanctions on individuals in addition to the firm (discussed in Section V). In most jurisdictions these amount to civil fines, but in some they can also take the form of criminal prosecution. The European Commission imposes no such sanctions on individuals.64 In a 2011 interview with World Competition, the Competition Commissioner, Joaquín Almunia, was asked whether he would consider complementary administrative measures, such as director disqualification orders or personal fines, to ensure that the incentives of companies and individuals were aligned. He responded by saying: ‘As laid down in the Treaty and Regulation 1/2003, the EU antitrust enforcement system provides for pecuniary sanctions on undertakings only. … While a number of EU Member States have introduced sanctions on individuals for competition infringements, the enforcement against undertakings remains the core principle at EU level.’65 This may suggest that there is neither the desire nor the means for the adoption of sanctions against individuals at the EU level. It is interesting to note that the Commission’s moves to improve the availability of damages within the tort regimes of Member States began with a deterrence objective but ended up focusing only on compensation.66 The extent to which cartel fines deter is also important to the effectiveness of the European Commission’s leniency programme. If leniency is to be successful at inducing the self-reporting of cartel infringements, the difference between immunity and the fine otherwise faced must be significant.

IV

THE EUROPEAN LENIENCY NOTICE

Leniency programmes have become the defining characteristic of antitrust enforcement. In contrast to regulatory controls on corporate behaviour, which are normally designed around some convenient control point, cartels can occur in many different parts of the economy and go to great lengths to avoid detection. As the OECD notes, ‘the challenge in attacking hard-core cartels is to penetrate their cloak of secrecy’.67 Leniency programmes are also justified on the grounds that competition authority investigations – even where guided by complaints from customers – can be very resource intensive. In offering immunity from fines to the first cartel member to reveal an infringement, leniency programmes induce cartel members into self-reporting and even providing the evidence necessary for the authority to complete its investigations and establish liability. The term ‘leniency’ has a wider meaning in the EU than it does in the US. Under US antitrust law, leniency refers only to the non-prosecution of the first firm to come forward and its employees. Subsequent cartel members can only be rewarded in return for their cooperation through reaching a plea bargain with the US Department of Justice.

64 Wouter PJ Wils, ‘Is Criminalization of Competition Law the Answer?’ (2005) World Competition 117, 151. 65 ‘José Rivas’ Interview with Commissioner Almunia’ (2011) 34 World Competition 1. 66 See Directive 2014/104/EU (n 4). For a discussion, see Andreas Stephan, ‘Does the EU’s Drive for Private Enforcement of Competition Law Have a Coherent Approach?’ (2018) 37(1) University of Queensland Law Journal 153. 67 OECD, ‘Fighting Hard Core Cartels: Harm, Effective Sanctions and Leniency Programmes’ (2002).

404  Research handbook on methods and models of competition law By contrast, EU leniency offers immunity to the first firm to come forward and fine discounts of up to 50 per cent to subsequent firms. The EU first introduced a leniency programme in 1996 and made important changes in 2002 and 2006.68 These changes were mainly aimed at making the process of applying for leniency more certain for prospective applicants. For example, the original 1996 notice stated that immunity would not be available after an investigation into an industry had been opened. As businesses typically had no way of determining if this was the case prior to a dawn raid, it may have had a chilling effect on some leniency applications. The 1996 notice also required that the immunity applicant provide ‘decisive evidence of a cartel’s existence’, which was thought to set the bar too high. This was changed to ‘information and evidence’ that would allow the Commission to ‘carry out targeted inspections’. Prior to 2006, applicants were also required to cease involvement in the cartel at the point at which they applied for leniency. This ran the obvious risk of the other cartel members suspecting that dawn raids were imminent (for example, where the reporting firm stopped attending cartel meetings or responding to communications), with the consequence that evidence might be destroyed. Considerable uncertainty was also created by the fact that the level of leniency discount would not be confirmed until the final decision of the authority. This was changed to a marker system that protects the applicant’s place in the leniency queue. Their immunity or fine discount is confirmed in writing once it has been determined that the evidence submitted meets the conditions of the leniency notice. Efforts to ease the process of applying for leniency have continued with a system of ‘eLeniency’, launched in March 2019. This allows firms to apply for leniency online without the need to travel to the Commission’s premises.69 Under paragraph 9 of the 2006 leniency notice, the ‘information and evidence’ that must be provided in return for immunity must include a detailed description of the alleged cartel arrangement, the products or services it effected, its geographical scope, and information on the dates, locations and content of meetings. In addition, the party must identify the individuals known to have been involved in the cartel and whether other competition authorities (including the national authorities of EU Member States) have also been approached. An immunity application may still fail, however, if at the time of its submission the Commission already has enough evidence to deliver a finding of guilt. Paragraph 12 contains additional conditions. The applicant must cooperate fully and on a continuous basis. It must provide the Commission with prompt replies to requests for relevant information and evidence relating to the alleged infringement; not destroy, conceal or tamper with evidence; not alert the other cartel members; and ‘make current (and, if possible, former) employees and directors available for interviews with the Commission’. This requirement of cooperation of employees warrants some consideration. It was noted above that the European Commission’s approach to cartel enforcement is to impose significant fines that not only punish and deter, but also encourage future compliance with the law. Therefore, where a business is subject to a competition law infringement decision because of the actions of a small number of employees, it is right that the business will want to discipline

68 Commission Notice on the non-imposition or reduction of fines in cartel cases [1996] OJ C207; Commission Notice on immunity from fines and reduction of fines in cartel cases [2002] OJ C45/3; Commission Notice on immunity from fines and reduction of fines in cartel cases [2006] OJ C298/17. 69 European Commission, ‘Antitrust: Commission Launches New Online Tool for Cartel Leniency and Settlements and Non-Cartel Cooperation’ (Media Release, 19 March 2019).

The EU method of antitrust enforcement  405 and possibly dismiss those responsible. The problem is that the best ‘information and evidence’ on the cartel infringement – needed by the business to successfully apply for leniency – will be held by those same individuals responsible. This is because cartels are not typically organized within the institutional framework of the firm and may be actively hidden from others within the organization. As Whelan observes, leniency programmes incentivize cartelists to preserve evidence of cartel activity because such evidence has an ‘economic and strategic value’, in the knowledge that an application for leniency may one day be necessary.70 As a consequence of this contradiction between the need to discipline an employee and the value of their cooperation, firms may find themselves having to soften their demands for disciplinary action, or possibly even let them fall by the wayside, in the interest of ensuring that the business is in the best possible position to apply for leniency. It should also be noted that the best evidence and information is likely to be held by the ringleader. Although the leniency programme excludes those who have coerced others, immunity has never been refused in Europe on that basis. This tension between the undertaking and its delinquent employees further highlights the need for individual sanctions as a complement to corporate fines. The availability of individual sanctions plays two important roles. First, it ensures that individuals face the threat of punishment for their actions, regardless of the conflict of interests identified above. Second, personal consequences for cartel conduct heighten the ‘race’ to the competition authority, as the immunity enjoyed by the first revealing firm usually extends to its employees. Under Part III of the European Leniency Notice, undertakings that miss out on the immunity prize can still benefit from a reduction in any fine imposed, so long as their cooperation represents significant ‘added value’ to the Commission’s investigation. This cooperation is of great importance to the authority. It allows them to corroborate the accuracy and completeness of the information provided by an immunity applicant and other cooperating firms. It also reduces the scope for abuse of the leniency programme by firms that induce a breach of the law only to come forward seeking penalties imposed on their competitors. Although such a scenario may seem far-fetched, cartel appeals to the CJEU have revealed allegations of whistle-blowing parties exaggerating evidence that was relied upon by the Commission. Examples of this include the Cartonboard and Pre-Insulated Pipes cartels, where it was alleged that a cooperating firm exaggerated the participation of other firms.71 In Copper Fittings, the fine of one infringing firm was increased by 50 per cent because it initially provided information through the leniency notice, which later proved to be inaccurate.72 The leniency programme identifies ranges of fine discounts available to subsequent firms that come forward (para 26): ●● first undertaking to provide significant added value: a reduction of 30–50 per cent; ●● second undertaking to provide significant added value: a reduction of 20–30 per cent; and

70 Peter Whelan, The Criminalization of European Cartel Enforcement: Theoretical, Legal, and Practical Challenges (Oxford University Press 2014) 133. 71 Cartonboard [1998] ECT II-2099; Pre-Insulated Pipes 2002/C 202/02: appeal brought on 21 May 2002 by Dansk Rørindustri A/S against the judgment delivered on 20 March 2002 by the Court of First Instance of the European Communities (Fourth Chamber) in Case T-21/99, Case C-189/02 P, Dansk Rørindustri A/S v Commission of the European Communities. This issue is discussed in Andreas Stephan, ‘The Direct Settlement of EC Cartel Cases’ (2009) 58(3) International and Comparative Law Quarterly 627. 72 European Commission (Media Release, IP/06/1222, 20 September 2006).

406  Research handbook on methods and models of competition law ●● subsequent undertakings that provide significant added value: a reduction of up to 20 per cent. These quite generous fine discounts (on top of the immunity prize to the first revealing firm) help us understand why the actual fines imposed by the European Commission average around 2.2 per cent of annual turnover, when they can be up to 10 per cent. Although leniency was in part introduced to free up resources that could be employed in ‘own-initiative’ investigations, the proportion of cases uncovered through leniency has steadily increased since its introduction in 1996 and in recent years has stood at around 75 per cent.73 So, in around three-quarters of cases (some of which involve as few as three or four firms), a party is receiving immunity. As Stephan and Nikpay point out, this may amount to an ‘over-reliance’ on leniency as a cartel detection tool, raising the possibility that there are not enough ‘own-initiative’ cases to maintain a credible threat of detection.74 The discounts for ‘significant added value’ are available in all cases – even those where the Commission has discovered a cartel through its own investigations and parties seek to cooperate following dawn raids. The availability of leniency in these cases is right, but they have the effect of significantly reducing the penalties imposed on cartelists, raising the further question of whether there is a sufficient and credible threat of punishment. Taken as a whole, the European Commission’s leniency programme has the effect of significantly reducing punishment, with implications for deterrence. Since 2008, these discounts have been further increased by the introduction of a settlement procedure, seen as a natural corollary of the leniency programme.75 Once the Commission has commenced proceedings in a particular case, it can invite parties to declare in writing whether they wish to engage in settlement discussions. These discussions lead to a common understanding of the ‘scope of the potential objections and the estimation of the range of likely fines to be imposed by the Commission’. The firms are then given 15 working days to make a settlement submission that largely reflects what was agreed during the course of the settlement discussions. Essentially, the settlement notice facilitates a streamlined procedure in which the firms are given early access to the case file. The process is meant to have the effect of reducing both challenges to the Statement of Objections and the propensity for firms to appeal the Commission’s final decision to the CJEU. It essentially reduces the length of cartel investigations by at least 12 months.76 In return, firms receive a fixed discount on the fine of 10 per cent (para 32). This is applied after the 10 per cent annual worldwide turnover cap and in addition to any leniency discounts. The settlement notice is therefore different from US plea bargaining, where parties can negotiate their penalty and enter a negotiated settlement in lieu of a criminal trial and waiving

73 This is based on the database of EU cartel decisions maintained by the author. It is also acknowledged in a report prepared by the European Commission: see European Commission, ‘Ten Years of Antitrust Enforcement under Regulation 1/2003’ (Commission Staff Working Document, SWD(2014) 230/2, 2014) para 16, http://​ec​.europa​.eu/​competition/​antitrust/​legislation/​swd​_2014​_230​_en​.pdf. 74 Andreas Stephan and Ali Nikpay, ‘Leniency Decision-Making from a Corporate Perspective: Complex Realities’ in Caron Beaton-Wells and Christopher Tran (eds), Anti-Cartel Enforcement in a Contemporary Age: The Leniency Religion (Hart Publishing 2015) 149. 75 Commission Notice on the conduct of settlement procedures in view of the adoption of Decisions pursuant to Article 7 and Article 23 of Council Regulation (EC) No 1/2003 in cartel cases, OJ C167, 2.7.2008, 1–6. 76 See Stephan, ‘The Direct Settlement of EC Cartel Cases’ (n 71) 634–5.

The EU method of antitrust enforcement  407 Table 17.1

Car and truck bearings cartel Reduction under the leniency notice

Fine (€)

Approximate fine absent

JTEKT

100%

0

86 037 000

NSK

40% (+10% settlement discount)

62 406 000

124 812 000

NFC

30% (+10% settlement discount)

3 956 000

6 593 333

SKF

20% (+10% settlement discount)

315 109 000

450 155 714

Schaeffler

20% (+10% settlement discount)

370 481 000

529 258 571

NTN

0% (+10% settlement discount)

201 354 000

223 726 667

953 306 000

1 420 583 285 (+467 277 285)

leniency and settlement (€)

Total

Source: European Commission, ‘Antitrust: Commission Fines Producers of Car and Truck Bearings €953 million in Cartel Settlement’ (Media Release, IP/14/280, 19 March 2014).

their rights to appeal.77 Indeed, when the settlement notice was introduced, the Commission was keen to stress that it ‘neither negotiates nor bargains the use of evidence or the appropriate sanction’78 to ensure that settlements were not seen as a ‘bazaar-like’ process.79 Despite this, the settlement discussions may have some bearing on factors such as the range of fine, the severity of the infringement, and its duration. So, in practice, the settlement notice may very well have the effect of providing a more generous discount than the headline 10 per cent figure.80 To show the extent to which leniency and settlement have exerted a downward pressure on fines, one can look at recent cartel cases completed by the European Commission. Between 2010 and 2015, the Commission fined 30 cartels. Of these, immunity was awarded to the revealing firm in 25 cases, a 30–50 per cent discount was awarded in 23 cases, and a discount of under 30 per cent in 19 cases.81 Table 17.1 takes a typical example of a case where the full range of leniency has been applied (Car and Truck Bearings, 2014). In that case, the combination of leniency and settlement discounts has reduced the level of corporate fines imposed by a third, but the reductions can be far more generous. Table 17.2 gives the same break-down for the recent Car Parts (2016) case, involving just three cartel members. In that case, where leniency was particularly generous, the penalty imposed on the cartel was reduced by two-thirds. These two examples demonstrate how the fewer the cartel members, the more generous leniency will be. This is ironic, as cartels with low numbers will, in principle, be easier to coordinate because there are fewer members who have to make decisions and so these may actually be the more successful cartel arrangements.82 The average decrease Ibid, 651. European Commission, ‘Antitrust: Commission Introduces Settlement Procedure for Cartels’ IP/08/1056 (Media Release, 30 June 2008). 79 OECD, ‘Plea Bargaining/Settlement of Cartel Cases’ (Directorate for Financial and Enterprise Affairs, Competition Committee. Working Party No 3 on Co-operation and Enforcement, DAF/COMP/ WP3(2006)3, 2003) 6. 80 For a discussion of the settlement discussions, see Flavio Laina and Elina Laurinen, ‘The EU Cartel Settlement Procedure: Current Status and Challenges’ (2013) 4(4) Journal of European Competition Law & Practice 302. 81 This is based on the database of EU cartel decisions maintained by the author. 82 See, for example, Iwan Bos and Joseph E Harrington, ‘Competition Policy and Cartel Size’ (2015) 56(1) International Economic Review 133. 77 78

408  Research handbook on methods and models of competition law Table 17.2

Car parts cartel Reduction under the leniency notice

Fine (€)

Approximate fine absent

Denso

100%

0

157 000 000

Hitachi

30% (+10% Settlement Discount)

26 860 000

44 766 666

Melco

28% (+10% Settlement Discount)

110 929 000

178 917 741

137 789 000

380 684 407 (+242 895 407)

leniency and settlement (€)

Total

Source: European Commission, ‘Antitrust: Commission Fines Car Parts Producers €137 789 000 in Cartel Settlement’ (Media Release, IP/16/173, 27 January 2016).

in penalty across all leniency cases (including settlement) stands at around 40 per cent, but this nevertheless constitutes a very significant cut in the penalty in return for cooperation. Generous discounts are given to the second and third firms to come forward and cooperate to avoid the danger of a chilling effect where parties are unsure whether the immunity prize is still available. A possible 50 per cent discount for coming second makes coming forward more likely. While cartel fines still far exceed financial penalties imposed in almost any other area of corporate wrongdoing, continued research is needed to determine the extent to which they are deterring prospective cartel arrangements. In addition to the possible failure of fines to exceed the profits earned by successful cartels, and the increasing downward pressures of generous leniency and settlement discounts, there are other sources of concern. Around a third of cartels uncovered in Europe formed after the introduction of leniency in 1996 and the European Commission continues to receive frequent leniency applications, suggesting that many potential infringements still are not being effectively deterred. In addition, empirical studies suggest that many cartels fail before the first firm steps forward to report them.83 This may indicate that leniency is being used strategically where cartels fall apart, rather than inducing active and successful cartels to break up. Just as with the uncertainties surrounding deterrent fines, the incentive to come forward and apply for leniency could be heightened by the additional use of individual sanctions. Leniency to the first revealing firm would extend to the employees of the revealing firm, giving them a strong incentive to come forward. It has already been explained how the European Commission does not wish to adopt additional sanctions against individuals. However, most EU Member States do have individual sanctions in their domestic antitrust regimes. The question is therefore whether these sanctions under national law can be used to complement enforcement on the EU level.

V

THE RELATIONSHIP BETWEEN EU LAW AND NATIONAL LAW

The relationship between EU and national competition law is one of continued evolution. For many years, the European Commission applied EU rules broadly, retaining the exclusive power to grant an Article 101(3) exemption, in part because competition law was underdeveloped at 83 Andreas Stephan, ‘An Empirical Assessment of the European Leniency Notice’ (2009) 5(3) Journal of Competition Law & Economics 537; Stephan and Nikpay (n 74).

The EU method of antitrust enforcement  409 the Member State level. For example, the UK did not fully bring its antitrust laws into line with those of the EU until the Competition Act 1998, and other Member States did not adopt leniency programmes until more recently. The introduction of Regulation 1/2003 and the creation of the ECN has encouraged harmonization and consistency of competition rules across Europe. However, it has not prevented national competition law regimes from diverging in their treatment of individuals responsible for cartel conduct. Few have followed the European Commission’s example of sticking doggedly to a corporate fines-only approach, but they have each instead taken diverging approaches to the issue. The UK, Ireland, France, Greece, Denmark, Slovenia and Romania have gone as far as to adopt criminal offences. Germany, Austria, Italy, Poland and Hungary apply criminal sanctions to bid-rigging only.84 With the exception of bid-rigging cases in Germany, the level of enforcement using these sanctions against individuals is very low.85 In addition to criminal sanctions, director disqualification orders can be applied in the UK, Lithuania and Sweden.86 Finally, individuals can be fined under a civil process in Germany, Lithuania, Portugal, Poland, Spain and the Netherlands.87 The relationship between EU and national competition law is governed by Regulation 1/2003. The key provision is Article 3, which states that the domestic equivalent of Article 101 of the TFEU cannot be applied more strictly than the EU provision, thereby ensuring consistency in their application. Article 3 also creates an obligation for Member States to apply Article 101 alongside national law where they are dealing with an arrangement that may affect trade between Member States. Indeed, unless they are dealing with a very local infringement, national competition authority decisions typically cite both in their infringement decisions. In addition, Article 11(6) of the Regulation gives the Commission exclusive competency where it begins proceedings against firms, even if a competition authority of a Member State is already acting on the case. Implicitly, the obligations created by Articles 3 and 11 mean that ‘national competition law’ cannot be applied where the Commission begins proceedings in a case. The only reference made to criminal sanctions in the Regulation is in recital 8 of the preamble, which states: ‘this Regulation does not apply to national laws which impose criminal sanctions on natural persons’. In addition, Article 3(3) states that the Regulation does not ‘preclude the application of provisions of national law that predominantly pursue an objective different from that pursued by Article [101]’. However, no mechanism exists whereby a European Commission investigation can be held up by criminal or civil proceedings on the national level. The only example of parallel EU and national proceedings is in the case of Marine Hoses. Three UK

84 See Norton Rose Fulbright, ‘The Criminal Cartel Offence around the World’, Competition World (Quarter 2, 2016), https://​www​.nortonrosefulbright​.com/​en​-gb/​knowledge/​publications/​10864243/​ competition ​ - world ​ - a ​ - global ​ - survey ​ - of ​ - recent ​ - competition​ - and​ - antitrust​ - law​ - developments​ - with​ -practica. The UK ceased to be a member of the European Union on 31 January 2020. 85 On Germany, see Florian Wagner-von Papp, ‘What If All Bid Riggers Went to Prison and Nobody Noticed? Criminal Antitrust Law Enforcement in Germany’ in Caron Beaton-Wells and Ariel Ezrachi (eds), Criminalising Cartels: Unexplored Dimensions and Unforeseeable Consequences (Hart 2011). 86 The UK’s first ‘Competition Disqualification Order’ was imposed on 1 December 2016 against Daniel Aston, with a duration of five years. See UK Government, ‘CMA Secures Director Disqualification for Competition Law Breach’ (Press Release, 1 December 2006). 87 See Florian Wagner-von Papp, ‘Compliance and Individual Sanctions in the Enforcement of Competition Law’ in Johannes Paha (ed), Competition Law Compliance Programmes (Springer 2016) fn 20.

410  Research handbook on methods and models of competition law nationals involved in the cartel were arrested in the US and entered into a plea agreement with the US Department of Justice in which they agreed to return to the UK and plead guilty to the cartel offence.88 The European Commission was carrying out a parallel investigation at the time, which resulted in a fine of €131 million on the undertakings concerned.89 Ireland and the UK have dealt with this enforcement framework in very different ways. Ireland decided essentially to criminalize Article 101 of the TFEU in its domestic law, bringing it firmly within the framework of Regulation 1/2003 and meaning that it is unable to complement cases investigated by the European Commission, even if they involve individuals responsible for cartel conduct who are based in Ireland or who are Irish nationals.90 By contrast, the UK has a less conventional arrangement. Cartel conduct by undertakings is unlawful under the civil prohibition in Chapter I of the Competition Act 1998 (the domestic equivalent of Article 101), while a separate criminal offence against individuals exists under section 188 of the Enterprise Act 2002. The UK’s competition regime was deliberately designed in this way to keep the cartel offence distinct from Article 101 of the TFEU.91 The UK government and its competition authority have always maintained that the dual civil/criminal regime was enough for the cartel offence to circumvent the obligations created by Regulation 1/2003. This was also tested in a pre-trial challenge heard by the English Court of Appeal in IB v The Queen.92 In agreeing with the authority, the Court took a narrow view that ‘national competition law’ referred to where the relevant objective was applying Article 101, not preventing anticompetitive practices more generally.93 However, this view is highly questionable, given that Article 3(3) of the Regulation refers to provisions that ‘predominantly pursue an objective different from’ that of Article 101, not the application of Article 101 itself. The objective of Article 101 is unquestionably punishment and deterrence. If one looks at the White Paper that preceded the UK cartel offence, the objective there also clearly appears to be punishment and deterrence.94 Even if it is possible for the UK or other Member States to apply their criminal offence alongside the European Commission’s investigations, there are practical challenges. Although Commission investigations are undertaken to a very high standard of evidentiary proof and safeguards, any ongoing parallel civil investigation by the Commission risks undermining criminal proceedings within a Member State. In addition, while national competition authorities can access confidential information via the ECN, under Article 12(3) of Regulation 1/2003 this information ‘cannot be used by the receiving authority to impose custodial sanctions’. This does not prevent the authority from using that information to guide its own investigations,

88 Office of Fair Trading (OFT), ‘Three Imprisoned in First OFT Criminal Prosecution for Bid-Rigging’ (Media Release, 11 June 2008); OFT, ‘OFT Brings Criminal Charges in International Bid Rigging, Price Fixing and Market Allocation Cartel’ (Media Release, 19 December 2007). 89 Commission Decision of 28 January 2009 relating to a proceeding under Article 81 of the Treaty and Article 53 of the EEA Agreement, Case COMP/39406, OJ L1, 4.1.2003, 1. 90 See Mark Furse, The Criminal Law of Competition in the UK and the US (Edward Elgar Publishing 2012) ch 5. 91 Department of Trade and Industry (DTI), A World Class Competition Regime (White Paper, Cm 5233, 2001) 7.30–7.31. 92 [2009] EWCA Crim 2575. 93 Ibid, para 35. 94 DTI White Paper (n 91), 7.13–7.18.

The EU method of antitrust enforcement  411 but those investigations are unlikely to enjoy any assistance from the authorities of other Member States.95 In principle, civil sanctions against individuals such as fines and director disqualification orders may more easily complement enforcement action by the European Commission. For example, the UK’s ‘Competition Disqualification Orders’ can be sought on the basis of an infringement decision by the European Commission where it relates to arrangements that had an actual or potential impact on the UK.96 These civil sanctions typically have a low standard of proof and can either be imposed directly by the competition authority or, as in the case of the UK’s disqualifications, with the simple approval of a court where it can be shown that the individual ought to have known about the breach of competition law. Even when it comes to civil sanctions against individuals, four key obstacles exist to employing those sanctions as a complement to enforcement action by the European Commission. The first is that the Commission’s investigations do not typically identify the role played by individuals and do not usually name them in their final infringement decisions. So, for example, the three UK nationals imprisoned in Marine Hoses were not named in the Commission’s final infringement decision.97 The second is that it is hard to see how a national competition authority could begin proceedings against individuals for a civil breach of competition rules without impinging on the European Commission’s exclusive right to investigate a case under Article 11(6), Regulation 1/2003. While the Regulation states in the preamble that it ‘does not apply to national laws which impose criminal sanctions on natural persons’, it makes no reference to the application of non-criminal sanctions. The third relates to director disqualifications. Even if it were possible to impose these on the back of a Commission decision, there is currently no EU-wide system for disqualification, meaning that the sanction would only apply within the jurisdiction of the Member State imposing it.98 However, the fourth and most serious obstacle relates to leniency. In 2005, Competition Commissioner Neelie Kroes called for a ‘one-stop-shop’ for leniency to enhance certainty.99 This would mean immunity being granted to the first firm to self-report to any of the EU’s competition agencies. More than a decade on, this still has not been achieved and the CJEU has ruled that even the workings of the Commission’s leniency notice are not binding on Member States.100 The ECN published a ‘model leniency programme’ in 2006 and 2012, which has encouraged some harmonization between leniency programmes. But, as the 2012 document explains, ‘an application for leniency to one authority is not to be considered as an application

95 See the discussion in Michael O’Kane, The Law of Criminal Cartels (Oxford University Press 2009) 7.04–7.07. 96 Competition and Markets Authority, ‘Director Disqualification Orders in Competition Cases’ (2010) 4.6–4.8, https://​www​.gov​.uk/​government/​uploads/​system/​uploads/​attachment​_data/​file/​324978/​ oft510​.pdf. Section 204 of the Enterprise Act 2002 amended the Company Directors Disqualification Act 1986, inserting new ss 9A–9E and creating the ‘Competition Disqualification Order’ with a maximum period of 15 years. 97 Commission Decision of 28 January 2009 (n 89). 98 See Andreas Stephan, ‘Disqualification Orders for Directors Involved in Cartels’ (2011) 2(6) Journal of European Competition Law & Practice 529, 535. 99 Neelie Kroes, ‘The First Hundred Days: 40th Anniversary of the Studienvereinigung Kartellrecht 1965–2005, International Forum on European Competition Law’ (Speech, 7 April 2005). 100 See, for example, Case C-360/09, Pfleiderer AG v Bundeskartellamt, Judgment of 14 June 2011, ECLI:​EU:​C:​2011:​389.

412  Research handbook on methods and models of competition law for leniency to another authority’.101 This means that where there is any uncertainty as to whether an alleged infringement will be dealt with by the European Commission or national competition authorities, the parties in question must file multiple leniency applications to all the relevant authorities. It also means that a guarantee of immunity from the European Commission does not necessarily extend to individual sanctions imposed by Member States. The greater use of such sanctions therefore risks undermining the EU leniency programme. The Commission does currently coordinate its leniency efforts with UK authorities. It ensures that where immunity is granted to the firm it extends to the protection of employees from criminal prosecution in the UK.102 Beyond this individual arrangement, there does not appear to be any Community-wide coordination of leniency programmes on this issue.

VI CONCLUSION This chapter has provided a critique of the EU’s method of antitrust enforcement, focusing on cartels and paying attention to its central aim of punishing and deterring the most damaging infringements of competition law. It began by explaining the design of the substantive cartel prohibition contained in Article 101 of the TFEU. This is deterrence-enhancing in that it creates a clear presumption of liability in relation to the most harmful categories of bilateral behaviour. Further, it punishes agreements, concerted practices and attempts to cartelize, regardless of whether they were implemented or were successful at restricting competition. This makes it very hard for firms to circumvent the prohibition by entering into looser forms of collusive behaviour. However, the boundaries of these restrictions ‘by object’ are unclear and it has been alleged that this presumptive prohibition is being applied to conduct that should be subject to a full ‘effects’ analysis. The wide scope of ‘by object’ also raises the danger of capturing novel categories of behaviour that are benign, but that do not generate the sorts of efficiency defences that would benefit from an Article 101(3) exception. Once liability has been established, the European Commission calculates fines according to their published guidelines. The extent to which these make the process of calculating fines transparent is questionable – especially in relation to adjustments made for aggravating and attenuating circumstances. Although cartel fines are substantial, it is unclear whether they are actually deterring prospective cartel infringements – especially as they may not exceed the potential illegal profits earned by successful cartels. Fines that exceed illegal profits may not be feasible in practice, but it is hard to discern what type of deterrence the Commission is pursuing if it is not one based on economic deterrence theory. For example, the Commission does not reward compliance initiatives undertaken by the firms (whether before or after they are detected) because it believes that corporate fines are deterrent enough to incentivize good compliance. Neither does it engage in much advocacy work in comparison to national regulators, such as the UK’s Competition and Markets Authority.

101 ECN Model Leniency Programme (November 2012), http://​ec​.europa​.eu/​competition/​ecn/​mlp​ _revised​_2012​_en​.pdf. 102 European Commission, Ten Years of Antitrust Enforcement under Regulation 1/2003: Achievements and Future Perspectives’ (9 July 2014) para 219, http://​ec​.europa​.eu/​competition/​antitrust/​legislation/​ antitrust​_enforcement​_10​_years​_en​.pdf.

The EU method of antitrust enforcement  413 Penalties are subject to significant reductions under the European Leniency notice, which constitutes the EU’s principal cartel detection tool. It provides immunity to the first firm to self-report and discounts in fines of up to 50 per cent for firms that subsequently come forward or that agree to cooperate after a Commission investigation has been opened. While it is clearly deterrence-enhancing in saving resources and uncovering cartels that might not otherwise have been detected, some questions remain. Immunity is now granted in around three-quarters of EU cartel cases and the overall effect of leniency and of the Commission’s settlement notice is to significantly reduce cartel fines. Whereas the maximum allowable fine is 10 per cent of annual worldwide turnover, final fines now stand at around 2.2 per cent. In addition, leniency creates an obstacle for businesses seeking to discipline employees who can be identified as responsible for the infringement. These individuals will typically hold the information and evidence needed for their employer to successfully apply for immunity or give sufficient ‘added value’ for a leniency discount. Indeed, it is a condition for immunity that the firm and its employees provide continued and complete cooperation. This may make it difficult for businesses to keep their houses in order, in the way the Commission appears to expect them to do. The EU approach to antitrust enforcement is unusual in not employing sanctions against individuals as well as undertakings. Indeed, such sanctions might alleviate concerns over the level of fines imposed and the effectiveness of leniency. Oddly the EU corporate fines-only approach takes precedence over the competition law regimes of EU Member States, many of which have criminal and civil sanctions against individuals at their disposal. Unfortunately, there is currently no mechanism through which these can effectively complement the sanctioning powers of the European Commission, not least because there is still no one-stop-shop or effective coordination of leniency programmes within the EU. As the discussion of Ireland and the UK has shown, this raises an unsatisfactory situation where individuals responsible for local, low-level cartel infringements are more likely to be punished for their conduct than those responsible for the more damaging international infringements that are investigated by the European Commission. The EU’s antitrust enforcement regime has undoubtedly enjoyed considerable success in terms of the number of cartels uncovered and the volume of fines imposed. It is rightly held as a model for other jurisdictions. However, this chapter has identified the organic nature of EU antitrust, which is still relatively new and undergoing a process of ongoing evolution. In critically analysing the methodology of EU antitrust, the chapter has questioned the extent to which EU enforcement can be said to have a deterrent effect. While it employs a clear prohibition against hard-core cartel conduct and is uncovering many instances of cartel behaviour, it is unclear whether corporate fines are deterring prospective future infringements. If corporate fines alone are not achieving effective deterrence, it may be that the problem is being compounded by very generous levels of leniency and settlement discounts. The better use of individual sanctions on the national level would appear to be an obvious way of boosting deterrence, but more work needs to be done to ensure the effective coordination of EU and national sanctions and leniency programmes. Further research is also needed on the deterrent effects of EU antitrust enforcement as the extent of its effectiveness has implications for similar enforcement systems around the world.

18. Cartel enforcement: critical reflections from the South African experience Simon Roberts

I

THINKING ABOUT CARTELS

Cartels are typically referred to as the most egregious form of anticompetitive behaviour – with good reason, as they directly remove competition by agreement. However, it is increasingly unlikely that sophisticated firms with compliance training in competition law reach cartel agreements that can simply be identified as such. Identifying collusive arrangements requires a combination of analysis to identify the likely areas where coordination is taking place, with the powers to obtain evidence. In developing countries, such as South Africa, this is compounded by the probability that international firms reach their understandings outside of the country. The arrangements can include information exchange between competitors in different forms, which may also have procompetitive effects. At the same time, there are good reasons why firms may need to collaborate in order to facilitate joint investments and to take initiatives to improve production capabilities. In the South African context, there is also the particular concern with facilitating the entry and growth of smaller businesses, including black entrepreneurs, where cooperation can assist in increasing rivalry to the main incumbents but may fall foul of the cartel provisions. The South African Competition Commission was established in 1999 to enforce the then new Competition Act 1998. The Commission investigates alleged anticompetitive conduct and refers it to the Competition Tribunal for adjudication.1 The Commission’s first five years up to 2004 were marked by almost no cartel cases. Indeed, an Organisation for Economic Co-operation and Development (OECD) review put the low level of cartel enforcement down to the fact that firms were so deterred by the new Competition Act that this conduct had largely ceased.2 This has proved entirely incorrect as a very large number of cartels were subsequently uncovered. The South African experience, in fact, demonstrates the prevalence of coordination in close business communities. From around 2006, the Commission did step up its enforcement activities against cartels and found very widespread collusive activity in the economy. There are a number of reasons for the Commission’s success. First, the Commission introduced a corporate leniency policy (CLP) in 2004 and made it work, including with a substantive revision in 2008.3 Second, there was a move to more proactive investigations to identify likely cartel activity, associated with rooting out anticompetitive conduct in the Commission’s pri-

There is also a specialist Competition Appeal Court, a division of the High Court. Michael Wise, ‘Competition Law and Policy in South Africa’ (OECD Global Forum on Competition Peer Review, 11 February 2003). 3 See Chantal Lavoie, ‘South Africa’s Corporate Leniency Policy: A Five-Year Review’ (2010) 33 World Competition 141. 1 2

414

Cartel enforcement in South Africa  415 ority sectors. Third, the Commission used settlements very effectively to identify additional conduct. Fourth, international cooperation has seen the Commission identify an increased number of international cartels that have effect in South Africa, including car parts, freight for motor vehicles, air cargo and foreign exchange trading. This chapter begins by considering what cartels are, using the South African experience to discuss the differing nature and extent of arrangements that have been considered collusive, as well as those that have not. The processes of screening and prioritization are then explored. The South African experience in two typical sectors of cartel enforcement – cement and construction – is reviewed to provide insights into how cartel conduct has evolved, as well as enforcement strategies for relatively young authorities. The chapter then examines leniency and settlements before drawing some brief conclusions.

II

WHAT ARE CARTELS? A CATEGORIZATION OF CARTELS IN SOUTH AFRICA

The South African Competition Act separates so-called ‘hard-core’ conduct (direct or indirect price-fixing, market division and bid-rigging), as per se contraventions, from other horizontal arrangements that would be judged in terms of their anticompetitive effects weighed against their efficiency-enhancing implications. There is a penalty of up to a maximum of 10 per cent of total turnover for cartel conduct. Until the amendment coming into force on 12 July 2019, the penalty only applied to hard-core conduct.4 The prohibition on hard-core conduct is also somewhat unusual in that it treats the three specified areas of collusion as separate types of conduct instead of potentially aspects of the same agreement or understanding. It is well established that monitoring market shares is an effective way to check if there is secret discounting from an agreed price. Similarly, allocating tenders will involve rigging the bids to effectively allocate customers. The findings of cartel conduct embodied in settlements confirmed by the Tribunal provide a profile of the identified cartels in South Africa from the Act coming into force in September 1999 to the end of 2016.5 The dramatic growth in the number of settlements per year from 2009 to 2013 reflects the effects of making leniency work.6 While there has been somewhat of a falling-off from the peak of 39 settlements in 2013, there was still a substantial number (20 settlements) in 2016 (Figure 18.1). The sectoral profile (Figure 18.2) indicates that the main sectors are very much in line with those found in jurisdictions around the world.7 These are: food and agriculture (including

4 A finding does mean that damages claims can be brought by customers and other parties that have been harmed. An amendment that came into force in 2016 introduced criminal sanctions for hard-core cartel conduct but, as of mid-2017, there had been no reported criminal cases. 5 Note that there can be a number of separate settlements with individual firms in the same cartel case. 6 Gertrude Makhaya, Wendy Mkwananzi and Simon Roberts, ‘How Should Young Institutions Approach Enforcement? Reflections on South Africa’s Experience’ (2012) 19 South African Journal of International Affairs 43. 7 See, for example, John M Connor and C Gustav Helmers, ‘Statistics on Modern Private International Cartels: 1990–2005’ (Department of Agricultural Economics, Purdue University Working Paper No 06-11, 2006).

416  Research handbook on methods and models of competition law

Figure 18.1

Settlements, number over time

grain storage); construction; and industrial products (such as steel). In agriculture, there were marketing boards which had organized the markets and after their liberalization in 1996 the role of coordinating was taken over in many markets by the industry association. In industrial products, including construction materials such as concrete and plastic pipes, and reinforcing steel, the conditions in a relatively small economy with a tight-knit business community are highly conducive to collusion, and so it has proved. Extensive bid-rigging of construction contracts has also been uncovered, including through ‘Construction Fast Track’ settlements. In the few cases since 2007 not settled, the matters have almost all related to the size of the penalty.8 Some of these also included decisions as to whether fringe market participants were involved in the arrangements. In addition, there have been a very few decided matters on what constitutes coordination, including market division arrangements between small local locksmiths, and on whether a group of firms can be colluding or constitutes a single firm.9 An

8 See Competition Tribunal decisions in the cycling cartel (Competition Commission and Mailot Juane Trading (Pty) Ltd [2015] ZACT 44 (29 April 2015)); the removals cartel (Competition Commission and Stanley’s Removals [2016] ZACT 89 (12 December 2016)); plastic pipes (Competition Commission and DPI Plastics (Pty) Ltd [2012] ZACT 27 (4 July 2012)), (Competition Commission and Fio-Tek Pipes and Irrigation (Pty) Ltd [2010] ZACT 72 (27 October 2010)), (Competition Commission and Marley Pipes System (Pty) Ltd [2010] ZACT 24 (31 March 2010)); mining roof bolts (Competition Commission and Aveng (Africa) Ltd t/a Duraset [2010] ZACT 56 (25 August 2010)); (Competition Commission and RSC Ekusasa Mining (Pty) Ltd [2012] ZACT 82 (19 September 2012)); and wire mesh cases (Competition Commission and Aveng (Africa) Ltd t/a Steeldale [2012] ZACT 32 (7 May 2012)). See also Tribunal and Competition Appeal Court in concrete pipes case (Competition Commission and Gralio Precast (Pty) Ltd [2011] ZACAC 7 (20 October 2011)). 9 Competition Commission and Delatoy Investment (Pty) Ltd [2016] ZACT 37 (14 April 2016).

Cartel enforcement in South Africa  417

Note: Total not 100% due to rounding.

Figure 18.2

Number of settlements by main sector/industry, 2004–2016

important case relating to what constitutes an agreement or concerted practice in the vehicle tracking market is discussed in more detail below. In terms of the magnitude of the penalties, a large proportion have been settled with low penalties of less than 3 per cent of turnover, and many of the penalties where a percentage of turnover is not specified are also relatively low (Table 18.1).10 All settlements specify the monetary value of the penalty, while only some express the amount as a percentage of a relevant measure of turnover. As discussed below, the ‘Construction Fast Track’ settlements were an ‘all-in’ settlement for numerous instances of bid-rigging by the firms concerned and so need to be considered separately. Only just over one-tenth of cartel settlements were at penalties of 7 per cent or more of turnover. This is complicated by the fact that different turnovers are referred to in settlements, typically with turnover of the relevant business entity or division within a large business group being referred to (and not just the specific good or service which was cartelized). The penalties have thus been relatively low when compared with the order of magnitude of the harm normally found from collusion. In other words, the Commission has generally made 10 See also Tapera Muzata, Simon Roberts and Thando Vilakazi, ‘Penalties and Settlements for South African Cartels: An Economic Review’ in Jonathan Klaaren, Simon Roberts and Imraan Valodia (eds), Competition Law and Economic Regulation: Addressing Market Power in Southern Africa (Wits University Press 2017).

418  Research handbook on methods and models of competition law Table 18.1

Settlements, 2004­–16, by level of penalty

Penalty, % of turnover

Number of settlements

Not specified

41

0