European Competition Law Annual 2013: Effective and Legitimate Enforcement of Competition Law 9781849467452, 9781782257813, 9781509900473

This volume contains papers presented at the 18th Annual EU Competition Law and Policy Workshop. The papers examine mean

304 120 8MB

English Pages [683] Year 2016

Table of contents :
Contents
List of Sponsors
List of Participants
Table of Cases
Effective and Legitimate Enforcement of Competition Law: A Riddle Wrapped In a Mystery Inside an Enigma?
1. Introduction
2. A riddle ... the deterrence trap
3. Wrapped in a mystery ... the compliance trap
4. Inside an enigma ... a legitimacy trap?
5. Conclusion: regulation and the politics of markets
From Regulation 1 to Regulation 2: Enforcement of EU Law by National Sanctioning Regimes and the Need for Further Convergence
1. Introduction
2. Effective and efficient cartel prosecution
3. Sanctioning cartel infringements in Germany-challenges
4. Framework for reform
5. Solutions
6. Outlook: towards a 'Regulation 2'
The Federal Trade Commission and Monetary Remedies
1. Introduction
2. Summary of the 2003 policy statement
3. The FTC's 2012 Withdrawal
4. The FTC's reasons for withdrawing the statement are not persuasive
5. Would a reduction in private antitrust enforcement justify additional efforts by public enforcers to pursue monetary remedies?
6. When is it appropriate for an enforcement agency to pursue disgorgement?
The Principle of Effectiveness, Competition Law Remedies and the Limits of Adjudication
1. Introduction
2. The principle of effectiveness and remedial discretion
3. Legitimate remedies and the limits of adjudication
4. Remedial discretion and commitment decisions: exploring the limits of adjudication
5. Conclusions
Negotiated Remedies in the Modernisation Era: The Limits of Effectiveness
1. Negotiated remedies in the modernisation era
2. The limits of effectiveness
3. Conclusion
Behavioural Remedies for Antitrust Infringements-Opportunities and Limitations
1. Introduction
2. The purposes of remedies
3. Practical considerations
4. Conclusion
Behavioural versus Structural Remedies in EU Competition Law
1. Introduction
2. Behavioural versus structural remedies
3. Conditions under Regulation 1/2003
4. Conclusion
Bibliography
The Antitrust Conversation (Continued)
1. Introduction
2. US amicus practice in general
3. American antitrust amici: an update
4. Outside the US
5. A look to the future
Enforcement of Article 106(1) TFEU by the European Commission and the EU Courts
1. Introduction
2. European Commission cases
3. 'Indirect enforcement' of Article 106(1) by the Court of Justice: case law on Article 106(1) TFEU arising from preliminary rulings
4. Conclusions
Hard Look Review of Anticompetitive State Action
1. The state action problem and leading responses
2. The hard look alternative
3. Comparing the three models
4. Conclusion
Parallel Proceedings in EU Competition Law: Rethinking Ne Bis In Idem as a Limiting Principle
1. The problem and the structure of the enquiry
2. Inadequacy of discretionary prosecutorial restraint as a limiting principle
3. Ne bis in idem as a limit to parallel proceedings
4. The function of the ne bis in idem principle in limiting multiple proceedings and decisions in EU competition law
Reflections on Parallel Enforcement, Fundamental Rights and the Rule of Law in the Competition Law Context
1. Legal succession in fines and fundamental rights
2. ThyssenKrupp Nirosta
3. The Swiss Publigroupe judgment and Regulation 1/2003
Creating a Respected Brand: How Competition Agencies Signal Quality
1. Introduction
2. Brands and public institutions
3. Forces that determine the quality of a competition agency's brand
4. Conclusion
International Cooperation in Antitrust Enforcement: A Canadian Perspective
1. Introduction
2. Canada's approach to international cooperation in antitrust enforcement
3. Legitimacy of information sharing
4. International coordination of merger control and merger remedies
5. International cooperation on criminal matters
6. Conclusion
Effectiveness of Enforcement Cooperation in Developing Countries: What Role Can Existing Institutions Play?
1. Introduction
2. The WTO effort to incorporate competition in trade agreements
3. The International Competition Network and the cooperation agenda
4. Major problems in antitrust enforcement in developing countries
5. The role of the SADC: some reflections on the experience of three African countries (Zambia, Zimbabwe and Tanzania)
6. What kind of cooperation for cross-border cases?
7. Conclusion
The Need for International Cooperation in Merger Enforcement
1. Introduction
2. The scope for disagreement on international mergers
3. The costs of international disagreements
4. Does it matter?
5. Will the situation get better or worse?
The UK Competition Regime: Developments and Further Proposals for Change
1. Introduction
2. Institutional changes
3. Private enforcement/CAT jurisdiction
4. EU-level developments
5. Regulatory appeals consultation
6. Judicial independence
What Is To Be Done?
1. The Hearing Officer
2. Other reforms
Quis custodiet ipsos custodes?
1. Competition law and the challenge for the judge
2. Theories and fashions change
3. Differences between countries
4. Differences in priorities
5. Does 'old' law help when considering 'new' cases?
6. The Court is robust and radical in some fields
7. The European Courts as locomotives of legal innovation and policy change
8. Where do we stand as to the ECHR?
9. The ECHR and competition decision-making
Interaction between Public and Private Enforcement of Competition Law
1. Full judicial review of decisions of the European Commission and of national competition authorities
2. Disclosure of evidence
3. Interim measures
4. Final decisions and remedial powers
5. Limitation periods and joint and several liability
6. Quantification of harm
Selected References

Recommend Papers

European Competition Law Annual 2001: Effective Private Enforcement of EC Antitrust Law 9781472559456, 9781841131986

The European Competition Law Annual 2001 is the sixth in a series of volumes following the annual workshops on EU Compet

213 102 3MB Read more

European Competition Law Annual 2011: Integrating Public and Private Enforcement of Competition Law - Implications for Courts and Agencies 9781474201797, 9781849463515

This volume contains the written contributions and oral proceedings of the “16th Annual EU Competition Law and Policy Wo

171 87 9MB Read more

European Competition Law Annual 2005: The Interaction between Competition Law and Intellectual Property Law 9781472560056, 9781841136455

This is the tenth in a series of volumes based on the annual workshops on EU Competition Law and Policy held at the Robe

170 77 4MB Read more

European Competition Law Annual 2004: The Relationship Between Competition Law and the (Liberal) Professions 9781472559883, 9781841136127

The European Competition Law Annual 2004 is ninth in a series of volumes following the annual workshops on EU Competitio

161 17 13MB Read more

European Competition Law Annual: 2006: Enforcement of Prohibition of Cartels 9781472560148, 9781841137513

This is the eleventh in the series on EU Competition Law and Policy produced by the Robert Schuman Centre of the Europea

166 65 4MB Read more

European Competition Law Annual 2008: Antitrust Settlements under EC Competition Law 9781472560650, 9781841139586

This is the thirteenth in a series on EU Competition Law and Policy produced under the auspices of the Robert Schuman Ce

162 59 4MB Read more

Alternative Enforcement of Competition Law 9789462746091, 9789462367074

Competition authorities are known for imposing enormous fines on companies that have infringed the law. However, most au

162 57 2MB Read more

European Competition Law Annual 1998: Regulating Telecommunications 9781472558954, 9781841130996

The 1998 Volume on the regulation of communications markets is the third in a successful series of European Competition

170 18 3MB Read more

European Competition Law Annual 2012: Competition, Regulation and Public Policies 9781474201803, 9781849465823

Every year, top-level market regulators, academics and legal and economic practitioners contribute to the Annual Competi

160 33 7MB Read more

European Competition Law Annual 1997: The Objectives of Competition Policy 9781472558817, 9781901362671

This volume of essays contains contributions by a group of specialists in the area of competition law, including heads o

198 107 33MB Read more

European Competition Law Annual 2013: Effective and Legitimate Enforcement of Competition Law
9781849467452, 9781782257813, 9781509900473

Author / Uploaded
Philip Lowe
Mel Marquis
Giorgio Monti (editors)

0 0 0
Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up

File loading please wait...

Citation preview

EUROPEAN COMPETITION LAW ANNUAL 2013: Effective and Legitimate Enforcement of Competition Law

ii

EUROPEAN COMPETITION LAW ANNUAL 2013 Effective and Legitimate Enforcement of Competition Law

Edited by

Philip Lowe, UK Competition and Markets Authority Mel Marquis, European University Institute and Giorgio Monti, European University Institute

OXFORD AND PORTLAND, OREGON 2016

Published in the United Kingdom by Hart Publishing Ltd 16C Worcester Place, Oxford, OX1 2JW Telephone: +44 (0)1865 517530 Fax: +44 (0)1865 510710 E-mail: [email protected] Website: http://www.hartpub.co.uk Published in North America (US and Canada) by Hart Publishing c/o International Specialized Book Services 920 NE 58th Avenue, Suite 300 Portland, OR 97213-3786 USA Tel: +1 503 287 3093 or toll-free: (1) 800 944 6190 Fax: +1 503 280 8832 E-mail: [email protected] Website: http://www.isbs.com © The editors 2016 The editors have asserted their right under the Copyright, Designs and Patents Act 1988, to be identified as the authors of this work. Hart Publishing is an imprint of Bloomsbury Publishing plc. 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, without the prior permission of Hart Publishing, or as expressly permitted by law or under the terms agreed with the appropriate reprographic rights organisation. Enquiries concerning reproduction which may not be covered by the above should be addressed to Hart Publishing Ltd at the address above. British Library Cataloguing in Publication Data Data Available ISBN: 978-1-50990-047-3 ISSN: 1467-744X Typeset by Compuscript Ltd, Shannon

CONTENTS List of Sponsors vii List of Participants xi Table of Casesxiii Effective and Legitimate Enforcement of Competition Law: An Overview

xxix

Mel Marquis PART I: Effective Enforcement of Competition Law 1. Effective sanctions and compliance I Christine Parker, Effective and Legitimate Enforcement of Competition Law: A Riddle Wrapped In a Mystery Inside an Enigma? II Konrad Ost, From Regulation 1 to Regulation 2: Enforcement of EU Law by National Sanctioning Regimes and the Need for Further Convergence 2. Effective remedies III Joshua D Wright, The Federal Trade Commission and Monetary Remedies IV Ioannis Lianos, The Principle of Effectiveness, Competition Law Remedies and the Limits of Adjudication V Damien MB Gerard, Negotiated Remedies in the Modernisation Era: The Limits of Effectiveness VI Giorgio Monti, Behavioural Remedies for Antitrust Infringements—Opportunities and Limitations VII Frank P Maier-Rigaud, Behavioural versus Structural Remedies in EU Competition Law

13

31

85

105 139 185 207

3. Agencies as amici curiae VIII Stephen Calkins, The Antitrust Conversation (Continued)231 4. Infringement procedures and public measures distorting competition IX José Luis Buendía Sierra, Enforcement of Article 106(1) TFEU by the European Commission and the EU Courts 279 X Daniel A Crane, Hard Look Review of Anticompetitive State Action 307

vi Contents PART II: Legitimate Enforcement of Competition Law I Renato Nazzini, Parallel Proceedings in EU Competition Law: Rethinking Ne Bis In Idem as a Limiting Principle II Wolfgang Kirchhoff, Reflections on Parallel Enforcement, Fundamental Rights and the Rule of Law in the Competition Law Context III William E Kovacic, Creating a Respected Brand: How Competition Agencies Signal Quality

353

383 393

PART III: Effectiveness and Legitimacy in International Enforcement Cooperation I Julie Soloway, Charles Layton and Eric Richmond, International Cooperation in Antitrust Enforcement: A Canadian Perspective 433 II Alberto Heimler, Effectiveness of Enforcement Cooperation in Developing Countries: What Role Can Existing Institutions Play? 447 III Antonio Capobianco, John Davies and Sean Ennis, The Need for International Cooperation in Merger Enforcement469 PART IV: Issues for Courts and Perspectives on the Judicial Role I Gerald Barling, The UK Competition Regime: Developments and Further Proposals for Change II James S Venit, What Is To Be Done? III Ian S Forrester, Quis custodiet ipsos custodes? Assessing the Judicial Role in a Lawful System of Competition Enforcement IV Mario Siragusa, Interaction between Public and Private Enforcement of Competition Law Conclusions

523 543

557 587 593

Philip Lowe Selected References599

SPONSORS OF THE EU COMPETITION LAW AND POLICY WORKSHOP

Baker Botts LLP Contact: James Rill, Esq 1299 Pennsylvania Ave, NW Washington, DC 20004 USA Tel.: +1 202 639 78 83 Fax: +1 202 639 78 90 E-mail: [email protected] Cleary, Gottlieb, Steen & Hamilton Contact: Mario Siragusa Piazza di Spagna, 15 I-00187 Rome Italy Tel.: +39 06 695 221 Fax: +39 06 692 00 665 E-mail: [email protected] Compass Lexecon Contact: Atilano Jorge Padilla Davidson Building 5 South Hampton Street B-1000 London WC2E 7HA Tel.: + 44 20 7632 5005 Fax: + 44 20 7632 5155 E-mail: [email protected] CRA International Contact: Cristina Caffarra 99 Bishopsgate London EC2M 3XD UK Tel.: +44 20 7664 3700 Fax: +44 20 7664 3998 E-mail: [email protected]

viii Sponsors Hengeler Müller Contact: Jochen Burrichter Benrater Strasse 18–20 D-40213 Düsseldorf Germany Fax: +49 211 83 04 222 E-mail: [email protected] Latham & Watkins LLP Contact: Sven Völcker Boulevard du Régent, 43–44 B-1000 Brussels Belgium Tel.: + 32 2 788 62 42 Fax: + 32 2 788 60 60 E-mail: [email protected] Martínez Lage, Allendesalazar & Brokelmann Contact: Santiago Martínez Lage Claudio Coello 37 ES-28001 Madrid Spain Tel.: +34 91 426 44 70 Fax: +34 91 577 37 74 E-mail: [email protected] RBB Economics Contact: Andrea Lofaro RBB Economics London The Connection 198 High Holborn London, WC1V 7BD UK Tel.: +44 20 7421 2414 Fax: +44 20 7421 2411 E-mail: [email protected] Skadden, Arps, Slate, Meagher & Flom LLP Contact: James Venit 523 Avenue Louise B-1050 Brussels Belgium Tel.: +32 2 639 03 00 Fax: +32 2 639 03 39 E-mail: [email protected]

Sponsors ix White & Case LLP Contact: Assimakis Komninos 62, rue de la Loi B-1040 Brussels Belgium Tel.: +32 2 219 16 20 Fax: +32 2 219 16 26 E-mail: [email protected] WilmerHale LLP Contact: John Ratliff Bastion Tower Place du Champ de Mars/ Marsveldplein 5 B-1050 Brussels Belgium Tel.: +32 2 285 49 08 Fax: +32 2 285 49 49 E-mail: [email protected]

x

18th ANNUAL EU COMPETITION LAW AND POLICY WORKSHOP EFFECTIVE AND LEGITIMATE ENFORCEMENT OF COMPETITION LAW

Organisers Philip Lowe, Mel Marquis and Giorgio Monti Contributing authors Gerald Barling High Court, London José Luis Buendía Sierra Garrigues, Brussels Stephen Calkins Irish Competition Authority, Dublin Antonio Capobianco OECD, Paris Daniel A Crane University of Michigan School of Law John Davies OECD, Paris Sean Ennis OECD, Paris Ian S Forrester White & Case, Brussels Damien MB Gerard Université catholique de Louvain Alberto Heimler Scuola Nazionale di Amministrazione, Rome Wolfgang Kirchhoff German Federal Supreme Court William E Kovacic George Washington School of Law, Washington, DC and UK Competition and Markets Authority, London Charles Layton Blakes, Toronto Ioannis Lianos University College London Philip Lowe UK Competition and Markets Authority, London Frank P Maier-Rigaud NERA, Berlin; Université catholique de Lille Mel Marquis European University Institute, Florence Giorgio Monti European University Institute, Florence Renato Nazzini LMS, Milan; King’s College, London Konrad Ost Bundeskartellamt, Bonn Christine Parker Monash University, Melbourne Eric Richmond Blakes, Toronto Mario Siragusa Cleary Gottlieb Steen Hamilton, Rome Julie Soloway Blakes, Toronto James S Venit Skadden Arps, Brussels Joshua D Wright George Mason University School of Law

xii

TABLE OF CASES 1. Judgments of the Courts of the European Union A Joined Cases C-89, 104, 114, 116, 117 and 125 to 129/85 A Ahlström Osakeyhtiö and others v Commission [1993] ECR I-1307�� 498, 572 Joined Cases C-204/00 P, C-205/00 P, C-211/00 P, C-213/00 P, C-217/00 P and C-219/00 P Aalborg Portland A/S and Others v Commission [2004] ECR I-123�� 34, 354 Case C-196/08 Acoset SpA v Conferenza Sindaci e Presidenza Prov Reg ATO Idrico Ragusa and Others [2009] ECR I-9913��300 Case 145/83 Adams v Commission [1985] ECR 3539��566 Case C-303/05 Advocaten voor de Wereld VZW v Leden van de Ministerraad [2007] ECR I-3633��386 Case T-128/98 Aéroports de Paris v Commission [2000] ECR II-3929��197 Case C-437/09 AG2R Prévoyance v Beadout Père et Fils SARL [2011] ECR I-973�� lii, 293 Case T-342/99 Airtours plc v Commission [2002] ECR II-2585�� 498, 548, 572 Case C-617/10 Åklagaren v Hans Åkerberg Fransson, EU:C:2013:105��356 Case C-62/86 AKZO Chemie BV v Commission [1991] ECR I-3359��192 Case C-97/08 P Akzo Nobel NV and Others v Commission [2009] ECR I-8237�� 5, 33 Case C-67/96 Albany International BV v Stichting Bedrijfspensioenfonds Textielindustrie [1991] ECR I-5751�� 264, 298 Joined Cases C‑628/10 P and C‑14/11 P Alliance One International Inc and Standard Commercial Tobacco Co Inc v Commission, EU:C:2012:479�� 5, 33 Case C-32/11 Allianz Hungária Biztosító Zrt and Others, EU:C:2013:160�� 120, 573 Case T-170/06 Alrosa Company Ltd v Commission [2007] ECR II-2601�� 62, 140, 213, 490 Case T-224/00 Archer Daniels Midland v Commission [2003] ECR II-2597�� 356, 379 Case C-397/03 P Archer Daniels Midland Co v Commission [2006] ECR I-4429��358 Case C-347/06 ASM Brescia SpA v Comune di Rodengo Saiano [2008] ECR I-5641��302 Case C-295/05 Asociación Nacional de Empresas Forestales (Asemfo) v Transformación Agraria SA (Tragsa) and Administración del Estado [2007] ECR I-299�� lii, 299 Case C-220/06 Asociación Profesional de Empresas de Reparto y Manipulado de Correspondencia v Administración General del Estado [2007] ECR I-12175��289 Case C-410/04 Associazione Nazionale Autotrasporto Viaggiatori (ANAV) v Comune di Bari and AMTAB Servizio SpA [2006] ECR I-3303��300

xiv Table of Cases Case C-457/10 AstraZeneca AB v Commission, EU:C:2012:770��362 Case T-395/94 Atlantic Container Line AB and Others v Commission [2002] ECR II-875�� 63, 114, 161, 193 Case T-24/90 Automec v Commission [1992] ECR II-2223�� 69, 161, 202, 220 Case T-9/92 Automobiles Peugeot SA v Commission [1993] ECR II-493��203 Case C-322/93 P Automobiles Peugeot v Commission [1994] ECR I-2727��203 C-280/06 Autorità Garante della Concorrenza e del Mercato v Ente tabacchi italiani—ETI SpA and Others [2007] ECR I-10893�� lix, 34, 387 B Case T-15/02 BASF AG v Commission [2006] ECR II-497�� 168, 575 Case C-185/95 P Baustahlgewebe GmbH v Commission [1998] ECR I-8417��492 Case C-387/02 et al Silvio Berlusconi and Others [2005] ECR I-3565��48 Case C-413/06 P Bertelsmann AG and Sony Corporation of America v Independent Music Publishers and Labels Association (Impala) [2008] ECR I-4951�� 490, 548 Case 30/87 Corinne Bodson v SA Pompes funèbres des régions libérées [1988] ECR 2479��572 Case 7-72 Boehringer Mannheim GmbH v Commission [1972] ECR 1281�� 332, 379 Case C-297/07 Klaus Bourquain [2008] ECR I-9425��369 Case C-310/93 P BPB Industries and British Gypsum v Commission [1995] ECR I-865��575 Joined Cases C-46 and 48/93 Brasserie du Pêcheur SA v Bundesrepublik Deutschland and the Queen v Secretary of State for Transport, ex parte: Factortame Ltd and others [1996] ECR I-1029��569 Case T-219/99 British Airways plc v Commission [2003] ECR II-5917; appeal dismissed: Case C-95/04 P, [2007] ECR I-2331��72, 507, 566 Case T-289/03 British United Provident Association Ltd (BUPA), BUPA Insurance Ltd and BUPA Ireland Ltd v Commission of the European Communities [2008] ECR II-81��303 Case C-7/97 Oscar Bronner GmbH v Mediaprint Zeitungs-und Zeitschriftenverlag GmbH [1998] ECR I-7791��572 Case C-681/11 Bundeswettbewerbsbehörde and Bundeskartellanwalt v Schenker & Co AG and Others, EU:C:2013:404�� 6, 47 C Case C-67/13 P Groupement des cartes bancaires (CB) v Commission, EU:C:2014:2204��573 Case 8/72 Cementhandelaren v Commission [1972] ECR 977��560 Case C-386/10 P Chalkor AE Epexergasias Metallon v Commission [2011] ECR I-13085��44, 77, 576 Case T-15/89 Chemie Linz v Commission [1992] ECR II-1275��365 Joined Cases T-25, 26, 30–32, 34–39, 42–46, 48, 50–65, 65, 68–71, 87, 88, 103 and 104/95 Cimenteries CBR and Others v Commission [2000] ECR II-491�� 169, 365

Table of Cases xv Case 283/81 CILFIT and Lanificio di Gavardo SpA v Ministry of Health [1982] ECR 3415��lix Joined Cases C-94 and 202/04 Cipolla and Macrino [2006] ECR I-11421��313 Joined Cases T-125 and 127/97 The Coca-Cola Company [2000] ECR II-1733��114 Case C-441/07 P Commission v Alrosa Company Ltd [2010] ECR I-5949�� 62, 126, 140, 186, 213, 490, 550 Case C-42/92 P Commission v Anic Partecipazioni SpA [1999] ECR I-4125�� 34, 365 Case C-301/04 P Commission v SGL Carbon AG [2006] ECR I-5915��45 Case C-308/04 P SGL Carbon AG v Commission [2006] ECR I-5977�� 332, 356 Case C-328/05 P SGL Carbon AG v Commission [2007] ECR I-3921�� 332, 356, 575 Case C-419/07 Commission v Sweden, EU:C:2008:387��290 Case C-12/03 P Commission v Tetra Laval [2005] ECR I-987��549 Case C-141/02 P Commission v T-Mobile Austria GmbH, formerly max.mobil Telekommunication Service GmbH [2005] ECR I-1283�� 265, 281 Case C-441/11 P Commission v Verhuizingen Coppens, EU:C:2012:778��365 Case T-342/11 Confederación Española de Empresarios de Estaciones de Servicio (CEEES) and Asociación de Gestores de Estaciones de Servicio v Commission, EU:T:2014:60��151 Case T-93/02 Confédération nationale du Crédit mutuel v Commission [2005] ECR II-143��284 Case C-198/01 Consorzio Industrie Fiammiferi [2003] ECR I-8055��317 Joined Cases 56 and 58/64 Établissements Consten SARL and Grundig-Verkaufs-GmbH v Commission [1966] ECR 299�� 114, 572 Case C-320/91 Paul Corbeau [1993] ECR I-2533��288 D Case T-38/02 Groupe Danone v Commission [2005] ECR II-4407�� 176, 575 Joined Cases C-189/02 P, C-202/02 P, C-205/02 P to C-208/02 P and C-213/02 P Dansk Rørindustri and Others v Commission [2005] ECR I-5425��151 Case 43/75 Defrenne v Société Anonyme Belge de Navigation Aérienne Sabena [1976] ECR 455��570 Case T-279/02 Degussa v Commission [2006] ECR II-897��575 Case C-280/08 P Deutsche Telekom AG v Commission [2010] ECR I-9555�� lxi, 362 Joined Cases T-107 and 354/08 Diamanthandel A. Spira BVBA v Commission, EU:T:2013:367��172 Case T-169/08 Dimosia Epicheirisi Ilektrismou AE (DEI) v Commission, EU:T:2012:448, set aside and remanded: C-553/12 P Commission v Dimosia Epicheirisi Ilektrismou AE (DEI), EU:C:2014:2083�� 263, 286, 287 E Case T-87/05 EDP—Energias de Portugal, SA v Commission [2005] ECR II-3745��490

xvi Table of Cases Case C-260/89 Elliniki Radiophonia Tiléorassi AE and Panellinia Omospondia Syllogon Prossopikou v Dimotiki Etairia Pliroforissis and Sotirios Kouvelas and Nicolaos Avdellas and others [1991] ECR I-2925��296 Joined Cases C-240/12 and T-211/13 Eni SpA v Commission, EU:T:2014:132��356 Joined Cases T-360/09 E.ON Ruhrgas and E.ON AG, EU:T:2012:332��81 Case C-125/07 P Erste Group Bank AG v Commission [2009] ECR I-8681�� 34, 575 Case T-153/06 European Association of Euro-Pharmaceutical Companies, EU:T:2008:54��42 Joined Cases T-374, 375, 384 and 388/94 European Night Services Ltd v Commission [1998] ECR II-3141��573 Case 6-72 Europemballage Corporation and Continental Can Company Inc v Commission [1973] ECR 215��513 Case C-266/06 P Evonik Degussa GmbH v Commission [2008] ECR I-81*��38 Case C-226/11 Expedia Inc v Autorité de la concurrence and Others, EU:C:2012:795�� 487, 573 F Joined Cases T-271 and 245/03 Fédération nationale des syndicats d’exploitants agricoles (FNSEA) v Commission [2004] ECR II-271��356 Case T-34/92 Fiatagri UK Ltd and New Holland Ford Ltd v Commission [1994] ECR II-905; appeal dismissed: Case C-8/95 P, [1998] ECR I-3175��188 Case C-469/03 Filomeno Mario Miraglia [2005] ECR I-2009�� lviii, 372 Joined Cases T-217 and T-245/03 FNCBV and Others v Commission [2006] ECR II-4987��575 Joined Cases C-68/94 and 30/95 France v Commission (Kali und Salz) [1998] ECR I-1375��548 Case T-339/04 France Télécom SA v Commission [2007] ECR II-521��42 Case T-340/04 France Télécom SA v Commission [2007] ECR II-573��42 Case C-189/95 Harry Franzén [1997] ECR I-5909��304 G Case C-58/12 P Groupe Gascogne SA v Commission, EU:C:2013:770��492 Case C-467/04 Giuseppe Francesco Gasparini [2006] ECR I-9199��lvii, 332, 363 Case T-370/09 GDF Suez SA v Commission, EU:T:2012:333��81 Case T-168/01 GlaxoSmithKline Services v Commission [2006] ECR II-2969��547 Joined Cases C-501/06 P, C-513/06 P, C-515/06 P and C-519/06 P GlaxoSmithKline Services and Others v Commission and Others [2009] ECR I-9291�� 547, 563 Case 293/83 Gravier v City of Liège [1985] ECR 593��571 Joined Cases C-187 and 385/01 Hüseyin Gözütok and Klaus Brügge [2003] ECR I-1345��369 Case T-38/02 Groupe Danone v Commission [2005] ECR II-4407�� 176, 575 Case C-67/13 P Groupement des cartes bancaires (CB) v Commission, EU:C:2014:2204��573 Case T-310/94 Gruber + Weber GmbH & Co KG v Commission [1998] ECR II-1043�� 63, 115, 161, 189

Table of Cases xvii Case C-385/07 P Der Grüne Punkt—Duales System Deutschland GmbH v Commission [2009] ECR I-6155��492 Joined Cases 18 and 35/65 Gutmann v Commission [1966] ECR 103��328 H Case C-438/02 Krister Hanner [2005] ECR I-4551��304 Case T-64/12 Henkel AG & Co KGaA and Henkel France v Commission, EU:T:2013:116��359 Case C-51/92 P Hercules Chemicals NV v Commission [1999] ECR I-4235��491 Case T-30/89 Hilti AG v Commission [1991] ECR II-1439��190 Case C-53/92 P Hilti AG v Commission [1994] ECR I-667��190 Case 85/76 Hoffmann-La Roche & Co AG v Commission [1979] ECR 461�� 512, 565 Case C-41/90 Klaus Höfner and Fritz Elser v Macrotron GmbH [1991] ECR I-1979��296 Case T-209/01 Honeywell v Commission [2005] ECR II-5527��548 Case C-199/92 P Hüls AG v Commission [1999] ECR I-4287��491 I Case T-36/91 Imperial Chemical Industries plc v Commission [1995] ECR II-1847��176 Case T-184/01 R IMS Health v Commission [2001] ECR II-3193��191 Case C-418/01 IMS Health GmbH & Co OHG v NDC Health GmbH & Co KG [2004] ECR I-5039��561 Case T-76/89 Independent Television Publications Ltd v Commission [1991] ECR II-575��114 Case C-231/14 P Innolux Corp v Commission, EU:C:2015:451��419 Case C-429/07 Inspecteur van de Belastingdienst v X BV [2009] ECR I-4833�� xxxv, 41, 255 Case C-583/11 P Inuit Tapiriit Kanatami and others v Parliament and Council, EU:C:2013:625��133 Joined Cases 6 and 7-73 Istituto Chemioterapico Italiano S.p.A. and Commercial Solvents Corporation v Commission [1974] ECR 223�� 67, 185 Case C-213/00 P Italcementi—Fabbriche Riunite Cemento v Commission, EU:C:2003:84��328 Case 41/83 Italy v Commission [1985] ECR 873��572 J Case 222/84 Marguerite Johnston v Chief Constable of the Royal Ulster Constabulary [1986] ECR 1651��571 K Case C-350/07 Kattner Stahlbau GmbH v Maschinenbau- und MetallBerufsgenossenschaft [2009] ECR I-1513��293 Case C-272/09 P KME Germany AG v Commission [2011] ECR I-12789�� 44, 576

xviii Table of Cases Case C-389/10 P KME Germany AG, KME France SAS and KME Italy SpA v Commission [2011] ECR I-13125��77 Case C-557/12 Kone AG and Others v ÖBB-Infrastruktur AG, EU:C:2014:1317��109 Case C-52/09 Konkurrensverket v TeliaSonera Sverige AB [2011] ECR I-527��262 Case C-367/05 Norma Kraaijenbrink [2007] ECR I-6619��362 Case C-288/05 P Jürgen Kretzinger [2007] ECR I-6441��362 Case T-223/00 Kyowa Hakko Kogyo Co Ltd v Commission [2003] ECR II-2553��379 L Case T-54/03 Lafarge SA v Commission [2008] ECR II-120*��176 Case T-7/93 Langnese Iglo GmbH v Commission [1995] II-1533�� 63, 161 Case C-279/95 P Langnese-Iglo v Commission [1998] ECR I-5609�� 114, 360 Joined Cases C-468 to 478/06 Lelos kai Sia EE v GlaxoSmithKline [2008] ECR I-7139��563 Case T-128/11 LG Display v Commission, EU:T:2014:88; appeal dismissed: Case C-227/14 P, EU:C:2015:258�� 330, 359 Joined Cases C‑238/99 P, C‑244/99 P, C‑245/99 P, C‑247/99 P, C‑250/99 P to C‑252/99 P and C‑254/99 P Limburgse Vinyl Maatschappij and Others v Commission [2002] ECR I‑8375�� 355, 575 M Case T-203/01 Manufacture française des pneumatiques Michelin v Commission [2003] ECR II-4071��507 Case C-62/00 Marks & Spencer plc v Commissioners of Customs & Excise [2002] ECR 6325��571 Case T-111/08 MasterCard, Inc and Others v Commission, EU:T:2012:260; appeals dismissed: Case C-382/12 P, EU:C:2014:2201��lxxii, 152, 330 Joined Cases C-295 to 298/04 Vincenzo Manfredi v Lloyd Adriatico Assicurazioni SpA [2006] ECR I-6619�� 108, 376 Case C-2/91 Wolf W Meng [1993] ECR I-5751��264 Case C-179/90 Merci Conventionali Porto di Genova SpA v Siderurgica Gabriella SpA [1991] ECR I-5889�� 296, 573 Case T-201/04 Microsoft Corp v Commission [2007] ECR II-3601�� 66, 152, 197, 561 Case C-567/07 Minister voor Wonen, Wijken en Integratie v Woningstichting Sint Servatius [2009] ECR I-9021��304 Case 395/87 Ministère Public v Jean-Louis Tournier [1989] ECR 2521��573 Case C-119/05 Ministero dell’Industria, del Commercio e dell’Artigianato v Lucchini SpA [2007] ECR I-6199��389 Case C-235/92 P Montecatini Spa v Commission [1999] ECR I-4539��354 Case C-49/07 Motosykletistiki Omospondia Ellados NPID (MOTOE) v Elliniko Dimosio [2008] ECR I-4863�� lii, 263, 290

Table of Cases xix N Case 322/81 Nederlandsche Banden Industrie Michelin v Commission [1983] ECR 3461�� 114, 558 O Case C-245/91 Ohra Schadeverzekeringen NV [1993] ECR I-5851�� liii Case T-402/13 Orange v Commission, EU:T:2014:991�� lvi, 360 Case C-7/97 Oscar Bronner GmbH v Mediaprint Zeitungs-und Zeitschriftenverlag GmbH [1998] ECR I-7791��572 P Case C-458/03 Parking Brixen GmbH v Gemeinde Brixen and Stadtwerke Brixen AG [2005] ECR I-8585��300 Case 294/82 Parti écologiste “Les Verts” v European Parliament [1986] ECR 1339��569 Case 267/86 Pascal Van Eycke v ASPA NV [1988] ECR 4769��lv Case 730/79 Philip Morris Holland BV v Commission [1980] ECR 2671��197 Case C-439/09 Pierre Fabre Dermo-Cosmétique SAS v Président de l’Autorité de la concurrence [2011] ECR I-9419�� 119, 573 Case 25-62 Plaumann & Co v Commission [1963] ECR 95��133 Case C-209/10 Post Danmark A/S v Konkurrencerådet, EU:C:2012:172�� 513, 573 Case T-353/94 R Postbank NV v Commission [1994] ECR II-1141��505 Case C-379/98 PreussenElektra AG v Schhleswag AG [2001] ECR I-2099��264 Case C-375/09 Prezes Urzędu Ochrony Konkurencji i Konsumentów v Tele2 Polska sp. z o.o., devenue Netia SA [2011] ECR I-3055�� 46, 330, 354, 386 Case C-68/12 Protimonopolný úrad Slovenskej republiky v Slovenská sporiteľňa as, EU:C:2013:71�� 6, 47 R Joined Cases C-241 P and 242/91 P Radio Telefís Éireann (RTE) and Independent Television Publications Ltd (ITP) v Commission (Magill) [1995] ECR I-743�� 185, 219, 561 Case C-163/96 Silvano Raso [1998] ECR I-533��297 Case 120/78 Rewe-Zentral AG v Bundesmonopolverwaltung für Branntwein [1979] ECR 649��570 Case 2/74 Jean Reyners v Belgium [1974] ECR 631��571 Case C-18/88 Régie des télégraphes et des téléphones v GB-Inno-BM SA [1991] ECR I-5941��296 Case T-342/07 Ryanair Holdings plc v Commission [2010] ECR II-3457�� 511, 546 S Case T-241/01 Scandinavian Airlines System v Commission [2005] ECR II-2917��576

xx Table of Cases C‑501/11 P Schindler Holding Ltd and Others v Commission, EU:C:2013:522�� 38, 499 Case T-310/01 Schneider Electric SA v Commission [2002] ECR II-4071�� 548, 572 Case C-573/07 Sea Srl v Comune di Ponte Nossa [2009] ECR I-8127��302 Case C-451/03 Servizi Ausiliari Dottori Commercialisti Srl v Giuseppe Calafiori [2006] ECR I-2941�� lii, 299 Case C-308/04 P SGL Carbon AG v Commission [2006] ECR I-5977�� 332, 356 Case C-328/05 P SGL Carbon AG v Commission [2007] ECR I-3921�� 332, 356, 575 Case C-289/04 P Showa Denko KK v Commission [2006] ECR I-5859��356 Case C-320/95 P Società Finanziaria Siderurgica Finsider v Commission [1994] ECR I-5697��575 Joined Cases T-68, 77 and 78/89 Società Italiana Vetro SpA, Fabricca Pisana SpA and Vernante Pennitalia v Commission [1992] ECR II-1403��572 Case T-30/91 Solvay SA v Commission [1995] ECR II-1775�� 176, 505 Case C-110/10 P Solvay SA v Commission [2011] ECR I-10439��505 Case C-345/89 Alfred Stoeckel [1991] ECR I-4047��571 Case T-241/97 Stork Amsterdam v Commission [2000] ECR II-309��360 Case C-53/03 Syfait and others v GlaxoSmithKline AEVE and GlaxoSmithKline plc [2005] ECR I-4609��563 T Case C-295/12 P Telefónica SA v Commission, EU:C:2013:619�� 152, 584 Case C-295/12 P Telefónica SA, Telefónica de España SAU v Commission, EU:C:2014:2062��584 Case T-5/02 Tetra Laval BV v Commission [2002] ECR II-4381��548 Case T-83/91 Tetra Pak International SA v Commission [1994] ECR II-755�� 115, 190 Joined Cases T-144, 147 to 150 and 154/07 ThyssenKrupp Liften Ascenseurs NV v Commission [2011] ECR II-5129��375 Case C-352/09 P ThyssenKrupp Nirosta v Commission [2011] ECR I-2359�� 39, 386 Case T-24/07 ThyssenKrupp Stainless AG v Commission [2009] ECR II-2309�� 355, 374 Case C-340/99 TNT Traco SpA v Poste Italiane SpA and Others [2001] ECR I-4109��303 Joined Cases T-236, 239, 244 to 246, 251 and 252/01 Tokai Carbon and Others v Commission [2004] ECR II-1181, upheld on appeal: Case C-328/05 P SGL Carbon v Commission [2007] ECR I-3921��575 Joined Cases T-71, 74, 87 and 91/03 Tokai Carbon and Others v Commission [2005] ECR II-10��575 Case C-301/04 P Tokai Carbon v Commission [2006] ECR I-5915�� 45, 361 Case T-155/06 Tomra Systems ASA and Others v Commission [2010] ECR II-4361; appeal dismissed: Case C-549/10 P, EU:C:2012:221�� 496, 566 Case C-145/88 Torfaen Borough Council v B&Q plc [1989] ECR 3851��571

Table of Cases xxi Case C-17/10 Toshiba Corporation v Úřad pro ochranu hospodářské soutěže, EU:C:2012:72�� 330, 354, 513 Case T-39/06 Transcatab SpA v Commission [2011] ECR II-6831��356 Case C-491/07 Vladimir Turanský [2008] ECR I-11039��369 U Case C-119/97 P Union française de l’express (Ufex), formerly Syndicat français de l’express international (SFEI), DHL International and Service CRIE v Commission [1999] ECR I-1341�� 187, 215 Case C-250/06 United Pan-Europe Communications Belgium SA and Others v Belgium [2007] ECR I-11135�� lii, 298 V Case C-436/04 Leopold Henri Van Esbroeck [2006] ECR I-2333��362 Case 26/62 NV Algemene Transport- en Expeditie Onderneming Van Gend en Loos v Netherlands Inland Revenue Administration [1963] ECR 3��569 Case C-150/05 Jean Leon Van Straaten v Staat der Nederlanden [2006] ECR I-9327�� lviii, 361 Case C-439/08 Vlaamse federatie van verenigingen van Brood- en Banketbakkers, Ijsbereiders en Chocoladebewerkers (VEBIC) VZW [2012] ECR I-12471�� xxxiv, 40, 109 Case 238/87 Volvo (UK) Ltd v Veng AB [1988] ECR 6211��561 W Case C-91/08 Wall AG v La ville de Francfort-sur-le-Main and Frankfurter Entsorgungs- und Service (FES) GmbH [2010] ECR I-2815��302 Case T-116/04 Wieland-Werke v Commission [2009] ECR II-1087��576 Case 14/68 Walt Wilhelm and others v Bundeskartellamt [1969] ECR 58�� lvi, 327 Case C-2/91 Wolf Meng [1993] ECR I-5751�� liii Z Case C‑212/08 Zeturf Ltd v Premier Ministre [2011] ECR I-5633��314 Case C-439/11 P Ziegler SA v Commission, EU:C:2013:513��580 2. Decisions of the European Commission Antitrust Commission Decision of 23 December 1993 in Case IV/32.745—Astra��65 Commission Decision of 13 July 1994 in Case IV/C/33.833—Cartonboard��188 Commission Decision of 10 April 2013 in Case AT.39727—CEZ�� 150, 215 Commission Decision of 16 July 2008 in Case COMP/C2/38.698—CISAC��181 Commission Decision of 13 March 2011 in Case COMP/39579— Consumer Detergents��328 Commission Decisions of 14 September 2007 in Case COMP/39.140—Daimler Chrysler��166

xxii Table of Cases Commission Decision of 22 February 2006 in Case COMP/B-2/38.381—De Beers��61 Commission Decision of 25 July 2001 in Case COMP/36.915—Deutsche Post AG (Deutsche Post II—Interception of cross-border mail)��198 Commission Decision of 11 October 2007 in Case COMP/37.966—Distrigaz��166 Commission Decision of 25 July 2013 in Case AT.39847—E-Books�� 163, 196 Commission Decision of 4 December 1991 in Case IV/33.157—Eco System/Peugeot��203 Commission Decision of 14 December 1985 in Case IV/30.698—ECS/AKZO��192 Commission Decision of 29 September 2010 in Case COMP/39.315—ENI��150 Commission Decision of 22 December 1987 in Cases IV/30.787 and 31.488—Eurofix-Bauco v Hilti��188 Commission Decision of 14 September 2007 in Case COMP/39.141—Fiat��166 Commission Decision of 3 December 2009 in Case COMP/39.316—Gaz de France��130 Commission Decision of 26 November 2008 in Case COMP 39.389—German Electricity Balancing Market�� 130, 220 Commission Decision of 26 November 2008 in Case COMP 39.388—German Electricity Wholesale Market�� 130, 220 Case AT.39740—Google [2013] OJ C120/22��163 Commission Decision of 2 December 1981 in Case IV/25.757—Hasselblad��200 Commission Decision of 13 May 2009 in Case COMP/37.990—Intel��147 Commission Decision of 21 December 2000 in Case COMP/35.918—JCB��200 Commission Decision of 10 May 2007 pursuant to Article 86(3) of the EC Treaty, on the special rights granted to La Banque Postale, Caisses d’Epargne and Crédit Mutuel for the distribution of the livret A and livret bleu, C(2007) 2110 final�� 266, 283 Commission Decision of 17 March 2010 in Case COMP/39.386—Long-term contracts France��174 Commission Decision of 19 December 2007 in Case COMP/34.579—MasterCard��154 Commission Decision of 20 June 2001 in Case COMP/36.041/PO—Michelin��66 Commission Decision of 31 May 2002 in Case COMP/36.041/PO—Michelin��199 Commission Decision of 28 March 2012 in Case COMP/39452— Mountings for windows and window-doors��80 Commission Decision of 16 December 2009 in Case COMP/39.530—Microsoft (tying)�� 163, 195

Table of Cases xxiii Commission Decision of 3 July 2001 in Case COMP/38.044—NDC Health/IMS Health: Interim Measures [2002] OJ L59/18, withdrawn by Commission Decision of 13 August 2003 in Case COMP/38.044—NDC Health/IMS Health: Interim Measures [2003] OJ L268/69��561 Commission Decision of 14 September 2007 in Case COMP/39.143—Opel��166 Commission Decision of 23 October 2001 in Case COMP/37133—La Poste��283 Commission Decision of 22 March 2006 in Case COMP/38.173—Premier League��163 Commission Decision of 9 December 2009 in Case COMP/38.636—Rambus�� 130, 163, 178 Commission Decision of 12 April 2006 in Case COMP/38.348— REPSOL CPP SA—Distribution de Carburants et Combustibles��149 Commission Decision of 18 March 2009 in Case COMP/39.402— RWE gas foreclosure�� 130, 150 Commission Decision of 27 April 2014 in Case AT.39939—Samsung— Enforcement of UMTS standard essential patents��130 Commission Decision of 18 June 2012 in Case COMP/39736— Siemens/Areva��153 Commission Decision of 7 October 2008 in Case COMP/39.562— Slovakian postal legislation relating to hybrid mail services�� 282, 314 Commission Decision of 15 November 2011 in Case COMP/39.592— Standard & Poor’s�� 130, 163 Commission Decision of 23 January 2013 in Case AT.39839— Telefónica/Portugal Telecom�� 147, 153 Commission Decision of 22 June 2011 in Case COMP/39.525— Telekomunikacja Polska�� 63, 147 Commission Decision of 24 July 1991 in Case IV/31043—Tetra Pak II��198 Commission Decision of 14 September 2007 in Case COMP/39.142—Toyota��166 Commission Decision of 19 October 1994 in Case IV/34.446— Trans-Atlantic Agreement��193 Commission Decision of 17 February 1992 in Cases IV/31.370 and 31.446—UK Agricultural Tractor Registration Exchange��188 Commission Decision of 5 June 1991 in Case IV/32.879—Viho/Toshiba [1991] OJ L287/39��561 Commission Decision of 26 February 2014 in Case AT.39398—VISA MIF��130 Commission Decision of 27 June 2012 in Case COMP/39611— Water Management Products��188 Mergers Commission Decision of 22 November 2012 in CASE COMP/M.6541— Glencore International plc/Xstrata plc��476

xxiv Table of Cases 3. Judgments of the European Court of Human Rights Case No 17862/91, Cantoni v France, 11 November 1996��391 Case No 5100/71, Engel v Netherlands, 8 June 1976�� 354, 577 Case No 10890/84, Groppera Radio AG and Others v Switzerland, 28 March 1990��391 Case No 70982/01, Horciag v Romania, 15 March 2005��370 Case No 73053/01, Jussila v Finland, 23 November 2006��578 Case No 43509/08, Menarini Diagnostics v Italy, 27 September 2011�� 44, 354, 499, 579 Case No 37591/97, Metzger v Germany, 31 May 2001��40 Case No 28034/04, Müller v Austria (No 2), 18 September 2008��370 Case No 70982/01, Niedermeier v Germany, 3 February 2009��370 Case No 50178/99, Nikitin v Russia, 20 July 2004��370 Case No 4837/06, Segame SA v France, 7 June 2012��581 Case No 75602/01, Sundqvist v Finland, 22 November 2005��370 Case No 29881/07, Sievert v Germany, 19 July 2012��371 Case No 47650/99, Silvester’s Horeca’s Service v Belgium, 4 March 2004��581 Case No 18139/91, Tolstoy Miloslavsky v United Kingdom, 13 July 1995��391 Case No 59493/00, Withey v United Kingdom, 26 August 2003��370 Case No 14939/03, Zolotukhin v Russia, 10 February 2009�� 331, 362 4. US cases (courts and the agencies) 3M v LePage’s Inc, 542 US 943 (2004)��244 American Express Co v Italian Colors Restaurant, 570 US ___ (2013)��251 American Medical Association v FTC, 638 F2d 443 (2d Cir 1980)��396 American Needle, Inc v National Football League, 560 US 183 (2010)�� 242, 245, 246 Andrx Pharmaceuticals, Inc v Kroger Co, 543 US 939 (2004)��244 Armstrong Surgical Center, Inc v Armstrong County Memorial Hosp, 530 US 1261 (2000)��244 Bell Atlantic Corp v Twombly, 550 US 544 (2007)��57, 93, 250 Brown v Pro-Football, Inc, 518 US 231 (1996)�� 226, 242 California Retail Liquor Dealers Ass’n v Midcal Aluminum, Inc, 445 US 97 (1980)�� 268, 310 Cascade Health Solutions v Peacehealth, 479 F3d 726 (2007)��247 Comcast Corp v Behrend, 569 US __ (2013)��532 Continental TV, Inc v GTE Sylvania Inc, 433 US 36 (1977)�� 101, 236, 272, 562 Copperweld Corp v Independence Tube Corp, 467 US 752 (1984)��246 Credit Suisse Securities (USA) LLC v Billing, 551 US 264 (2007)�� 58, 94 CSU, LLC v Xerox Corp, 531 US 1143 (2001)��244 Dee-K Enterprises, Inc, v Heveafil SDN BHD, 539 US 969 (2003)��244 Dell Computer Corp, 121 FTC 616 (1996)�� 57, 91 Eastman Kodak Co v Image Technical Servs, Inc, 504 US 451 (1992)�� 226, 242 Ferguson v Skrupa, 372 US 726 (1963)��311

Table of Cases xxv Ford Motor Co v United States, 405 US 562 (1972)��161 FTC v Actavis, Inc, 570 US 756 (2013)�� 242, 247, 249, 569 FTC v The Hearst Trust, No 1:01CV00734 (DDC 2001)�� 56, 87 FTC v H.J Heinz, 246 F3d 708 (DC Cir 2001)��403 FTC v Lundbeck Inc, 650 F3d 1236 (8th Cir 2011)�� 56, 90 FTC v Mylan Labs, Inc, 62 F Supp 2d 25 (DDC 1999)�� 55, 86 FTC v Perrigo Co, No 1:04CV01397 (RMC) (DDC 2006)��56, 90, 102 FTC v Phoebe Putney Health System, Inc, 568 US ___ (2013)�� 226, 245 FTC v Schering-Plough, Inc, 548 US 919 (2006)��244 Greater Boston Television Corp v FCC, 444 F2d 841 (DC Cir 1970)�� 268, 308 Hartford Fire Ins Co v California, 509 US 764 (1993)�� 226, 242 F Hoffman La Roche Ltd v Empagran SA, 542 US 155 (2004)�� 242, 247 Illinois Tool Works Inc v Independent Ink, Inc, 547 US 28 (2006)�� 242, 250 Jaffee v Redmond, 518 US 1 (1996)��240 Joblove v Barr Laboratories, Inc, 551 US 1144 (2007)��244 Leegin Creative Leather Products, Inc v PSKS, Inc, 551 US 877 (2007)�� 101, 226, 245, 250, 508 Lochner v New York, 198 US 45 (1905)�� liv, 317 McFarling v Monsanto Co, 552 US 1096 (2008)��244 Dr Miles Medical Co v John D. Park & Sons Co, 220 US 373 (1911)��101 Monsanto Co v Spray-Rite Serv Corp, 465 US 752 (1984)�� 226, 242 Motor Vehicle Mfrs Ass’n v State Farm Mut Auto Ins Co, 463 US 29 (1983)��322 Motorola Mobility LLC, No 121-0120, 2013 WL 124100 (FTC 3 January 2013)��92 Neonatology Associates, PA v Commissioner of Internal Revenue, 293 F3d 128 (3d Cir 2002)��239 Nynex Corp v Discon, Inc, 525 US 128 (1998)��250 Pacific Bell Telephone Co v linkLine, 555 US 438 (2009)�� 57, 93, 246, 250 Parker v Brown 317 US 341 (1943)�� 268, 310 Pikes Peak Broad Co v FCC, 422 F2d 671 (DC Cir 1969)�� 268, 308 Rambus, Inc, No 9302 (FTC 5 February 2007)�� 57, 91 Robert Bosch GmbH, No C-4377 (FTC, 21 November 2012)�� 57, 91 Ryan v Commodity Futures Trading Commission, 125 F3d 1062 (7th Cir 1997)��238 SEC v Chenery Corp, 318 US 80 (1943)��323 Southern Motor Carriers Rate Conf, Inc v US, 471 US 48 (1985)��312 Statoil ASA v Heeremac VOF, 534 US 1127 (2002)��244 Texaco Inc v Hasbrouck, 496 US 543 (1990)�� 226, 242 Texas Indus Inc v Radcliff Materials, Inc, 451 US 176 (1980)�� 226, 242 Union Oil Co of Cal, 140 FTC 123 (2005)�� 57, 91 United States v Arnold, Schwinn & Co, 388 US 365 (1967)��101 United States v Carolene Prods Co, 304 US 144 (1938)��311 United States v Grinnell Corp, 384 US 563 (1966)��85 United States v Microsoft, No 98 Civ 1232 (CKK), 2002 (DDC 12 November 2002)��139 United States v Virginia, 518 US 515 (1996)��318

xxvi Table of Cases Verizon Communications Inc v Law Offices of Curtis V Trinko, 540 US 398 (2004)�� 57, 93, 242, 250 Voices for Choices v Illinois Bell Telephone Co, 339 F3d 542 (7th Cir 2003)��238 Weyerhaeuser Co v Ross-Simmons Hardwood Lumber Co, 549 US 312 (2007)��250 5. Other national jurisdictions Australia TPC v TNT Australia Pty Ltd and Ors [1995] ATPR 41-375��17 France Conseil Constitutionnel—Decision No 2013-331 QPC, 5 July 2013�� 500, 580 Decision of the Conseil de la concurrence No 07-D-31—Citroën, 9 October 2007��166 Germany Case 1 BvR 2172/96 of 26 February 1997��45 BGH, judgment of 29 June 2012—2 C_484/2010—Publigroupe, Bundesgericht Entscheidungen (BGE) 139 I 72��390 OLG Düsseldorf, judgment of 17 December 2012—V-1 Kart 7/12 (Owi), Wirtschaft und Wettbewerb 2013, 749—Silostellgebühren II�� 334, 385 BVerfG, decision of 30 August 2013—2 BvR 2752/11, WuW/DE-R 4081—ThyssenKrupp Nirosta��389 Case 2 BvR 2752/11—ThyssenKrupp Nirosta, Order of 30 October 2013��39 BGH, decision of 10 August 2011—KRB 55/10, Wirtschaft und Wettbewerb 2012, 81—Versicherungsfusion�� 333, 384 Ireland Doherty v South Dublin County Council [2007] 1 IR 246��256 Fitzpatrick v FK [2007] 2 IR 406��257 HI v Minister for Justice Equality and Law Reform [2003] IR 197��256 O’Brien v Personal Injuries Assessment Board (No 1) [2005] 3 IR 328��257 United Kingdom Case No 1178/5/7/11, 2 Travel Group PLC (in liquidation) v Cardiff City Transport Services Limited [2012] CAT 19��526 Case No 1166/5/7/10, Albion Water Limited v Dŵr Cymru Cyfyngedig [2013] CAT 6��526

Table of Cases xxvii American Cyanamid v Ethicon [1975] 2 WLR 316��204 Case No 1098/5/7/08, BCL Old Co Limited & Ors v BASF SE & Ors��526 Co-operative Insurance Society Ltd v Argyll Stores Ltd [1998] AC 1��204 Case No 1173/5/7/10, Deutsche Bahn AG & Ors v Morgan Crucible Company PLC & Ors��526 Devenish Nutrition Ltd v Sanofi-Aventis SA (France) [2009] 3 All ER 27, [2008] EWCA Civ 1086��353 Emerald Supplies Limited & Anr v British Airways plc [2010] EWCA Civ 1284; on appeal from [2009] EWHC 741 (Ch)��530 Case No 1077/5/7/07, Emerson Electric Co & Ors v Morgan Crucible Company PLC��526 English Welsh & Scottish Railway Ltd v Enron Coal Services Ltd [2009] EWCA Civ 647��526 Enron Coal Services Ltd (in liquidation) v English Welsh & Scottish Railway Ltd [2011] EWCA Civ 2��525 Ladd v Marshall [1954] EWCA Civ 1��538 Case No 1107/4/10/08, Merger Action Group v Secretary of State for Business, Enterprise and Regulatory Reform��539 Case No 1147/5/7/09, Moy Park Limited & Ors v Evonik Degussa GmbH and Degussa Limited��526 Napp Pharmaceutical Holdings Ltd v Director General of Fair Trading (Napp No 3) [2001] Comp AR 33, [2002] ECC 3��354 Napp Pharmaceutical Holdings Ltd v Director General of Fair Trading (Napp No 4) [2002] ECC 13��354 Pernod Picard/Campbell Distillers [2004] CAT 10��135 Skyscanner Limited v CMA [2014] CAT 16��131 Case No 1201/5/7/12, Vion Food Group Limited & Ors v Tessenderlo Chemie NV. and Britphos Limited��526 Wanadoo UK Plc v Office of Communications [2004] CAT 20��132

xxviii

Effective and Legitimate Enforcement of Competition Law: An Overview

Mel Marquis* This volume presents contributions prepared for the 18th edition of the Annual EU Competition Law and Policy Workshop, held on 19–20 July 2013 at the European University Institute in Florence. On that occasion the Workshop focused on the closely connected themes of effective enforcement and legitimate enforcement of competition law in national, European and international settings. The subjects are intertwined in the sense that they are mutually dependent. If a competition authority is granted far-reaching enforcement powers in the name of the public interest and uses them aggressively, for example, but if those powers are not tempered by adequate due process constraints protecting undertakings, including but not limited to independent and demanding ex post scrutiny, these enforcement arrangements would hardly be legitimate—there is no ‘output’ legitimacy in this context if output is taken to be the only relevant criterion. The legitimacy of process is thus an essential complement.1 Indeed, if the quality of process is sacrificed to ensure effectiveness and output, then trust in the regime is apt to break

* Mel Marquis is Part-Time Professor of Law at the European University Institute in Florence and Co-Director of the EU Competition Law and Policy Workshop. He is the CUFE Chair Professor at the Chinese University of Finance and Economics in Beijing and Professore a contratto at LUMSA University in Rome. Many thanks are due to all the Workshop participants for their efforts and contributions to this volume. As always, thanks also go to Philip Lowe and Giorgio Monti for their valuable contributions in relation to the planning and organization of the discussions, and to Claus-Dieter Ehlermann for his continuous support of the program. Valuable assistance has also been provided by Gonçalo Miguel Banha Coelho, who helped with organizational details, and by Jotte Mulder, who kindly prepared first drafts of the transcriptions. Gracious secretarial support has been provided by Laurence Duranel and Annick Bulckaen. Finally: a very bon vent to Annick as she departs from the EUI for the well-earned benefits of retirement. 1 Schmidt has introduced the notion of ‘throughput’ legitimacy, although this requires a certain corruption of the term throughput. See Vivien Schmidt, Democracy and Legitimacy in the European Union Revisited: Input, Output and ‘Throughput’, 61 Political Studies 2, 5–9 (2013). Of course, ‘throughput’ legitimacy is not a new concept: it refers to mechanisms of, for instance, efficacious operations (as opposed to efficient output), accountability, openness/transparency, inclusiveness, and in general the quality of governance processes and of (‘constructive’) interactions among relevant actors. It is however a new, and perhaps dubious, nomenclature. A better umbrella term for this variety of elements would be process legitimacy.

xxx Introduction down; this in turn is likely to generate perverse consequences and raise the risk of long-term damage to the cause of enforcement.2 For example, any ‘compliance’ with decisions (leaving aside incentives to circumvent them or to comply with them in bad faith) would in those circumstances be involuntary, a decidedly unsatisfactory situation in a world where a voluntary, internalized compliance culture is badly needed. Furthermore, if ‘users’ of the system as well as the broader public lose confidence in the competition authority and its powers, there may be a distinct risk that the authority will be substantially emasculated, and that over-enforcement will be supplanted by its opposite. In short, the effectiveness of enforcement can only be assured if legitimacy is assured. On the other hand, applying a counterpart logic, if a regime fails to prevent and sanction anticompetitive conduct effectively—if it fails to protect the public interest due to design or execution flaws or due to flawed policies, decision-making or institutions—that regime will likewise have no claim to legitimacy. As the present initial chapter will discuss, the rest of this book addresses the themes of both effectiveness and legitimacy from a variety of perspectives and by reference to a diverse range of dimensions. Background. The EU Competition Law and Policy Workshop is an ongoing program that explores topical policy and enforcement issues in the area of competition law, economics and governance.3 Each year the Workshop brings together a group of top-level EU and international policy makers, judges, legal practitioners, economic experts and scholars to take part in intensive debates that explore specific competition-related issues in an informal and non-commercial environment. Our hope is to stimulate critical reflection on the part of both the Workshop participants and the broader public. Structure of this chapter. This introductory chapter briefly presents the various contributions provided by the authors that participated in the 2013 Workshop. It is divided into the following sections: (1) (2) (3) (4)

Effective enforcement of competition law; Legitimate enforcement of competition law; Effectiveness and legitimacy in international enforcement cooperation; Issues for courts and perspectives on the judicial role.

2 Discussing this and related themes, see, eg Tom Tyler, ‘The Psychology of Self-Regulation: Normative Motivations for Compliance’, in Christine Parker and Vibeke L ehmann Nielsen, eds., Explaining Compliance: Business Response to Regulation, Edward Elgar, 2011, chapter 4. On the links between the rule of law and accountability mechanisms where power has been delegated to independent agents, see also Imelda Maher, ‘Functional and Normative Delegation to Non-Majoritarian Institutions: The Case of the European Competition N etwork’, 7 Comparative European Politics 414, 419–410 (2009), with references. 3 In this series, published by Hart of Oxford (see http://www.hartpub.co.uk/Series Details.aspx?SeriesName=European+Competition+Law+Annual), we have discussed, to name only a few topics: regulation and public policies, the interaction of public and private enforcement, judicial review by the EU Courts, and cartel settlements and commitment decisions. The next volume will address the subject of institutional change as numerous competition authorities see their structure, missions and powers being transformed.

Introduction xxxi The remainder of this chapter discusses the written contributions of the authors, and the standard caveat applies: the remarks made are selective and cannot substitute for a reading of the chapters themselves. Furthermore, no attempt is made to synthesize the transcriptions of the oral debates. The reader is invited to consult those discussions, also included in this book, and to read the Conclusions of Philip Lowe, for a more complete picture of the proceedings.

PART 1: Effective Enforcement of Competition Law A. Effective sanctions and compliance Christine Parker, ‘Effective and Legitimate Enforcement of Competition Law: A Riddle Wrapped in a Mystery Inside an Enigma?’ Parker begins her chapter with the observation that a competition agency bears the quasi-Sisyphean burden of having to face at least one new challenge for every challenge overcome. As she says, the enforcement strategies that must be employed to escape what she calls the ‘deterrence trap’ lure the agency into a ‘compliance trap’, which also implies a ‘legitimacy trap’. In describing this cascading process, Parker draws on previous empirical research focused on the Australian enforcement experience. The above-mentioned traps are explained as follows. A deterrence trap occurs when, in order to be truly deterrent, fines would actually have to be fixed at a level that exceeds an infringer’s ability to pay.4 The argument would also apply where the penalty constraint is imposed not by inability to pay but by a statutory cap, which may only approximate inability to pay or proportionality in a very imperfect way. To the extent this dilemma occurs, it throws serious doubt on systems such as that of the EU which are heavily dependent on the effectiveness of corporate fines.5 As Parker points out, however, some of the concerns associated with the deterrence trap may be mitigated or addressed by the reputational effects of condemnation and by—where applicable— individual responsibility. And of course, insofar as an agency can succeed in enhancing the compliance ethos of the business community, the need to rely

4 This proposition rests on a variety of factual predicates and should not be taken as universally assured. 5 Parker’s deterrence trap is one among several possible weaknesses of a system that relies heavily or exclusively on corporate fines. For a survey of issues, see Damien G eradin, Christos Malamataris and John Wileur, ‘The EU Competition Law Fining System’, in Ioannis Lianos and Damien Geradin, eds., Handbook on European Competition Law: Enforcement and Procedure, Edward Elgar, 2013, chapter 6.

xxxii Introduction on imperfect deterrence mechanisms recedes.6 However, as Australia’s ACCC (and its pre-1995 predecessor, the Trade Practices Commission) found ways to make compliance a pillar of its enforcement policy (including by means of including mandatory compliance terms such as corporate training and audits) to make up for an overly soft penalty regime, the agency found itself caught in a compliance trap. The essence of this problem seems to be that the acceptance of compliance conditions and mandated systems does not guarantee that a company will internalize a compliance ethic—that its compliance efforts will be motivated by a desire to be a good corporate citizen, and not by a hedonistic instinct to avoid pain and stigma. It seems that the agency may in fact be capable of contributing to this inauthentic form of compliance if it fails to respect the limits of its own powers—that is, if it becomes a bully. If the perception takes hold that compliance is being extracted in this way, one might expect the business community to activate hostile political powers, triggering a political crisis for the agency that could have long-term adverse consequences for its effectiveness. While Parker lays out the links from deterrence policy to compliance policy to the political problem, it may be added that the risk of an agency’s wings being clipped by politicians defending industrial interests may arise even without the second link in the chain: it has not been uncommon to see, in the age of the great cartel crusade, attempts to curtail the powers or resources of agencies in direct reaction to their aggressive use. Competition agencies in this regard resemble other organs operating in the complex environment of public policy and politics. As Parker explains, the strategy of the ACCC in response to the compliance trap was to advocate the criminalization of cartel conduct, which did in fact lead to the introduction in Australia of criminal penalties. But as many jurisdictions have discovered, the adoption of criminal laws does not of itself transform shared norms, values and practices. The roots of resistance may be deep, which can lead to the uncomfortable situation where an institution of doubtful legitimacy is enforcing laws of doubtful legitimacy. This is the legitimacy trap, and manifestations of it are drawn from interviews conducted by Parker and her fellow researchers. Again, the legitimacy trap is presented in terms of the Australian experience yet it would also be relevant in contexts where there is no criminal regime strictly speaking but where acceptance of the enforcement system is lacking, for example because (as alluded to at the beginning of this chapter) severe sanctions are unbalanced by procedural fairness and accountability. Such a legitimacy problem seems particularly corrosive for both the goal of deterrence and the ultimate goal of compliance. Indeed, as Parker explains, the legitimacy trap amplifies the other traps described above. Overall, the model is helpful in that it highlights what have 6 It is difficult to share the conviction of some agencies that the best means to ensure effective compliance is to rely predominantly on firm deterrence mechanisms such as, specifically, a leniency program and high fines.

Introduction xxxiii been called in a different context ‘equilibrating tendencies’, and it shows how a wide range of stakeholders, including the social structure itself, participate, in varying ways, in the design, execution and dissemination of competition enforcement. On the other hand, it will also be important to evaluate these processes applying a significant time dimension so that it can be gauged, for example, whether the traditional tendency—to view price fixing as the province of regulation and technocracy and not of criminal law—is capable of evolving.7 If so, and if the risks to legitimacy are managed in the meantime, then the triple trap model will be a valuable admonition but it will not seal the fate of aggressive enforcers if they employ a judicious policy mix in which infringements are prosecuted firmly and deterrence plays a part but voluntary compliance is incentivized to the point where the need for prosecution is minimized. Konrad Ost, ‘From Regulation 1 to Regulation 2: Enforcement of EU law by National Sanctioning Regimes and the Need for Further Convergence’. The chapter provided by Ost, as summarized in its title, suggests that competition enforcement in the EU is ready to build on the foundations introduced by Regulation 1/2003, and indeed that this is becoming a pressing need.8 In particular, Ost advocates a strengthening of the powers of NCAs since they are sometimes hamstrung by the peculiarities of national law, including

7 Andreas Stephan has considered how attitudes in the UK have changed or remained constant since 2007. (He also considers current attitudes in Germany, Italy and the United States, but diachronic analysis with regard to the UK is made possible by his earlier survey, where the UK was the focus.) On the one hand, price fixing continues to be perceived by most respondents as less serious than other corporate wrongdoing such as fraud, tax evasion or insider trading (although a significant part of the sample responded otherwise). However, public support for sending price fixers to prison has grown significantly (11% of respondents in favor in 2007; 27% in favor in 2014), a finding which Stephan thinks might be explained by the global financial crisis and the corporate misconduct that likely exacerbated it. It is easy to accept that corporate scandals have not endeared corporate executives to the public but there may be more general processes of evolving attitudes at play, as the UK public over time assimilates messages (such as: price fixing is repugnant) transmitted by public policy actors and various media. While core societal beliefs tend to be very resistant to change, norms and attitudes are more malleable and may change depending on a variety of shifting conditions. On the latter point, see Jingyuan Ma and Mel Marquis, ‘Business Culture in East Asia and Implications for Competition Law’, forthcoming in 51 Texas International Law Journal ___ (2016). 8 The recognition of need for broad legislative action at the level of the EU is growing. For example, see ‘Editorial comments’, 52 Common Market Law Review 1191, 1195 (2015). The editorial points out that legislative action would also provide an opportunity to overcome fragmented enforcement (in light of the Commission’s limited resources) by means of attributing to each NCA the power to sanction undertakings for effects beyond the national borders of the NCA concerned. See ibid, 1198. No hint of such a move emerges from the pending consultation launched on 4 November 2015 by the European Commission aimed at enhancing NCA powers and independence. (For information, see http://ec.europa. eu/competition/consultations/2015_effective_enforcers/index_en.html.) For an early reaction to the consultation process, see Giovanni Pitruzzella, ‘The Public Consultation on Regulation 1/2003: A Stronger Institutional Infrastructure for Fostering the EU Common Competition Culture’, 7 Journal of European Competition Law and Practice 1 (2016).

xxxiv Introduction national constitutional law as interpreted by national courts. The case study presented in this regard is German competition law enforcement, for which primary responsibility resides with the Bundeskartellamt. As Ost points out, the German enforcement framework has been defined largely by Regulation 1/2003 and by, in particular, the 2005 amendment to the German competition law, the GWB. Within this framework, it has become increasingly apparent that German enforcement is not as effective as it could be.9 Indeed, Ost refers to a number of ‘huge challenges’ to the integrity of enforcement at the national level. The first problem, briefly, concerns the increased number, duration and administrative burden of cases, which have stretched the BKA’s capacities. A second problem, with a more specifically legal dimension, arises from tensions between, on the one hand, the EU concept of an ‘undertaking’, which permits the Commission to sanction single economic entities as a whole and thus generally to ‘score’ higher fines while also holding parent companies jointly and severally liable; and on the other hand, the more restricted sanctioning powers of the BKA under German law. German administrative penal law recognizes a thick corporate veil, and the actions of a representative of one legal person generally cannot be attributed to another one. Strictly speaking this is true even if, for example, the latter entity exercises decisive influence over the former, although the BKA has partially maneuvered around this impediment by applying a theory of supervisory duty (as distinct from liability for breach of competition law), which is allowed under German law. Of course, where Article 101 is applied there can be no derogation from the EU definition of an undertaking, but from a German law perspective that definition is a matter of substantive law, whereas the sanctions that may be imposed by the BKA are regarded as belonging to national procedural law.10 One may add that under EU law such a national law would have to be disapplied if it were such as to compromise the effectiveness of EU law. The VEBIC judgment of the ECJ appears to be at least indirectly relevant in this respect,11 as does the case law linking the effectiveness of penalties at national

9 Some of the themes discussed by Ost are also raised in the chapter by Wolfgang irchhoff. Like Ost (and others—see the editorial above n 8), Judge Kirchhoff considers K further harmonization of rules such as those on sanctions (especially as regards legal succession) to be justified and necessary (see below). 10 If the ECJ were to hold that parental liability flowing from the single economic entity doctrine is to be considered substantively as part of the concept of an undertaking, it would be immaterial that the issue may traditionally have been conceived of at the national level as a rule of enforcement or procedure, and not of substance. The primacy of EU law—which includes primacy of the interpretations given to EU law by the Court—would be decisive. 11 Case C-439/08 Vlaamse federatie van verenigingen van Brood- en Banketbakkers, Ijsbereiders en Chocoladebewerkers (VEBIC) VZW [2012] ECR I-12471 (where the facts concerned the anomalous inability of the Belgian competition authority to effectively defend its own decisions in appellate proceedings).

Introduction xxxv level to the coherent application of the EU competition rules.12 Nevertheless, greater clarity with regard to the apparent clash of legal perspectives would be helpful. Ost also points to the traditionally conservative German approach to successor liability, which left open multiple paths to impunity for creative wrongdoers. This problem has been addressed in some respects by the 2013 amendment of the GWB, but certain loopholes remain, such as where an infringer’s assets are transferred to another entity, leaving a shell company behind.13 Still another idiosyncrasy relates to the application of the 10 per cent fining cap provided for in the GWB. Here the problem stems from the way the application of this cap is understood by the German Federal Court of Justice (Bundesgerichtshof), in particular because it diverges from the way the 10 per cent cap contained in Article 23 of Regulation 1/2003 is understood and applied at the level of the EU. According to the Bundesgerichtshof, the 10 per cent figure cannot be treated as a cap but rather must be construed as the top end of a range, this top end being reserved for the most egregious violations. To any who might propose that weaknesses in the German fining system could be overcome if criminal sanctions (strict sensu) were introduced, arguments are provided in opposition to such a move (section IV.3), not least because in Germany it would imply even longer, more cumbersome and more complex procedures. All of this leads Ost to call for action, including action at the level of the EU, to achieve greater harmonization of national procedures and sanctions. Such action is also made necessary, he says, by the fact that, if an investigation is reallocated within the ECN, the sanctions and procedures that apply to an undertaking may differ considerably and yet the undertaking has no right to challenge the reallocation. It is also apparent that Ost would favor harmonization in the form of a regulation as opposed to a directive, as a regulation could presumably establish greater clarity of obligations. National courts in this case would apply the regulation directly, and would not be applying, in the first instance, transposed laws of national character, which draw naturally on national concepts.

12 In the context of tax deductibility proceedings, see Case C-429/07 Inspecteur van de Belastingdienst v X BV [2009] ECR I-4833, paras 36–37 (‘To dissociate the principle of prohibition of anti‑competitive practices from the penalties provided for where that principle has not been observed would therefore deprive of any effectiveness the action taken by the authorities responsible for monitoring compliance with that prohibition and punishing such practices. Thus, the provisions of Articles 81 EC and 82 EC would be ineffective if they were not accompanied by enforcement measures provided for in Article 83(2)(a) EC. As the Advocate General stated at point 38 of his Opinion, there is an intrinsic link between the fines and the application of Articles 81 and 82 EC. The effectiveness of the penalties imposed by the national or Community competition authorities on the basis of Article 83(2)(a) EC is therefore a condition for the coherent application of Articles 81 EC and 82 EC.’ (emphasis added)). 13 Ost expects that ultimately the fate of the German rules on succession may be decided in the context of a preliminary reference to the ECJ.

xxxvi Introduction The European Commission has been exploring procedural and institutional issues and suggesting there is some scope for action.14 However, it is not clear that the Commission is being sufficiently bold in this regard.

B. Effective remedies Joshua Wright, ‘The Federal Trade Commission and Monetary Remedies’. The background and impetus for this chapter relate to the ambiguity of the FTC’s official position on the conditions in which it is appropriate to seek the equitable remedy of disgorgement in competition cases. The FTC had adopted a Policy Statement in 2003 providing guidance in this respect. According to that document, the FTC was to consider three criteria when deciding whether to pursue disgorgement, namely: whether the case involved a clear violation of the law; whether there was a reasonable basis for calculating the amount to be disgorged; and whether remedies in separate litigation brought against the defendant are likely to fail to fulfill the aims of the antitrust laws. However, in 2012, before Wright became an FTC Commissioner, and notwithstanding the dissent of Commissioner Ohlhausen, the FTC withdrew the Statement. The majority of Commissioners found the announced criteria to be too limiting and insufficiently robust. They also felt that the Statement had made the FTC too hesitant to seek disgorgement, and they intimated that the more rigorous constraints the Supreme Court had imposed on private plaintiffs in recent years justified more aggressive pursuit of this type of remedy. In this chapter, Wright files his own informal dissent against the FTC’s decision to retract the said Statement. In sections 4 and 5 he refutes the grounds given by the FTC, arguing, inter alia, that the available theory and evidence do not support the FTC’s assumptions that private suits have become difficult to win, and that the state of US law results in under-deterrence of anticompetitive conduct. If the Supreme Court has adopted anti-plaintiff decisions, Wright says, those decisions are ‘motivated by economic learning and an acceptance of the error-cost approach to designing liability rules in antitrust and not out of a concern related specifically to private antitrust suits’.15 Turning to the circumstances in which Wright considers disgorgement might be appropriate, he points to the following. First, he suggests that disgorgement should only be sought in hard core cartel cases or where a monopolist’s 14 As of this writing the Commission’s public consultation in this regard is still pending (see above n 8). See also Commission Staff Working Document, Enhancing competition enforcement by the Member States’ competition authorities: institutional and procedural issues (9 July 2014), SWD(2014) 231/2, http://ec.europa.eu/competition/antitrust/ legislation/swd_2014_231_en.pdf. The accompanying press release is at http://europa.eu/ rapid/press-release_IP-14-800_en.htm. 15 Page 101.

Introduction xxxvii restrictive conduct has no plausible efficiency justification. Vertical restraints cases, for example, would thus be excluded.16 He also cautiously embraces the criteria that were established by the FTC in the Statement. Above all, he thinks it is necessary to provide the business community with guidance on when the FTC might pursue disgorgement, without which there may be a risk that efficient activity will be deterred. Not long ago, Commissioner Ohlhausen and then-Commissioner Wright dissented formally from the FTC’s proposed consent order whereby, in addition to injunctive relief, Cardinal Health agreed to disgorge about 27 million dollars of ‘ill-gotten gains’ derived from the monopolized sale and distribution of low-energy radiopharmaceuticals in various metropolitan areas across the United States.17 In rather vague terms, the majority of Commissioners offered the following statement as regards its policy on disgorgement: ‘As always, the Commission will continue to exercise responsibly its prosecutorial discretion in determining which cases are appropriate for disgorgement. We regard disgorgement as one of many remedial tools at our disposal in competition cases, and will employ it judiciously to protect consumers and promote competition.’18 Ioannis Lianos, ‘The Principle of Effectiveness, Competition Law Remedies and the Limits of Adjudication’. This chapter explores the relationship between the principle of effectiveness, the use of remedies and the issue of the (il)legitimate exercise of public authority. It also offers normative views, from the point of view of legitimacy, as to how broadly the arena of adjudication should be understood, since the radius of interests directly and indirectly affected by a dispute can extend far beyond the interests of the ‘parties’. 16 See also Dissenting Statement of Commissioner Wright in Cardinal Health, Inc, File No. 101-0006 (17 April 2015), https://www.ftc.gov/system/files/documents/public_stateme nts/637771/150420cardinalhealthwright.pdf, page 3 (‘I would support a limitation on the Commission’s ability to pursue disgorgement only against naked price fixing agreements among competitors or, in the case of single-firm conduct, only if the monopolist’s conduct violates the Sherman Act and has no plausible efficiency justification. This latter category would include a monopolist’s fraudulent or deceptive conduct, or tortious activity such as burning down a competitor’s plant if such conduct violates the Sherman Act. I would also provisionally support disgorgement in a case if there were evidence demonstrating that a particular category of conduct shown to harm consumers was not adequately deterred through private suits and public enforcement actions seeking injunctive relief. This case does not belong in that category. Declining to pursue disgorgement in most cases involving vertical restraints has the virtue of taking the remedy off the table—and thus reducing the risk of over-deterrence—in the cases that present the most difficulty in distinguishing between anticompetitive conduct that harms consumers and procompetitive conduct that benefits them […].’). 17 See Statement of the FTC in the matter of Cardinal Health Inc, FTC File No. 1010006 (17 April 2005), https://www.ftc.gov/system/files/documents/public_statements/63 7781/150420cardinalhealthcommstmt.pdf. For other recent disgorgement cases, see FTC v AbbVie, Inc, No. 2:14-cv-05151-HB (E.D. Pa., filed 8 September 2014); Brief for FTC in Opposition to Cephalon’s Motion to Dismiss for Lack of Subject Matter Jurisdiction, FTC v Cephalon, Inc, No. 2:08-cv-2141 (E.D. Pa., filed 18 November 2013). 18 Ibid, page 6.

xxxviii Introduction The concept of the chapter is described in its introduction. A starting point is the fact that remedies tend to involve a certain degree of discretion on the part of the decision-maker; they may be molded according to particular circumstances, and thus remedy types are not necessarily prefabricated. This may open the door to risks associated with ‘discretionary remedialism’.19 Naturally, discretionary remedialism may find some justification on the basis of the principle of effective application of the law,20 yet it is equally obvious that this discretion cannot be unbounded. Some of the limits to which it is subject arise from the imperatives of fundamental rights protection, from other general principles of (EU) law, and from the correlativity of private law disputes (that is, the correlativity between the rights and duties of litigants).21 In the present contribution Lianos discusses a further limit: the legitimacy of the authority deciding on the remedy, given that some types of remedial intervention would stretch beyond the boundary of legitimate public action. As intended in this chapter, legitimacy refers to legitimacy-building mechanisms which guarantee that an undertaking will consider the remedies imposed upon it to be politically acceptable even if, presumably, it would prefer to do without them. Those mechanisms of legitimacy may differ depending on which form of social ordering is in play. It is suggested that, if the exercise of authority is such that the public agency’s functions ‘trespass’ upon those of another type of social ordering, then in order to avoid serious legitimacy problems the agency should submit to the legitimacy-building tools that govern that other sphere. The type of social ordering to which the imposition of remedies (and thus remedial discretion) belongs is (structural) ‘adjudication’,22 as opposed 19 Such risks may be particularly acute in the context of Article 9 commitment procedures, where the Commission enjoys significant discretion in a number of respects. (National practice sometimes appears quite different, as NCAs may in some cases be subject to firmer constraints; Lianos provides examples including, as noted below in the main text, the interesting Skyscanner case in the UK.) See for example pages 132–134. Similar concerns have been expressed by many, although the Commission’s procedure has also been defended by observers arguing that criticisms have been overstated. Among the body of literature on the subject, see Claus-Dieter Ehlermann and Mel Marquis, eds, European Competition Law Annual 2008: Antitrust Settlements under EC Competition Law, Hart Publishing, 2010; and see also Damien Gerard’s contribution to this volume (summarized below). 20 The term ‘principle of effectiveness’ may also evoke the notion of ‘optimal enforcement theory’. The latter theory, assuming for the sake of discussion that a consensus as to its meaning exists, may inform a normative understanding of the principle of effectiveness, and when the effectiveness principle is vindicated this may at the same time advance the interests embraced by that theory. But this would merely be an overlap since, by definition, they are distinct. The principle of effectiveness, in particular, is a multifaceted and potent legal concept. See, eg the remarks of Lianos at pages 108–112. 21 See page 117. 22 As noted below in the main text, Lianos later distinguishes ‘structural’ adjudication from the more classical ‘dispute resolution’ adjudication. (The words ‘limits of adjudication’ in the chapter title are a reference to this classical model.) The fact that the structural adjudication model in some ways resembles the managerial/administrative (regulatory) model leads Lianos to consider it suitable for resolving matters affecting a wide range of interests. See pages 123–124.

Introduction xxxix to contracting/bargaining, managerial/administrative discretion or legislation. Lianos mentions the example of managerial/administrative discretion, where an important legitimacy-building mechanism is the participation of affected interests in the remedial process. This type of discretion may often be seen in the sphere of economic regulation, and it is thus relevant here. The concern described above with regard to participation in the remedial process points to one of the chapter’s recurring issues: whether and the extent to which it is necessary to enhance participatory rights and to broaden the circle of those involved in the remedial process to include a community of diverse affected interests. To illuminate the implications of this community of diverse interests, Lianos revisits the web-like ‘polycentric’ dimension of disputes, as elaborated in the 1970s by Lon Fuller and as developed and glossed by later scholars. Polycentricity and the problems it raises for the standard process of adjudication—including the legitimacy problem where competition law remedies ‘trespass’ into areas traditionally inhabited by regulation—are discussed in section 3 of the chapter. It is observed in this context that competition law has features that accentuate the polycentric nature of disputes. For example, if a dominant firm and a rival engage in competition law litigation, other competitors as well as upstream suppliers and intermediate and final consumers may all be affected. Indeed, the outcome of the dispute may have repercussions in sectors and industries well beyond the relevant market. Lianos puts forward two sets of examples: (i) the Microsoft and Google cases, given their impact on innovation and competition in the global IT sector; and (ii) network industry cases, such as energy, telecoms or transport, which again tend to have far-reaching effects across sectors. At this point, Lianos explains that the adjudication model encompasses two sub-types. One of these, the ‘dispute resolution’ adjudication model, refers to a dispute that essentially concerns the arguments and evidence of the litigants themselves. But of more relevance for this discussion is the ‘structural’ adjudication model. The latter model allows for a broader spectrum of interests and affected persons, some of whom may be heard by the decision-maker—as in the case of amicus representations. And where the diversity of interests is prominent, it may in fact become increasingly difficult to distinguish this model from the regulatory model. In this light, Lianos describes the potential legitimacy problem as follows: Put simply, the more EU competition law moves towards the regulatory/managerial model, and ‘structural’ adjudication comes close to that, the more it should integrate the legitimacy-building mechanisms of that model, with the enhanced participation of the entities subject to the remedies as well as of all those whose interests may be affected (ie consumers, competitors in related markets and interests vicariously represented by organizations and citizen’s groups, [such as] environmental associations).23 Page 124.

23

xl Introduction Section 4 of the chapter extends the problems of discretion and of its limits from a legitimacy perspective to the use of the commitment procedure under Article 9 of Regulation 1/2003.24 The background to this discussion is the occasional characterization of this procedure by commentators and by the ECJ itself as exemplifying a ‘consensual’ (or ‘contract law’) model. The position taken here is that the more proper classification is the structural adjudication model, which in some ways blurs with the regulatory (managerial/ administrative) model, and which in normative terms requires a distinctive set of participation-oriented constraints. As Lianos contends in detail, the fact that the latter constraints are not embedded in the EU law on commitments (at least not to an adequate degree) may make Article 9 procedures almost irresistibly attractive in non-cartel cases from the Commission’s point of view. Whereas Article 7 procedures entail significant constraints,25 in an Article 9 case the Commission may be in a position to ‘use its bargaining power in order to achieve remedies that would not only attempt to reverse the situation to the status quo ante but would also aim to establish a new, allegedly more competitive, equilibrium […]’.26 This desired equilibrium may also involve the pursuit of wide liberalization and regulatory objectives, and the theory of harm in such cases may not be evidently solid enough to sustain an Article 7 claim of infringement. Lianos adds to this by criticizing the supposed dichotomy between a public law paradigm under Article 7 and the above-mentioned ‘contract law’ paradigm under Article 9. Clearly rejecting the proposition that the Article 9 procedure is consensual, he refers to the ‘psychological pressure’ imposed by the shadow of Article 7, which ‘enables the Commission to extract disproportionate remedies’ under Article 9.27 Since, for these reasons, 24 For further extended analysis of the Article 9 commitments procedure, see the contribution by Damien Gerard (summarized below). Several features of the commitment procedure noted by Lianos and mentioned in the paragraph corresponding to this footnote are also discussed critically by Gerard, above all in section 2 of his chapter. 25 See the contribution by Giorgio Monti (summarized below). 26 Page 128. Concerning the subject of the Commission’s superior bargaining power as well as the potential for its abuse and the consequences for substantive outcomes—on which views are not universal—see also the remarks made below in reference to the chapter by Damien Gerard. 27 Page 129. Such remedies must legally conform to the principle of proportionality— but it is the more diluted form of that principle as established in Alrosa. This is merely, as Lianos notes, a type of rationality test. However, the Commission deserves credit for holding itself to a standard somewhat higher than the one adopted by the ECJ. In the Commission’s Notice on best practices for the conduct of proceedings concerning Article 101 and 102 TFEU, [2011] OJ C308/6, para 115, the Commission states: ‘In light of the principle of proportionality, the Commission must verify that the commitments address the identified competition concerns and that the commitments offered do not manifestly go beyond what is necessary to address these concerns. When carrying out that assessment, the Commission will take into consideration the interests of third parties. However, it is not obliged to compare such voluntary commitments with measures it could impose under Article 7 of Regulation (EC) No 1/2003 and to regard as disproportionate any commitments which go beyond such measures.’ (emphasis added; citation omitted) The latter statement would be enforceable against the Commission in an action for annulment before the General Court on the basis of the principle of legitimate expectations.

Introduction xli Article 9 cases become a means of addressing issues—not necessarily confined to competition—which have wide implications for a variety of interests, they epitomize the notion of polycentric procedures. And again, they fit the model of structural adjudication.28 Yet interested third parties, including complainants, do not have extensive participatory rights in Article 9 cases. For Lianos, this constitutes a major weakness that needs to be addressed through procedural reform. He punctuates his argument with a discussion of the stricter standard to which the then-OFT was held by the UK Competition Appeal Tribunal in the Skyscanner case. Damien Gerard, ‘Negotiated Remedies in the Modernization Era: The Limits of Effectiveness’. A summary of this chapter may be prefaced by a word as to the meaning of the ambiguous term ‘effectiveness’, which appears in Gerard’s title. The intuitive meaning is arguably the successful vindication of the public interest in the enforcement of competition law, even if it is true that the content of the term ‘public interest’ in a given jurisdiction is typically contested. Gerard’s chapter, however, is based on an understanding of effectiveness rooted in the more subjective calculations of the competition law enforcer, which will have an interest in maximizing its own administrative resources when it can do so.29 The subjective and dynamic interests of a public authority thus do not ipso facto equate to the interests of the public. A possible example relates to fines: subjectively, a competition enforcer might derive great utility from the successful imposition of a massive financial penalty; yet it is not clear, and for the sake of discussion let us say it is far from clear, that aggressive fining policies have succeeded as an instrument of effectiveness from a public interest perspective. Another example is provided by a source cited in Gerard’s chapter, according to which an agency might select a case not in order to fulfill its public mission to the fullest extent (the source assumes this mission to be securing benefits for consumers), but because the agency expects the case will be relatively easy to settle.30 It is by no means unusual to regard the enforcer as a self-interested actor subject to legal and institutional constraints that limit (or fail to limit) bureaucratic drift,31 but it is perhaps helpful to emphasize the point because the use of terminology will be immediately understood, and because the manner in which ‘effectiveness’ is understood has implications for the way its relationship with legitimacy is conceptualized and evaluated.

Page 131. In section 2.1.2 of the chapter, Gerard also refers to a separate concept, the ‘substantive effectiveness’ of antitrust principles. 30 See page 143, footnote 20 (citing Ginsburg and Wright). Gerard proceeds, at page 145, to draw a distinction between effectiveness (again, from the perspective of the enforcer) and efficiency (welfare maximization). It appears from that discussion that the achievement of efficiency through enforcement constitutes, in Gerard’s view, the mission of the European Commission. 31 There is no shortage of literature on this, within and beyond the legal field. 28 29

xlii Introduction That being said, aside from the ‘limits of effectiveness’, the title also refers to ‘negotiated remedies’, by which Gerard means remedies in the form of commitments made binding, whether at EU or national level. This remedial tool is of a piece with leniency and cartel settlements, which in turn exemplify procedural modernization, one of the three sisters of the ‘legal and cultural revolution’, along with institutional and substantive modernization. The focus on commitments in this chapter is thus shared in common with the chapter by Lianos (see above). While Gerard ostensibly may seem to take a different view inasmuch as he stresses a paradigm of negotiation while Lianos rejects the consensual/contractual nature of commitments, both authors agree that: the implied threat of an infringement procedure makes the ‘voluntary’ nature of commitments suspect or illusory;32 and that the abundant use of commitments in non-cartel cases implies the expansive and to a large extent unchecked discretion of the European Commission. Gerard’s chapter is rich in detail and argumentation; this short tour will unavoidably be incomplete. As an initial matter, readers interested in the incentives driving both the authority and the undertakings concerned to engage in ‘negotiated solutions’ will want to pay particular attention to section 1.1.2, where numerous factors affecting those incentives are highlighted. To select just one example, a linkage with substantive modernization is discussed since, on the side of the enforcer, the additional costs and complexity of an infringement procedure that arise from the turn to an ‘effects-based’ approach can be substantial. In this sense, the advantages created by procedural modernization allow it ironically to cannibalize the potential gains (ie, potentially more efficient outcomes) that substantive modernization was intended to capture. In section 1.2, Gerard expresses concern with regard to the fact that commitment procedures focus on designing remedies rather than on establishing an infringement. In that setting, the theory of harm emerges from the mutual discussion of possible remedies and from the possible influence of third parties who are ‘rarely motivated by purely benevolent interests’.33 To that extent, the theory of harm may be biased, and—as already discussed—the applicable principle of proportionality under the relevant case law is a loose one that

32 As far as Gerard is concerned see, in particular, pages 157–158. Monti takes a different view and suggests that the leverage the Commission wields based on the constant possibility that it might withdraw from the commitment procedure under Article 9 and revert to the standard Article 7 infringement procedure is more apparent than real. As Monti notes, the Commission’s Notice on best practices (above n 27) indicates the Commission’s willingness to re-submit commitments if the market test is negative, thus keeping its Article 7 powder dry pending further efforts to resolve the case under Article 9. See Giorgio Monti, ‘Alrosa and Commitment Decisions in Perspective’, in Barry Hawk, ed, International Antitrust Law and Policy: Fordham Competition Law 2014, Juris Publishing, 2015, chapter 17. 33 Page 156.

Introduction xliii might not provide a sufficient safeguard against procedures and outcomes gone astray.34 The critique of this process leads to a call, in section 2, for a redoubled effort from the Commission to pursue ‘optimal’ outcomes and to address the existing ‘legitimacy gap’. The fundamental idea behind section 2, which is the heart of the chapter, is that the tension between procedural and substantive modernization as manifest in the commitments context can be reconciled by restoring (i) the principle of proportionality and especially (ii) the protection of fundamental rights, to their proper strength. Doing so, it is argued, would likewise restore the balance between effectiveness (maximization of administrative resources) and efficiency (maximization of welfare). With regard to proportionality and fundamental rights, Gerard suggests that due process may serve as a proxy for proportionality in terms of ensuring optimal outcomes, and that this would be a way to deal with the uncomfortable case law in this area. The discussion in section 2 covers a lot of ground. For example, it describes the shift to increasingly ‘regulatory’ (as opposed to corrective) remedies, and it underlines the problematic decline of solid, predictable new case law and the rise of a very influential but potentially error-inducing body of commitment decisions—not to mention the near-disappearance of judicial control by the EU Courts, which has further perverse consequences. As Gerard maintains, ‘commitment decisions tend to stretch the boundaries of antitrust legal standards with the paradoxical effect of affecting the overall predictability of the scope of Articles 101 and 102 TFEU’.35 In this way, the Commission pursues the (subjective) effectiveness of enforcement as defined above, but in reality the factors skewing its enforcement toward suboptimal commitment decisions have the effect of undermining the stability and ‘substantive effectiveness’ of antitrust principles.36 With a view toward closing the legitimacy gap arising from this troubling state of affairs, Gerard in section 2.2 enters into a detailed discussion of the need for greater respect for proportionality and due process (right to be heard, access to file, effective judicial protection). As he explains, and apart from their other virtues, due process helps to ensure that the Commission will identify the relevant issues and tailor its analysis while proportionality ensures that the substance of the adversarial process is translated into the appropriate outcome. Yet both principles are relaxed or truncated in the context of commitment procedures.37 In this part of the essay Gerard discusses, and registers 34 See above n 27 (the latter footnote recalling, however, that the Commission holds itself to an intermediate standard of proportionality falling somewhere between the Alrosa rationality test and the traditional standard that applies under Article 7). 35 Page 165. 36 Ibid. See also page 167. 37 Again, Monti has expressed doubts as to the criticisms lodged by Gerard and others regarding proportionality and due process (and hence the seemingly lopsided bargaining power of the Commission and the risk of suboptimal remedies), and has described them

xliv Introduction his disappointment with, the ECJ’s Alrosa judgment, for example because in his view it leaves ample scope for suboptimal remedies and lowers the bar as far as due process is concerned. On the other hand, it is acknowledged that the Commission has not run amok with its long leash but rather has cleaved to an intermediate proportionality principle (ie, a ‘manifestly disproportionate’ test) by virtue of its own scruples, typically (though not always) by limiting the duration of commitments made binding. Yet Gerard questions whether time limits have been too onerous in light of the pace and cycles of the industry concerned; and ultimately, he doubts that this self-discipline (even if the Commission’s soft law statements create legitimate expectations) will be satisfactory as a sustainable safeguard. Given the state of the case law, it seems that in order to counterbalance a relatively weak proportionality standard, a more robust application of due process requirements is needed (section 2.2.2).38 In Gerard’s view, the weaknesses of the commitment procedure call for a series of corrective measures. First of all, two systematic and enforceable obligations should be imposed on the Commission to ameliorate the ‘due process deficit’: (i) firms that propose commitments should be provided with full non-confidential versions of the observations of third parties who respond to the market test; and (ii) they should furthermore be provided with a reasoned decision if the Commission opts to reject market-tested commitments. Second, in order to restore the voluntary character of commitments, Gerard advocates immunity from fines for firms admitted to commitment proceedings if the commitment process fails and the Commission ultimately finds an infringement.39 Removing the sword from above the head of the undertaking concerned would in his view ‘radically contribute to closing the legitimacy gap … because it would force the Commission to “weigh carefully” the optimal character of negotiated solutions and would create a more balanced

as overstated—in particular given the enforceable soft law statements the Commission has made as regards the manner in which it conducts the commitment procedure. See Monti, ‘Alrosa and Commitment Decisions’, above n 32. 38 With regard to the principle of proportionality, Gerard accepts that it may be justified to apply a more relaxed standard in commitment cases; but not necessarily to the extent that the case law suggests. And his willingness to accept a looser proportionality standard is conditioned on the reforms he proposes as far as due process is concerned (see the main text below). He thus states, at page 172, that if such reforms are properly implemented, it is ‘not per se unacceptable for the Commission to benefit from reasonable flexibility in the application of the proportionality principle in commitment cases’. 39 This proposal was endorsed by Jenny in Frédéric Jenny, ‘Worst Decision of the European Court of Justice: The Alrosa Judgment in Context and the Future of Commitment Decisions’, in Barry Hawk, ed, International Antitrust Law and Policy: Fordham Competition Law 2014, Juris Publishing, 2015, chapter 16. Also published as 38 Fordham International Law Journal 701 (2015).

Introduction xlv framework for the negotiation of (the final set of) commitments’.40 Finally, Gerard calls for the further strengthening of judicial review by the EU Courts of the Commission’s infringement decisions so that full appellate jurisdiction is effectively exercised at least where the Commission has imposed a fine or periodic penalty payment. (In this he joins several others among the Workshop participants.) Beyond the other merits of robust review, this would help to re-establish a genuine choice for firms between offering to settle by way of commitments and, by contrast, fighting to the end.41 To the above-described proposals one could add other conceivable reforms as well, even if some of them might be unrealistic for the time being. Considering that in the next few years we are liable to witness a significant revision of Regulation 1/2003, an interesting opportunity for improvements may present itself. Will it be missed, is the question. Giorgio Monti, ‘Behavioural Remedies for Antitrust Infringements— Opportunities and Limitations’. As the title suggests, Monti’s essay concerns behavioral remedies imposed by the European Commission in the specific scenario where an infringement has been found under Article 7 of Regulation 1/2003.42 The aim of the discussion is to put more focus on the role of behavioral remedies in this setting, their scope, the (judge-enforced) principles that the Commission follows in designing them, and their practical and procedural dimensions. With regard to their role, Monti points out that they are to be understood as restoring the possibility of competition by giving the undertaking concerned the guidance needed to secure compliance—guidance

40 Page 181 (citation omitted). The need for a specific bar is however questioned by Monti: ‘[A] party who is fined after an unsuccessful, good faith attempt to make commitments may have a basis for a successful appeal against the fine. The issue remains to be explored and it may well be that in the future the Court will take the view that accepting to go down the commitment path creates a legitimate expectation that there will not be a fine’. Monti, above n 32. 41 Monti (above n 32) suggests that the oft-stated proposition that judicial review of Commission infringement decisions has been weak is overstated. He explains that such claims are largely based on the Commission’s ‘winning streak’ in Article 102 cases, and that the Commission’s success may be attributed above all to the expansive substantive scope of the prohibition. This latter observation is undoubtedly true, although one might still find it significant that the substantially greater financial impact of an infringement decision that emerged in the last dozen years did not of itself seem to prompt a more exacting review of Commission findings to temper the consequences of the prohibition. 42 Commitments in the context of the Article 9 procedure are discussed incidentally, that is, by reference to their systemic relationship to remedies imposed under Article 7, and by wondering whether some of the constraints governing remedies under each of the two different procedures (the term ‘remedies’ is used here in a non-technical sense since there is no finding of infringement in the Article 9 context) might provide inspiration—in terms of discipline in the case of Article 9, or in terms of transparency in the case of Article 7. For his recent and more direct commentary on Article 9 commitments as an enforcement tool, see Monti, ‘Alrosa and Commitment Decisions’, above n 32 (casting doubt on many of the criticisms that have been leveled against the commitment procedure). The subject of commitment decisions is also discussed in other contributions to this volume, specifically in the chapters of Ioannis Lianos and Damien Gerard (see above).

xlvi Introduction ensuring that the undertaking does not re-offend.43 This role is distinct from, and may to some extent be complementary to, a role of pure deterrence, general or specific. It is also distinct from a different vision of remedies which goes farther and gives the enforcer the latitude to address anticompetitive effects caused by the impugned conduct (other than the effects that would persist but for the cessation of the infringement), and to shape the dynamics of a market in pursuit of a welfare-optimizing outcome.44 Due in part to these distinctions, Monti observes that the Commission’s use of Article 7 remedies do not fit within either the ‘crime/tort’ model or the ‘administrative’ model of enforcement, using the terminology adopted by Crane in his work. Since there is no handy name available, Monti simply describes these remedies as being of a hybrid nature. With regard to the principles to which the Commission must adhere in designing a remedy under Article 7, they tend to follow from the above conceptualization. The first is that such a remedy must have a direct link to the infringement; Article 7 may not be used to pursue broader agendas. The second (which, as applied, may impose constraints similar to those implied by the first) is that the remedy must comply with the rights of defense. That is to say, the infringement decision must make clear how the remedy addresses the infringement; it may not address, for example, competition concerns that the decision leaves unidentified. The third principle is that of proportionality: the remedy must not exceed what is necessary to restore the opportunity for competition in the market. The fourth and final principle is equal treatment, which would be relevant, in particular, where access or supply remedies are to be designed. If the remedy consists of permission to cross a unique bridge, for example, that permission may have to be given on a non-discriminatory basis. This principle may in some cases require a partial derogation from the others but, as Monti explains, it mainly concerns competitive conditions downstream. Overall, the foregoing principles appear to constrain the Commission’s remedial powers—which may perhaps come as some relief when considered in the context of the concerns raised in other chapters of this book. On the other hand, Monti also points out that these constraints may

43 In section 2 of the chapter Monti breaks down this conception into three related functions: an effective remedy under Article 7 will achieve the termination of the offending practice, prevent its recurrence and restore the opportunity for competition in the market. The significance of the terms ‘opportunity for competition’ and, similarly, ‘possibility for competition’ (as opposed to restoring competition) is that illegal conduct may well have caused various effects that are unlikely to be cancelled merely by ensuring that it ceases and does not reappear. Monti gives the examples of the financial strength, market position and rate of innovation of parties to a cartel, or advantages that may have been gained from exclusionary conduct by a dominant firm. In the latter case, the remedy is aimed at restoring the excluded rival, so far as possible, to its ex ante position. 44 This particular observation regarding the way remedies under Article 7 are understood may be contrasted with the normative preference for a welfare-optimizing approach to remedies that is evident in the next chapter by Frank Maier-Rigaud (see below).

Introduction xlvii be significant factors in the preference for the looser discipline of Article 9 procedures where broader, market-shaping remedies are desired.45 Finally and briefly, Monti discusses several practical considerations and considers the benefits and costs of introducing improvements to the applicable procedures. These include, among other issues, the occasional need for a defendant to cooperate with the Commission even as it continues to fight its corner and deny any wrongdoing, or the use of certain techniques to facilitate monitoring and compliance with imposed remedies. In the context of his observations on practical considerations, he addresses a misunderstanding, on the basis of a certain narrow reading of the CFI’s case law, as concerns the Commission’s authority to impose affirmative obligations such as entering into contractual supply obligations under Article 101. Monti shows that a remedy of this kind is possible so long as the relevant principles, including in particular the principle of proportionality, are respected. While it could be objected that harmed parties may pursue such remedies through private law means and that the imposition of those remedies by the Commission is thus disproportionate, Monti suggests that significant advantages might potentially be gained if the affirmative remedies as well as any necessary compensation orders were rolled into the public law procedure. In other words, the public authority would protect relevant private interests in parallel with its defense of the common interest of the EU. Frank Maier-Rigaud, ‘Behavioural versus Structural Remedies in EU Competition Law’. If one were to consider competition law remedies from behind a veil of ignorance, perhaps no type of remedy would appear to have inherent superiority over others. Instead of assuming a hierarchy the innocent observer might conclude that the best approach for an economic infraction would be to impose whatever remedy or mix of remedies is meet for the case. However, it would seem that most lawyers tend to regard remedies through two inter-related and socially constructed lenses which create an aversion to structural remedies in the non-merger context. The first lens derives from the Lockean and classical liberal tradition that attaches high value to property rights, which are to be protected, as a general rule, from public intervention—irrespective of whether the intervention provides for compensation or outright confiscation, though the latter is certainly beyond the pale. The much longer tradition of coercive taxation by the state as well as other principles of sovereign power have always made private property rights relative rather than absolute, but nevertheless such rights are generally strong and dearly cherished. The second lens derives from being socialized as lawyers. A non-lawyer, at least if she is also innocent of the liberal tradition,

45 On this latter point, see page 196. A variety of factors favoring resort to the commitment procedure are also discussed in section 4 of the chapter by Lianos (see above), for example at pages 130–131.

xlviii Introduction might be more inclined to accept the social function of property rights;46 but for many lawyers, rights are, again, to be vigorously defended. The hierarchy that appears so natural through these lenses has been reinforced through the long and chequered past of structural remedies in the antitrust experience of the United States.47 In the EU context, it also appears to be reinforced by the wording of Article 7(1) and Recital 12 of Regulation 1/2003. Maier-Rigaud, an economist, challenges this point of view that assumes structural remedies to be remedies of last resort. Unsurprisingly, his analysis, which finds odd quirks in the EU approach (also taking account of the practice under Article 9 of the said Regulation), is framed by the assumption that EU competition policy should be determined primarily on the basis of economic effects. From that vantage point he endeavors to ‘rehabilitate’ structural remedies; but beyond establishing the desirability of using them in appropriate cases, he also argues that the legal path to the use of structural remedies is not as laden with obstacles as may be assumed. His first order of business is to arrive at a definition of a structural remedy, and he proposes that it should be understood as a measure that changes the structure of a firm by way of a transfer of tangible or intangible property rights, following which transfer the prior and present owners of those rights have no ongoing r elationship.48 Such a remedy removes the incentive or the means to revert to the anticompetitive conduct in question, and therefore requires no ex post monitoring. In contrast to the regulatory, market-constraining nature of behavioral remedies,49 the affirmative case for structural remedies stems from the fact that they are ‘within the logic of the market system and allow an efficient adaptation to changing market conditions. … [To an observer] they cannot be distinguished from the usual workings of a market system in which mergers and divestitures are a characteristic feature of normal market developments. In that sense they take full advantage of market allocation dynamics …’50 46 The social function of property, a concept originally propagated by the Bordeaux scholar Léon Duguit (1859–1928), is mainly associated, through expansive and politicized interpretation, with Latin American constitutionalism. See, eg Thomas Boccon-Gibod, ‘Duguit, et après? Droit, propriété et rapports sociaux’, 28 Revue Internationale de Droit Economique 285 (2014); MC Mirow, ‘Origins of the Social Function of Property in Chile’, 80 Fordham Law Review 1183 (2011). 47 The literature is voluminous. To cite only two authorities, see William Kovacic, ‘Designing Antitrust Remedies for Dominant Firm Misconduct’, 31 Connecticut Law Review 1285 (1999); Richard Epstein, ‘Monopolization Follies: The Dangers of Structural Remedies under Section 2 of the Sherman Act’, 76 Antitrust Law Journal 205 (2009) (the title being, as Epstein admits, ‘a tad unkind’). 48 See pages 209–210. 49 As Maier-Rigaud states, ‘it is surprising to see how widespread behavioural remedies are in a competition law context as they are fundamentally at odds with the idea that a decentralised competitive process … alone allows the efficient allocation of resources and the highest welfare’. Page 212, footnote 16. 50 Page 211. For observations regarding the advantages of behavioral remedies (in particular given the limitations of a system such as that of the EU in which fines are a central pillar), while recognizing their disadvantages, see the chapter by Giorgio Monti, for example at pages 186–187.

Introduction xlix Moreover and ironically, a structural remedy may be ‘advantageous’ for a firm deprived of its assets in the sense that it can get on with its business unencumbered by ongoing regulatory commitments. Section 3 of the chapter is devoted to a textual analysis of Article 7(1) of Regulation 1/2003, the aim of which is to derive an ‘economically sound interpretation’ of that provision which is also consistent with the original intent behind it. The starting point is the notion that an infringement can only be brought effectively to an end if no significant incentive to re-offend remains—and in this regard it is not assumed that the threat of future fines in case of infringement obviates any concern about incentives.51 Maier-Rigaud examines the particular elements of Article 7(1), including proportionality, necessity and ‘equal effectiveness’, and then arranges them into a graphically depicted filtering process. His discussion of the successive filters leads him to the conclusion that Regulation 1/2003 does not establish any preference for behavioral over structural remedies. ‘On the contrary’, he says, ‘the structural or behavioural nature of a remedy is immaterial as long as such remedies are not equally effective’.52 He concludes that the meaning of the key proviso would be clearer if the wording were adapted to the effect that behavioral remedies may be imposed only if there is no more effective structural remedy or where any equally effective structural remedy is equally or more burdensome than the behavioral remedy for the undertaking concerned.

C. Agencies as amicus curiae Stephen Calkins, ‘The Antitrust Conversation (Continued)’. The ‘conversation’ in this context is one between the judiciary and amici curiae, a category which typically includes civil society but which in some contexts might also include actors such as sovereign states or supranational institutions. One immediate observation that may be made here is that there is no guarantee that amici curiae represent the public interest, any more than it is guaranteed that the interests of the litigants themselves will coincide with the public interest. Nevertheless, and on the optimistic side, where a court permits a range of amici to participate, this may open up the process of deliberation to a broader spectrum of legal and policy views which may enhance the quality of the marketplace of ideas within the boundaries of the litigation. To a certain extent, a conscientious court can also efficiently screen amici to reduce the risks of being diverted by rent-seekers and time-wasters. Inevitably, a cost-benefit analysis is made as to whether an amicus brings marginal value to the dispute.

See pages 216–217, footnote 29. Page 220.

51 52

l Introduction The bulk of this chapter concerns US practice, and above all the practice of the US Supreme Court. Calkins provides a rather thorough overview of amici interventions in the most prominent US antitrust disputes, but also of other types of cases (chiefly those raising constitutional issues) decided in the last dozen years. Do these interventions sway the Justices? While one cannot speak with certainty in terms of causation,53 Calkins shows that amicus briefs may be of distinct value to a court. First, he points out that amicus briefs are routinely used by the Court and its clerks to sift through the tiny percentage of cases in respect of which petition for certiorari will be granted. Furthermore, once a case is selected for litigation, amicus briefs on the merits may influence the Court if, for example, it uncovers a latent issue of significance; or if it demonstrates the wider implications of a judgment beyond the interests of the parties, possibly by highlighting a risk of unintended consequences and a corresponding need to restrict (for example) the scope of a judgment’s ratio decidendi. Calkins also explains the role amicus briefs have played as litigation played out in specific Supreme Court cases; and he emphasizes the privileged status enjoyed by briefs received from the US Solicitor General (both at the certiorari stage and on the merits), which reflect the position of the incumbent executive branch of the federal government.54 Having discussed a variety of aspects of the US practice, Calkins proceeds in section 4 of the chapter to add a summary of practice elsewhere. In general, the role of amici curiae outside the US has been substantially more limited. In certain European settings, for example, including in litigation before the EU Courts, the vehicle for presenting arguments to judges is formal intervention as an interested third party. In the EU, leaving aside the privileged positions of the Member States and the EU Institutions, the possibility of intervening formally is restrictive.55 (Of course, if an organization or individual wishes to make its own public statements and present them to the public as observations relevant for the EU Courts, nothing prevents it or him from doing so, but this is merely informal practice.56) One is therefore left to wonder whether, given the relatively

53 For example, as a general matter, the fact that an amicus brief is cited in a judgment is hardly a sound basis, at least absent additional corroborative indicia, for hard and fast conclusions as to whether the brief has been decisive. 54 The Solicitor General is by no means a member of the Supreme Court (and may thus be distinguished from an Advocate General in the EU context) but is the third-highest ranking official of the US Department of Justice. As a frequent repeat player, representing the US government in numerous cases each year, he is essentially in a continual dialogue with the Court on issues affecting the government. For further information, see the SG’s web site, http://www.justice.gov/osg/about-office-1. 55 As Calkins points out, intervention before the Court of Justice is governed fundamentally by Article 40 of the Statute of the Court of Justice, available at http://curia.europa.eu/ jcms/upload/docs/application/pdf/2012-10/staut_cons_en.pdf. 56 For an example from Amnesty International in the context of human rights litigation (minimum standards for classification and status of third-country nationals as refugees under Council Directive 2004/83/EC), see Observations by Amnesty International and

Introduction li c umbersome nature of intervention, the Court of Justice and other courts would benefit from a reform giving them the (non-mandatory) flexibility to open the door to amicus submissions. It may be true that the Advocate General already provides the Court with a well-considered and unbiased point of view, and that is certainly a valuable service; but he can scarcely be regarded as a substitute, as a general matter, for amici curiae. At the level of the courts of the Member States, while US-style amicus submissions by private parties are not yet rapidly gaining traction, Calkins recalls that the European Commission is, and certain NCAs are, developing an interesting and fairly active role as ‘amici’ in private litigation by virtue of Article 15 of Regulation 1/2003.57 As far as the Commission and NCAs are concerned he rightly expects this trend to continue as private litigation in Europe continues to grow. But he also predicts that global trends toward larger-scale, institutionalized litigation will eventually lead to greater amicus involvement by private parties. On the basis of the US record, he would welcome such a development: ‘Just as one hopes that Europe can learn from the US’s experience and achieve the best of private enforcement, one hopes it can learn from the US’s experience with amici.’58

D. Dealing with public measures that restrict competition José Luis Buendía Sierra, ‘Enforcement of Article 106(1) TFEU by the European Commission and the EU Courts’. In this chapter, Buendía Sierra reviews a number of issues in connection with the enforcement of Article 106(1), which, as EU lawyers know, is typically applied in tandem with Article 102, the prohibition of abuse of dominance. One significant part of the chapter59 is Buendía’s endorsement of the judgment of the Court of Justice in Greek Lignite,60 one of a small number of Article 106(1) cases that

the International Commission of Jurists on the case X, Y and Z v Minister voor Immigratie, Integratie en Asiel (C‑199/12, C‑200/12 and C‑201/12) following the Opinion of Advocate General Sharpston of 11 July 2013, https://www.amnesty.org/en/documents/ POL33/003/2013/en/. 57 For useful information concerning the Commission’s ‘amicus’ practice, see http:// ec.europa.eu/competition/court/antitrust_amicus_curiae.html. 58 Page 259. 59 The chapter addresses several issues not discussed here, including for example the Commission’s case practice, or the ECJ’s application of the threshold matter of whether the entity holding exclusive or special rights is an undertaking, or again the ECJ’s application (or non-application) of Article 106(1) in conjunction with Article 49 TFEU and with Article 56 TFEU. 60 Case C-553/12 P Commission v Dimosia Epicheirisi Ilektrismou AE (DEI), EU:C:2014:2083 (setting aside the earlier judgment of the General Court and remanding the case to that court).

lii Introduction the Court has had the occasion to decide in the last decade.61 The paucity of cases may be related in part, though not entirely, to what Buendía suggests is an untoward passivity on the part of the European Commission, which could presumably seek out more cases if it had the appetite for it. As it transpires, however, the Commission—acting with extraordinary and anomalous discretion in this context—seldom invites Article 106(1) to the ‘family table’.62 As he has done elsewhere,63 Buendía on the one hand questions the propriety of this judge-approved discretion and on the other criticizes the Commission for failing to use it to bring more cases.64 With regard to Greek Lignite, a case that did arise from a Commission investigation, Buendía regards the ECJ’s judgment as consistent with the understanding, correct in his view, that if the Commission can establish that a state measure leads to effects equivalent to an abuse—for example because it effectively extends a dominant position from one market to another—the Commission (and by implication a plaintiff) is not required to show any actual or potential abuse to which the privileged undertaking has been led or could be led by the state measure.65 The contested measure in such a case creates ‘inequality of opportunity’: the privileged undertaking is favored, while other undertakings are placed at a disadvantage. In essence, this egalitarian position divorces the responsibility of the Member State from the scope of 61 Other than Greek Lignite, the most significant Article 106 case during that recent time frame is probably Case C-49/07 Motosykletistiki Omospondia Ellados NPID (MOTOE) v Elliniko Dimosio [2008] ECR I-4863. For cases where the joint application of Article 106(1) and Article 102 TFEU was of some (at least potential) relevance but either did not constitute the core issue or was left undetermined for lack of admissibility, see also Case C-437/09 AG2R Prévoyance v Beadout Père et Fils SARL [2011] ECR I-973, paras 68–72; Case C-250/06 United Pan-Europe Communications Belgium SA and Others v Belgium [2007] ECR I-11135, paras 18–23; Case C-451/03 Servizi Ausiliari Dottori Commercialisti Srl v Giuseppe Calafiori [2006] ECR I-2941, paras 21–26; Case C-295/05 Asociación Nacional de Empresas Forestales (Asemfo) v Transformación Agraria SA (Tragsa) and Administración del Estado [2007] ECR I-299, paras 39–45. 62 Page 280. In this regard the objection is that, in the Article 106(3) context, it is practically impossible for a party complaining about a state measure to challenge the Commission’s decision not to act on the complaint. 63 See José Luis Buendía Sierra, ‘Exclusive or Special Rights under Article 106 TFEU: An Overview of EU and National Case Law’, eCompetitions, article N° 44436, 20 March 2012. 64 On the other hand, if one harks back to around 1991, when Article 106(1) TFEU (then Article 90 EEC) began to emerge as an important competition law provision, the formative cases and the later cases in which the Court developed the application of that provision materialized as part of the preliminary ruling procedure (or in Buendía’s terms, indirect enforcement), and hence as part of national litigation. Preliminary references in this context have become relatively more scarce, possibly for want of relevant disputes (a surprising lull, given that Member States have certainly not abstained from adopting or maintaining measures with implications for competition) and perhaps in part because cases that should be referred are not. Further, as Buendía points out, some cases dealing with Article 106(3) directives are at bottom also Article 106(1) cases given that such directives are meant to vindicate the policy expressed in the latter provision. 65 Similarly, see the Opinion of Advocate General Wathelet in Case C-553/12 Commission v DEI, EU:C:2013:807, paras 55–65.

Introduction liii the prohibition contained in the relevant competition rule, Article 102.66 And, as Buendía explains, the judgment of the ECJ is not merely a vindication of its old case law; it features an innovative and rather expansive move. The Court followed the effects doctrine to its logical consequence. Indeed, the judgment states that any state measure producing anticompetitive consequences would infringe Articles 106 and 102 TFEU. Contrary to previous judgments, this statement is not confined to the ‘extension of a dominant position’ from one market to another caused by an exclusive or special right. It also applies to the original grant of a stand-alone exclusive or special right to a company.67

In Buendía’s view, the consummation of the effects doctrine makes sense because it results in a harmonization of the application of Article 106(1) and the application of the Treaty’s free movement rules—where the scope of the prohibitions tends to be broad and where the burden thus often rests with the Member State to justify its actions.68 For my own part I have been more skeptical of the ECJ’s judgment in this case,69 but such criticism is now academic. The Court has taken a firm stand, and as Buendía notes, Article 106 is pregnant with ‘enormous possibilities’ and ‘far-reaching content’.70 What fruit and how much of it the provision will bear are to be seen. Daniel Crane, ‘Hard Look Review of Anticompetitive State Action’. Crane’s chapter extends the subject of public restraints of competition by introducing a semi-abstract and normative evaluation of different possible models of judicial antidotes to abusive regulation. The first two models—the ‘representation reinforcement’ model (ie a particular form of political accountability model, drawn from democratic theory and mapped onto the Parker-Midcal experience in the US) and the ‘substantive review’ model (ie the EU model)— are found to be flawed, and accordingly they are rejected.71 Crane instead advocates a third model, the hard look review model, which is based on the

66 There is thus an asymmetry here compared to the ECJ’s jurisprudence relating to rticle 4(3) TEU and Article 101 TFEU. See in particular Case C-2/91 Wolf Meng [1993] A ECR I-5751 and Case C-245/91 Ohra Schadeverzekeringen NV [1993] ECR I-5851. 67 Page 288. 68 Buendía has long argued that the old Corbeau case—although it was ostensibly an Article 106(2) case, and though the Court did not enter into an analysis under Article 106(1)—essentially signaled a convergence between the competition rules in this context and the free movement rules. 69 See Mel Marquis, ‘The State of State Action in EU Competition Law and a National Competition Strategy for China’, in Niels Philipsen, Stefan Weishaar and Guangdong Xu, eds., Market Integration: The EU Experience and Implications for Regulatory Reform in China, Springer Verlag, 2015, chapter 2. 70 Pages 288–289. 71 The two basic models are discussed in further detail in Daniel Crane, ‘Judicial Review of Anticompetitive State Action: Two Models in Comparative Perspective’, 1 Journal of Antitrust Enforcement 418 (2013).

liv Introduction ractice of US administrative courts imposing a certain procedural discipline p on regulatory agencies under their control. The principal features of hard look review are described in section 2 of the chapter. They include first of all the requirement that an agency build a public record in its decision-making process, on the basis of which it must prepare an adequately reasoned, internally consistent decision. In explaining its decision the agency must explain why obvious alternative solutions were passed over and must show how ‘significant’ public comments were taken into account. However, the nature of the scrutiny remains procedural and not substantive—the court does not itself balance interests, and it does not substitute judgment (for example, by insisting on the least restrictive alternative). The core objective of this approach is to ensure that the decision-making process is publicly minded, and that it is not driven by private interests to the detriment of the common good. Underlying Crane’s analysis are his assumptions that the representation reinforcement model and the substantive review model are deficient. One could perhaps raise objections with regard to the assumed weaknesses of the substantive review model in certain institutional contexts. In particular, the analogy between the era of ‘Lochnerism’ and substantive due process in the early 20th Century in the United States, on the one hand, and the relatively rigorous scrutiny by the ECJ of anticompetitive legislation in the EU context on the other hand, may be deceiving. The legitimacy deficit, as it were, that tainted the judicial activism of the Lochner era72 was closely linked to the fact that substantive due process was developed according to a certain ideological interpretation of the US Constitution by the US courts.73 By inferential reasoning, it was held in a number of cases that regulations imposing, inter alia, health and safety standards in various sectors were contrary to the constitutional clauses prohibiting the government from depriving

72 Lochner v New York, 198 U.S. 45 (1905). The case is reviewed in, among others, Daniel Crane, ‘Lochnerian Antitrust’, 1 New York University Journal of Law and Liberty 496, 498–501 (2005). 73 While the word ‘ideological’ can appear to be pejorative, it may be useful to recall that most courts of significant jurisdiction, duty-bound as they are to decide consequential disputes, are seldom capable of escaping entirely from their conscious and unconscious ideological inclinations. To a certain extent, something of a ‘third man’ argument may apply here: if a judge (assuming she is not a perfect Hercules) seeks, for the sake of scruple, to distance herself from her own conscious ideology-driven views, a certain ideology—albeit a different one—may creep back into the process of deciding how to achieve ‘objectivity’ and overcome the biases of which she is aware. Despite the reference to US courts in the age of Lochner espousing a certain ideology, one may find ideology permeating the choices of the courts that later repudiated Lochner, not to mention, undoubtedly, those of innumerable other courts. It is fine for Justice Holmes to say that ‘a constitution is not intended to embody a particular economic theory, whether of paternalism and the organic relation of the citizen to the State or of laissez faire’, and that it is ‘made for people of fundamentally differing views’, but an ideological perspective may be unearthed here too without much difficulty. The validity of the statement itself may also perhaps be questioned, depending on how narrowly or broadly it is construed, but that discussion may be omitted here.

Introduction lv persons of their liberty and property without due process of law. The hermeneutic approach of the Supreme Court in that context relied on something akin (speaking anachronistically) to penumbras and emanations—there is certainly no explicit constitutional rule forbidding the federal or state governments from adopting anticompetitive measures, and even the long history of the ‘dormant’ Commerce Clause was based on an extrapolative reading of the Constitution.74 One may therefore plausibly argue that the Lochner saga bears little resemblance to the judicial control of anticompetitive state measures in Europe, in particular because Article 106 TFEU is quite explicit that Member States must avoid adopting or maintaining (unjustified) anticompetitive measures where a public or privileged undertaking is concerned, and because Article 107 TFEU explicitly prohibits (unjustified) state aid.75 In the contexts of these Articles, it is submitted, there is no countermajoritarian difficulty with judicial intervention or, as the case may be, with intervention in the first instance by the European Commission, as the relevant specific competences have been built into primary law.76 Admittedly and perhaps ironically, the ECJ’s Van Eycke77 line of case law rests on inferential reasoning,78 but the subsequent limitations resulting from cases such as Ohra and Meng79 may be interpreted as a move of judicial conscience. Ultimately, the judicially 74 Specifically, the relevant clause is Article I, Section 8, which in pertinent part provides that: ‘The Congress shall have the Power … To regulate Commerce with foreign Nations, and among the several States, and with the Indian Tribes’. It is of course recognized that local protectionism was an important catalyst behind the Philadelphia Convention in 1787. See, eg Julian Eule, ‘Laying the Dormant Commerce Clause to Rest’, 91 Yale Law Journal 425, 429–435 (1982). 75 The twice-appearing word ‘unjustified’ in this sentence punctuates the general point being made about the competence and hence the legitimate authority conferred by the Treaty upon the Court of Justice to review public restraints of competition—and in so doing to arbitrate if necessary between potentially conflicting policy interests, and potentially (certainly not automatically) to displace legislative or regulatory choices. This feature of supranational governance implies a judicial role which causes discomfort for many both within and outside Europe. See, for example, Crane’s discussion at pages 317–319 (which, inter alia: draws a distinction between competition as a value of somewhat lesser significance and other interests of ‘moral principle’ such as gender classifications and limitations on free speech; and maintains that judges are ill-suited ‘to engage in a sort of open-ended cost-benefit analysis that would sometimes result in the judges substituting their view of what interests are important for those in the executive or legislative branches’). 76 While the prohibition contained in Article 106 is clearly enshrined in the Treaty, the direct effect of Article 106 (with the possible exception of the second sentence of Article 106(2), since the direct effect of that sentence has not yet been established with certainty) is a separate issue. The ability of claimants in national proceedings to invoke Article 106 has been confirmed by the hermeneutical practice of the ECJ, and indeed the very foundations for the direct effect of sufficiently clear and precise Treaty provisions were a product of the judicially driven constitutionalization process of the 1960s. To that extent, the debate about the ECJ’s judicial activism, documented in many classic commentaries especially since the 1980s (and originally often linked to what was in those days the diminishing control of the Council and the rise of supranational governance), has never really subsided. 77 Case 267/86 Pascal Van Eycke v ASPA NV [1988] ECR 4769. 78 The ironic aspect is that while the Treaty is explicitly concerned with public measures in scenarios involving public and privileged undertakings, there is no such explicit apprehension with regard to cartels. 79 Both cases are cited above n 66.

lvi Introduction determined expansive nature of the duty of sincere cooperation (Article 4(3) TEU), which serves inter alia as a gap-filler, may still invite charges of judicial activism, but these days the main battlegrounds in relation to public restraints (as exemplified by the above discussion of the chapter by Buendía Sierra) are Article 106 and the state aid regime. Having said all of that, it must be recalled that Crane’s discussion is semi-abstract; he recognizes that the hard look approach may not fit all institutional environments.80 Likewise, the above objection relating to Europe’s ‘constitutional’ decision to regulate public restraints as a matter of explicit primary law is jurisdiction-specific and would be irrelevant, for better or worse, in many actual and future contexts.

PART 2: Legitimate Enforcement of Competition Law Renato Nazzini, ‘Parallel Proceedings in EU Competition Law: Rethinking Ne Bis In Idem as a Limiting Principle’. The problems of concurrent and consecutive competition enforcement actions have been latent in Europe since the 1960s,81 but they have intensified since the adoption of Regulation 1/2003 and the related maturation of Europe’s multi-level enforcement network.82 The fact that Article 50 of the EU Charter of Fundamental Rights (ne bis in idem) is binding primary EU law merely punctuates the matter.83 Accordingly, the relationship between effective and legitimate enforcement is here once again front and center, and it is not surprising to find that Nazzini believes that the two objectives are inadequately balanced. Discretionary prosecutorial restraint, he says, does not sufficiently protect against the risk of multiple proceedings and decisions against the same undertakings for the same conduct. If the ne bis in idem principle were (de lege ferenda) construed properly, however, it would effectively limit (ex post) the ‘vexatious and inefficient exercise of concurrent jurisdiction in competition matters within the Union’.84 See page 325. See Case 14-68 Walt Wilhelm and others v Bundeskartellamt [1969] ECR 1. 82 See, eg Giorgio Monti, ‘Managing Decentralized Antitrust Enforcement: Toshiba’, 51 Common Market Law Review 261 (2014). 83 Besides the protection that follows from the general principles of EU law and indirectly from the ECHR (Protocol 7, Article 4), as Nazzini points out, the ne bis in idem principle is expressed in various other legal instruments such as Chapter 3 of the Convention implementing the Schengen Agreement and Article 3(2) of the Council Framework Decision of 13 June 2002 on the European Arrest Warrant. 84 Page 358. In some measure the risk of multiple proceedings is compounded by the ECJ’s holding in Tele2 Polska that NCAs are precluded under Regulation 1/2003 from finding positively that an undertaking has not committed an infraction under the EU competition rules. The argument that the holding in that case justifies a more flexible interpretation of ne bis in idem was rejected by the General Court in Case T-402/13 Orange v Commission, EU:T:2014:991. The latter judgment was not appealed. 80 81

Introduction lvii The Court of Justice should thus substantially modify the approach it has taken to ne bis in idem in connection with competition law proceedings in the EU’s system of parallel competences, and should tighten up the applicable limiting principles in that regard.85 To suggest how this could be done, Nazzini lays out a six-part survey of the role of ne bis in idem in a variety of contexts, mainly as it pertains to preclusion or non-preclusion of multiple proceedings within the EU.86 A number of points emerge from this survey. For example, Nazzini shows that the ECJ’s restrictive reading of ne bis in idem in the context of parallel or multiple competition law proceedings has in fact resulted in a fragmentation of the Court’s own case law, when compared to the jurisprudence in the European Arrest Warrant and the Schengen contexts.87 In the latter contexts, the Court does not impose the high hurdle of demonstrating that a second prosecution is driven by the same ‘legal interest’; identity of the relevant conduct is the core issue. More fundamentally than the fragmentation issue, it may well be that the bar has been placed so high that the ne bis in idem principle no longer effectively serves its function of securing repose and protecting fundamental rights. Nazzini suggests that ne bis under Article 50 of the EU Charter and as a general principle of EU law should be interpreted in such a way that two infringements constitute the same ‘offense’ when they arise from the same facts or facts inextricably linked in terms of space, time and subject matter. Drawing inspiration from Advocate General Sharpston and from the ECJ’s case law in the Schengen Agreement context, Nazzini also offers a practical criterion by which the ne bis in idem principle might be applied. According to this criterion, a distinction should be made between decisions to close a case that are based on a preliminary and summary investigation, which should not bar further proceedings [in the ECN context], and decisions to close a case that follow a significant consideration of the merits of the case and are adopted in a procedure where the defendant has had a full opportunity to exercise his rights of defence, which should bar further proceedings whatever the formal classification of the final decision.88

85 The arguments are also presented in Renato Nazzini, ‘Fundamental rights beyond legal positivism: rethinking the ne bis in idem principle in EU competition law’, 2 Journal of Antitrust Enforcement 270 (2014). 86 Nazzini also addresses the scenario where a non-EU authority has already condemned (or ‘acquitted’) an undertaking, and where a competition authority within the EU investigates the same undertaking in relation to the same facts. However, in this global context— taking account of public international law—there is no precise equivalent to the ne bis in idem principle, and EU law does not recognize ne bis protection in such scenarios. Nazzini accepts and endorses the state of EU law on this point. 87 For further discussion of the ne bis principle in these contexts, see John Vervaele, ‘‘Schengen and Charter-related ne bis in idem protection in the Area of Freedom, Security and Justice: M and Zoran Spasic’, 52 Common Market Law Review 1339 (2015). 88 Page 360 (citing the Opinion of Advocate General Sharpston in Case C-467/04 Giuseppe Francesco Gasparini [2006] ECR I-9199, para 96 (suggesting preclusion where

lviii Introduction In Nazzini’s view, this two-part test (assessment of the merits; full exercise of the rights of defense) should also be applied for the purpose of determining whether a commitment decision under Article 9 of Regulation 1/2003 precludes later proceedings on ne bis in idem grounds. Under this approach, and in cases where the test is met, fundamental rights would effectively override the scheme of the Regulation and its recitals, which assume that Article 9 has no such preclusive effects. The test would also have implications for scenarios where the EU Courts invalidate an infringement decision of the Commission on purely procedural grounds, which opens the possibility for the Commission to reopen its proceedings, correct the defects and readopt a decision. Here Nazzini suggests an intermediate position whereby a second round of proceedings could ensue if necessary to secure effective enforcement, and if the procedural errors were either not attributable to the competition authority or excusable, which is to say they must not have been caused by the authority’s gross negligence or bad faith. Applying a nuanced version of ne bis in this way would seem likely to encourage an authority to be more meticulous the first time around. That might also conceivably (and all else being equal) lengthen the average duration of proceedings but, if so, the tradeoff may be worthwhile. Nazzini’s conclusions on these points and others are summarized conveniently in section 4 of his chapter. Wolfgang Kirchhoff, ‘Reflections on Parallel Enforcement, Fundamental Rights and the Rule of Law in the Competition Law Context’. As noted earlier,89 Germany is something of a poster child as regards the need for harmonized rules on sanctions. Perhaps the most prominent issue in that context is the partially divergent position on parental liability. In this chapter, Judge Kirchhoff does not discuss parental liability as such but he discusses the related issue of the extent to which the liability of an undertaking to pay fines survives or is extinguished when the undertaking is transformed or restructured in the sense of German company law. Other issues relevant to the relationship between effective enforcement and fundamental rights protection are also discussed, such as the principle of nullum crimen, nulla poena sine praevia lege poenali and the variant of that principle contained in Article 7(1) ECHR. With regard to successor liability following a transformation of an undertaking, Kirchhoff refers to the restrictive jurisprudence of the German F ederal Court of Justice (Bundesgerichtshof), and in particular its interpretation of Section 30 of the German Act on Regulatory Offences.90 According to that

the first proceedings involved ‘significant consideration’ of the merits); and citing Case C-469/03 Filomeno Mario Maraglia [2005] ECR I-2009, paras 30 and 34 and Case C-150/05 Jeon Leon Van Straaten v Netherlands [2006] ECR I-9327, paras 56–60). See above the summary of the chapter by Konrad Ost. Section 30 governs regulatory fines imposed on legal persons and associations. For an English-language version of the Act, see http://www.gesetze-im-internet.de/englisch_owig/ englisch_owig.html#p0143. 89 90

Introduction lix jurisprudence, where a company that has been fined by the Bundeskartellamt merges with another company, or where the company’s assets are sold, the ability of the Bundeskartellamt to pursue the successor in interest becomes very limited. Successor liability was accepted only where the old and the new assets were identical or nearly so from a commercial perspective.91 A broader approach was considered contrary, in the Federal Court’s view, to the nulla poena principle as expressed in Article 103(2) of the German Basic Law. It was believed that this rather wide gap in the German law on successor liability is for the legislator, not the Federal Court, to close. The legislator did later address the issue, but as Kirchhoff points out, the solution was merely partial and may not suffice to forestall determined efforts to circumvent the collection of significant fines. In national litigation in Germany the argument has been raised that Article 5 of Regulation 1/2003 may imply a directly applicable rule of successor liability that would be independent of the Act on Regulatory Offenses. However, the Appellate Court of Düsseldorf rejected this proposition on a variety of grounds including, unsurprisingly here again, the principle of nulla poena. In my view, a plain-language reading of Article 5 seems to support the conclusion that the provision does not have the purpose or effect of extending the Commission’s sanctioning powers to the NCAs. Nevertheless, since national laws may not deprive the EU competition rules of their effectiveness, a preliminary reference to the ECJ would have been welcome even assuming under Article 267 TFEU it was not obligatory since the Court of Appeal is a lower court. On appeal, it appears that the Federal Court of Justice would in principle be required to make such a reference, provided the acte clair doctrine does not apply.92 Judge Kirchhoff doubts that Article 5 and its proper interpretation should be classified as ‘acte clair’, though he circumspectly refrains from taking a position on the merits of the case. The discordance between EU law and national law and the corresponding tension between fundamental principles lead Kirchhoff to an incidental but

91 This is quite a contrast compared to the Commission’s powers under Article 23 of Regulation 1/2003. As the Grand Chamber of the ECJ has held (in a case where Italian law self-consciously aligned itself on the parallel provisions of Regulation 1/2003), ‘the legal forms of the entity that committed the infringement and the entity that succeeded it are irrelevant. Imposing a penalty for the infringement on the successor can therefore not be excluded simply because, as in the main proceedings, the successor has a different legal status and is operated differently from the entity that it succeeded.’ Case C-280/06 Autorità Garante della Concorrenza e del Mercato v Ente tabacchi italiani—ETI SpA and Others [2007] ECR I-10893, para 43. 92 Ost affirms flatly in his chapter (page 48) that ‘the ECJ will have to deal with that question’. For the classic jurisprudence on the acte clair doctrine (ie, that a national court need not refer where there are objective grounds of obviousness), see Case 283/81 CILFIT and Lanificio di Gavardo SpA v Ministry of Health [1982] ECR 3415, paras 16–20. A key criterion for the application of acte clair is that, ‘[b]efore it comes to the conclusion that [a referral is unnecessary], the national court or tribunal must be convinced that the matter is equally obvious to the courts of the other Member States and to the Court of Justice’.

lx Introduction significant observation regarding the harmonization of enforcement powers and procedures in Europe. As he says: In the past, compelling political reasons may well have prevented the European legislator from stipulating uniform powers of investigation and penalties for NCAs equal to those of the Commission. But in an internal market which requires a level playing field for undertakings, serious differences in investigation powers and sanctions between the Commission and NCAs as well as among NCAs are unacceptable in the long run and contradictory to the principles of a system of parallel enforcement of EU competition law.93

Kirchhoff adds that the need for harmonization is further accentuated by the need for the ECN to function properly, given that an authority hobbled by national law limitations cannot be regarded as ‘well placed’ from a Unionwide point of view to handle cases. This perspective evokes, once again, the neofunctionalist nature of the modernization enterprise. Two other cases highlighted by Kirchhoff may be mentioned here briefly. First, there is Outokumpu’s complaint before the German Federal Constitutional Court essentially challenging the lawfulness of the EU fining system and of its application, notwithstanding an earlier judgment by the ECJ to the contrary. In regard to this case Kirchhoff recalls the classic jurisprudence of the Constitutional Court, which normally (in theory) abstains from interfering on condition that the EU institutions ensure the protection of fundamental rights to a standard equivalent to that guaranteed by the German Basic Law. Conceivably, an indirect challenge might be brought before the ECtHR, although theissue of whether the Member States may be held collectively and vicariously responsible for the acts of the Court of Justice remains rather unclear—one would think that attributability poses problems in this regard.94 The other case to which Kirchhoff refers is relevant by analogy to the basic

Page 388. The residual responsibility of the Member States for the acts of the Union (again, if attributability can be established) seems to be accepted in principle by the ECtHR. See Case 24833/94, Matthews v United Kingdom, 18 February 1999, para 32 (‘The Court observes that acts of the [European Community] as such cannot be challenged before the Court because the [Community] is not a Contracting Party. The Convention does not exclude the transfer of competences to international organisations provided that Convention rights continue to be “secured”. Member States’ responsibility therefore continues even after such a transfer.’ (emphasis added)). See also the aborted Case 56672/00, DSR Senator Lines GmbH, 10 March 2004 (where an application against all of the then-15 Member States of the Community—on the ground that the applicant’s obligation to pay a fine of 273 million euros imposed by the Commission before the appellate process played out infringed Article 6 ECHR—was found to be moot and inadmissible by the ECtHR given the annulment of the fine on other grounds by the Court of First Instance); Case 51717/99, Guérin Automobiles EURL v The Member States of the European Community, 4 July 2000 (where a judgment of the ECJ was indirectly challenged; application inadmissible not because it indirectly contested an act of the ECJ but because the claim lodged concerning a right of information with regard to judicial remedies fell outside the scope of Articles 6 and 13 ECHR). 93 94

Introduction lxi principle of self-assessment that is inherent to the EU’s post-modernization universe. Are the requirements of competition law so indefinite and approximate as to be unconstitutionally vague? More precisely, does the system of self-assessment respect the requirements of accessibility and foreseeability under Article 7 of the ECHR?95 As Kirchhoff points out, the Swiss Federal Court considers that Swiss competition law contains indefinite legal concepts such as the ‘abuse of a market-dominating position’ prohibited by Article 7 of the Kartellgesetz. It emerges from an a contrario reading of the case law that the Swiss court would regard such a concept as impermissible were it not for the possibility to notify the Swiss Competition Authority in advance before implementing a practice that might raise competition concerns,96 a procedure that provides a measure of ex ante certainty and thus ‘compensates’ for the indefinite nature of Article 7. Kirchhoff (correctly) considers, however, that it would be difficult to attack the very foundations of the post-modernization edifice in the EU context, and that the more conceivable scenario that could arise would be a challenge to the particular application of sanctioning powers in circumstances where there is insufficient experience and positive law and guidance—that is to say, the sanctioning of ‘grey area’ practices where the boundaries of legitimate conduct are unclear and cannot be discerned despite reasonable efforts to comply. With regard to past practice, as far as certain individual cases are concerned, opinions have differed. William Kovacic, ‘Creating a Respected Brand: How Competition A gencies Signal Quality’. This chapter begins with the simple observation that ‘[a]n agency with a strong brand stands a greater chance of being effective than one with a weak brand’,97 the term ‘brand’ being a metaphor for reputation and perceived organizational identity. Brands, Kovacic says, can perform two positive functions for an enforcement agency: they can provide information about what the agency does; and they can signal institutional quality. The foundations for a good brand in this context include: strong substantive programs and judicious selection of which programs to operate; sound procedures; solid proficiency and capacities; and a ‘healthy’ organizational culture, the latter implying qualities such as integrity, courage and serious c ommitment

95 In EU law, one cannot say that the system of self-assessment is inherently contrary to the principle of legal certainty (such a position would seem quite radical), but that principle does constrain, for example, the methods employed to determine whether an undertaking’s conduct is anticompetitive. See Case C-280/08 P Deutsche Telekom v Commission [2010] ECR I-9555, para 202 (‘[assessing the legality of the defendant undertaking’s pricing solely on the basis of its own costs and prices is consistent with legal certainty insofar as it] allows that undertaking, in the light of its special responsibility under Article 82 EC, to assess the lawfulness of its own conduct’). 96 See Article 49a of the Kartellgesetz. If the notification is made a minimum of five months in advance and if the competition authority does not open an investigation, immunity from fines for the practice in question can be obtained. 97 Page 393.

lxii Introduction to continuous, long-term improvement. If the agency is successful in signaling quality on the basis of those foundations, its brand becomes a significant capital asset that can help the agency achieve its goals. By virtue of its strong brand the agency enjoys to some extent a ‘halo’ effect with respect to a variety of interlocutors and stakeholders such as judges, legislators and any relevant supervisory committees or paymasters, other public regulators, and civil society—not to mention the broader domestic public and international audiences. The foregoing considerations suggest a virtuous circle: by establishing effective enforcement and effective organizational activities, the agency earns a strong brand which allows it to further strengthen its effectiveness. Moreover, the process of branding and the definition and evolution of a brand must certainly be key to the agency’s self-understanding, self-expectations and identity, all of which are of great significance. Section 3 of Kovacic’s chapter provides a detailed set of factors that contribute to the definition of a brand and to the level of brand quality. Many of these could be described as superior agency practices and (long-term) good governance techniques. These will not be enumerated in full here; suffice it to note a few brief points that emerge. One is that Kovacic understandably cautions against entrusting an enforcer with a mission that is too variegated. Where the agency is put in that position, the process of priority setting may be confused; and messaging, both externally and internally, may be too diluted. In those circumstances, brand quality and hence effectiveness may be jeopardized. Of course, the risks of mission fragmentation and paralysis tend to increase as the agency’s assigned tasks grow increasingly incompatible (leading, in the worst case, to ‘schizophrenia’), and where the diverse demands placed on the agency stretch the limits of its capacity. Another point raised, and a consistent message in Kovacic’s voluminous body of work, concerns the sometimes deceptive link between an agency’s quantitative output––more infringement decisions and prohibitions (provided they generally withstand appeals), more investigations, ever-higher fines–– and its quality, effectiveness and true impact on the economy. The ‘more is better’ assumption exerts a strong influence on brand image, and one therefore would think this implies a tradeoff. On the one hand, an agency’s efforts to be visibly active can represent an opportunity cost: the agency might refrain from more meaningful investments in projects such as advocacy programs, market studies, self-evaluations and ‘policy R&D’, and as Kovacic has often said, there is often too much focus on activity levels as a performance indicator. On the other hand, as already noted, the enhancements of brand strength provide the agency with advantage. Taking a pragmatic perspective, Kovacic points out that there may be circumstances––eg where there is new leadership at the agency (which might be linked to a change in the incumbent political party) or where the agency itself is new––where the authority is enabled, by pursing a visibly active agenda with flashy output, to accumulate ‘capital’ to be spent later, for example in the form of a (pro-competitive but perhaps c ontra-populist)

Introduction lxiii decision not to prosecute. Notwithstanding this acknowledgement, however, Kovacic comes back to the fallacy of easy measurables: ‘To say that an agency is bringing lots of cases, or collecting substantial fines, does not establish that its program is improving economic performance––the genuine test of effectiveness’.98 He goes on to discuss how the tendency to overvalue cases and fines while neglecting the deeper question of effectiveness is reinforced by the confluent professional interests of various stakeholders who benefit from high levels of agency intervention. The final selected point to be mentioned here concerns quality control mechanisms, both at the level of the agency and its investigatory and decision-making processes, and at the level of courts reviewing the agency’s decisions—assuming that judicial review is a reality.99 The challenge of quality control (including the challenge of an agency’s confirmation bias), in terms of the rigorous testing of inculpatory evidence and the appearance of procedurally fair decision-making, is particularly acute where a number of institutional duties—deciding when to prosecute, carrying out the prosecution, determining liability and deciding the sanction—are combined within a single body. In those circumstances, the adequacy of checks and balances is of vital importance. Judicial review is one obvious form of ensuring quality control, and one may expect optimistically that if the agency is held to a rigorous standard on appeal, then it will internalize that standard in its future activities. As an example of this positive dynamic, Kovacic recalls the prominent 2002 annulments of Commission merger decisions by the Court of First Instance. On the other hand, if a court, for statutory or other reasons, fails to ensure that the relevant evidence is robust or that due process is respected, quality control may be missing at both levels. All of this is crucial for institutional design when regimes are being introduced or reformed. It can also have important implications for the perceived quality of the brand of the agency, not to mention that of the relevant judicial system.

PART 3: Effectiveness and Legitimacy in International Enforcement Cooperation Alberto Heimler, ‘Effectiveness of Enforcement Cooperation in Developing Countries: What Role Can Existing Institutions Play?’ In sections 2 and 3 of his chapter, Heimler reviews the star-crossed romance between trade and

Page 404. This is not the case in all jurisdictions, as illustrated notably by the Chinese enforcement system. 98 99

lxiv Introduction competition within the framework of the WTO, and the not coincidental rise of the ICN as a relatively successful locus of international cooperation in competition matters.100 Section 3 begins with some of the ICN’s history but Heimler shows some of the difficulties the ICN has faced in its evolution. His main criticism, among others, is that the ICN fails to function as a motor of convergence which could support and provide useful guideposts for the agencies of developing countries. For all the apparent but ambiguous success of some of its Recommended Practices (particularly in relation to merger notification and review procedures), Heimler notes a ‘progressive abandonment’ of the recommendation approach, and a turn toward a mere ‘encyclopedic’ reporting of the variety of practices prevailing worldwide as regards given fields of enforcement and practice. At the end of section 3, Heimler suggests a set of initiatives which in his view would get the ICN back on track with its proper role as he sees it. But the core of Heimler’s chapter, presented especially in sections 4 and 5, concerns the tribulations developing countries have experienced in their attempts to establish and maintain effective competition enforcement regimes. The starting point in this regard is that, while competition laws as such now encircle the globe, the factors necessary to breathe life into them––money, adequate human capital, enforcement capacity and so on––are all too often lacking. There were many reasons to be optimistic about the potential capacity of regional cooperation, including the conclusion of competition chapters in regional trade agreements and in some cases the establishment of supranational authorities, to fill some of the enforcement gaps. But as Heimler underlines, most of the attempts to achieve a functioning system of regional competition enforcement have met with little success or none at all. To substantiate this, Heimler provides a useful summary survey of the main experiences with regional cooperation efforts,101 and he points out a number of flaws in their design, governance and implementation. Depending on the region, these flaws include, to name only a few: the absence of adequate cross-border merger control provisions; jurisdictional uncertainty, eg due to overlapping memberships; insufficient independence from national interests; or a lack of flexible case allocation mechanisms (which Heimler refers to, in a non-technical sense, as a principle of ‘subsidiarity’). More generally and crucially, the political will necessary to establish and maintain (and fund) an effective system of law and of enforcement is highly variable across numerous jurisdictions

100 For recent discussion of the links between these stories, see Mel Marquis, ‘Idea erchants and Paradigm Peddlers in Global Antitrust’, 28 Pacific McGeorge Global BusiM ness and Development Law Journal 155 (2015). 101 Specifically, Heimler discusses Mercosur, the Southern African Customs Union, the Andean Community, COMESA, CARICOM, and WAEMU. Special attention is given in section 5 of the chapter to the Southern African Development Community.

Introduction lxv worldwide. In section 6, Heimler holds out the relatively deep and innovative ‘ANZCERTA’ arrangements between Australia and New Zealand in relation to competition cases as an example of how developing countries could cooperate effectively without the need for elaborate, costly and ultimately ineffective supranational authorities and frameworks. In light of this chapter and other similar findings one might say that, despite all of the strides made over the last 20 years, the global competition project— insofar as enforcement in most developing countries is concerned—has faltered. And it is up to the competition community to imagine more creative ways to address the substantial asymmetries among agencies in their credibility as enforcers. Experience suggests that regional cooperation agreements cannot by themselves stimulate the necessary mix of capacity, incentives and political will to advance and achieve effective competition enforcement in developing countries. For Heimler, the more promising strategy is a combination of deeper bilateral agreements for case practice issues and, at the level of established international organizations, renewed efforts to achieve closer procedural and substantive convergence. Antonio Capobianco, John Davies and Sean Ennis, ‘The Need for International Cooperation in Merger Enforcement’. The need for cooperation arises from the Cambrian proliferation of merger regimes around the world, from extraterritorial claims of jurisdiction and from the rarity of one-stop shop mechanisms. In those circumstances, the risk of divergent outcomes is substantial.102 Of course, one may maintain that a criss-crossing web of regimes is better than none, but the question is how to secure the gains of the worldwide quilt of merger laws while minimizing costs and risks. In this chapter the authors begin by breaking down the reasons why two or more jurisdictions might reach different regulatory results when reviewing essentially the same merger, bearing in mind that differences among merger decisions may be completely justified by different circumstances. Notably, these reasons include: differences in legal tests, or in attitudes toward efficiency or in underlying goals; varying conditions of competition; differences in evidence; and the happenstance of different authorities making different judgment calls. This range of challenges implies that convergence on substantive standards provides little assurance of convergent outcomes. 102 This theme is discussed in expanded form, not only relation to mergers but to cartels as well, in OECD Background Paper, Challenges of International Co-operation in Competition Law Enforcement (2014), http://www.oecd.org/daf/competition/ChallengesCompetition-Internat-Coop-2014.pdf. The Background Paper is also available from the SSRN site: see Antonio Capobianco, John Davies and Sean Ennis, ‘Implication of Globalisation for Competition Policy: The Need for International Co-operation in Merger and Cartel Enforcement’ (20 June 2014), http://papers.ssrn.com/sol3/papers.cfm?abstract_ id=2450137. For further discussion, with more specific focus on experiences in non-OECD countries, see UNCTAD Secretariat, International Cooperation in Merger Cases as a Tool for Effective Enforcement of Competition Law (27 April 2015), http://unctad.org/meetings/ en/SessionalDocuments/tdrbpconf8d4_en.pdf.

lxvi Introduction The authors also discuss some of the costs of disagreement among authorities, which include negative externalities imposed on other jurisdictions in scenarios where either the jurisdiction with the strictest approach de facto blocks a deal with global effect despite the desirability of the merger in other jurisdictions, or where for example structural remedies affecting undertakings’ upstream operations have similar spillover effects. There is good cause for concern since the multiplication of veto players has led to ever greater scope for externalities and irrationality in the global regulatory (non-)framework. At least in theory, the odds that a global merger will be prohibited increase, and the most interventionist veto player may become, so to speak, a hostage taker. Beyond the direct costs occasioned by this fragmentation, one can also speculate that market operators may be deterred from undertaking desirable global transactions, or that they may structure a transaction in suboptimal ways to mitigate risk. There was a time when perhaps an optimistic perspective prevailed, in particular because GE/Honeywell led to redoubled efforts between the EU and the US to avoid conflicts. But particularly given the emergence of China’s MOFCOM as a potent merger regulator and given its tendency to tread its own path—sometimes prohibiting deals on puzzling grounds—the global situation as regards large international mergers appears to have regressed. At the same time, there do not currently seem to be either de jure or de facto mechanisms in place that can adequately address the problems described. As the authors point out, there is a glaring and unfortunate gap in governance. Following the completion of this chapter by Capobianco, Davies and Ennis, the OECD Council adopted its Recommendation on International Co-operation on Competition Investigations and Proceedings,103 but—notwithstanding its generally salutary approach—good cooperation practices among OECD countries cannot fundamentally address that gap.

PART 4: Issues for Courts and Perspectives on the Judicial Role Gerald Barling, ‘The UK Competition Regime: Developments and Further Proposals for Change’. As the title suggests, Barling’s chapter was prepared midstream prior to the implementation of a number of profound reforms to the UK competition law system. The essay discusses private and public enforcement in the UK, and the position of the Competition Appeal Tribunal (CAT) in the midst of these fundamental changes. Putting the reforms in context, he underlines first of all the flawed and illogical arrangements http://www.oecd.org/daf/competition/2014-rec-internat-coop-competition.pdf.

103

Introduction lxvii e stablished when the Competition Act 1998 was amended by the Enterprise Act 2002, specifically in relation to the CAT’s inability in a private damages suit to decide whether a practice constituted an antitrust infringement. This limitation in turn stunted the growth of private enforcement in the UK. Barling in his essay welcomed the Consumer Rights Bill (now the Consumer Rights Act—see below) for introducing many new features including stand-alone claims for damages before the CAT, powers of injunctive relief (essential, for example, in stand-alone cases of alleged abuse of dominance) and for the provisions on both opt-in and opt-out collective actions (whereas, previously, only opt-in actions were possible—with anemic results). With specific regard to the collective action provisions, Barling found them in general to be a positive step forward. He was surely justified in saying that ‘[t]here is simply no point in giving citizens rights if there are no effective remedies and procedures enabling those rights to be enforced’.104 (As an aside, I would transpose this view beyond the UK: if a Member State simply adopts an optin collective action procedure without adding some controlled possibilities for an opt-out procedure, it is akin to taking no action whatsoever.105) Having welcomed the new regime in principle, Barling reacted to the envisaged provisions, pointing out for example that reasonable judges may fiercely disagree about whether, for certification purposes, multiple claims are the same, similar or related. With regard to the BIS consultation on streamlining appeals in regulatory and competition law cases,106 Barling notes his opposition to any weakening of the CAT’s standard of review (ie, by substituting a legality standard for the current merits review standard); he further observes fallacies with respect to the possibility floated by the Government of limiting the use of new evidence in appeals against the CMA’s decisions.107 In a coda to his

Page 530. To be more explicit, this is a criticism I respectfully lodge with the European Commission and the opt-in constituencies that dissuaded the Commission from treading the bolder path. However, with regard to the Commission’s position, Barling’s assessment is more charitable. ‘Understandably’, he says, the Commission ‘does not ask all [the Member] States, with very differing levels of sophistication in their justice systems, to attempt to introduce opt-out procedures’. Page 536. This may be a fair observation, but it does not follow that the best course is to recommend that all the Member States should presumptively, absent a justification, adopt an opt-in procedure. To address the concern regarding variable levels of capacity, the Commission could at the very least have left that choice to each Member State rather than establishing a hierarchy favoring opt-in schemes. 106 BIS, Streamlining Regulatory and Competition Appeals: Consultation on Options for Reform https://www.judiciary.gov.uk/publications/qbd-response-bis-consultationstreamlining-regulatory-competition-appeals/ (19 June 2013), https://www.gov.uk/ government/uploads/system/uploads/attachment_data/file/229758/bis-13-876-regulatory-andcompetition-appeals-revised.pdf. 107 See now the Government’s Response to the Consultation on Revising the CAT Rules of Procedure (September 2015), https://www.gov.uk/government/uploads/system/uploads/ attachment_data/file/460442/BIS-15-357-competition-appeal-tribunal-rules-of-proceduregovernment-response.pdf, pp 15–18. CAT Rule 9(4)(h) now requires the appellant’s notice of appeal to identify any new substantive evidence that was not before the CMA or the rel104 105

lxviii Introduction chapter, Barling issues a robust warning against attempts by the government, without proper justification and possibly with a view to indulging corporate interests, to clip the wings of the courts—specifically, his. As he remarks pointedly: ‘It is crucially important that courts, particularly small specialist ones, whose judicial personnel are few in number and well-known to their users, should not have to expect that giving a judgment to this or that effect might well lead to intense lobbying for jurisdictional and procedural changes, with the aim of lessening the scrutiny to which certain decisions would be subject in the future.’108 Following the completion of Barling’s chapter, the UK did in fact adopt the Consumer Rights Act 2015 (‘CRA’),109 thereby amending the Competition Act 1998 (as modified by the Enterprise Act 2002). The relevant provisions, contained in Schedule 8,110 as further supplemented by Part 5 of the CAT Rules 2015,111 are generally in line with the features of the Bill discussed by Barling. A number of details could be added, but suffice to note here that non-UK claimants are not automatically included in an opt-out action and must opt in to such an action if they wish to participate. After some early controversy, law firms can at least potentially be approved as a representative for the purpose of obtaining a collective proceedings order, provided of course that there is no conflict of interest between the firm and the class of claimants; and provided, more generally, that the CAT deems such representation to be just and reasonable. As Barling noted, the CRA also provides for collective settlement orders as a means of early resolution. For completeness one may also call attention to the voluntary redress system introduced by the CRA.112 The relevant rules provide that the CMA may approve a redress scheme, possibly on a conditional basis, during an investigation or upon its completion. The CRA required the CMA to publish guidance to be approved by the Secretary of State, and to that effect the CMA consulted

evant regulator. A similar counterpart rule applies to the CMA or regulator. Rule 15(3)(c) requires the CMA or regulator to explain in detail any objections to the admission of the appellant’s new evidence. Page 540. http://www.legislation.gov.uk/ukpga/2015/15/pdfs/ukpga_20150015_en.pdf. 110 http://www.legislation.gov.uk/ukpga/2015/15/schedule/8/enacted/data.pdf. 111 Like the CRA 2015, the new CAT Rules, which amend the 2003 version, became effective on 1 October 2015. The CAT Rules are available at http://www.catribunal.org. uk/240/Rules-and-Guidance.html. See also the CAT’s revised Guide to Proceedings (2015), www.catribunal.org.uk/files/Guide_to_proceedings_2015.pdf. For discussion of the private enforcement aspects of the CRA and the CAT Rules, see, eg Barry Rodger, ‘The Consumer Rights Act 2015 and Collective Redress for Competition Law Infringements in the UK: A Class Act?’, 3 Journal of Antitrust Enforcement 258 (2015); Roger Gamble, ‘Not a class act (yet): Europe moves softly towards collective redress’, 37 European Competition Law Review 14, 21–22 (2016). 112 See pages 10–13 of Schedule 8 (link provided above n 110). Following the CRA amendments, the relevant provisions are Sections 49C to 49E of the Competition Act 1998. 108 109

Introduction lxix on a draft text113 and in the summer of 2015 adopted the final document.114 Under this approval system, the CMA will assess the effectiveness of the scheme by considering the value of the (monetary or non-monetary) compensation and the governance of its disbursement to victims of the anticompetitive behavior in question. A company whose redress scheme is certified by the CMA can earn forgiveness of up to 20 per cent of the fine the CMA otherwise would have imposed.115 Once approved, the CMA may enforce its terms, including by way of court actions.116 It emerges from the final guidance that the CMA may even approve a redress scheme following an infringement decision of the European Commission. An obvious risk in relation to redress schemes is that a company might seek to lowball victims with paltry offers, but in theory the CMA will be vigilant in its screening responsibilities. Another potential point of concern might be the use of the system to preempt collective damages actions, in particular opt-out actions. The possibility that some victims will go uncompensated is always present but—to the extent that such a scheme has a genuine incentive effect, and to the extent it stimulates serious offers and at least partially restores victims to their ex ante position— the interplay of the two systems might prove salutary. Coming back finally to the overhauled system of collective proceedings in the UK under the CRA 2015, much of the detailed functioning of this system will depend not on what the provisions of the CRA or the CAT Rules say, but on how particular issues are managed as the cases and appellate judgments accumulate over time. James Venit, ‘What Is To Be Done?’. The title of this chapter can be either crystal clear or cryptic, depending on the eye of the beholder; readers familiar with the long-running criticisms of the EU enforcement architecture will immediately understand. Venit has often underlined the deficiencies of the EU system, and he recapitulates the core complaint here. As he says, ‘the credibility of the EU competition enforcement regime is itself on trial. … [T]he combination of an inquisitional (as opposed to adversarial) administrative system which confers substantial enforcement powers on the European Commission with a system of limited judicial review is too one-sided and calls into question the legitimacy of the EU regime of antitrust enforcement.’117

113 See CMA, Draft Guidance on the CMA’s Approval of Voluntary Redress Schemes (2 March 2015), https://www.gov.uk/government/uploads/system/uploads/attachment_ data/file/408326/Consultation_-_draft_guidance_-_voluntary_redress_schemes.pdf. 114 CMA, Guidance on the Approval of Voluntary Redress Schemes for Infringements of Competition Law (14 August 2015), https://www.gov.uk/government/uploads/system/ uploads/attachment_data/file/453925/Voluntary_redress_schemes_guidance.pdf. 115 See the final guidance document, above n 114, para 3.30. 116 See revised Section 49E of the Competition Act 1998. 117 Page 543.

lxx Introduction According to this critical perspective, the flaws tainting the Commission’s administrative investigations and the seeming half-measures of the EU Courts (notwithstanding some improvements) to address a weak system of judicial review—which stop short of a comprehensive commitment to close scrutiny of Commission decisions—are troubling not just on fundamental rights grounds but also because they may skew substantive outcomes. Most of the chapter discusses the EU’s system of judicial review, and expounds the view that Article 263 TFEU, which is the textual basis for ‘control of legality’,118 does not present any obstacle to robust and extensive review.119 But Venit draws attention to the awkwardness of the fine distinctions that have developed in the concept of judicial review (outside the ‘unlimited jurisdiction’ context where fines are scrutinized) as the EU Courts understand it. As he argues: In theory, the distinction between the Court’s rigorous assessment of the correctness, completeness and reliability of the facts and the appropriateness of the conclusions drawn from them, on the one hand, and the requirement that it not substitute its assessment for that of the Commission, on the other hand, is coherent in its articulation and can be neatly encapsulated in the notion that the Court can only annul where the Commission has committed a manifest error of assessment by drawing implausible or unwarranted conclusions from the facts. However, we are on a very slippery slope, and it is far from self-evident how, in practice, the Court can easily draw the line between thoroughly reviewing whether the facts are sufficient to support the conclusions the Commission has drawn from them while at the same time respecting the obligation not to substitute its judgment for that of the Commission. Respecting this distinction is likely to present serious problems in practice.120

The chapter discusses relevant jurisprudence, including from the merger context, and develops Venit’s case for more rigorous judicial review—which can be achieved, he says, without resort to Treaty amendment. He also adds,

118 While one may acknowledge the commonly accepted and historically grounded notion that the jurisdiction of the Court under Article 263 TFEU is a derivation of French administrative law principles (see eg Jürgen Schwarze, ‘Judicial Review of European Administrative Procedure’, 68 Law and Contemporary Problems 85, 86 (2004)), and while Article 263 illuminates the types of acts that are reviewable, it is worth recalling that the provision makes no explicit reference to any particular standard of scrutiny such as manifest error of assessment, nor to any injunction against a substitution of the Court’s appraisal for that of the administrative authority. 119 Indeed, it is clear that the EU concept of legality control must be construed in a manner consistent with the principle of full jurisdiction as elaborated by the ECJ in its Chalkor and KME judgments of December 2011 and subsequent case law. This jurisprudence seems to require a mixed (and in my view unsatisfactory) approach: a general standard of correctness applies to the Commission’s fact-finding and appraisals of the facts as well as the legal conclusions drawn; but the Commission still enjoys a margin of discretion with respect to questions of technical or economic complexity. This mixed approach is given a name by Ian Forrester, who calls is ‘legality plus’ or ‘merits minus’. See page 576. 120 Pages 549–550.

Introduction lxxi without going into detail, a number of other more incidental suggestions for the EU system that would help to address issues of legitimacy, including, among others, a substantial strengthening of the Hearing Officer (to the point where the current tendency to seek help from the EU’s Ombudsman is rendered unnecessary) and the establishment of a system of ‘complete file access’ based on protective orders that would ‘eliminate the scandalously wasteful process of redaction, ensure fair file access and remove one of the procedural bases on which Commission decisions can be delayed and ultimately annulled’.121 Ian Forrester, ‘Quis custodiet ipsos custodes? Assessing the Judicial Role in a Lawful System of Competition Enforcement’. Who are the Guardians in this context? The idea could perhaps apply at two levels. A first guardian is the European Commission, indeed it is common to refer to the Commission as the ‘Guardian’ of the Treaties.122 More specifically to our purposes, it is also the primary guardian of the EU’s competitive market order. At the next level, another guardian is the EU judicature. Who shall guard this guardian? One would have to think that this answer is not a ‘who’ but the Rule of Law itself, a concept which necessarily incorporates the core fundamental protections recognized in the Treaties, in the Charter and the ECHR, and in the general principles of EU law. Perhaps Forrester intended that his title be associated mainly with the question of who shall guard the Commission, but it seems more likely that he meant to raise the question with regard to each of these guardians. The concerns expressed in his essay apply to both. It is a chapter of significant breadth; this summary will only scratch the surface. One of the premises of the essay, a theme touched on several times in this book and which Forrester has discussed at length in his other works, is that the Commission is a distinguished, righteous and proficient institution but it is marked by the fatal and structural flaw of improperly circumscribed powers and procedures.123 Is the flaw cured by proper guardianship at the level of the EU Courts? Much of the essay is devoted to laying out the reasons why it is not. A pertinent irony that is highlighted is that, whereas the power of public authority in the context of competition law enforcement by the Commission is insufficiently or at least inconsistently checked, the Court of Justice in other contexts—above all where sensitive questions of the relationship

Page 555. One may also recall the textual basis for the Commission’s great responsibility, Article 17(1) TEU, of which the first three sentences provide that ‘[t]he Commission shall promote the general interest of the Union and take appropriate initiatives to that end. It shall ensure the application of the Treaties, and of measures adopted by the institutions pursuant to them. It shall oversee the application of Union law under the control of the Court of Justice of the European Union.’ 123 For a summary of the criticism against the Commission’s administrative procedures, see Forrester’s oral remarks at pages 498–499. See also pages 577–578. 121 122

lxxii Introduction between supranational and national law are concerned—has heroically established itself as a Marbury-style constitutional and progressive court. One is also struck by the palpably different roles assumed by the Court according to whether the case before it is an action for annulment of a Commission decision or a request for a preliminary ruling. There are unmistakable differences in these procedures, but whether the Court should feel so constrained when performing its appellate function may be doubted. In short, the judicial Guardian has not shown itself capable of constraining the enforcer to an appropriate extent, and with sufficient predictability. Forrester thus logically appeals to the Rule of Law. That is to say, he attempts to show that judicial scrutiny should be strengthened so that the enforcer’s decisions are reviewed consistently with proper rigor—including as regards issues of fact, which clearly can be and often are determinative in competition cases. Some progress has been made in that regard, he concedes: the Court of Justice has made efforts to tighten up judicial review.124 This is a step in the right direction, he allows, but it is insufficient.125 One can agree with Forrester in particular with regard to the perplexing persistence of the ‘complex economic/technical assessments’ doctrine. Its justification remains fuzzy. One also hopes that the principles governing judicial review in this context will develop to the point when the articulation of standards is more closely matched with the intensity of scrutiny that is actually applied. Ultimately these issues lead back, as Forrester observes, to the question of legitimacy. In his words, ‘[t]he availability of rigorous, consistent and effective judicial

124 The position of the Court of Justice with regard judicial review by the General Court is summarized in, for example, Case C-382/12 P MasterCard and Others v Commission, ECLI:EU:C:2014:2201, paras 155–156:

155. As regards the extent of judicial review, it is apparent from EU case-law that where the General Court is seised, in accordance with Article 263 TFEU, of an action for annulment of a decision applying Article 81(1) EC, the General Court must as a general rule undertake, on the basis of the evidence adduced by the applicant in support of the pleas in law put forward, a full review of the question whether or not the conditions for the application of that provision are met (see, to that effect, judgments in Remia and Others v Commission, EU:C:1985:327, paragraph 34; Chalkor v Commission, C‑386/10 P, EU:C:2011:815, paragraphs 54 and 62; and Otis and Others, C‑199/11, EU:C:2012:684, paragraph 59). The General Court must also establish of its own motion that the Commission has stated reasons for its decision (see, to that effect, judgments in Chalkor v Commission, EU:C:2011:815, paragraph 61 and the case-law cited, and Otis and Others, EU:C:2012:684, paragraph 60). 156. In carrying out such a review, the General Court cannot use the Commission’s margin of discretion, by virtue of the role assigned to it in competition policy by the EU and FEU Treaties, as a basis for dispensing with the conduct of an indepth review of the law and of the facts (see, to that effect, judgments in Chalkor v Commission, EU:C:2011:815, paragraph 62, and Otis and Others, EU:C:2012:684, paragraph 61). Among other passages, see pages 582–583.

125

Introduction lxxiii review is not something “desirable” and “worthy”, but an indispensable element in the acceptability of the whole system’.126 Mario Siragusa, ‘Interaction between Public and Private Enforcement of Competition Law’. Siragusa here provides his reflections on certain issues regarding the subject in the title. Not limited to a discussion of courts or perspectives on the role of judges, Siragusa refers specifically to: the indispensability of effective judicial review in light of Article 9 of the EU Damages Directive (the so-called ‘binding effect’ of decisions); the disclosure of evidence; interim measures at EU and national level; the remedial powers (including structural remedies) of NCAs; limitation periods and joint and several liability; and quantification of harm. The chapter is brief and can be summarized selectively in a few lines here. One observation to highlight is Siragusa’s concern that the Directive’s rules on disclosure, although generally to be welcomed to the extent that it opens up broader possibilities in some Member States, may be unduly restrictive in its formulation that evidence sought should be described as precisely and as narrowly as possible in the reasoned justification presented to the judge. Also notable is the discussion of the possibility of NCAs imposing not only behavioral remedies but also structural remedies, where appropriate. Siragusa contends that such power, while not granted explicitly by Article 5 of Regulation 1/2003, is nevertheless implicit and that its contours are shaped, or logically should be shaped, by those of the Commission’s analogous power under Article 7. With regard to private enforcement, the Damages Directive certainly has its weaknesses.127 If there is cause for optimism it is because it is a first generation instrument, and as such it is a milestone, however modest and ambivalent it may be.

Page 586. Among several other critical commentaries, see Mel Marquis and Giorgio Monti, eds, Shaping Private Antitrust Enforcement in Europe, forthcoming, Hart Publishing. 126 127

lxxiv

PART I Effective Enforcement of Competition Law Philip Lowe: Good morning, and a very warm welcome to everyone. On behalf of Giorgio and Mel, many thanks to those who are returning to the Workshop, and a special welcome to those joining us here for the first time. There are some territorial imperatives at play, but in general new blood comes in to every edition of this event, and the new blood provides us with fresh and stimulating input—which is not to overlook the important contributions of the old blood. [Laughter] We have a title for the Workshop this year, which is very broad in the sense that effective and legitimate enforcement of competition law has many components, many aspects. However, it is certainly crucial to the credibility of every authority and competition law system. So we are going to have a series of sessions today and tomorrow where we will go into depth about what constitutes effective and legitimate enforcement of competition law. In this regard we will consider how competition authorities establish or should establish priorities for enforcement, and we’ll also discuss how authorities can establish conditions that promote compliance with the law. So let us begin. I will now hand the floor over to Giorgio, who is chairing the first session. DD

1 EFFECTIVE SANCTIONS AND COMPLIANCE Chair: Giorgio Monti Speakers: Simon Bishop Frédéric Louis Konrad Ost Christine Parker Giorgio Monti: Let me add my words of welcome to you as well. We begin with a marathon session. That is to say, we have 10 s peakers before the coffee break. The speakers have an injunction to speak for 10 minutes each. DD

2 Part I: Effective Enforcement of Competition Law The way I will try to enforce this injunction is that when you will have about a minute left you will see a yellow card and when you are done there is a red card. There is a further mechanism for effective enforcement that I have as well but that I will not disclose. Now you will have legitimate expectations and my enforcement will be legitimate and hopefully effective. Speaking first is Christine Parker. Christine Parker: Thank you very much and yes, I am a rookie. Thank you for the opportunity and invitation. What I want to do in my presentation is to draw together the themes of my research, which has been mainly on empirical research on enforcement and compliance in competition law. It is great to be able to do that in the presence of the bright minds that are together here. I will finish my presentation with a couple of comments on the potential translation of my empirical research in what I see as some of the substantive issues that I have read about in the papers. Obviously, I come from Australia, and I am actually not a competition lawyer. I am a socio-legal empirical researcher who has had the opportunity, over the last 15–20 years, to study in depth the enforcement practices of the Australian Competition and Consumer Commission (ACCC) and how businesses respond to them. I think the same issues face every competition law enforcement agency—and, indeed, any business regulation enforcement agency—and I see that in the papers here. One other thing I should point out is that the ACCC is basically a prosecution agency. They don’t generally make decisions except about mergers and certain authorisations, but for most of the offences they are a prosecution-only agency. My main interest is the crucial question of compliance. As I said, my main research is empirical and the main themes of my research have been that it is too easy to conceptualise business response as something that will happen automatically and to the right degree if we get optimal deterrence right. I think it is provocative—our organisers have asked me to be—to say that it would be better if we got rid of the words ‘deterrence’ and ‘compliance’ and talk about what they really mean, because businesses and individuals respond to and interact with the law in complex and plural ways. The ways they respond include economic calculations about their self-interest, but also judgements about the substantive legitimacy of the law. That is the procedural legitimacy of the enforcement agency and an operator’s own capacity to notice and understand the law in the first place, and then a sort of scanning of the environment to see what other businesses and stakeholders are doing about the law. In my paper I have described what this plural model of compliance and deterrence means. In practice, what this means is that a competition law agency like the ACCC has to overcome a series of escalating or embedded challenges or n egotiations DD

Part I: Effective Enforcement of Competition Law 3 with business and society to make its enforcement both effective and legitimate in achieving the goals it is supposed to pursue. The paper that I’m proposing tells a story of the ACCC’s attempts to overcome each of what I describe as three ‘traps’ of effective enforcement and compliance and business responses to them, and I use my own research to tell that story. The first one, the deterrence trap, occurs where penalties for non-compliance cannot be big enough to deter rational misconduct without being so large that they exceed the capacity of firms to pay.1 What happened was that the ACCC, under the chairmanship of Professor Allan Fels, overcame that trap by m aking the process of investigation and sanctioning very painful in other ways than penalties by moralising in the media about cartel misconduct and other types of competition law misconduct while targeting individuals as well as firms. They address the deterrence issue, but the problem with that is that the c rucial issue of what we might call optimal deterrence had not been addressed, so the ACCC achieved an apparent enforcement success but without actually addressing the substantive moral issues about what the limits of competition law might be and therefore they appeared to be acting against procedural justice and potentially achieving over-compliance, but really they were probably creating the seeds of their own political illegitimacy. The ACCC has achieved apparent enforcement success, but without gaining the commitment of all businesses to the substantive moral basis for its enforcement activity. The ACCC had not gained acceptance and commitment to its own moral and economic evaluation of what ‘optimal’ deterrence required in terms of the process of enforcement and the substantive goals of competition policy. This is what I call the compliance trap: businesses criticise and resent the enforcement mechanisms even though they are settling most of their cases. So it appears compliance is improving but perhaps, covertly, the whole system is being undermined. What I argue is that the compliance trap is a symptom of a social polity that is still very ambiguous about the basis for its laws for businesses and whether business misconduct should be treated as a crime, and what should be allowed within which limits. That is where optimal deterrence lies. In that kind of ambiguity a regulatory agency is basically given an impossible task. They appear to be either over-enforcing or going soft. The ACCC tried to overcome this issue by building up substantive political support for its competition law enforcement by arguing in favour of the criminalisation of cartels in law. That happened and we now in Australia have a criminal cartel offence. Now, therefore, there appears to be a legitimate

1 See also John C Coffee Jr, ‘“No Soul to Damn: No Body to Kick”: An Unscandalized Inquiry into the Problem of Corporate Punishment’, 79 Michigan Law Review 386 (1981).

4 Part I: Effective Enforcement of Competition Law basis for this very strong moralising approach of anti-cartel enforcement. However, arguably, as our research shows, even cartel criminalisation, with cartel conduct being essentially a per se offence, is just too stark. It does not match business people’s moral or economic sense of what makes sense in the marketplace. I think the law and its enforcement are still in grave danger of falling into what I call the legitimacy trap. This is a situation where there is ambiguity and controversy about whether the law is substantively justified or not, and on what basis. It means that we do not really know whether any criminal prosecutions can ever succeed, or what effect they will have. Basically, what I am pointing to with my research and the paper is what we might call the elephant in the room. Whenever we talk about the techniques and processes of competition law enforcement we have this unspoken issue of ambiguity about both the substantive rationale for competition law and the procedural nature of the regulatory enforcement task in criminal competition law. We have this continuing ambiguity regarding to what extent competition should be balanced or tempered with other social and political goals, and we also have an ongoing ambiguity whether competition law fulfils simply an administrative regulatory function or whether it is something that should be thought of as amounting to a crime. These debates and ambiguities can be used and abused by businesses and their lawyers as part of their own strategy, to pursue their own economic self-interest, and also as rationalisations and excuses for wrongdoing. To conclude, I will make two broad comments. One is that I notice a lot of talk in the papers about optimal deterrence, and what I want to suggest is that the discussion about optimal deterrence is really a cipher for ambiguity or indecision about the substantive legitimacy of competition law enforcement. It’s really about not having addressed what the fundamentals are of what we think competition is and how it interacts with social and political goals, and whether it helps make our markets and societies the way we want them to be. It would be interesting to see if we can get rid of the term ‘optimal deterrence’ and talk about something else. The other comment is on procedural legitimacy, which I think is really a cipher for confusion about the nature of the regulatory enforcement task and whether it is criminal or administrative. I’ll end here. Konrad Ost: As I see that the 10-minute rule is strictly enforced, I will hurry up. Thank you very much for inviting me today. The focus of my paper is about the enforcement of antitrust rules by the national competition authorities. In the dawn of the modernisation process, the Bundeskartellamt (BKA) was rather reluctant to follow the mainstream move towards more harmonisation. Today we find ourselves more on the other side of the table, asking for even more. In my paper I argue that in order to reach an effective enforcement system, national systems need more alignment with the EU system. I have to admit that if the German system worked smoothly there DD

Part I: Effective Enforcement of Competition Law 5 would be less urgency to think about convergence, but effective sanctioning is at stake in Germany on two issues. One regards the addressee of a fine under German law and the other is the German procedural background. From an enforcer’s point of view, these are shortcomings that have to be changed. The European substantive provisions are of course applied by national competition authorities. The BKA has to enforce Article 101 TFEU against infringing ‘undertakings’ as defined under EU law. The German sanctioning law, however, provides for the imposition of sanctions on legal entities whose representatives commit the infringement. The legal entity is liable for any offence committed by a representative of that specific legal entity. In this logic, the action of a representative of another legal entity (even if it is a member of the same economic unit in European terms) may not be attributed to another legal person (eg the parent company). The result is a certain divergence between the enforcement of the substantive rules on the one side and the sanctioning law on the other. For example, there is no sufficient legal basis in German law to impose fines on parent companies for infringements committed by subsidiaries even though the substantive prohibition, as is clear from the case law, is addressed to the parent companies as part of the undertaking.2 We therefore have an important divergence on this issue of parental liability. The other issue is the succession in fines.3 It is possible, through internal restructuring within a group of undertakings, to evade sanctions. So this is more or less a cost–benefit analysis, which puts effective sanctioning at risk. Hence, effective enforcement of Article 101 in Germany is not entirely safeguarded. From the perspective of the national competition authority, there is a breach of EU law if it is not possible to fine the infringer of the EU competition rules effectively. From an undertaking’s point of view it is interesting as well, because the extent of the sanction depends on whether it is the Commission or the BKA that will impose a fine. At least there is a potential here to argue that there is a violation of rule of law principles.

2 On parental liability see, eg Joined Cases C‑628/10 P and C‑14/11 P Alliance One International Inc and Standard Commercial Tobacco Co Inc v Commission, EU:C:2012:479, para 42 (citing Case C‑90/09 P General Química and Others v Commission [2011] ECR I-1, paras 34–36; Case C‑521/09 P Elf Aquitaine v Commission [2011] ECR I-8947, para 53). See also Case C-97/08 P Akzo Nobel NV and Others v Commission [2009] ECR I-8237, para 56. 3 See Cases KRB 2/10 Fusion Transportbetonhersteller and 55/10 Versicherungsfusion, both of 10 August 2011. According to the Federal Court of Justice, it was only possible to hold a successor entity liable where the following conditions were cumulatively met: universal legal succession of the infringing entity by the successor entity, ie all assets and liabilities of the infringing entity must have been transferred to the succeeding entity in a single legal act; the infringing entity must have ceased to exist (an example would be the merger of one company into another); and economic identity between the infringing entity and the successor entity, which is only the case where the assets of the infringing entity are used in the same way by the successor entity (ie where there is a continuation of the same economic activity) and where they constitute the overwhelming part of the assets of the successor entity.

6 Part I: Effective Enforcement of Competition Law Now a brief look at the procedural side. In Germany, the rights and guarantees in antitrust proceedings are largely the same as in criminal cases. Compared to the EU system, judicial review in Germany is extreme. A ntitrust proceedings are treated as normal criminal cases, with 20, 40 or sometimes even a 100 days in court and procedures lasting more than a year. This is due to the focus on witnesses and oral procedures, which is explained by the fact that the individual can be fined in addition to the undertaking. The problem is to find the right balance. The procedures have become lengthy. The court has to repeat more or less everything in the administrative procedures. In the end, it is too burdensome. As I read in at least one paper that argues a little against my own, many of you that defend undertakings might prefer this German system at first glance, but ask German practitioners whether the German system is really preferable. I doubt it. My conclusion is that we need more convergence in both procedural and sanctioning law in order to create a level playing field that guarantees effective enforcement and sufficient foreseeability as to procedures and sanctions. What are the solutions at hand? One could have a look at the law as it stands and argue for more convergence on the basis of Regulation 1/2003. The question here is how much procedural alignment is already contained in the Regulation itself. My interpretation is that, in contrast with its original meaning, Article 5 of Regulation 1/2003, in the interpretation of the EU Courts, determines and marks the limits of national enforcement law. I argue that recent case law hints about the existence of wider effects on national procedural law than was thought before. This includes in particular the empowerment of national procedural law to impose fines on undertakings. Article 5 is in that regard incomplete. Recent case law of the ECJ in Schenker4 and Slovenskà sportitel’na5 on cases handled by NCAs shows that, from a German perspective, the ECJ ruled on issues of sanctions usually governed by national law. I suppose that the whole notion of an undertaking, including the issues of parental liability and succession of fines, is determined by EU law. Here I have to admit that we are on shaky ground, and there are neighbouring fields of law involved that are not determined by Article 5. The conclusion, therefore, is that one cannot stop at Regulation No 1, and therefore I call for a ‘Regulation No 2’ that could define the degree of harmonisation that is necessary for the effective enforcement of Article 101 TFEU on the procedural side. On the procedural side there are, of course, advantages of general features of the EU sanctions system from the enforcer’s point of view. However, it is clear that this cannot be a blueprint. Effective sanctioning is possible even with higher

4 Case C-681/11 Bundeswettbewerbsbehörde and Bundeskartellanwalt v Schenker & Co AG and Others, EU:C:2013:404. 5 Case C-68/12 Protimonopolný úrad Slovenskej republiky v Slovenská sporiteľňa as, EU:C:2013:71.

Part I: Effective Enforcement of Competition Law 7 standards of protection of fundamental rights, and a Regulation 2 could form the start for further alignment from both sides and from both extremes. Maybe there are other alternatives: we could just leave it as it is and come to effective sanctioning by focusing more on individual deterrence. However, as I have argued, the criminalisation of antitrust that we have with regard to bid rigging in Germany is not so successful. I do see that there is a certain need for introducing new sanctions, but I argue in my paper that the best-placed entity to do so is not the enforcer but the undertaking itself. Individual deterrence could be achieved through an effective compliance programme. A driver for effective compliance, as was seen in the Thyssen-Krupp case, where a company sued its ex-manager for 100 million euros, could be high corporate fines.6 The conclusion is that effective sanctioning of EU law requires enforcement by the NCAs, and for that to be effective and legitimate more convergence on both substantive and procedural grounds is needed. After 10 years of Regulation No 1, it is time to reflect on a Regulation No 2. Simon Bishop: I need to start by making two apologies. The first one is that as an economist I am not really sure what the difference is between effective and legitimate sanctions. To me, if it is legitimate it is also going to be effective and vice versa. The second apology is that I will have a narrow focus on monetary sanctions, fines and their appropriate levels. Although in practical terms all economics literature is focused on cartel behaviour, I will also comment on the appropriate level of fines in Article 102 TFEU cases. To give you the punchline to the latter in advance: we should not have fines for infringements of Article 102. In relation to cartel behaviour and fines, there is a lot of literature that speaks about ‘optimal deterrence’.7 Whenever we hear this, we should be rather wary. There are a lot of simple economic models out there and a lot of models provide a very good framework for thinking about how one might want to devise policy, but there are also a lot of economic models that get implemented and taken at face value that raise a lot of problems. Why do I say that? Optimal deterrence is based on two factors: overcharge benefit; and the probability of detection. Obviously, if you can raise the likelihood of detection, that is a good thing. From that perspective, the rationale for the leniency programme makes a lot DD

6 See AR Fabisch, ‘Managerhaftung für Kartellrechtsverstöße’, [2013] Zeitschrift für Wettbewerbsrecht—ZWeR 91–119. 7 See, eg William Landes, ‘Optimal Sanctions for Antitrust Violations’, 50 University of Chicago Law Review 652 (1983); Wouter Wils, ‘Optimal Antitrust Fines: Theory and Practice’, 29 World Competition 183 (2006); John Connor and Robert Lande, ‘How High do Cartels Raise Prices: Implications for Optimal Cartel Fines’, 80 Tulane Law Review 513 (2005).

8 Part I: Effective Enforcement of Competition Law of sense. But one thing that we will never know is the actual level of detection. No one will ever know this and any attempt to do so is doomed to fail. Therefore, if one of your crucial inputs into the model for optimal deterrence is something you will never know, how good is that for justifying any given number? In some sense, I am saying that whether it is 10 per cent of turnover or 30 per cent of turnover, these numbers are just plucked out of thin air. If we turn to the other input for the model, which is overcharge benefit, again I think we need to be really careful. There is a study going round by Connor which has sort of stated that the average overcharge is 25–30 per cent.8 Really? Where do these numbers come from? What is the underlying data? Numbers such as these very soon get overused before they have been sufficiently critically scrutinised. Are these numbers really robust? There is also a further problem with Connor’s estimates: even if, on average, they are correct, we still need to know the distribution, because for some cartels there could be an overcharge from 5–10 per cent and for others it could be 100 per cent. This would give you the average number of 25–30 per cent. I think this is a plausible outcome since, as we know, economic conditions differ drastically from market to market and one can therefore question whether it is really appropriate to apply an average figure for fines to different firms in completely different industries. Therefore, although I do not have any solutions as to the appropriate number, I think we need to be sceptical about how far our knowledge actually takes us.9 Now let me turn to Article 102 TFEU, the abuse of a dominant position, and specifically the appropriate monetary sanction for infringements in this field. I think there is a key difference between collusion and conduct held to infringe Article 102. Collusive conduct is always bad. Collusion might have a negative to zero impact on final prices but we know that it is never actually pro-competitive, whereas practically all conduct which has been held to infringe Article 102 also has a potential pro-competitive rationale. Therefore, this raises a big issue regarding the risk of over-intervention. That is because intervention against conduct that is pro-competitive will be harmful to consumer welfare. This element of potential harm through over-intervention has not featured enough in the public policy debate. We need to be more conscious of the value of doing nothing. I think the difficulties of enforcing Article 102 is particularly important because, when it comes to compliance, how do firms know whether their conduct is going to be anticompetitive or not? Often people say, well firms will know whether they are dominant or not, but that is not true in a lot of cases. They may have a high market share

8 For a breakdown of Connor’s findings, see recently John Connor, ‘Price-Fixing Overcharges’ (revised 3rd edn 2014), http://papers.ssrn.com/sol3/papers.cfm?abstract_ id=2400780. 9 See also Christian Ehmer and Francesco Rosati, ‘Science, Myth and Fines: Do Cartels Typically Raise Prices by 25%?’, Concurrences No 4-2009, Art No 28832, October 2009.

Part I: Effective Enforcement of Competition Law 9 but they are not dominant, or at least dominance is uncertain. One client asking compliance advice had a 50 per cent market share and wanted to see if it could implement a loyalty-based rebate scheme: the answer is no. Their response is that this is ridiculous. Two years ago they had an 80 per cent market share and now all their competitors are using an important competitive feature—loyalty rebates—that they are not allowed to use. You can say, as the EU Commission said in the old British Airways case, ‘Don’t worry—as long as you are no longer dominant you can start using them again’. But this seems to miss the point. There are real compliance dangers with this kind of interventionist application of Article 102. And that brings me back to the purpose that a fine serves for firms that are held to infringe that provision. A fine does not restore competition; that is not its purpose. Next, detection is not a factor because you cannot hide large dominant firms, and complainants will bring questionable behaviour to the attention of the authorities. Therefore, if a fine does not serve to restore competition, is it then about deterrence? Yes and no. The problem here is that if you start to impose very large fines for Article 102 infringements, this can have deterrent effects on legitimate p ro-competitive strategies in different industries. You can say in specific cases that it is a one-off, but over time such cases have the power of precedent, and spillover effects in other industries will arise. Finally, does a fine also serve to make media headlines? It can be good because it raises education and makes business p eople aware of competition policy in the same way it has been done in cracking down on cartel behaviour. But at the same time it should be recognised that it has a negative side. Politics with a small ‘p’ comes into this. Therefore, the reasons for the application of fines in Article 102 cases are from an administrative perspective close to zero. The people who suffered can also go through private enforcement and recover losses that way. Fred Louis: In the context of this discussion, I will speak only about cartels. I want to start with the issue of optimal deterrence, or whatever you want to call it. I think Simon has shown that this concept should play no part but that it is necessary instead in sanctioning to look more closely at the concrete impact of conduct on a market than is the case today. When you want to explain to an economist what is the difference between effectiveness and legitimacy, one would say ‘Well, imposing a 500 million fine on a cartel which was never implemented and had zero effect on the market may be effective in the sense that it scares people so that it may help deterrence, but is it legitimate’? Are concepts that were devised at the time when fines consisted of a thousand EU monetary units really legitimate for the current level of the fines? One cannot impose fines in the hundreds of millions and feel that you can skip the entire debate on whether or not the conduct you are fining has had a concrete impact on the market. This does not imply a full effects-based application or a quantification of the harm; but the current situation, where enforcers think that they can get away with imposing the current levels of DD

10 Part I: Effective Enforcement of Competition Law fines without demonstrating any effect, is problematic. I submit that one of the reasons that the EU is so reluctant to discuss criminalisation is a demonstration of the legitimacy gap because they have not got the burden of proof as to what exactly is the level of harm to society at large for cartels. I think we can all agree with Simon that there are no pro-competitive features to be found in cartels, although in crisis cartels there may be some discussion, but we can all agree with this as a basic starting point. However, does it justify the current level of fines? I think you really need to do better than this. The second point is that if you look at the fines today, you see from a practitioners perspective that the fines are high enough already and probably too high. If you look at the EU fining guidelines, which are increasingly seen as the model standard for enforcers in Europe, we are told that we can fine cartel conduct up to 30 per cent. In theory, you have a range between 0 per cent and 30 per cent. In practice, very little of that range is used. There has been, except for Telefónica/Telecom Portugal,10 no fine below 15 per cent and there has been no fine above 25 per cent. In most cases the fines are around 17–18 per cent. When you look at the cases you will see that the different cases between 15 and 18 per cent are sometimes indistinguishable from each other. Sometimes the conduct of a 15 per cent case seems to be worse that the conduct of an 18 per cent case. There is very little explanation as to what determines a choice of a certain percentage. The difference between a choice of 15 per cent or 18 per cent can be massive, and it translates into really high numbers when you apply the duration multiplier, which sends most cases nowadays into the 10 per cent fining cap category. One can ask whether that is the reason why you can see so little differentiation in the basic amounts of fines because you are already reaching astronomical numbers with this. F urther differentiation would only show more clearly what I would qualify as the disrespect for the law that is in the current system. Whether you consider the 10 per cent as a cap or a ceiling, as in Germany, the fact of the matter is that the legislator has set this limit. The enforcer then says ‘Well, this guy deserves a fine of 600 million euros and it is only because I respect the law that the fine is reduced to 120’. We are in an area of compliance where the regulator says that my system of calculating the fines has been devised irrespective of the limits put by the legislator. It is only at the last moment that I correct the amount and the fine is reduced. That does not show respect for the fact that the legislator has said that fines should not be over 10 per cent. This system deliberately flouts the 10 per cent rule and as an undertaking you have to find a way to evade the 10 per cent limit as much as you can. Contrary to prior practice, the Commission nowadays always involves the corporate group in a fine, and the justification is: ‘I can do this because the judges have authorised me to do it’. The fine

10 Commission Decision of 23 January 2013 in Case COMP/39839—Telefónica/Portugal Telecom.

Part I: Effective Enforcement of Competition Law 11 will be assessed against the group irrespective of whether the parent company has been involved in the practice or had any knowledge, or irrespective of whether the parent company has taken any steps to make sure that any subsidiary and business division complies with the law. The reasoning is to look at the ultimate parent because of a need to negate the effect of the 10 per cent limit if the enforcer is to have any chance of applying the first two steps of the fining guidelines. It is a situation that is blatantly unfair and creates huge problems of legitimacy. Because of a legal fiction we can hit the entire group, and we do not need to look at whether the group had involvement in this. This is something that no one outside the enforcers’ circles understands and it makes it hard for people to accept the system. Last point: convergence is in theory desirable, but please do not choose companies as a pretext to say that we therefore need to all have the EU system because, although there may be problems with the German system, if you ask the seven partners in my antitrust group which system they would choose, they would say: ‘Give me the German system any day!’ On the other hand, if you could analyse the dreams of policemen everywhere in the world, you would see that the dream system for them is the administrative decision-making system of the EU Commission and the light touch manifest error judicial review of the EU Courts in Luxembourg. Thank you.

12

Christine Parker* Effective and Legitimate Enforcement of Competition Law: A Riddle Wrapped In a Mystery Inside an Enigma?

1. Introduction Winston Churchill famously commented: ‘I cannot forecast to you the action of Russia. It is a riddle, wrapped in a mystery, inside an enigma …’1 So with regulatory enforcement, every time an enforcement agency overcomes one challenge, a deeper one rears its head. As this chapter demonstrates, the ‘deterrence trap’ is solved, but only by enforcement strategies that raise a ‘compliance trap’, which in turn raise a ‘legitimacy trap’. I briefly outline each of these on the basis of 15 years of empirical research on the enforcement strategies of the Australian Competition and Consumer Commission (ACCC) and the response of businesses.2 ‘But perhaps there is a key,’ Churchill went on to say. For Churchill, that key was Russian national interest. In the case of effective and legitimate enforcement of competition law, the ‘key’ is to recognize the political nature of competition regulation and enforcement to shape markets. That is: what is at stake is not competition in some pure or ‘blindly’ economic sense, but rather a political debate over the very shape of capitalism at the broad level

* Centre for Regulatory Studies and Faculty of Law, Monash University, Melbourne, Australia. This chapter summarises and reflects on some the main findings about effective and legitimate anti-cartel enforcement from two major empirical socio-legal research projects over the last 15 years: the Australian Competition and Consumer Commission Enforcement and Compliance Project (2005–2011) and the Cartel Project (2010–2012). I am grateful to the Australian Research Council for funding both projects and to my various collaborators on the two projects—Professor Vibeke Lehmann Nielsen, who was my co-investigator on the first project; and Professors Caron Beaton-Wells, Fiona Haines and David Round and research assistant Ms Janette Nankivell on the second project. I am also grateful for the support of Professors John Braithwaite and Imelda Maher and the C entre for Competition and Consumer Policy, Regulatory Institutions Network, A ustralian National University for initiating and supporting the ACCC Enforcement & Compliance Project. Further details of these projects can be found at http://www.law.unimelb.edu.au/cartel/ related-projects/the-australian-competition-and-consumer-commission-enforcementand-compliance-project and http://www.law.unimelb.edu.au/cartel/cartel-project. Footnote citations and references are omitted from the text below. 1 Made in a radio broadcast in October 1939. 2 The Appendix provides a list of the publications emanating from these two research projects.

14 Christine Parker and therefore at the microlevel about how competition law and regulation interacts with business people’s everyday lives. Various players (and wouldbe players) use competition law to contest and defend markets for particular products and services in ways that reflect their own interests and positions. This process of contestation and defence is one of the ways in which markets are transformed. Competition law is also used to constitute, contest, rationalise and transform capitalism at a broader level, and to set the boundaries of interaction between the social and justice policies of the broader socio-polity and the market.3 The key, then, to overcoming the cascading challenges of regulatory enforcement must lie in democratic engagement in relation to the central social, economic and justice questions raised by competition law, namely: —— The substantive values and interests served by competition law: how should markets and capitalism to operate? In the language of economics, this is the question of what ‘consumer welfare’ means. A perennial issue is to what extent competition law should be used to encourage and protect small business, or to what extent a competitive market can be allowed to produce dominant market players.4 —— The processes of investigation, adjudication, sanctioning and enforcement, and how they are experienced by individual business people, firms and other stakeholders. In the language of economics, this is the question of ‘optimal deterrence’. Perennial conundrums include the wisdom of going after easy scalps or high-publicity cases, larger versus smaller players, how to use immunity and leniency policies to reward whistleblowing or genuine attempts at compliance, and when to settle and when to keep going for greater sanctions. These questions are often framed in economists’ language as questions of ‘consumer welfare’ and ‘deterrence’. Criminologists and socio-legal scholars have noted that to these questions, we need to add the question of ‘compliance’.5 That is, to what extent do businesses actually respond to the law in the way anticipated? This tradition of research points out that businesses and individuals respond to and interact with the law in complex and plural ways which

3 See Brayden King and Nicholas Pearce, ‘The Contentiousness of Markets: Politics, Social Movements, and Institutional Change in Markets’, 36 Annual Review of Sociology 249 (2010). 4 See Tony Freyer, Regulating Big Business: Antitrust in Great Britain and America 1880–1990, Cambridge University Press, 1992; Tony Freyer, Antitrust and Global Capitalism, 1930–2004, Cambridge University Press, 2006. 5 See Christine Parker and Vibeke Lehmann Nielsen, Explaining Compliance: Business Responses to Regulation, Edward Elgar, 2011; Christine Parker, ‘Criminalisation and Compliance: The Gap Between Rhetoric and Reality’ in Caron Beaton-Wells and Ariel Ezrachi, eds, Criminalising Cartels: Critical Studies of an International Regulatory Movement, Hart Publishing, 2011, chapter 11.

Effective and Legitimate Enforcement of Competition Law 15 include economic calculations about their self-interest, judgements about the substantive legitimacy of the law, the procedural legitimacy of the enforcement agency and the businesses’ own capacity to notice and understand the law in the first place, and their scanning of the environment to see how other businesses and stakeholders treat the law.

2. A riddle … the deterrence trap The ‘deterrence trap’ occurs where penalties for non-compliance cannot be big enough to deter rational misconduct without being so large that they exceed the capacity of firms to pay.6 This means that there may be no ‘optimal’ financial penalty to deter cartel conduct. Economists generally consider that the amount of the financial penalty should be based on an amount that is greater than the gains of the misconduct and is then increased to take into account the fact that not all cartels are swiftly detected and prosecuted. Since the risk of discovery and enforcement for cartels is quite low, most commentators and policy-makers believe that fines for cartel behaviour should be many times the amount of the gain from the cartel. However, fines that are large enough to be an effective deterrent may be too large for some firms to bear. This means that enforcement agencies and courts might baulk at imposing financial penalties so heavy that a firm’s innocent employees and investors lose their jobs and savings. Moreover, fines that are so large as to be effective might also deter economically or socially desirable market behaviour, leading some commentators to suggest that there is a ‘sweet spot’ where fines are optimally deterrent but do not over-deter.7 It has been suggested that regulators should nonetheless project an image of invincibility in order to elicit compliance.8 Under the period in which Professor Allan Fels was its Chairman, the ACCC did indeed manage to a large degree to project an appearance of tough, swift and sure enforcement action.9 My research for the ACCC Enforcement and Compliance Project10 6 John C Coffee, ‘“No Soul to Damn: No Body to Kick”: An Unscandalized Inquiry into the Problem of Corporate Punishment’, 79 Michigan Law Review 386 (1981). 7 See Parker, ‘Criminalisation and Compliance’, above n 5, 241–242; Wouter Wils, ‘Optimal Antitrust Fines: Theory and Practice’, 29 World Competition 183 (2006); Peter Whelan, ‘A Principled Argument for Personal Criminal Sanctions as Punishment Under EC Cartel Law’, 4 The Competition Law Review 7 (2007). 8 Ian Ayres and John Braithwaite, Responsive Regulation: Transcending the Deregulation Debate, Oxford University Press, 1992, 44. 9 See Christine Parker and Vibeke Lehmann Nielsen, ‘The Fels Effect: Responsive Regulation and the Impact of Business Opinions of the ACCC’, 20 Griffith Law Review 91 (2011); see also Fred Brenchley, Allan Fels: A Portrait of Power, John Wiley & Sons, 2003. 10 The ACCC Enforcement & Compliance Project was conducted by Professor Christine Parker (then at University of Melbourne) in collaboration with Professor Vibeke Lehmann

16 Christine Parker showed that under the leadership of Fels the ACCC’s main cartel enforcement successes occurred via two strategies for ‘leveraging’ deterrence—and thus overcoming the deterrence trap:11 First, the ACCC leveraged up the deterrence value of the modest financial penalties available under Australian competition law with the administrative and personal cost and inconvenience of the investigation process and the reputational damage caused by publicity. Businesses and their lawyers came to see the process of ACCC investigations, and the publicity associated with ACCC enforcements, as more effective motivators of compliance than penalties. Second, the ACCC targeted a range of individuals and organisations (such as industry associations, compliance professionals and potential whistleb lowers) with the capacity to improve business compliance to put in place a web of controls that reduce the opportunities for, or the advantages of, offences and embed norms and practices for avoiding non-compliance in the social structure of industry. The ACCC developed a policy of holding as many parties—individual and corporate—legally liable as actually contributed to the conduct, rather than just focusing on the main corporate participants in the cartel. Even if penalties against firms are too small relative to the firm’s size and profits to be of deterrent value, this strategy spreads the deterrent threat to individuals who are likely to be more sensitive to smaller penalties, or even just to the shame of having a finding of liability and an injunction against reoffending against them. ACCC enforcement action also almost always included a requirement that the firm implement or upgrade a trade practices compliance programme—institutionalisation of a form of corporate conscience. Finally, the ACCC often forced or encouraged other parties who had some capacity to regulate the market practices of the relevant industry or individual firms in the case to do so. For example, in a number of cases, as a result of enforcement action, relevant industry associations became committed to educating and empowering business members to take competition law compliance seriously. Simple deterrence will often fail to produce compliance commitment because it does not directly address business perceptions of the morality of regulated behaviour—it merely puts a price on non-compliance and the ability of that price to deter misconduct will depend on the operation of the deterrence trap. Under Fels, the ACCC became expert at o vercoming

Nielsen. This research empirically tested major theories of how businesses respond to regulatory enforcement and why they do or do not comply with the law using survey data from nearly 1,000 larger businesses in Australia, as well as qualitative interviews with ACCC staff, trade practices lawyers and business people. A full list of papers is provided in the Appendix. 11 Christine Parker, ‘The “Compliance” Trap: The Moral Message in Regulatory Enforcement’, 40 Law & Society Review 591, 596–602 (2006).

Effective and Legitimate Enforcement of Competition Law 17 the deterrence trap by communicating a potent moral message about the seriousness of cartel conduct. The ACCC ‘leveraged’ the deterrence impact of their relatively small financial penalties by using the process and publicity to underline their own campaign about the social unacceptability of cartel conduct and enrolling other parties to enforcing compliance. However, this leveraged deterrence strategy also led to conflict between business and the ACCC that showed the ACCC was in a ‘compliance trap’. That is, the ACCC has achieved apparent enforcement success but without gaining the commitment of all business to the substantive moral basis for its enforcement activity; it had not gained business agreement and commitment to its own moral and economic evaluation of what ‘optimal’ deterrence required in terms of the process of enforcement and the substantive goals of competition policy.

3. Wrapped in a mystery … the compliance trap Under Fels and beyond, business criticised and openly expressed its resentment for the ACCC’s enforcement methods, even while settling most cases with the ACCC and apparently improving cooperative compliance. Thus, for many years, the ACCC struggled with the problem that business would cave in to enforcement action because of the harsh, moralising aspect of the sanctions and investigation process as opposed to the fines. However, a business would not necessarily admit that they had broken the law or done anything wrong. This created the conditions in which some businesses could politically attack the ACCC and seek to undermine its success. For example, in the Freight Case,12 the ACCC’s first major cartel enforcement success which ran from 1992 to 1994, the ACCC believed that TNT settled because the weight of evidence collected by the regulator indubitably demonstrated its guilt. However, company insiders put a slightly different slant on it. Although TNT had admitted to the misconduct for the purposes of the settlement, the company had also stated publicly at the time that: I wish to emphasize that the withdrawal of our defense was taken purely for commercial reasons and under no circumstances does TNT accept guilt or liability, nor do we have any evidence that these alleged activities ever took place. In fact, if we had continued our defense we believe we could have won.13

12 TPC v TNT Australia Pty Ltd and Ors [1995] ATPR 41-375. For a fuller description of this case and the methodology used as a basis for this analysis, see Parker, ‘The “Compliance” Trap’, ibid, 599–602. 13 Quote attributed to Mr Fred Millar, General Manager TNT Australia, in ‘Record Fine for TNT, Ansett over Price-fixing’, The Age, 8 November 1994.

18 Christine Parker Another insider stated a number of years later: [W]hen the litigation was proceeding there was at least one story per week in the Australian Financial Review about [the] litigation. The company and its advisors had no doubt that the TPC kept feeding the media. This exposure generated a negative sentiment about the company within the financial sector and make no mistake this may have been a bigger incentive to settle than the costs of the litigation. There is equally no doubt that the costs of the litigation were a major issue for the TNT group. Despite what the TPC may say about its belief in the success of the claim there is no doubt that the defendants were anything but convinced … But again make no mistake that management becomes substantially diverted by such proceedings to the detriment of the performance of the business. It is of course difficult to place amounts upon such matters but it was accepted by the CEO that this detriment far exceeded the legal costs. On a straight financial analysis there could be little extra needed to justify settlement.14

As Chairman of the ACCC between 1991 and 2003, Fels was said to have created ‘Australia’s competition watershed … he electrified business into taking competition law seriously’.15 Yet, at the same time, business, including the chief executive officers of a number of Australia’s biggest companies, regularly and openly criticised the ACCC for adversarialism that they perceived as procedurally unfair and publicity that was stigmatising. Essentially, they claimed that the ACCC was over-reaching: business criticised the ACCC, and Professor Fels himself, for being ‘unfair, unjust and immoral’, and having no ‘line of accountability at all’. The ACCC’s behaviour was criticised for being ‘a corruption of administration of the Trade Practices Act’, ‘false and misleading behaviour on the part of the cop’, and for using publicity ‘in a way that damages companies before they are proven guilty’. One chief executive said that ‘Fels’ use of the media … smacks of the Gestapo’. Another called Fels a ‘smiling assassin’ who had inflicted ‘irreparable harm to the Australian economy’. Professor Fels was described as ‘a maverick autocrat whose overzealous application of merger law retards companies’ ability to grow big enough at home to compete in a globalised marketplace’; and big business argued that the ACCC needed to give Australian businesses much more freedom to merge with one another in order to achieve ‘scale’, so that they could export overseas and Australia could avoid the fate of becoming a ‘branch economy’.16

14 Anonymous informant interviewed by Christine Parker. See Parker, ‘The “Compliance” Trap’, above n 11, 600–601. 15 Brenchley, A Portrait of Power, above n 9, 278–279. 16 Ross Gittins, ‘Perhaps This Is Why Big Business Is Ganging Up on Allan Fels’, The Age, 10 July 2002, 13; Brenchley, A Portrait of Power, above n 9, 220; Cameron Stewart, ‘Making Markets Add Up’, The Weekend Australian, 8–9 June 2002, 21; Christian Catalano, ‘Gloves Off as Retailer Hits Out’, The Age, 30 June 2003, 2. See also Parker, ‘The “Compliance” Trap’, above n 11, 603; Christine Parker and Vibeke Lehmann Nielsen, ‘What do Australian Businesses Really Think of the ACCC and Does it Matter?’, 35 F ederal Law Review 187, 187–188 (2007).

Effective and Legitimate Enforcement of Competition Law 19 Professor Fels argued that big business would always criticise an effective regulator: ‘There is always deep resentment when we prosecute or block mergers or cut prices, no matter how strong the justification’.17 According to him, the ACCC was merely bringing competition and consumer law enforcement strategies ‘into line with North American standards’, and debate should be less about ‘the restrictions that should exist to protect business against an allegedly zealous regulator’ and more about ‘the greater danger that there are few safeguards for business or the public against a tame regulator’.18 Where fulsome political and moral support for the enforcement regime is lacking, the ‘compliance’ trap is set. Responsive regulators find themselves in a dilemma: overcome the deterrence trap by making morally tough demands that not only undermine business commitment to compliance in the longer term (because they lack political legitimacy) but also undermine their own political support (because business will respond by lobbying government to emasculate the regulatory enforcement agency), or avoid conflict with businesses by not making any difference at all. Since political support for a tough, moralising enforcement approach to business regulation is often lacking, active regulatory enforcement agencies who want to be effective often find themselves having to choose whether to jump out of the deterrence trap frying pan into the compliance trap fire.19 It is a ‘compliance’ trap because it occurs only when regulators are actively seeking to improve business compliance, and commitment to compliance, through their enforcement activity. A regulator that engages in formalistic enforcement activity for its own sake will not face this dilemma. It is a ‘trap’ because, in the absence of external political support, there is nothing the regulator can do to escape. The regulator must either choose weakness (no compliance impact) or have weakness thrust upon it (lack of legitimacy leading to emasculation).20 Regulators are thus often forced into a situation where they feel they have to avoid the conflict engendered by strong enforcement action and strong statements of moral responsibility. This may be one reason why regulatory

17 Professor Allan Fels in Malcolm Maiden, ‘The Bell Tolls for Fels at ACCC Kennel’, Business and Money, The Age, 21 June 2003, 1. Elsewhere, Professor Fels was reported as having commented about Gerry Harvey’s criticisms of him (quoted above): ‘Professor Fels said Mr Harvey had been ‘totally uncooperative’ during every step of the ACCC’s proceedings against his company, and had dragged the inquiry to exhaustive lengths’. Catalano, ‘Gloves Off’, ibid. 18 Allan Fels, ‘ACCC Needs Support From the Top’, The Australian Financial Review, 30 June 2003, 55. Professor Fels was also quoted as saying: ‘Some business people are throwing a trial-by-media slogan at practices that are quite normal in the field of law enforcement’. Damon Kitney and Katharine Murphy, ‘Big Business Steps Up Attack on ACCC’, The Australian Financial Review, 13 May 2002, 1. See also Parker and Nielsen, ‘What do Australian Businesses Really Think’, above n 16, 188–189. 19 See Parker, ‘The “Compliance” Trap’, above n 11, 593. 20 Ibid.

20 Christine Parker enforcement agencies often seem to seesaw between formal and punitive, or soft and facilitative, enforcement approaches. It is so difficult to escape the two traps that most regulators most of the time take comfort in a formalism that verges on industry capture—capitulating to business most of the time and taking only the most egregious cases to court in simple deterrence mode the rest of the time.21 The Fels ACCC tried to avoid this approach by using ‘responsive regulation’. Ayres and Braithwaite’s responsive regulation theory suggests that regulators should use multiple enforcement strategies in contextually sensitive ways.22 This theory sees the capacity for regulators to persuade regulated businesses to voluntarily comply in a social interaction as the baseline for a pyramid of enforcement strategies that can be escalated to deterrence where necessary. Deterrence or punishment are considered to be most effective where they are held in reserve, ‘threatening in the background but never threatened in the foreground’,23 and used only in the most egregious cases at the tip of the pyramid. Responsive regulation theory further suggests that a regulatory enforcement agency is only likely to be effective at promoting compliance where it is able to project quite a sophisticated set of messages about itself and its behaviour to those who it regulates: ‘[i]t must have an image of invincibility at the same time as it has an image of mercy and forgiveness’.24 The regulatory agency needs to be both procedurally and substantively just, while being simultaneously accommodating and flexible, and known to be capable of tough and effective enforcement action when a breach occurs. Therefore, we expect that regulators need to try to make sure that businesses (and the public) hold together a quite complex set of potentially contradictory o pinions and expectations about the way that the regulators do their job. It is the combination—the handling of tension between the different dimensions of opinions—rather than the separate dimensions in and of themselves that is likely to be of most practical importance for regulators.25 Nevertheless, the ACCC’s responsive regulation strategy led to conflict between business and the ACCC over the substance of the law, its application in business contexts and the moral seriousness of the social harm caused by its breach. This led to the government commissioning an independent review of competition law and to Professor Fels’s final term of appointment being

Ibid. Ayres and Braithwaite, Responsive Regulation, above n 8. See also Christine Parker, ‘Twenty Years of Responsive Regulation: An Appreciation and Appraisal’, 7 Regulation & Governance 2 (2013). 23 John Braithwaite, Restorative Justice and Responsive Regulation, Oxford University Press, 2002, 119. 24 Ibid. 25 See Parker and Nielsen, ‘The Fels Effect’, above n 9. 21 22

Effective and Legitimate Enforcement of Competition Law 21 shorter than might otherwise have been expected—despite his success and popularity with the general public.26 The compliance trap is a symptom of a socio-polity that has not reached agreement about whether and what business misconduct should be treated as a crime. Debate over what is the ‘optimal’ level of deterrence is really a cipher for this lack of clarity and agreement about the social, moral and political basis of anti-cartel law. ‘Optimal’ deterrence that would allow a certain degree of cartel conduct would make little sense if cartels really are morally reprehensible conduct akin to fraud. Debate about optimal deterrence indicates that it is not clear to what extent anti-cartel law is simply economically ‘regulating’ behaviour or to what extent it is intended to prohibit and stop criminal conduct. That is, it reflects a lack of agreement about the limits of proper business conduct and where the sweet spot of optimal deterrence lies. In the absence of a full arsenal of enforcement tools, and political and cultural support for the law that they are required to enforce, business regulators are being given an impossible task to promote compliance. The compliance trap can only be resolved politically, external to any particular regulatory enforcement encounter. That is, there must be political commitment to the seriousness of the law and its enforcement, or at least a clear acceptance of how much deterrence, and therefore compliance, is optimal. One would expect to see this reflected both in the substance of the law and in material and symbolic support for the regulatory enforcement agency. But this, in turn, may lead to the third and final trap, the legitimacy trap.

4. Inside an enigma … a legitimacy trap? The ACCC responded to the compliance trap by campaigning hard for the criminalisation of cartel conduct and extra resourcing for anti-cartel enforcement.27 (Previously, cartel conduct could be prosecuted as a civil offence only.) The Australian Government did indeed, eventually, provide moral support for the ACCC’s enforcement activities by criminalising cartel conduct and setting jail penalties for individuals involved.28

Parker, ‘The “Compliance” Trap’, above n 11, 612. Caron Beaton-Wells and Fiona Haines, ‘The Australian Conversion: How the Case for Cartel Criminalisation Was Made’, 1 New Journal of European Criminal Law 499 (2010); Caron Beaton-Wells and Fiona Haines, ‘Making Cartel Conduct Criminal’, 42 Australia and New Zealand Journal of Criminology 218 (2009). 28 See ibid, ‘The Australian Conversion’; Caron Beaton-Wells and Brent Fisse, Australian Cartel Regulation: Law, Policy and Practice in an International Context, Cambridge University Press, 2011. 26 27

22 Christine Parker However, the empirical research undertaken by my colleagues and I for the Cartel Project29 found that the legitimacy of criminalising cartel conduct was still questioned by both the general public and the business community. It has been suggested that the moral status of white collar crimes is often ambiguous and controversial. That is, whereas the criminalisation of street crimes often reflects widespread public censure of the conduct involved, when it comes to crimes in the suites, there are not necessarily widely shared norms about what conduct should count as a criminal offence, a civil, administrative or purely regulatory offence, or should be excused or even encouraged. In relation to cartel conduct, it is difficult to pinpoint any unambiguous substantive moral justification for criminalising cartel conduct in Australia apart from the goal of supercharging deterrence. There is still considerable debate in Australia as to whether the rationale for prohibiting cartel conduct is a fairly narrow Chicago School idea of efficiency (popular in Australia since the 1980s) or the older ‘workable competition’ ideas that seeks to balance social welfare against competition. The desire to protect small business has always been—and remains—a particularly strong strand in practical policy debates about Australian competition law. This mirrors some debates in Europe.30 Gerber points out that the goal of ‘economic efficiency’ has been ‘marginal’ in Europe because its ‘abstract economic language of wealth maximization does not comport easily with other [more substantive] goals or with administrative discretion’.31 Anti-cartel law and its enforcement in Australia have fallen into a ‘legitimacy trap’ because of this ambiguity and controversy about whether the law is substantively justified or not—and, if so, on what basis. Lack of a greement about

29 The Cartel Project was based at the University of Melbourne and led by Professor Caron Beaton-Wells, in collaboration with Professors Fiona Haines, Christine Parker and David Round. Full details are available at http://www.law.unimelb.edu.au/cartel/cartelproject. Many of the main empirical findings of the project on public opinion about cartel criminalisation in Australia and its likely impact on deterrence and compliance are summarised in Caron Beaton-Wells and Christine Parker, ‘Justifying Criminal Sanctions for Cartel Conduct: A Hard Case’, 1 Journal of Antitrust Enforcement 198 (2013). A full list of publications from the project can be found in the Appendix. 30 Gerber’s analysis of the origins and development of twentieth-century European competition law and its enforcement identifies four main sets of underlying objectives, none of which relates to efficiency in the Chicago School sense. These four sets of objectives have been: the protection of the freedom of economic actors (appealing to a classical liberal conception of freedom); the prevention of an excessive concentration of economic resources and restraining private economic power in order to achieve and preserve a democratic polity after the Second World War (especially in Germany); ensuring fairness and justice for small and medium-sized enterprises which might be unfairly disadvantaged in competition, and a fair go for the common man; and, as a matter of economic policy, to counter inflation and assist in price control, and later to promote economic development and increase international competitiveness. See David Gerber, Law and Competition in Twentieth Century Europe, Oxford University Press, 1998. 31 Gerber, ibid, 420.

Effective and Legitimate Enforcement of Competition Law 23 the substantive basis for the legitimacy of anti-cartel law means that business people who engage in cartel conduct are able to rationalise away the illegality (and immorality) of their conduct by reference to the ambiguity around the legitimacy of anti-cartel law itself. They do this in three different ways: (i) by avoiding or distancing themselves from knowledge of the anti-cartel law and the fact that it applies to them; (ii) by resisting and d enying the moral legitimacy of the law; and (iii) by knowingly playing games with the law so as to avoid its application to them.

4.1 Avoiding or distancing themselves from knowledge of the anti-cartel law and the fact that it applies to them In the Cartel Project survey of business people, less than half could identify that a very clear cut case of price fixing in a short scenario was in fact a criminal offence and less than a quarter could identify that jail was available as a penalty.32 Accurate knowledge of the law correlated with substantive agreement with the law—that is, business people who thought cartel conduct ought to be criminal were more likely to know or believe that it was in fact criminal.33

4.2 Resisting and denying the moral legitimacy of the law The Cartel Project team interviewed people who had been subject to cartel enforcement in the past. It found that small business people and managers lower down corporate hierarchies generally held a resistant view to the expansion of completion through anti-cartel law. They understood themselves to be ‘innocent’ and ‘naive’ about competition law. That is, they claimed that they did not know much about the law when they engaged in the conduct later found to be illegal, they did not consider that the law might apply to them and they felt that the law should not apply to them. They understood their own businesses as too small for the law to apply to and their own conduct as harmless. They understood themselves to be acting according to alternative values that they had to put into action in order to survive politically, socially and economically. In some cases, the judge had indeed acknowledged that these interviewees showed little moral blameworthiness. However, although this might be relevant to the quantification of penalties, it was not relevant 32 See Christine Parker, ‘The War on Cartels and the Social Meaning of Deterrence’, 7 Regulation & Governance 174 (2013); Beaton-Wells and Parker, ‘Justifying Criminal Sanctions’, above n 29. 33 Ibid, ‘The War on Cartels’.

24 Christine Parker to liability since anti-cartel offences are per se breaches.34 In the quotations below, each paragraph comes from a different interviewee.35 If I knew that it was a contravention of the Act at the time I wouldn’t have done it … Perfectly sensible undertaking. Perfectly reasonable. No immoral connotation to that at all … It just didn’t occur to me there’d be a problem … All this Trade Practices stuff, as far as I can tell, does criminalise conduct which is not criminal. It’s just normal—this is all normal commercial behaviour. And cartels are too, I’m sure … I was sort of going on my sense of morality. I shouldn’t say they’re not aware [of the anti-cartel law]. They’d probably be aware from the big corporations, alright, that’s as far as the extent goes. But to talk candidly to one of your competitors, what did you go in for, or what do you reckon it’s worth, and blah, blah, blah. And you could go and cop a big fine for that. No one’s aware of that … because the legislation is a pile of books like that [indicating a very high pile with his hand]. A lawyer doesn’t even know what actually specifies every sub-clause in every paragraph, how are we [as normal workers] supposed to know? If you look at something like the Pratt situation [referring to a very well known price fixing cases involving two large cardboard box manufacturers] then well that’s probably up there with the criminal issues because it really is a deliberate attempt to extract more money … on a fairly massive scale. But for the little blokes who are sort of plodding along essentially trying to find the margin, and we’re talking margins of two or three per cent [this should not be criminal]. You get four or five big concrete blokes sitting around a room like this and they say right we’re going to set the price—the concrete at such and such a price, and then you’re going to have that job and that job. That’s collusion at its worst … We didn’t make heaps of money out of it. That’s the criminalisation, if we were ripping off people and all driving around in the latest vehicles and going overseas. Mate! But we weren’t, we’re not doing that. Probably in a case like ours [all small operators], you could probably separate— create a separation. Well, yeah, we were just merged in this contravening act, the same as someone who’s a corporation that had contravened the act and ripped off bloody millions of people. Now, you know, lumped into the same category.

4.3 Knowingly playing games with the law so as to avoid its application to them On the other hand, executives higher up the hierarchy in larger businesses, and external and in-house lawyers, tend to embrace the official rationale for

34 Parker, ‘The War on Cartels’, above n 32; Christine Parker, ‘Economic Rationalities of Governance and Ambiguity in the Criminalization of Cartels’, 52 British Journal of Criminology 974 (2012). 35 All quotations are from the two papers in the previous footnote.

Effective and Legitimate Enforcement of Competition Law 25 anti-cartel law at least in their public statements. However, at the same time they are often able to take advantage of the fact that they operate in large organisations and there is ambiguity about individual versus organisational responsibility for cartel conduct in the law to avoid personal and organisational responsibility for cartel conduct. Instead, they are able to shift liability to individuals and units down the line.36 It was a one-off incident, by one person, with nothing anywhere that would show that it was part of a broader policy within the business, that this was the way things were done. I don’t think there’s anything else we could have done to prevent the conduct at the time. This was a rogue senior manager in charge of a rogue department. He was trained, he had a compliance program. He knew there was an issue and that’s the reason why he sought legal advice. He, on the record, had told his people of having inappropriate discussions with competitors. He knew the law. He made a conscious decision.37

4.3.1 The legitimacy trap, the deterrence trap and the compliance trap The ACCC fell into the legitimacy trap after escaping both the deterrence and the compliance traps. The legitimacy trap, however, feeds into and amplifies both the deterrence trap and the compliance trap. The legitimacy trap amplifies the deterrence trap because, as we have seen, where people do not substantively agree with the law, they are more likely to exhibit bias in hearing and understanding what the law is and the fact that it applies to them at all. Hence they will tend to underestimate the application of deterrence to their own conduct. This, in turn, requires larger and larger penalties and more stigmatising enforcement and publicity to bring it to their attention, and thus further amplifies the deterrence trap. The legitimacy trap amplifies the compliance trap because, where people do not agree with the substantive justification given for the law or do not feel that there is any clear substantive justification, they are also more likely to find ways in which they believe the procedural application of the law is at fault. Thus their outrage at the immorality of both the means and goals of the regulator is likely to be heightened. For example, the ACCC Enforcement and Compliance Project found that regulatees’ assessments of strategic sophistication—an opinion about how well the ACCC’s procedures and strategies contributed to good substantive policies—are significantly related to normative commitments to compliance. These findings suggest that we need

36 Ibid. See also Jeffrey Sonnenfeld, ‘Executive Apologies for Price Fixing: Role Based Perceptions of Causality’, 24 The Academy of Management Journal 192 (1981). 37 Quote from Parker, ‘The War on Cartels’, above n 32, 188.

26 Christine Parker to weave examination of perceptions of substance and procedure together in understanding how opinions of regulators affect compliance. For example, Australian businesses’ assessments of whether the ACCC behaves in a procedurally ‘unreasonable’, ‘aggressive’, ‘unfair’ or ‘uncooperative’ way appear to be influenced by how businesses perceive the substantive goals of the ACCC. The more they see competition and consumer policy as morally important, the more likely they are to see tough public enforcement action as reasonable and fair. The more they disagree with the priorities of the competition law and the ACCC’s interpretation of it, the more likely they are to label any attempts by the ACCC to enforce the law as unfair.38 This will lead the regulator further into the compliance trap as it either tries to argue ever more strongly that the law it enforces is morally justified and its own enforcement actions significant and important or it backs off.

5. Conclusion: regulation and the politics of markets Ambiguity about the legitimacy of white collar or business crime is endemic because conflict about the limits and responsibilities of markets, commodification and business is fundamental to the nature of modern democratic capitalist societies. Conflict about competition policy and law, what competition should and should not mean, to what extent it should be limited or balanced against social and political goals, or to what extent it should be a sole policy goal are necessarily at the centre of democratic debate in contemporary regulatory capitalism. Many arguments over the procedural legitimacy or illegitimacy of competition law enforcement stem from fundamental conflict about whether breaches of competition law represent conduct that should be considered criminal and immoral (and therefore deserve punishment) or whether they are merely conduct that is economically non-optimal in certain circumstances (and therefore need to be regulated). The former, more substantive, moral rationale for competition law would also imply that higher procedural fairness standards should be followed in its enforcement. Yet the political choices behind the substantive rationale for competition law and the procedures for its enforcement are often hidden. The point is that this substantive debate about the place of business, competition, markets, commodification, economisation in our societies and polities cannot be bracketed from a discussion of the procedural or legal-technical

See Parker and Nielsen, ‘The Fels Effect’, above n 9.

38

Effective and Legitimate Enforcement of Competition Law 27 effectiveness and legitimacy of competition law enforcement and sanctioning. People’s substantive understanding of and agreement or disagreement with the law will inevitably seep into their responses to attempts at enforcement and sanctioning. In other words, effective enforcement and sanctioning is not just a technical skill; it also requires the support of a robust democratic sociopolity which has the opportunity and takes the time to deliberate and contest the meaning and shape of the law to be enforced on an ongoing basis.

Appendix: references to research on which this chapter is based The Australian Competition and Consumer Commission Enforcement and Compliance Project http://www.law.unimelb.edu.au/cartel/related-projects/the-australian-competitionand-consumer-commission-enforcement-and-compliance-project

Online reports Ainsworth, P, Parker, C and Stepanenko, N (2004) ACCC Enforcement and Compliance Project: The Impact of ACCC Enforcement Activity in Cartel Cases (Centre for Competition and Consumer Policy, Australian National University). Nielsen, VL and Parker, C (2005) The ACCC Enforcement and Compliance Survey: Report of Preliminary Findings (Centre for Competition and Consumer Policy, Australian National University). Parker, C and Nielsen, VL (2008) ACCC Enforcement and Compliance Project: Explanation of Project and Methodology (Centre for Competition and Consumer Policy, Australian National University). Parker, C and Stepanenko, N (2003) ACCC Enforcement and Compliance Project: Preliminary Research Report (Centre for Competition and Consumer Policy, Australian National University). Parker, C, Braithwaite, J and Stepanenko, N (2004) ACCC Enforcement and Compliance Project: Report on ACCC Compliance Education and Liaison Strategies (Centre for Competition and Consumer Policy, Australian National University). Sharpe, M and Parker, C (2006) ACCC Enforcement and Compliance Project: Assessment of the Impact of ACCC Regulatory Enforcement Action in Unconscionable Conduct Cases (Centre for Competition and Consumer Policy, Australian National University).

Book Parker, C and Nielsen, VL (eds) (2011) Explaining Compliance: Business Responses to Regulation (Edward Elgar Publications).

28 Christine Parker

Book chapters Parker, C and Gilad, S (2011) ‘Compliance Management Systems: Structure, Agency and Culture‘, in C Parker & V Nielsen (eds), Explaining Compliance: Business Responses to Regulation (Edward Elgar Publications) 170–95.

Journal articles Parker, C (1999) ‘Compliance Professionalism and Regulatory Community: The Australian Trade Practices Regime’ Journal of Law & Society 26(2), 215. Parker, C (1999) ‘Evaluating Regulatory Compliance: Best Practice and Standards’ Trade Practices Law Journal 7(2), 62. Parker, C (2003) ‘Regulator-required Corporate Compliance Program Audits’ Law and Policy 25(3), 221–244. Parker, C and Nielsen, VL (2006) ‘Do Businesses Take Compliance Seriously? An Empirical Study of Implementation of Trade Practices Compliance Systems in Australia’ Melbourne University Law Review 30, 441. Parker, C (2006) ‘The Compliance Trap: The Moral Message in Responsive Regulatory Enforcement’ Law & Society Review 40(3), 591. Parker, C and Nielsen, VL (2007) ‘What do Australian Businesses Think of the ACCC and Does it Matter?’ Federal Law Review, 35, 187. Parker, C and Nielsen, VL (2008) ‘How Much Does it Hurt? How Australian Businesses Think About the Costs and Gains of Compliance with the Trade Practices Act’ Melbourne University Law Review 32(2), 554. Parker, C and Nielsen, VL (2007) ‘To What Extent do Third Parties Influence Business Compliance Management Behaviour?’ Journal of Law & Society 35(3), 309. Parker, C and Nielsen, VL (2009) ‘Corporate Compliance Systems: Could They Make Any Difference?’ Administration & Society 41(1), 3. Parker, C and Nielsen, VL (2009) ‘Is Anyone Out There Listening?’ Trade Practices Law Journal 17, 106. Parker, C, Rosen, RE and Nielsen, VL (2009) ‘The Two Faces of Lawyers: Professional Ethics and Business Compliance with Regulation’ Georgetown Journal of Legal Ethics 22, 201. Parker, C and Nielsen, VL (2011) ‘Deterrence and the Impact of Calculative Thinking on Business Compliance with Competition and Consumer Regulation’ The Antitrust Bulletin 56(2), 377. Parker, C and Nielsen, VL (2011) ‘The Fels Effect: Responsive Regulation and the Impact of Business Opinions of the ACCC’ Griffith Law Review 20(1), 91. Parker, C and Nielsen, VL (2012) ‘Mixed Motives: Economic, Social and Normative Motivations in Business Compliance’ Law & Policy 34, 428. Parker, C (2013) ‘Twenty Years of Responsive Regulation: An Appreciation and Appraisal’ Regulation & Governance 7(1), 2. Nielsen, VL and Parker, C (2009) ‘Testing Responsive Regulation in Regulatory Enforcement’ Regulation and Governance 3, 376. Rosen, R.E., Parker, C and Nielsen, VL (2012) ‘The Framing Effects of Professionalism: Is There a Lawyer Cast of Mind? Lessons from Compliance Programs’ Fordham Urban Law Journal XL, 297.

Effective and Legitimate Enforcement of Competition Law 29 Sharpe, M and Parker, C (2007) ‘A Bang or a Whimper? The Impact of ACCC Unconscionable Conduct Enforcement’ Trade Practices Law Journal 15(3), 139.

The Cartel Project http://www.law.unimelb.edu.au/cartel/cartel-project

Online reports Beaton-Wells, C, Haines, F, Parker, C and Platania-Phung, C (2010) Report on a Survey of the Australian Public Regarding Anti-Cartel Law and Enforcement (December). Parker, C and Platania-Phung, C (2012) The Deterrent Impact of Cartel Criminalisation: Supplementary Report on a Survey of Australian Public Opinion Regarding Business People’s Views on Anti-Cartel Laws and Enforcement (12 January). Parker, C, Haines, F, Kotey, J, Nankivell, J and Round, D (2011) Report on Interviews with Civil Respondents in Cartel Cases (December).

Books Beaton-Wells, C and Ezrachi, A (eds) (2011) Criminalising Cartels: Critical Studies of an International Regulatory Movement (Hart Publishing). Beaton-Wells, C and Fisse, B (2011) Australian Cartel Regulation: Law, Policy and Practice in an International Context (Cambridge University Press).

Book chapters Beaton-Wells, C (2011) ‘Australia’s Criminalisation of Cartels: Will It Be Contagious?’ in C Beaton-Wells and A Ezrachi (eds), Criminalising Cartels: Critical Studies of an International Regulatory Movement (Hart Publishing). Beaton-Wells, C (2011) ‘Cartel Criminalisation and the Australian Competition and Consumer Commission—Opportunities and Challenges’ in C Beaton-Wells and A Ezrachi (eds), Criminalising Cartels: Critical Studies of an International Regulatory Movement (Hart Publishing). Beaton-Wells, C (2012) ‘Criminal Sanctions for Cartels—The Jury is Still Out’ in A Ezrachi (ed), International Research Handbook in Competition Law (Edward Elgar Publications). Beaton-Wells, C (2012) ‘The Billionaire, Prime Minister and Chairman: ACCC v Visy Ltd’ in B Rodger (ed), Landmark Cases in Competition Law (Kluwer International (Law)). Parker, C (2011) ‘Criminalisation and Compliance: The Gap between Rhetoric and Reality’ in C Beaton-Wells and A Ezrachi (eds), Criminalising Cartels: Critical Studies of an International Regulatory Movement (Hart Publishing).

30 Christine Parker

Journal articles Beaton-Wells, C and Fisse, B (2010) ‘Broadening the Definition of Collusion: A Call for Caution’ 38 Federal Law Review, 71. Beaton-Wells, C and Fisse, B (2011) ‘The Competition and Consumer Amendment Bill (No 1) “(Exposure Draft)—A Problematic Attempt to Prohibit Price Signalling”’ Australian Business Law Review 39(1), 28. Beaton-Wells, C and Fisse, B (2011) ‘US Policy and Practice in Pursuing Individual Accountability for Cartel Conduct: A Preliminary Critique’ Antitrust Bulletin 56(2), 277. Beaton-Wells, C and Haines, F (2009) ‘Making Cartel Conduct Criminal’ Australia and New Zealand Journal of Criminology 42(2), 218. Beaton-Wells, C and Haines, F (2010) ‘The Australian Conversion: How the Case for Cartel Criminalisation Was Made’ New Journal of European Criminal Law 1(4), 500. Beaton-Wells, C and Haines, F (2012) ‘Ambiguities in Criminalising Cartels: A Political Economy’ British Journal of Criminology 52, 953. Beaton-Wells, C and Parker, C (2012) ‘Education before Enforcement? Key Insights from Cartel Research’ Australian Business Law Review 40(1), 1. Beaton-Wells, C and Parker, C (2013) ‘Justifying Criminal Sanctions for Cartel Conduct’ Journal of Antitrust Enforcement 1(1), 198. Beaton-Wells, C and Platania-Phung, C Anti-Cartel Advocacy: How Has the ACCC Fared?’ (2011) ‘Sydney Law Review, 367. Parker, C (2012) ‘Economic Rationalities of Governance and Ambiguity in the Criminalization of Cartels’ British Journal of Criminology 52, 974. Parker, C (2013) ‘The War on Cartels and the Social Meaning of Deterrence’ Regulation & Governance 7(2), 174. Round, D and Agarwal, M (2010) ‘The Criminalisation of Serious Cartel Conduct in Australia’ The CPI Antitrust Journal 1, May.

Konrad Ost* From Regulation 1 to Regulation 2: Enforcement of EU Law by National Sanctioning Regimes and the Need for Further Convergence

1. Introduction Regulation 1/2003 and the 2005 Amendment to the German Competition Act (Gesetz gegen Wettbewerbsbeschränkungen, GWB) have changed the legal and political framework conditions for antitrust enforcement in Germany immensely. A key aspect of this change is the decentralisation of enforcement due to the closer integration of the national competition authorities (NCAs) and courts in the application of European law. The new changes found their expression at the legislative level in a certain convergence of national enforcement regimes with the EU enforcement system and at the institutional level in the focus on prosecuting hardcore cartels. In this particular area, the main tool of public enforcement are large fines, which are intended to reduce the economic incentives of companies to engage in cartel behaviour. Antitrust enforcement in Germany is currently faced with huge challenges. In view of the comparably manageable number of cases and moderate fines up until 2000, the German enforcement framework did not pose significant problems to cartel prosecution. Today, however, it can hardly cope with the number and scope of the cases. Whereas, for example, during the period from 1994 to 1997 a total of seven cartel proceedings were initiated, five concluded and fines amounting to less than 200 million euros imposed, from 2006 to 2009, 20 cartel proceedings were initiated, 14 concluded and fines amounting to almost a billion euros imposed. In 2012 alone, 12 cartel proceedings were * Director of General Policy, Bundeskartellamt, Bonn. All opinions expressed herein are personal. This chapter is partly based on the paper ‘Kartellbußgeldverfahren zwischen deutschem Systemdenken und europäischer Konvergenz’, drafted by K Ost/St Häfele and presented by the Bundeskartellamt at the annual meeting with German academics on 4 October 2012. See also K Ost, ‘Antitrust Procedure and the European Convention of Human Rights—The Right to an Independent and Impartial Tribunal and the Right not to Incriminate Oneself—With Special Regard to the Situation in Germany’, in C Baudenbacher, ed, Current Developments in European and International Competition Law, 14th St Gallen International Competition Law Forum 2007, Helbig & Lichtenhahn, 2008, 201 et seq. I am grateful to Daniel Dittert, Steffen Häfele, Gero Meeßen, Jörg Nothdurft, Wouter Wils and the participants at the 18th Annual EU Competition Law and Policy Workshop for stimulating discussions and/or for comments on the draft on which this chapter is based.

32 Konrad Ost concluded, with fines amounting to 312 million euros. Along with the increase in the absolute number of cases, the administrative burden and duration of cartel proceedings have reached a critical level. Key problems here can be found in the areas of sanctions and procedures. It is therefore time to take a critical look at the enforcement of EU law by the German system and seek constructive solutions to keep cartel prosecution both effective and efficient. Almost 10 years after the entry into force of Regulation 1/2003, the perspective on the German anti-cartel enforcement system is necessarily a European one. In 2002, Céline Gauer asked: ‘Does the effectiveness of the EU network of competition authorities require a certain degree of harmonisation of national procedures and sanctions?’1 After more than nine years of experience with Regulation 1/2003, this chapter argues in the affirmative.

2. Effective and efficient cartel prosecution Cartel prosecution fluctuates in an area of conflict between fundamental economic and legal requirements. On the one hand, it has to be effective and efficient both from an economic as well as a legal perspective; on the other hand, it has to guarantee a human rights standard of protection for those suspected of cartel behaviour. As far as the ‘optimal sanctions’ are concerned—in brief and simplified terms—one can conclude that, from an economic perspective, the optimal fine should exceed the expected gain from the infringement multiplied by a factor of probability of detection and punishment.2 The costs of anti-cartel enforcement, which have to be borne by the state and the market players, also have to be taken into account. However, it has been argued repeatedly that optimal antitrust enforcement is not about a simple equation3 but that, once a certain level of sanctions is 1 C Gauer, ‘Does the Effectiveness of the EU Network of Competition Authorities Require a Certain Degree of Harmonisation of National Procedures and Sanctions?’, in C-D Ehlermann and I Atanasiu, eds, European Competition Law Annual 2002: Constructing the EU Network of Competition Authorities, Hart Publishing, 2005, 187 et seq. 2 For a more elaborate equation see N Calviňo, ‘Public Enforcement in the EU: Deterrent Effect and Proportionality of Fines’, in C-D Ehlermann and I Atanasiu, eds, European Competition Law Annual 2006: Enforcement of Prohibition of Cartels, Hart Publishing, 2007, 317 et seq, text accompanying footnote 2: pU (Y − c − F) > (1 − p)U(Y), where p = probability of detection and enforcement of sanctions; U = utility function of the offender; Y = net income from illegal activity; c = costs incurred in case of detection, excluding fines (lawyers, stigmatising effect, etc.); and F = level of fines. 3 W Wils, Efficiency and Justice in European Antitrust Enforcement, Hart Publishing, 2008, chapter 3, paras 199 et seq, text accompanying footnote 63.

From Regulation 1 to Regulation 2 33 to be expected, the dominant factor of deterrence is the expectation of being discovered and fined. Assuming that most national enforcement regimes provide for harsh sanctions from the perspective of a perpetrator who is actually caught, the key issue of deterrence is the probability of being detected and punished. The latter depends largely on the procedural framework, the resources available and prioritisation by the enforcing bodies. As resources are not unlimited and anti-cartel enforcement is not the sole priority of competition agencies, a major factor for effective enforcement is the procedural framework. Even if the economic perspective cannot be the sole guiding principle for shaping cartel prosecution, it is nevertheless a necessary basis for analysing and assessing its effectiveness and efficiency. This analysis has to be supplemented by a legal analysis, ie an assessment of the guarantees of due process and requirements of European law.

3. Sanctioning cartel infringements in Germany—challenges Cartel prosecution in Germany is burdened with a number of problems in respect of not only its efficiency, but also to a greater extent its effectiveness. They relate to the areas of sanctions and procedural law.

3.1 Sanctions 3.1.1 Addressee of Article 101 TFEU—the concept of an ‘undertaking’ The prohibition of cartels applies to ‘undertakings’. According to the interpretation of the European Court of Justice (ECJ), the prohibition of cartels applies to a company or a group of companies as a single economic unit. The concept of an undertaking covers any entity engaged in an economic activity, regardless of its legal status and the way in which it is financed. That concept must be understood as designating an economic unit even if in law that unit consists of several natural or legal persons. When such an economic entity infringes the competition rules, it is for that entity, according to the principle of personal responsibility, to answer for that infringement.4

4 Joined Cases C‑628/10 P and C‑14/11 P Alliance One International Inc and Standard Commercial Tobacco Co Inc v Commission, EU:C:2012:479, para 42 (citations omitted). See also Case C-97/08 P Akzo Nobel NV v Commission [2009] ECR I-8237, para 56.

34 Konrad Ost The undertaking commits the substantive violation of law. As the parent company and subsidiary together form the undertaking,5 this results in a joint objective violation of law by the parent company and subsidiary. In this sense, the parent company also violates the substantive prohibition of cartels if its subsidiary commits an infringement.6 As a consequence, if punished by the Commission, the parent company may be held liable and the fine may be imposed jointly on the parent company and its subsidiary. This wide notion of parental liability explains the fact that problems of legal succession are rarely an issue in Commission proceedings. Furthermore, due to the wide notion of undertaking as interpreted by the European courts, it is almost impossible for the undertakings to escape liability by any form of internal restructuring within the group.7 Furthermore, the general concept of legal succession encompasses most issues of whether liability survives the existence of the undertaking. Under EU law, both legal8 and economic succession are covered to a broad extent and liability is transferred to the successor.9

3.1.2 Parental liability General principles According to the current understanding of German administrative offences law, this leap from substantive law to sanctions law would be one too far. Under German administrative penal law, the addressee of a fine is a legal person. This legal entity is liable for any offence committed by a representative of that specific legal entity. According to this logic, the action of a representative

5 J Kokott and D Dittert, ‘Die Verantwortlichkeit von Muttergesellschaften für Kartellvergehen ihrer Tochtergesellschaften im Lichte der Rechtsprechung der Unionsgerichte’ [‘The Liability of Parent Companies for Cartel Law Violations Committed by their Subsidiaries in the Case Law of the Courts of the European Union‘], [2012] Wirtschaft und Wettbewerb—WuW 670, 672 et seq. 6 In terms of sanctions law, this means that, under European law, fault can be attributed to the parent company as part of the corporate group. See Kokott and Dittert, ibid, 675. At 672, they state: ‘Both bear the responsibility under cartel law for their own actions; the fault cannot be attributed to another party and cannot be understood merely as a liability for organisational faults or an omission of duty on the part of the parent company vis-à-vis its subsidiary’ (author’s translation). 7 Joined Cases C-204/00 P, C-205/00 P, C-211/00 P, C-213/00 P, C-217/00 P and C-219/00 P Aalborg Portland A/S and Others v Commission [2004] ECR I-123, paras 356–359; Case C-280/06 Autorità Garante della Concorrenza e del Mercato v Ente tabacchi italiani—ETI SpA and Others [2007] ECR I-10893, para 48. 8 Case C-42/92 P Commission v Anic Partecipazioni SpA [1999] ECR I-4125, para 145; Aalborg Portland, ibid, para 359; Case C-125/07 P Erste Group Bank AG v Commission [2009] ECR I-8681, para 78. 9 For a recent account of the case law, see A Brown and M Schonberg, ‘Widening the Net: The General Court Extends the Principle of Successor Liability in EU Competition Law’, 34 European Competition Law Review 1 (2013).

From Regulation 1 to Regulation 2 35 of another legal entity (even if it is a member of the same economic unit in European terms) may not be attributed to another legal person (eg the parent company). Concept of supervisory duty However, in interpreting German administrative penal law, the Bundeskartellamt may nevertheless hold the parent company indirectly liable for an infringement committed by its subsidiary. German administrative penal law bestows a duty on the representative of a company to supervise its workforce to make sure that it does not commit any offences in relation to the legal obligations of the company. Whether this concept of supervisory duty may be applied to the relationship between a parent company and its subsidiary is much debated. The answer is given in the European concept of undertaking: it is up to European law to define the scope of violation of substantive law ‘by the undertaking’. If a violation of European competition law is to be expected from an undertaking, the undertaking is obliged to cease such action or to prevent such infringements within its own economic unit. In other words, if, under substantive European law, a cartel law violation by a subsidiary is to be automatically regarded as an infringement by the parent company,10 the parent company can only fulfil its obligation under the provision by preventing a potential infringement by carrying out supervisory measures. These supervisory duties follow directly from European law. Comparison The fining of a parent company failing to fulfil supervisory obligations within a group of undertakings has a similar function to the European concept of parental liability. The Bundeskartellamt is currently imposing fines based on such an assumed duty to supervise. Nevertheless, the German law still falls short of the European enforcement system. A sanction imposed under section 130 of the German Administrative Offences Act holds the company in question liable for breach of a supervisory duty, not for participating in the infringement. It may be more difficult to prove a breach of a supervisory duty than to establish the existence of an economic unit. Sister companies may have no supervisory duty at all, even if they might be held liable under certain circumstances for the acts of other members of the economic unit. Compliance programmes which do not influence the level of fines under European law may exonerate the parental company under the concept of supervisory duty. In addition, the Bundeskartellamt’s interpretation of the law in this

Kokott and Dittert, above n 5, 672.

10

36 Konrad Ost context has repeatedly been challenged. The courts have not yet had the opportunity to state their views on the issue. A glance at the jurisdictions of other NCAs within the European Competition Network (ECN) shows that there is no uniform approach to parental liability. There are reports that, for example, Polish and Danish competition law provide for even less parental liability than German law.

3.1.3 Succession German case law In two cases, Germany’s Bundesgerichtshof (Federal Court of Justice, FCJ) decided in 2011 that, for legal succession reasons, no fines could be imposed.11 The court’s interpretation led to significant enforcement gaps: where an infringing entity was merged into another entity in a merger between equals, it was not possible to fine the successor entity. This could even apply where two participants in the same cartel merged after having been fined. Also, corporate groups could avoid liability for antitrust infringements by corporate restructuring: an infringing entity within a group could be merged into another group entity to avoid the fine. Alternatively, an infringing entity’s assets could be transferred to another group entity, leaving the infringing entity as a mere shell without means to pay the fine. The 2013 reform of the GWB The judiciary called quite openly for the legislator to close this loophole. The 2013 amendment of the GWB provides that, in all cases of universal succession of the infringing legal entity by another entity, a fine may be imposed on the successor entity. The fine imposed on the successor entity may in that case not exceed the value of the assets that the successor entity acquired by absorbing the infringing entity. This amendment will remedy some but not all of the problems described in the context of legal succession. In particular,

11 See Cases KRB 2/10 Fusion Transportbetonhersteller and 55/10 Versicherungsfusion of 10 August 2011. According to the FCJ, it was only possible to hold a successor entity liable for an infringement committed by the infringing entity where the following conditions are cumulatively met. First, universal legal succession of the infringing entity by the successor entity, ie all assets and liabilities of the infringing entity must have been transferred to the succeeding entity in a single legal act. Second, the infringing entity must have ceased to exist. An example would be the merger of one company into another company. And third, there must be economic identity between the infringing entity and the successor entity. This is only the case where the assets of the infringing entity are used by the successor entity in the same way they were used by the infringing entity (ie there must be a continuation of the same economic activity) and constitute the overwhelming part of the assets of the successor entity.

From Regulation 1 to Regulation 2 37 the amendment does not take account of those cases of purely economic succession where the infringing entity still exists but its assets are transferred to another (group) entity. In such a case, the Bundeskartellamt would still be limited to imposing a fine on an entity that is a mere shell without sufficient means to pay a fine. In addition, it would still not be possible to tackle cases where a fined entity divests part of its assets by spinning off a separate legal entity (Abspaltung and Ausgliederung, according to the German company transformation law).12 These loopholes have led to undertakings and their lawyers continuously debating options for avoiding a fine or having a fine reduced. Again, a glance at the ECN shows a variety of approaches. The majority of jurisdictions probably follow the EU model. However, major jurisdictions such as Poland are reported to have similar loopholes.

3.1.4 Maximum level of fines: the 10 per cent rule Legislative convergence The 2005 reform of the GWB aligned the German law with Article 23 of Regulation 1/2003. An undertaking may be fined up to 10 per cent of its annual turnover. However, as far as coherent enforcement under Regulation 1/2003 is concerned, the most recent development in case law in Germany is interesting. Interpretation by the German Federal Court of Justice With its decision of 26 February 2013, the FCJ confirmed that the central fining provision of German antitrust law is in line with the German constitution. The decision also confirmed the fines imposed on the members of a cement cartel. It was the largest total amount of fines ever imposed by G erman competition courts.13 The main argument raised by the parties was that the 10 per cent rule is against the principle of legality enshrined in the German Constitution. In the previous case law of the Bundesverfassungsgericht (Federal Constitutional Court, FCC) it was held that, under the principle of nulla poena sine lege, a fines provision needs to provide an absolute frame within which a fine must be set and give some guidance on how the fine to be imposed on the perpetrator should be assessed. The FCJ ruled that the current interpretation of the 10 per cent rule as a mere capping threshold violates the German constitution. Nevertheless, the court rejected the appeal against 12 K Ost, ‘Die Regelung der Rechtsnachfolge und weitere Neuerungen im Kartellordnungswidrigkeitenrecht durch die 8. GWB-Novelle’, in F Bien, ed, Das deutsche Kartellrecht nach der 8. GWB-Novelle, Nomos, 2013, 305 et seq, 310 et seq. 13 Together with the fines not appealed, the total amounted to approximately 400 million euros.

38 Konrad Ost the decision, reasoning that an alternative interpretation of the 10 per cent rule as an upper limit of a framework/range of fines would be in line with the principle of legality. The main difference to the current methodology is that the 10 per cent must be reserved for the ‘worst possible offence’. Any ‘bottom up’ calculation would be questionable. The fine has to be assessed with the lowest and highest possible fines in mind. An interpretation as a capping threshold would offer no orientation for setting the amount. This is required of any criminal offences in Germany (for example, 1–10 years’ imprisonment for kidnapping provides for the relevant ‘scale for penalty assessment’). The gravity of the offence has to be assessed within the scale provided for in the law and an adequate fine has to be set. The 2013 Fining Guidelines Even if this seems only a minor shift in its enforcement policy, the Bundeskartellamt has had to change its methodology for calculating fines. On 25 June 2013 it issued new guidelines14 aimed at combining the new approach of the judiciary with the concept of ‘affected turnover’ of the European fining guidelines. The new system will not dramatically change the landscape of fining. However, it is likely that ‘single-product’ offenders will benefit from the new system, while ‘multi-product’ offenders will potentially face higher fines. This may lead to some discussion about the Commission’s enforcement practice. Outlook The FCJ avoided creating a new loophole in the enforcement system. However, it is notable that the highest court on competition matters in Germany, while making only brief reference to the ECJ’s judgment in Degussa,15 ruled on the same issue in quite the opposite direction. On this aspect, there is a clear divergence between German and European case law even on identical texts. Within the ECN, most jurisdictions provide for a system based on some kind of a 10 per cent rule. However, a huge variety of methodologies are used by the jurisdictions to calculate the actual fine. The German Federal Constitutional Court may have the opportunity to rule directly on the European approach in the near future. The question of the legal certainty of Article 23 Regulation 1/2003 is pending at the FCC

14 2013 Guidelines for the Setting of Fines in Cartel Administrative Offence Proceedings, available at www.bundeskartellamt.de. 15 Case C-266/06 P Evonik Degussa GmbH v Commission [2008] ECR I-81*, paras 45 et seq; Opinion of Advocate General Kokott in Case C‑501/11 P Schindler Holding Ltd and Others v Commission, EU:C:2013:522, para 148 et seq; K Lenaerts, ‘Due Process in Competition Cases’, [2013] NZKart—Neue Zeitschaft für Kartellrecht 175, 180.

From Regulation 1 to Regulation 2 39 following a constitutional complaint against a fines decision in the alloy surcharge cartel that was upheld by the ECJ.16

3.2 Procedure 3.2.1 The procedural framework in Germany German fines procedure The German fines procedure in competition matters follows the general rules for regulatory offences. German penal law was reformed in the late 1960s as part of a movement to decriminalise minor offences such as traffic misdemeanours. The procedure is very similar to a criminal procedure, with one important difference: the preliminary proceedings take place before an administrative authority. The regulatory offence is punishable by an administrative decision imposing a fine. In the case of appeal, the authority then transfers the file to the public prosecutor, who will submit the case to the court and thereupon assume the function of the prosecuting authority. The court will then usually hold a hearing and deliver a first instance judgment, which may include a more severe sentence. The decision of the competition authority is not subject to review but is replaced in its entirety. Rights of defence and procedural formalities The persons subject to the procedure have a maximum standard of rights of defence. The procedure is mainly oral. Direct witness evidence often prevails over documentary evidence. Experience shows that this type of procedure, which is perfectly suitable for traditional criminal proceedings, is rather ill-suited to complex economic litigation. This can be seen, for example, from the fact that there are almost no time limits for the presentation of new facts. The addressee’s appeal does not even have to be reasoned. German procedural rules provide for a very extensive right not to incriminate oneself (right to remain silent). In general, companies and individuals under investigation are not obliged to cooperate. In a first small step, the legislator introduced a limited duty of cooperation for legal entities with regard to providing data on turnover (new section 81a 16 See S Berra, ‘German Court to Consider Review of EU Judicial Process’, Global Competition Review, 25 February 2013. See also the decision of the Federal Constitutional Court declining to grant interim measures, Case 2 BvR 2752/11—ThyssenKrupp Nirosta, Order of 30 October 2013, cited by BeckRS 2013, 57257. For the ECJ’s judgment, see Case C-352/09 P ThyssenKrupp Nirosta v Commission [2011] ECR I-2359.

40 Konrad Ost of the GWB). The wide notion of the right to remain silent makes fact finding comparably difficult. The strict formality of the procedure results in rather bizarre days in court where expert opinions or final pleadings (all distributed in written form) are read aloud for several days. Formalities and the reliance on witness evidence make court proceedings extremely complex.17 For a medium-sized cartel in the paper wholesale market, the proceedings lasted for weeks, with about 20 days in court. For the large cement cartel referred to above, they took 37 days in court. The liquid gas cartel fined with over 240 million euros took over 100 days in court and more than three years in total. These are only the first of many complex cartel cases currently under appeal. One further drawback is the limited status of a cartel authority within the court proceedings: it has mere expert status; therefore, the public prosecution has to work through the whole material again in order to defend the authority’s case.18 This process is to be measured in several months or even years. As explained above, only then does the court take a decision. It is questionable whether this weak position of the cartel authority is in line with the call by the ECJ to strengthen the position of the cartel authorities in court proceedings.19

3.2.2 The problem of lengthy proceedings There are several reasons why lengthy proceedings have a negative impact on cartel prosecution. One is that they entail the risk of loss of evidence (for example, because the witnesses no longer have a clear recollection of all the details). The increasing length of cartel proceedings also stands in conflict with the case law of the European Court of Human Rights, which demands that enforcement proceedings be concluded within a reasonable period of time.20 Most of all, the long interval between the offence and the sanction is likely to significantly diminish the deterrent effect of the fine; this is especially true if, after close to a decade, the persons responsible no longer work in their previous functions or have left the undertaking altogether. Some cases may even become impossible to prosecute because they have become timebarred. It also has to be taken into account that lengthy proceedings can be

17 The burdensome procedures have also already been criticised by the OECD, and demands for reform have been made. See OECD Economic Department Working Paper No 507, ECO/WKP(206)35 (2006), para 15. 18 For further discussion, see P Klocker and K Ost, ‘Nach der Novelle ist vor der Novelle—Themen einer 8. GWB-Novelle’, in I Brinker, ed, Recht und Wettbewerb, Festschrift für Rainer Bechtold zum 65. Geburtstag, Beck, 2006, 229 et seq, 244; Ost, above n 12, 323. 19 Case C-439/08 Vlaamse federatie van verenigingen van Brood- en Banketbakkers, Ijsbereiders en Chocoladebewerkers (VEBIC) VZW [2012] ECR I-12471. See below Section 4.5. 20 Case No 37591/97, Metzger v Germany, 31 May 2001, para 42.

From Regulation 1 to Regulation 2 41 a mitigating factor in the calculation of fines, which can further diminish the deterrent effect.

4. Framework for reform 4.1 Effet utile The enforcement of European antitrust law must be effective. The principle of effectiveness in European law (effet utile) requires that national law provides effective and adequately deterrent sanctions for cartel law infringements.21 This means that, in designing their own national cartel procedural law, the Member States have to ensure at least a degree of effectiveness necessary to prevent any gaps in the decentralised system of cartel law enforcement. The prohibition of cartels under European law applies to undertakings as economic units which may consist of several legal entities (see Section 2.1.1 above). According to the established case law of the courts of the European Union, the substantive definition of an ‘undertaking’ as laid down by EU law is reflected on the sanctioning side by the fact that the parent company is liable for the offences of its subsidiaries22 and that a legal or economic successor of a company directly involved is to a large extent liable as well. Under Regulation 1/2003, substantive European cartel law must also be (effectively) enforced by the cartel authorities of the Member States. Only recently, then-Director General Alexander Italianer rightly stated in an (open) letter to the President of the Bundeskartellamt: the principle of effective antitrust enforcement requires that the national law be equipped with effective and sufficiently deterrent sanctions for violations of the law. Since the concept of what constitutes an ‘undertaking’ is a uniform concept under the law of the Union, a national regime will only meet this requirement if it … allows for an effective fining of the undertaking that is liable under substantive law aspects and in accordance with the Articles 101 and 102 TFEU, which are directly applicable.23

It follows that if a cartel law violation occurs, the undertaking (as an economic unit consisting of several legal persons) has to be fined under the Case C-429/07 Inspecteur van de Belastingdienst v X BV [2009] ECR I-4833, para 37. For more detail see Kokott and Dittert, above n 5, 670. 23 See the letter of Director General Italianer to the President of the Bundeskartellamt of 18 June 2012, published as an attachment to the Bundeskartellamt’s comment on the government draft bill for the 8th amendment to the GWB of 22 June 2012, printed paper of the German Bundestag 17(9)865 (translation by the author). See also T Ackermann, ‘Kartellgeldbußen als Instrument der Wirtschaftsaufsicht’, [2012] Zeitschrift für Wettbewerbsrecht—ZWeR 3, 16. 21 22

42 Konrad Ost respective provisions. Since the group of undertakings is the addressee of the European cartel law provisions, the liability under cartel law must not be limited to the legal person that is directly and actively involved in the cartel law violation. As a consequence, the requirement of effectiveness under European law becomes ultimately a requirement of congruence, which means that the substantive addressee of the provision, as defined under the European law, must be punishable under the national regime as well. In its initial interpretation, Regulation 1/2003 does not conclusively regulate how the European prohibition of cartels should be enforced and, if breached, penalised. Recently, there have been intense debates on the question of the direct applicability of Article 5 of Regulation 1/2003 (see Section 4.2 below).

4.2 Regulation 1/2003 as a driver for convergence The actual antitrust enforcement regime in Europe as established by Regulation 1/2003 is a system of parallel competences of the European Commission and the national competition authorities. With the introduction of Regulation 1/2003, the European Competition Network (ECN) was founded. The Commission Notice on Cooperation within the ECN stipulates certain criteria for case allocation. Accordingly, a case should be allocated as early as possible in the proceedings, but may be reallocated up until the conclusion of the proceedings. The grounds for a reallocation can be the effects of a cartel law violation, possibilities for the presentation of evidence and chances to effectively end the violation. A referral may (also) simply occur for resource reasons. Ultimately, the decision on which a competition authority applies European competition law under its respective (national) rules of procedure can therefore depend on criteria that have little to do with the case in question. The procedural laws of the Member States and of the European Commission differ in the range of fines, types of proceedings and rights of defence. The urgency of the question of whether the system of parallel enforcement requires similar sanctions and procedures depends to a certain degree on whether the allocation of a case is subject to judicial review. The European courts have had to deal with these issues in two cases, but unfortunately have only been able to decide in one of them. In France Télécom/Wanadoo, the Court of First Instance showed no sympathy for the idea of granting standing to the undertaking concerned in respect of case allocation.24 24 The France Télécom/Wanadoo case (Case T-339/04 France Télécom SA v Commission [2007] ECR II-521; Case T-340/04 France Télécom SA v Commission [2007] ECR II-573) concerned a Commission request for information in a case which had previously been handled by the French Conseil de la Concurrence. The EAEPC case (Case T-153/06 European Association of Euro-Pharmaceutical Companies, EU:T:2008:54) concerned a Commission decision of 10 April 2006 in which the Commission had rejected three complaints against

From Regulation 1 to Regulation 2 43 Even in the absence of clear case law, it is obvious that the argument for granting standing to undertakings regarding case allocation is weak. The system of parallel enforcement is designed as a flexible system. Apart from the turnover thresholds of undertakings involved in a merger, it is hard to imagine ‘hard’ criteria to delineate the competences of the EU on the one hand and of the NCAs on the other. Experience with the current system shows that the reallocation of cases is rarely an issue among the authorities, nor is it often called into question by the undertakings concerned. One may conclude that this system leads to an efficient allocation of cases and furthers effective cartel prosecution throughout Europe. However, if it is true that there is no acceptable alternative to the current system of parallel competences where case allocation does not accord rights for the undertakings concerned, this system loses its legitimacy in the long run if there is no further alignment in terms of sanctions and procedure.25 For the undertaking concerned, it is difficult to accept that a decision on case allocation which is not subject to appeal can decide on fundamental questions such as parental liability or succession for purposes of the application of fines.

4.3 Human rights framework 4.3.1 The institutional setting The case law of the Strasbourg Court considers an administrative procedure as compatible with the due process clause (Article 6 European Convention on Human Rights), subject to one basic requirement: the possibility of appeal against the final administrative decision before a judicial body with full jurisdiction, including the power to quash the challenged decision in all factual and legal respects.

GlaxoSmithKline for an infringement of Article 82 EC. The claimant had subsequently demanded that the Commission should initiate an abuse proceeding and, by initiating the proceeding under Article 11(6) of Regulation 1/2003, withdraw the case from the Greek competition authority, which was handling the case. Unfortunately, the complaint was withdrawn before a hearing took place. The Federal Republic of Germany acted as an intervener in the case on behalf of the Commission and argued that the provisions of Article 11 of Regulation 1/2003 existed solely to serve the public interest and that undertakings were not entitled to demand that the Commission withdraw a case from a national authority. Only the Member States were entitled to bring a claim, eg if the Commission opened an investigation and in doing so violated the duty of sincere cooperation or the principle of subsidiarity. 25 D Waelbroeck, ‘Twelve Feet All Dangling Down and Six Necks Exceeding Long. The EU Network of Competition Authorities and the European Convention on Fundamental Rights’, in C-D Ehlermann and I Atanasiu, eds, European Competition Law Annual 2002: Constructing the EU Network of Competition Authorities, Hart Publishing, 2005, 465 et seq.

44 Konrad Ost Both the EU procedure and the German procedure meet these requirements. The Düsseldorf Court of Appeal hearing the Bundeskartellamt’s cases has full and independent jurisdiction over factual and legal issues, including the fine. It is obvious that the German fines procedure, with its extensive scrutiny by the courts, which ultimately makes it an adversarial rather than administrative type of procedure, is in line with the Strasbourg case law. From the Menarini judgment,26 it is clear that even the much criticised EU system is compatible with the European Convention on Human Rights (ECHR) and Article 47 of the Charter of Fundamental Rights of the EU.27 Recently the Commission has raised transparency and increased safeguards for procedural rights,28 and the courts of the European Union have intensified their review of the Commission’s decisions.29 It is important to note that not every right of defence which is guaranteed by ‘secondary’ legislation in Germany is unchangeable due to constitutional concerns, as these rights often surpass the basic human rights standard guaranteed by the German constitution and the ECHR. The same is true for European law, where one has to analyse in detail what is an interpretation of Regulation 1/2003 or what is derived from general principles, and what stems from the fundamental rights standard.30

4.3.2 Human rights for undertakings? In order to clarify the fundamental rights standard, it is important to differentiate between the imposition of a criminal sanction against a natural person and the punishment of undertakings. The prospect of a custodial sanction justifies a particularly high standard. Fining undertakings is a different matter. The punishment of undertakings follows a different concept. Its main impetus is the governance of economic behaviour.31 Therefore it has long been argued that fining undertakings belongs more to the sphere of economic administrative law than to criminal law. Unlike individuals, companies can

Case No 43509/08, Menarini Diagnostics v Italy, 27 September 2011. Case C‑272/09 P KME Germany AG and Others v Commission [2011] ECR I-12789, paras 91 et seq; Case C‑386/10 P Chalkor AE Epexergasias Metallon v Commission [2011] ECR I-13085. 28 Additional information in the statement of objections regarding the setting of the fine, state-of-play meetings extended to cartel cases, enhanced access to the file, revised terms of reference of the hearing officer. 29 See, eg KME Germany, above n 27, paras 94 et seq. 30 See V Skouris and D Kraus, ‘Die Bedeutung der Grundfreiheiten und Grundrechte für das europäische Wettbewerbsrecht’, in G Hirsch, F Montag and FJ Säcker, eds, Münchener Kommentar zum Europäischen und Deutschen Wettbewerbsrecht (Kartellrecht), Beck, 2007, vol 1, 89 et seq, Einleitung, paras 383, 394–397. 31 K Schmidt, ‘Zur Verantwortung von Gesellschaften und Verbänden in Kartellordnungswidrigkeiten’, [1990] wistra 131, 133; Klocker and Ost, above n 18, 242; T Ackermann, above n 23. 26 27

From Regulation 1 to Regulation 2 45 change their identity, be dissolved or be re-established. Since the protection of human dignity does not apply to legal entities, the FCC stipulates that the basic right that no one can be obliged to incriminate himself (nemo tenetur) does not apply to legal entities.32 The same question arises with regard to the applicability of other fundamental rights to companies. Within the realm of the ECHR, it is settled case law that undertakings can invoke certain human rights. But here we seem to have the same situation, ie the Strasbourg court is hesitant to provide the same degree of protection for legal entities as for individuals.33 Even if there is no doubt that, generally speaking, Article 6 of the Convention, stating the essential procedural rights of the persons concerned, is applicable, it is by no means certain that the ECJ will apply the full extent of Article 6 to undertakings.34

5. Solutions 5.1 Autonomous convergence of the enforcement regime As far as efficiency and effectiveness are concerned, the European model has a number of attractive features. As already mentioned, there is strong criticism in respect of fundamental rights. Even if it is obvious that there are no problems concerning a fundamental right standard,35 it might be desirable to keep higher standards in some respect. This is arguable for both the procedural rules and the rules on sanctions, which, due to their general phrasing, have to be refined to a large extent by the case law of the European courts. More precise European legislation would be desirable. However, the law as interpreted by the courts leads to comparably effective enforcement by the Commission. To achieve further convergence, the German legislator explicitly transformed some rules of Regulation 1/2003 into German law in 2005. Whether this has led to real convergence is, however, doubtful. As explained above, the 10 per cent rule as a ‘European’ transplant into the German enforcement

See Case 1 BvR 2172/96 of 26 February 1997. W Wils, Principles of European Antitrust Enforcement, Hart Publishing, 2005, chapter 5, paras 516 et seq; W Wils, ‘EU Antitrust Enforcement Powers and Procedural Rights and Guarantees: The Interplay between EU Law, National Law, the Charter of Fundamental Rights of the EU and the European Convention on Human Rights’, 34 World Competition 189 (2011). 34 This can be inferred, for example, from the Court’s reasoning on the protection of private homes. See, eg the Opinion of Advocate General Geelhoed in Case C-301/04 P Commission v SGL Carbon AG [2006] ECR I-5915, para 64. 35 See Lenaerts, above n 15, 175. 32 33

46 Konrad Ost regime did not lead to a European interpretation. As long as the major pillars of sanctions and procedure remain enshrined in the German system of administrative fines governed by general principles of German law, convergence is at risk.

5.2 How far is national enforcement determined by European law? Here I analyse whether European law itself offers a possible solution to the problems of national enforcement systems.

5.2.1 The Tele2 Polska judgment In the ECJ’s Tele2 Polska judgment of 3 May 2011, the Court confirmed the directly applicable nature of the competence conferred by Article 5 of Regulation 1/2003 upon national authorities to decide that there are no grounds for action even in the absence of any such attribution of competence under national law.36 The Opinion of Advocate General Mazák37 further illustrates the idea of direct applicability of Article 5. He cites other examples of actions taken by NCAs on that basis (such as interim measures). It has been argued that direct applicability is acceptable for administrative actions like commitment decisions or interim measures. According to this view, however, Article 5 could not be directly applied with regard to the imposition of fines and other penalties due to the principle of nulla poene sine lege

36 Case C-375/09 Prezes Urzędu Ochrony Konkurencji i Konsumentów v Tele2 Polska sp z oo, devenue Netia SA [2011] ECR I-3055. 37 Opinion of Advocate General Mazák in Tele2 Polska, ibid, paras 57 et seq. As stated by the Advocate General: ‘[I]t is logical that Article 5 of Regulation No 1/2003 should be directly applicable—it allows all relevant NCAs to have as of 1 May 2004 consistent decision-making power, without the need to wait for the implementation of Article 5 into national law. Finally, to give a few concrete examples, it may be noted that the NCAs in Germany, Italy and Belgium have already directly applied Regulation No 1/2003. First, the Bundeskartellamt relied on Article 5 of Regulation No 1/2003 already before the 7th Amendment of the German Law on competition came into force. Secondly, the Autorità Garante della Concorrenza e del Mercato directly applied Article 5 of the regulation and took a decision ordering interim measures which was at the time not provided for under Italian law. That decision was upheld by the TAR Lazio on 6 December 2005. Finally, the Conseil de la concurrence took the view that it should be able to exercise enforcement power even in the absence of national provisions conferring these powers upon it, and notwithstanding national provisions to the contrary. That was the position in relation to the competence to accept commitments, one of the types of decision listed in Article 5, but not provided for under Belgian law. The Conseil de la concurrence inferred an “Article 9-type” competence for itself from the direct effect of Regulation No 1/2003. It follows from all the foregoing considerations that Article 5 of Regulation No 1/2003 is directly applicable and that the NCA may close its proceedings with a procedural decision finding that there are no grounds for action on its part’ (footnotes omitted).

From Regulation 1 to Regulation 2 47 certa.38 It seems convincing that if no sanction is foreseen for the breach of European competition law in a national jurisdiction, Article 5 cannot be used to fill this lacuna. To adopt Article 23 would be too far-reaching.39 However, the argument is more difficult when the national system provides for sanctions for the breach of European competition law.

5.2.2 The concept of classification and the difficulty of differentiating between substance and procedure As has been shown above, there is always a complicated interplay between the European ‘substantive’ law (Article 101 TFEU) and the national enforcement system. Within this interplay, it is difficult to delineate the scopes of European law and national law. One approach would be to delineate between substantive and enforcement rules. This does not really help, however, as there is no generally accepted definition of ‘substantive’ law. The recent case law of the ECJ in Schenker40 and Slovenskà sportitel’ňa41 on cases handled by NCAs shows that, from a German perspective, the ECJ ruled on issues of sanctions usually governed by national law. This is obvious with regard to the concept of error, intent and negligence,42 and it is likely in respect of the issue of who is acting as the agent for an undertaking and the preconditions for a valid representation,43 and, lastly, in respect of the idea of succession,44 which should be interpreted as a concept of union law to be applied in a consistent manner in national procedures throughout the EU.45 But how should the spheres be delineated? A persuasive approach could borrow the concept of classification used in private international law. The rules have to be classified as being on the substantive side and governed by European law or on the enforcement side and governed by national law. Due to the primacy of European law, the 38 E Gippini-Fournier, ‘The Modernisation of European Competition Law: First Experiences with Regulation 1/2003—Community Report to the FIDE Congress 2008’, in HF Koeck and MM Karrollus, eds, The Modernisation of European Competition Law, Nomos, 2008, vol 2, 357 et seq at 459. 39 The option of extending Article 23 to all NCAs was discussed by the European legislator but that option was declined—a clear legislative decision. See the proposal of the European Parliament, A5-0229/2001 of 21 June 2001, http://www.europarl.europa.eu/sides/ getDoc.do?pubRef=-//EP//NONSGML+REPORT+A5-2001-0229+0+DOC+PDF+V0// EN, 14. 40 Case C-681/11 Bundeswettbewerbsbehörde and Bundeskartellanwalt v Schenker & Co AG and Others, EU:C:2013:404. 41 Case C-68/12 Protimonopolný úrad Slovenskej republiky v Slovenská sporiteľňa as, EU:C:2013:71. 42 Schenker, above n 40, paras 33 et seq. 43 Slovenská sporiteľňa, above n 41, paras 28 et seq. The Austrian Supreme Court based its reasoning concerning the civil liability including issues of representation directly on European law. See Oberster Gerichtshof, Order of 2 August 2012, Case 4 Ob 46/12m. 44 See AGCM v ETI, above n 7, paras 40 et seq. 45 European Commission’s written observations pursuant to Article 15(3) of Regulation No 1/2003 to the Supreme Court of Slovakia in case 1/Szhpu/2/2008 Cargo Slovakia, http:// ec.europa.eu/competition/court/amicus_2010_cargo_slovakia_sk.pdf.

48 Konrad Ost classification is to be determined not by the lex fori but by European law. In other words: if national law foresees the possibility of punishing a violation of European law, there may still be issues traditionally seen as part of the enforcement regime which are classified by the ECJ as ‘substantive’ and therefore as governed by primary or secondary EU law. In this respect, Article 5 of Regulation 1/2003 has direct applicability as a bridge between the substantive provision of European law and the national enforcement rule. By including infringements of European cartel law in the GWB in 2005, the legislator provided for the necessary complementary rules for the application of this regulation.

5.2.3 Primacy of EU Law As the two bodies of European and German law together form the basis of enforcement, there is no nulla poena issue at stake. What happens if, as argued here, both parts provide for a diverging solution to specific problems (eg legal succession)? The solution would be that national law can only fill the gaps if European law allows it a certain scope for manoeuvre.46 With respect to legal succession, European law prevails over national law in applying the concept of an undertaking in substantive law (Article 101 TFEU). In the area of legal succession, this goes far beyond German law.47 The same is true for a presumed error in law.48 The Düsseldorf Higher Regional Court has flatly rejected this approach brought forward by the Bundeskartellamt without a detailed analysis of the case law of the European Court of Justice. This ruling is under appeal. The decision now rests with the Federal Court of Justice. As there is no acte clair, the ECJ will have to deal with that question.

5.3 Criminalisation is no solution 5.3.1 Effectiveness of corporate fines An assessment of the German fining system casts some doubts on its effectiveness. It has been argued repeatedly that a system of optimal sanctions 46 It could be argued that this situation is similar to the ECJ’s Berlusconi judgment rejecting the primacy of an EU Directive over national sanctioning law. Case C-387/02 et al Silvio Berlusconi and Others [2005] ECR I-3565, para 77. The main difference here is that a part of the Regulation, ie Article 5, is directly applicable. 47 See C Heinichen, Unternehmensbegriff und Haftungsnachfolge im Europäischen Kartellrecht, Nomos, 2011. 48 On German law, see C Vollmer, in G Hirsch, F Montag and FJ Säcker, eds, Münchener Kommentar zum Europäischen und Deutschen Wettbewerbsrecht (Kartellrecht), Beck, 2008, vol 2 (GWB), § 81, para 61. On European law, see Schenker, above n 40, paras 33 et seq.

From Regulation 1 to Regulation 2 49 has to be complemented49 by (or must even focus exclusively on) individual custodial sanctions. An argument in favour of criminalisation in the form of custodial sentences has been that even the presumably well-functioning European system, with its substantial fines imposed on undertakings, is not seen as sufficient to deter individuals that are responsible for such offences. Another argument brought forward is the phenomenon of recidivism. Available data shows that there are quite a number of undertakings which have been subject several times to substantial EU fines. However, empirical data on recidivism falls short of giving clear answers as to whether the significant increase in fines witnessed in the last 10 years has had an effect or not.50 Only those cartels which began after the increase in fines would give evidence of the deterrent effect, and such data is rare. The trend towards increasingly strict compliance programmes, even in the absence of a mitigating effect in the European fining system, might even prove that the landscape is truly changing and that the fining system is effective. If the assumption that the effectiveness of the German fining system is at stake is right, is criminalisation the answer to the problems discussed above? Following the structured approach Wouter Wils presented in his famous EUI Workshop paper of the same name (‘Is Criminalisation the Answer?’),51 it is clearly not. Wils argues that a strong focus has to be on corporate fines which should be complemented by custodial sanctions provided there is a procedural framework for their effective enforcement and a willingness on the part of the stakeholders to use this framework.52 To the contrary, one would assume that a further criminalisation would in Germany lead to an even less effective enforcement regime.

5.3.2 Bid-rigging as a criminal offence in Germany Bid-rigging as a specific agreement which restricts competition already constitutes a specific criminal offence under section 298 of the German Criminal Code. The practical impact of that provision is negligible. There are very few cases reported where a custodial sanction has been imposed and there are official statistics showing more convictions where section 298 of the German

See Wils, Efficiency and Justice, above n 3, 155 et seq. See W Wils, ‘Recidivism in EU Antitrust Enforcement: A Legal and Economic Analysis’, 35 World Competition 5 (2012), text accompanying footnote 85; W Wils, ‘Antitrust Compliance Programmes and Optimal Antitrust Enforcement’, 1 Journal of Antitrust Enforcement 52, 71 et seq (2013). 51 W Wils, ‘Is Criminalization of EU Competition Law the Answer?’, in C-D E hlermann and I Atanasiu, eds, European Competition Law 2006: Enforcement of Prohibition of Cartels, Hart, 2007, 267. 52 Wils, Efficency and Justice, above n 3, 188, paras 586 et seq, text accompanying footnote 155. 49 50

50 Konrad Ost Criminal Code has been cited.53 However, at a recent meeting of public prosecutors from all over Germany with a special competence for prosecuting bid-rigging, not one of them could remember any conviction including the imposition of a custodial sanction. Most proceedings on bid-rigging are stayed, and sometimes settled if the perpetrator agrees to pay a fine. The main reason for this is seen in the burdensome procedure and the scarce resources of the judicial system. Due to the necessary procedural guarantees, criminal proceedings are even more burdensome than fines proceedings (in particular, with regard to evidence law). Further arguments against the assumption that criminalisation would lead to effective deterrence are that German law does not provide for short custodial sentences, and longer sentences are often suspended. Finally, with a criminalisation of cartel law, a leniency policy would have to be provided for by the law, which in turn would give rise to political debate and also systemic legal problems.

5.3.3 Criminalisation may lead to less deterrence Apart from general concerns about an extended criminalisation of economic offences, there are doubts as to whether it would make sense to introduce a genuine criminal cartel law in view of the above conflict with an effective and efficient enforcement of cartel law. Fast and effective enforcement is crucial for a truly deterrent effect. Criminalisation would have the opposite result. Bo Vesterdorf, when President of the European Court of First Instance, rightly argued that criminalisation leads to more procedural complexity and maybe less enforcement.54 In addition—at least in Germany—a social consensus about criminalising antitrust offences does not seem to have emerged. One reason may be that the main profit earned from antitrust infringements is usually kept by the undertaking and not the individual.

5.3.4 The future of fining individuals In contrast to European law—and apart from bid-rigging—in German law, natural persons may be subject to a personal fine for cartel offences. The involvement of natural persons even outside the criminal sphere justifies a high standard of rights of defence. It could be argued that the German law should perhaps abandon the punishment of individuals altogether in order

53 F Wagner-von Papp, ‘What if all Bid Riggers Went to Prison and Nobody Noticed? Criminal Antitrust Law Enforcement in Germany’, in C Beaton-Wells and A Ezrachi, eds, Criminalising Cartels: Critical Studies of an Interdisciplinary Regulatory Movement, Hart Publishing, 2010, 157 et seq. 54 B Vesterdorf, roundtable remarks in B Hawk, ed, International Antitrust Law and Policy: Fordham Corporate Law 2004, Juris Publishing, 2005, chapter 28, 756 and 762.

From Regulation 1 to Regulation 2 51 to streamline the antitrust procedure against undertakings.55 If it is true that undertakings are generally best placed to prevent antitrust infringements in the most cost-efficient way,56 and if fast and broad enforcement is a more effective form of deterrence than a few cases with exemplary fining, there is a clear case for concentrating on sanctions for undertakings. Measures to make corporate sanctions more effective, eg by switching to a more administrative type of proceeding, justify forgoing the punishment of individuals.

5.4 The role of compliance If criminalisation is not the answer, maybe compliance is. Scepticism on sanctions for individuals is mainly due to enforcement problems. However, it is convincing that individual responsibility is an important factor for the deterrent effect of an enforcement system. If, however, public enforcement is difficult and ineffective as far as individuals are concerned, private sanctions would seem both feasible and effective. Many companies claim to have an effective compliance policy. However, until quite recently there were few reported actions of undertakings against their managers. As it is now clear that under German law it is a breach of the supervisory duties of a board not to sue the former or current managers involved in a cartel, there has been public reporting on this kind of ‘private enforcement’.57 Compliance programmes may provide for harsh private sanctions and take over the function of public sanctions for individuals. The company is best placed to get the message of compliance across to its employees and managers, eg by making sure that the individual sanctions for infringing competition law or for tolerating an infringement by others exceed the incentives to commit offences. Private leniency programmes may be tailored to the specific needs within an undertaking. They are much more flexible than public leniency programmes in a criminal law context could ever be. Compliance programmes will not always prevent infringement; however, the right combination of private sanctions and incentives to cooperate may create the necessary individual incentives to complement the other basic mechanisms of the enforcement system. The driver for this development is a high level of credible corporate fines. 55 Klocker and Ost, above n 19, 244. See also the critique of J Biermann, ‘Neubestimmung des deutschen und europäischen Kartellsanktionenrechts: Reformüberlegungen, Determinanten und Perspektiven einer Kriminalisierung von Verstößen gegen das Kartellrecht’, [2007] Zeitschrift für Wettbewerbsrecht—ZWeR 1, 44. 56 Wils, ‘Antitrust Compliance Programs’, above n 50, 58. 57 AR Fabisch, ‘Managerhaftung für Kartellrechtsverstöße’, [2013] Zeitschrift für Wettbewerbsrecht—ZWeR 91. See also Süddeutsche Zeitung of 10 December 2012, ‘Thyssen-Krupp verklagt Ex-Manager auf 100 Millionen Euro’ [‘Thyssen-Krupp Sues Ex-manager for 100 MillionEuro’],http://www.sueddeutsche.de/wirtschaft/korruptionsaffaerebei-stahlkonzern-thyssen-verklagt-ex-manager-auf-millionen-euro-1.1546105.

52 Konrad Ost

5.5 Procedural convergence 5.5.1 The framework Regulation 1/2003 is an example of a rather reluctant approach to aligning the procedural systems of the Member States. The notion of the ‘procedural autonomy’ of the Member States is an important doctrine of European law. However, there are limits to such autononomy. The ECJ ruled in the VEBIC judgment on a purely procedural issue of national law. The ECJ considered that the Belgian competition authority was not vested with sufficient rights during the appeals procedure to protect the interests of an effective enforcement of European law. As the Bundeskartellamt has almost no rights during the court proceedings, the same issue could be discussed in the context of German law.58 As undertakings are the main addressees of the rules of conduct under cartel law, it seems reasonable to take a look at administrative law for a conceptual solution to the procedural shortcomings of German administrative penal law. Should the EU system serve as a model for national jurisdictions? There are a number of features of both the EU administrative setting and the Luxembourg judicial review that have been heavily criticised.59 It is clear that the EU system would not and should not serve as a blueprint for a convergence model in its entirety. However, it would be possible to take over many features of the EU model while still achieving effective enforcement and safeguarding in a reasonable and appropriate manner the rights of the undertakings.60

58 K Ost and T Brenner, ‘Das deutsche Kartellbußgeldverfahren nach der Entscheidung VEBIC: Vereinbar mit Unionsrecht?’, in Kartellrecht in Theorie und Praxis, Festschrift für Canenbley, Beck, 2012, 369 et seq. 59 See, eg F Montag, The Case for a Radical Reform of the Infringement Procedure under Regulation 17’, 17 European Competition Law Rreview 428 (1996); J Temple Lang, ‘Three Possibilities for Reform of the Procedure of the European Commission in Competition Cases under Regulation 1/2003’, CEPS Special Report (November 2011); J Schwarze, R Bechtold and W Bosch, ‘Deficiencies in European Community Competition Law: Critical Analysis of the Current Practice and Proposals for Change’, Gleiss Lutz Rechtsanwälte, September 2008. For a balanced approach, see W Wils, ‘The Increased Level of EU Antitrust Fines, Judical Review, and the European Convention on Human Rights’, 33 World Competition 5 (2010); M Jaeger, ‘The Standard of Review in Competition Cases Involving Complex Economic Assessments: Towards the Marginalisation of the Marginal Review?’, 2 Journal of European Competition Law & Practice 295 (2011); Lenaerts, above n 15, 175. 60 See R Wesseling and M van der Woude, ‘The Lawfulness and Acceptability of Enforcement of European Cartel Law’, 35 World Competition 573, 598 (2012), arguing that it is not the EU legal framework which has to change to accommodate concerns about due process but that ‘the different players in this system will have to review their conduct’.

From Regulation 1 to Regulation 2 53

5.5.2 Oral evidence An example of deficits on both sides may be the issue of hearing witnesses during court proceedings. In Luxembourg, there are almost no reports on the hearing of witnesses. The freedom of the judges to concentrate on written evidence, the reluctance of the lawyers to ask for witness interrogations and the wide scope of the rights of investigation with a limited right to remain silent during the Commission’s procedure may explain this fact. Due to the weight attached to oral evidence forseen by national criminal procedural law, the German experience is on the other side of the extreme—in one case, more than 6o witnesses were interrogated. The priority German criminal procedural law gives witness evidence over documentary evidence leads to extensive hearings even on facts where there is little doubt that the document is more reliable than the witness. The deficits are obvious if the evidence concerns data or a former behaviour which is not itself part of the offence. In general, however, the memory of witnesses years after the incident or relevant fact may be weak (and/or under undue influence of the parties concerned). The answers are often no more than repetitions of a reading of the document by the witness himself. As is often the case, the ideal solution may lie between these extremes. While accepting that most evidence may be adduced in written form, the main witnesses of the infringement (if there are any) should be heard in court.

5.5.3 From criminal to administrative procedures The paradigm shift from a criminal to an administrative procedure could also be based on the original function of cartel sanctions, in which the dimension of effective governance of the economy is clearly more important than the repressive dimension.61 Cartel law cases are intrinsically very complex due to the economic processes on which they are based and the large number of actors involved. In this respect, the legal framework provided for administrative proceedings is more adequate than that for proceedings in a criminal procedural setting: the focus on the written procedure is important; expert evidence can be provided in written form; and the parties involved can refer to briefs instead of having to make oral statements. Instead of being forced to (re-)establish the facts of the whole case in oral proceedings and issuing a stand-alone decision, the court may limit itself to assessing the decision made by the authority on the basis of the grounds of appeal.

See above n 31 and accompanying text.

61

54 Konrad Ost

6. Outlook: towards a ‘Regulation 2’ There are two drivers for reform in Germany. One consists of the shortcomings of the enforcement system in terms of both sanctions and procedure. The other is the system of parallel competences within the ECN and the European imperative of the effet utile doctrine. The possibility of resolving the issues by way of internal harmonisation is not excluded, following the path taken in the last 10 years which has clearly resulted in far more convergence than was expected 10 years ago. However, the problems of the German system can be taken as pars pro toto for many shortcomings and divergences in the national enforcement systems. The German example shows that even if the legislator transplants European rules into a national law setting, convergence is not guaranteed. There are additional issues to be discussed on a European rather than national level: there is an obvious need for further elaboration of the practical cooperation if two or more national authorities fine the same or a similar cartel. The experience of the recent mills cases, handled in parallel by the German, French, Dutch and Belgian authorities, sheds some light on the potential areas of conflict.62 After 10 years of experience of working together within the ECN, the EU should start with a new ambitious project, ie the further alignment of the enforcement systems within the EU. After 10 years of Regulation No 1, it is time to reflect on a Regulation No 2.

62 Issues such as ability to pay, multiple use of the 10% rule, the relevance of turnover realised in other Member States, common proposals to settle, etc.

2 EFFECTIVE REMEDIES Chair: Giorgio Monti Speakers: Rafael Allendesalazar Jochen Burrichter Stephen Calkins John Davies Ian Forrester Damien Gerard Alberto Heimler Alexander Italianer

Bill Kovacic Ioannis Lianos Philip Lowe Frank Maier-Rigaud Renato Nazzini Konrad Ost Savvas Papasavvas Joshua Wright

Joshua Wright: I will not provide any insights in the dreams of policemen in this presentation. [Laughter] Thank you for the invitation. I am told by my minders at the Federal Trade Commission (FTC) that I must begin every talk, even those overseas—perhaps especially those overseas—with the disclaimer that nothing that I say can be used against the FTC or any of the other Commissioners. There you have it. They can hold any of it against me, though. [More laughter] I want to talk about the FTC and its use of monetary remedies, namely disgorgement in competition cases. A bit of background is necessary. There has been some action in this area in the FTC as of late and, for those with a lower powered incentive to follow these day-to-day developments, I will give some necessary background. The FTC can ordinarily only pursue equitable remedies in competition cases, such as injunctions or divestitures, but not monetary or corporate remedies, including disgorgement or restitution.1 The focus of my remarks will be on whether the FTC should seek to impose disgorgement in competition cases and when. The FTC didn’t pursue disgorgement until 1999,2 though it has long sought disgorgement in its consumer protection cases. In 2003, after accepting comments from interested DD

See 5 USC § 53(b). FTC v Mylan Labs, Inc, 62 F Supp 2d 25 (DDC 1999).

1 2

56 Part I: Effective Enforcement of Competition Law parties, the FTC adopted a policy statement on the use of monetary remedies in competition cases that was designed to give some guidance after these cases in the late 1990s on when it would pursue disgorgement and when it would not.3 That statement outlined three factors the FTC would consider when deciding to pursue disgorgement in any given competition case: (i) whether the violation was clear; (ii) whether there was a reasonable basis to calculate a disgorgement penalty payment; and (iii) whether remedies through other litigation were likely to fail to accomplish fully the purpose of the antitrust laws. That policy statement, adopted after a notice and comment procedure in the US, was endorsed without dissent by the Antitrust Modernization Commission. In 2012, without public comment, the FTC decided to withdraw that statement on several grounds that are less than satisfying, for r easons that I will outline. To be clear, the withdrawal occurred before I joined the FTC and over the dissent of my colleague, Commissioner Ohlhausen. The FTC gave a few reasons for withdrawing its policy statement.4 One was the e lement of consideration of rarity or clarity of the violation. This was not an element considered by courts in disgorgement requests and the withdrawal of the statement contemplated that deciding whether or not an issue was a rare or clear violation could become an unnecessary side issue in follow-on litigation. The policy reasons the FTC gave are more interesting. One policy reason is that the Commission considered that the policy statement had chilled the pursuit of monetary remedies in appropriate cases, and second the Commission noted that Supreme Court Jurisprudence in the US had increased the burden on plaintiffs in private cases, which thus warranted a greater use of disgorgement by the FTC. I want to talk a bit about the statement and the policy reason the Commission pointed to in arguing that the policy statement itself had chilled the pursuit of monetary remedies in appropriate cases. That policy reason in itself is wholly unsupported by empirical data. The Commission sought disgorgement in two cases during the nine-year period before the policy statement was in effect,5 and also sought disgorgement in two cases during the nine-year period while the policy statement was in effect.6 There are many things one can do with four data points over eighteen years. However, something that 3 Policy Statement on Monetary Equitable Remedies in Competition Cases, 68 Fed Reg 45820 (4 August 2003). 4 Statement of the Commission, Effecting the Withdrawal of the Commission’s Policy Statement on Monetary Equitable Remedies in Competition Cases, 77 Fed Reg 47070 (31 July 2012) (hereinafter Withdrawal), http://www.ftc.gov/os/2012/07/120731commissio nstatement.pdf. 5 FTC v The Hearst Trust, No 1:01CV00734 (DDC 2001) (settlement resulting in defendant paying $19 million in disgorgement); Mylan Labs, above n 13 (holding that the FTC could sue for monetary damages as well as injunctive relief, and resulting in a $100 million settlement). 6 FTC v Lundbeck Inc, 650 F3d 1236 (8th Cir 2011) (affirming the District Court’s ruling denying any relief on antitrust grounds); FTC v Perrigo Co, No 1:04CV01397 (RMC) (DDC 2006) (resulting in a settlement that included a $6.25 million disgorgement).

Part I: Effective Enforcement of Competition Law 57 one definitely cannot do is point to this data and conclude that there was a decrease of disgorgement cases. However, this seems to be the argument that the FTC has made. The picture in terms of the data gets worse if one includes the SEP cases that the Commission had brought, which required a defendant to license its relevant patents on a royalty-free basis.7 One might conceptualise those remedies a bit like disgorgement in the sense that any profit the defendant would make, had it been able to reach a licensing agreement, it would now be deprived of. When we include these cases the picture gets a little bit worse in terms of relevant data points. One such case was before the statement and three came after the statement. Therefore, if anything, one can point to the total now of eight data points in eighteen years, and once can even see a slight increase that is huge in percentage terms of cases after the policy statement was adopted. If the data don’t get us there, I think the second policy reason is the most interesting to reflect upon and I will spend the rest of my time discussing this. The FTC’s rationale was that there was a reduction in private antitrust enforcement as a response to Supreme Court cases that either restricted the scope of Section 2 of the Sherman Act or increased the burdens for plaintiffs. This warranted an increase in the use of disgorgement by the FTC. Former Chairman Jon Leibowitz and former Commissioner Thomas Rosch, two of the four Commissioners that voted to withdraw the statement, have stated that ‘concern over class actions, triple damages awards and costly jury trials are what has caused our Supreme Court in the US to limit the reach of antitrust and that limit has given some would-be antitrust defendants a “free pass”’.8 To make up for any loss from the loss of deterrent value of these private suits, those Commissioners and the majority that withdrew the statement argued that the FTC ought to pursue disgorgement with more frequency. Again, I am going to the data a bit here to see if these statements are supported. For these statements to make sense, two testable hypotheses would need to be confirmed. First, private antitrust lawsuits would have had to become more difficult to win during the relevant time period. The second hypothesis is that anticompetitive conduct in the US is currently undeterred by the antitrust law. I will discuss both briefly. The most commonly cited cases in the US to support the proposition that life has become more difficult for plaintiffs is a line of cases that runs from Trinko9 to Twombly,10 linkLine11 7 Union Oil Co of Cal, 140 FTC 123 (2005); Dell Computer Corp, 121 FTC 616 (1996); Robert Bosch GmbH, No C-4377 (FTC, 21 November 2012); Opinion of the Commission on Remedy in Rambus, No 9302 (FTC, 5 February 2007), http://www.ftc.gov/os/adjpro/ d9302/070205opinion.pdf. 8 Statement of Chairman Leibowitz & Commissioner Rosch, Intel Corp, No 9341, 2010 WL 4542454 (FTC, 2 November 2010), http://www.ftc.gov/os/adjpro/d9341/ 091216intelchairstatement.pdf. 9 Verizon Communications Inc v Law Offices of Curtis V Trinko, LLP, 540 US 398 (2004). 10 Bell Atlantic Corp v Twombly, 550 US 544 (2007) 11 Pacific Bell Telephone Co v linkLine, 555 US 438 (2009).

58 Part I: Effective Enforcement of Competition Law and Credit Suisse,12 among others. Trinko is the first of those cases and was brought in 2004. It is a common rhetorical move to look at the 2004 case and say that defendants have been winning a lot of cases since then. It is more rare to point to any of them and say that any of the individual cases was decided incorrectly. But if one were to go back to a longer stretch of cases, going back to, say, 1967 rather than looking at just the last 10, one would find in the time period before the defendants’ win streak that plaintiffs had an impressive win streak in their own right. Indeed, they won more than 80 per cent of the cases from 1967 to 1976.13 Whereas defendants won all 13 of the a ntitrust cases decided by the Supreme Court from 1997 to 2006, defendants won only 16 of 44 cases decided by the Supreme Court from 1967 to 1976. Thus the dates show that, even if the Supreme Court is making life more difficult for plaintiffs in one way or another in the cases over the last eight years, the record over the last 50 years is clearly much more of a mixed picture. If one looks at all of the private law suits of the US in the relevant time period, that is probably a better approach to get a sense of what is going on, and this means it is useful to look at the district court level. One testable implication of the hypothesis that the recent decisions have made life more difficult for antitrust plaintiffs is the response in terms of case filings. Case filings are an imperfect proxy for plaintiff win rates for a variety of reasons, but they are worthy of examining as a first cut because a plaintiff’s decision to file a case will undoubtedly be affected by its expectations of outcomes as a result of Supreme Court decisions. Luckily, other folks have done the heavy lifting, in particular Paul Godek, in counting the data.14 Godek found that private filings have actually increased and not deceased since Trinko in 2004. Most recently, there was a dramatic surge from 2011 to 2012. I do not want to draw too strong a statement or conclusion from this, but it is difficult to construct from this data a picture of plaintiffs that are scared to file a case or having a low expectation of winning cases after these Supreme Court cases. The last of my remarks is on whether the current state of law is such that one can conclude in the US that the type of conduct folks are seeking to stop in monopoly cases, and where the Commission might seek disgorgement, is currently undeterred. Luckily, here I can piggyback on the comments made by Simon before. Simon provided a lot of the economic insights on over- or under-deterrence in the Article 102 context. In this context, and in S ection 2 cases in the US, the probability of detection of much of this conduct is already equal to 1. For example, patent hold-up is hold-up. One has to raise one’s hand and hold up the other party. Hence, the probability of detection is high.

Credit Suisse Securities (USA) LLC v Billing, 551 US 264 (2007). See Leah Brannon and Douglas Ginsburg, ‘Antitrust Decisions of the US Supreme Court, 1967 to 2007’, 3(2) CPI Journal 1, 3 (2007). 14 See Paul Godek, ‘Does the Tail Wag the Dog? Sixty Years of Government and Private Antitrust in the Federal Courts’, Antitrust Source, December 2009, 1. 12 13

Part I: Effective Enforcement of Competition Law 59 The case for disgorgement on top of the available treble damages is therefore fairly weak. Lastly, the economic literature surrounding vertical conduct in terms of the empirics should make one very wary that the problem we face in that area is much more likely to be over-deterrence than under-deterrence. Because of that, it is ultimately fairly unwise to seek disgorgement in such cases and unwise to walk away from policy guidance as to when a public agency is going to seek disgorgement in those types of cases. I will stop here ahead of Giorgio’s red card. Ioannis Lianos: I was asked to talk about effective remedies. The first question I asked myself is what we mean by effectiveness. There is a clear trend in competition law legislation and jurisprudence that remedies should be sufficient to ensure effective legal protection in the fields covered by EU law. We are all familiar with Article 7(1) of Regulation 1/2003, where the word ‘effective’ appears in the sense that the remedies should bring the infringement effectively to an end. Today we should also look at Article 19(1) TEU on effective remedies and Article 47 of the Charter on Fundamental Rights, which provides a right to an effective remedy. We are also familiar with the jurisprudence of the EU Courts, which requires Member States to establish a system of legal remedies and procedures which ensure the right to effective protection of individual rights under EU law, and therefore requires them to put in place effective remedies for violations of EU law. The meaning of effectiveness in the EU case law is specific in the sense that it is focused on national remedies and the rule that national procedural rules should not make the enforcement of EU rights impossible or excessively difficult. However, in a quite interesting Opinion of AG Jääskinen in Donau Chemie,15 he examined the scope of Article 19(1) TFEU and concluded that the standard of effective judicial protection for EU-based rights seems to be more demanding than the classical formula of effectiveness and refers to practical impossibility.16 In his opinion, this means that national remedies must be accessible, prompt and reasonably cost-effective. We see here that a new meaning is given to the concept of effectiveness or effective remedies in the sense of cost-effectiveness. However, although this is an interesting point of view and important, it is quite clear that the concept of effective remedies is not really clear and what we can actually only really get from this is that the concept of effective remedies does not entail the power to impose any remedy. This is interrelated to the question of remedial discretion of competition authorities and the judiciary in competition cases. Therefore, when we look at economic analysis we know that there is a lot of analysis as regards damages and fines, and we DD

15 Opinion of Advocate General Jääskinen in Case C-536/11 Bundeswettbewerbsbehörde v Donau Chemie AG and Others, EU:C:2013:67. 16 Ibid, para 47.

60 Part I: Effective Enforcement of Competition Law are familiar with optimal enforcement theory, which is focused on deterrence. At the same time, looking from a legal perspective, lawyers tend to focus more on the limits to remedial power, whereby private lawyers will focus on corrective justice and public lawyers on the principle of proportionality. That does not necessary imply the complete absence of any deterrence considerations because deterrence might also be an objective of corrective justice. Therefore, it is not only in the context of optimal enforcement theory that one thinks of deterrence; –it can be important for other lawyers as well. It is therefore necessary to investigate the inherent limits to remedial discretion coming from an understanding of the concept of remedy in public and private law. I take this perspective because I think that competition law should try to connect with a more general debate over remedies in public and private law, taking into account that such a systematic analysis of the topic is quite recent while in common law jurisdictions the topic of remedies started to be examined systematically in the 1970s and many civil law jurisdictions actually recognise a proper law of remedies. It is part of the substantive law. I will examine two further issues in my presentation: first, the taxonomy of remedies; and second, the issue remedial discretion. Remedies in EU law and competition law are intrinsically related to the concept of bringing an infringement to an end. They are clearly distinguished in Regulation 1/2003 from penalties, sanctions and private enforcement, although remedies in the form of damages or injunctions may also be imposed by the courts. The distinction that is made between remedies and fines is therefore artificial. Remedies are related to a concept of bringing the infringement to an end. This was interpreted broadly in the case law to mean not only the termination of an infringement, but also the means that would prevent the occurrence of the infringement in the future if they were at the root of the litigation. I refer here to the Commercial Solvents case law.17 The concept of remedies has also been used to describe the various civil law consequences in national courts, such as nullity and damages. Therefore, I say that we should take a broader perspective on remedies. There has been a lot of work in private law, in particular on the taxonomy of remedies. What I would like to do here is to take a functional approach to the concept of a remedy, in view of the fact that there are very different perceptions of this concept in common and civil law traditions. There are three main functions of remedies: remedies are available (i) to cure the violation; (ii) to restore the plaintiff’s rightful position as if the plaintiff had never violated the law; and (iii) to restore the defendant’s position that he occupied prior to the violation. The third is the most problematic in that it seeks to ensure that there remain no practices that infringe competition law in the future. However, this can also

17 Case 7/73 Istituto Chemioterapico Italiano SpA and Commercial Solvents Corporation v Commission [1973] ECR 357, paras 45–46.

Part I: Effective Enforcement of Competition Law 61 be a side-effect of the first two restorative or punitive remedies. Preventive remedies aim directly at specific or general deterrence. Specific deterrence can be defined as the impact on the incentive of the infringer to adopt illegal behaviour in the future. General deterrence focuses on the public at large. Second, remedies may be of a purely prophylactic function. What I call ‘prophylactic’ remedies can be distinguished from deterrence in the sense that they focus on the ability, and not the incentive, to commit infringements in the future. That is, they focus on specific facilitators of potential infringements. These are not illegal practices in themselves, but in the specific circumstance of a case they may facilitate illegal conduct. Eliminating these practices does not exactly deter—rather, it reduces the ability to carry out illegal conduct. I make this point because it is interesting to develop a broader perspective and not just focus on injunctions or behavioural and structural remedies in public enforcement. Assuming that this taxonomy on remedies is accepted, the limits of discretion become important. This may depend on the capacity of each institution, but it is quite important to think about it more in general terms. My view would then be that competition authorities and courts should have discretion to define the appropriate remedy in individual cases rather than being limited to specific predefined remedies. The optimal or appropriate approach may be influenced by the views of economists, and somehow it might lead them to develop remedies that will improve the market equilibrium compared to that existing prior to the infringement. From a private law and public law perspective, it may seem to be contrary to corrective justice and proportionality. We have seen some case law of the EU Courts trying to limit the discretion of competition authorities by linking the remedy to the theory of harm that has been advanced. This is an important point to keep for the time being. I will end here. Damien Gerard: On 22 February 2006, the European Commission issued a remarkable decision. It was a decision that made binding on De Beers, the diamond conglomerate, a unilateral commitment not to purchase any rough diamonds from the Russian company Alrosa,18 as of the beginning of 2009 and forever after. These were and still are the two largest producers of rough diamonds in the world. Since 1995, De Beers and Alrosa agreed on a purchase agreement that allowed De Beers to purchase a significant share of output from Alrosa. According to the Commission, that agreement contributed to the entrenchment of De Beers’ dominant position in the market for rough diamonds. It reduced access to a viable source of supply of diamonds for downstream customers and prevented Alrosa from competing fully with De Beers. After adopting the decision, the Commission declared that, by obtaining the purchase commitments, for the first time in the history of the DD

Commission Decision of 22 February 2006 in Case COMP/B-2/38.381—De Beers.

18

62 Part I: Effective Enforcement of Competition Law diamonds market there was an opportunity for genuine competition. This disregarded the fact that Alrosa was not happy at all about the deal reached between De Beers and the Commission. Alrosa needed to secure the sales to De Beers in order to fund its expensive distribution network and, until some months before the decision, it had been fully involved in the negotiation of an alternative solution. In January 2003, Alrosa received a Statement of Objections from the Commission alleging that the agreement constituted a restriction of competition. In December 2004, it submitted, with De Beers, a joint proposal to the Commission to cap De Beers’ purchases at 35 per cent of the original amount within six years. This proposal was market tested in June 2005. According to the Commission, the outcome of the market test of the joint commitments had been negative. A majority of the respondents to the market test, who were mostly De Beers customers, considered that there should be no purchasing relationship at all. These observations of third parties together with the Commission’s own analysis led the Commission to suggest amendments to the commitments. De Beer decided to comply. Alrosa was not happy at all and brought an action for annulment to the General Court, alleging, first, infringement of freedom of contract and the principle of proportionality and, second, an infringement of the right to be heard and hence a breach to due process. Surprisingly, the General Court upheld both grounds and quashed the Commission decision.19 The Commission brought an appeal to the ECJ, which overturned the judgment on both grounds.20 Interestingly, the ECJ first defined the nature of the Commission’s proceedings as a new mechanism introduced by Regulation 1/2003 that was intended to ensure that the competition rules laid down in the Treaty are applied effectively. On proportionality, the Court found that the reach of the principle was limited in the context of Article 9 proceedings to the extent that the commitments have to address the theory of harm and that the parties have not proposed less far-reaching commitments that would also address the concerns. On due process, the ECJ essentially considered that the Commission has a wide degree of discretion to reject or accept commitments, and it was therefore not bound to give reasons for denying the original proposal. So ended the Alrosa saga. What explains this succession of twists and turns, which is rather uncommon in the EU context? The story concerns unanticipated effects of the process known as the modernisation of EU competition law, initiated by the effectiveness paradigm that drives negotiated antitrust enforcement and its limits when it comes to devising efficient remedies. These unanticipated effects caused a legitimacy gap that has still not been overcome today. Modernisation can be apprehended as an attempt to experiment with a utilitarian approach to the regulation of economic competition, with substantive, institutional and

Case T-170/06 Alrosa Company Ltd v Commission [2007] ECR II-2601. Case C-441/07 P Commission v Alrosa Company Ltd [2010] ECR I-5949.

19 20

Part I: Effective Enforcement of Competition Law 63 procedural dimensions. From a substantive perspective, it led to the acceptance of the efficiency consensus; from an institutional perspective, it led to the development of a network-based form of enforcement; and from a procedural perspective, it led to an increasing reliance on negotiated procedures in the enforcement of EU competition law: most prominently, the leniency, settlement and commitment procedures. The advent of negotiated enforcement is driven by effectiveness, including deterrence considerations. That is a willingness to maximise the allocation of enforcement resources with a view to strengthening compliance and the social impact of competition policy. Over time, the Commission decisions adopted pursuant to Article 9 of Regulation 1/2003 have been the most representative form of this negotiated form of enforcement. In the 2005–2013 period, the Commission adopted 31 commitment decisions and seven non-cartel infringement decisions. In recent years, the trend toward negotiated enforcement has become even more pronounced: between January 2008 and July 2013, only three non-cartel prohibition decisions were adopted,21 compared to 19 commitment decisions. The Commission therefore favoured a negotiated outcome in 85 per cent of non-cartel cases in that five-year period.22 The focus on effectiveness and the promotion of a negotiated procedure have had the effect of putting a greater emphasis on remedies. Negotiated procedures are completely focused on the design of remedies. C ommitment proceedings, by nature, focus on devising remedies rather than finding an infringement. The advent of a negotiation procedure and the design of negotiated remedies are also driven by effectiveness considerations. The EU Courts have allowed the imposition of remedies that were found to go beyond the restoration of the status quo ante, notably because they were designed to prevent the possible recurrence of an infringement in the future.23 Interestingly, Regulation 1/2003 follows the same approach by expressly granting the Commission the ‘power to impose any remedy, whether behavioural or structural, which is necessary to bring the infringement effectively to an end’.24 Remedies are not only designed to address prevention concerns, but also to

21 Commission Decision of 13 May 2009 in Case COMP/C-3/37.990—Intel; Commission Decision of 22 June 2011 in Case COMP/39.525—Telekomunikacja Polska; Commission Decision of 23 January 2013, Telefónica/Portugal Telecom, above n 11. 22 By comparison, it is estimated that ‘over 60 percent of antitrust disputes are resolved … by means of consent decrees’. See Richard Epstein, Antitrust Consent Decrees in Theory and Practice, AEI Press, 2007, 1 (or ‘roughly 70%’—see Michael Weiner, ‘Antitrust and the Rise of the Regulatory Consent Decree’, Antitrust, Fall 1995, 4). 23 See, eg Case T-395/94 Atlantic Container Line AB and Others v Commission [2002] ECR II-875, paras 398 et seq; Case T-310/94 Gruber + Weber GmbH & Co KG v Commission [1998] ECR II-1043, paras 177-178; Case T-7/93 Langnese Iglo GmbH v Commission [1995] ECR II-1533, para 205. 24 Regulation 1/2003, Recital 12 and Article 7.

64 Part I: Effective Enforcement of Competition Law ensure that they are workable and easy to implement. Moreover, because it is the remedy discussion that largely shapes the underlying theory of harm, the negotiated remedies are designed to address more loosely defined concerns. This also exposes them to a greater risk of restructuring by third party interests in market-testing processes. Negotiated remedies therefore raise the question of the relationship between effectiveness and efficiency. The promotion of negotiated procedures as part of a utility-maximising approach to competition law enforcement was designed to increase the effectiveness of enforcement. However, insofar as it leads to negotiated procedures becoming the default enforcement mechanism, that approach has the reverse effect of blurring the contours of the law and of leading to a loss in the predictability of antitrust principles, thereby leading to a loss in (substantive) effectiveness. The paradox goes full circle once one considers that modernisation as a whole was premised on the stabilisation of competition law principles. How to reconcile maximised enforcement resources with the promise of competition law as a tool of welfare maximisation? How to design optimal remedies in a negotiated context? Two legal concepts can assist in reconciling effectiveness end efficiency in this context: proportionality and due process. In the Alrosa case they were both unfortunately interpreted in a way that makes them unable to perform that mediation function properly. By relaxing both concepts, the ECJ created a legitimacy gap that threatens the legitimacy of competition law enforcement and its ability to achieve its social objective. There is a case for relaxing proportionality requirements in commitment cases, but doing away with both concepts creates a gap that threatens the legitimacy of competition law itself. In turn, filling that gap requires a review of the sequencing of commitment proceedings and associated p rocedural rights, as well as the (re)development of a credible alternative thereto, in order to salvage the voluntary character of commitments as a guarantee of proportionality and to turn commitment proceedings into a collaborative process capable of delivering optimal outcomes from both an effectiveness and an efficiency point of view. The Commission has taken a few steps in that direction, as reflected in the best practice guidelines, which explain that parties to commitment proceedings have access to the Hearing Officer at any time. The General Court, furthermore, made two useful suggestions in Alrosa: first, by providing non-confidential copies of the full responses to the market test, which already exists on a national level at least in some countries; and second, by requiring the Commission to formally motivate the rejection of market-tested commitments. To conclude, I put forward a third proposal, which is more systemic in nature in the sense that it could also restore the relevance of the voluntary nature of commitments as a means of achieving optimal remedies. I would propose that the Commission should not be able to impose fines in infringement procedures once it has engaged previously in the commitment path. In such cases, the sanction would be the infringement decision itself, and no additional monetary penalty would apply.

Part I: Effective Enforcement of Competition Law 65 Monti: OK, so now it is my turn. The idea of my paper is to look at the question of effective remedies and ask: what does this mean? I will focus on Article 7 remedies, not Article 9 remedies. There is a general agreement as to the role of remedies, which is threefold: (i) to terminate the infringement; (ii) to prevent it from occurring again; and (iii) to re-establish effective competition in the market. I mainly have a problem with the third: to re-establish competition in the market. Is this a legitimate objective to pursue for competition authorities? I will be looking at EU law only. In its decision against the Trans-Atlantic Agreement (TAA), the Commission had imposed an obligation whereby the undertakings who had infringed Article 101 had to ‘inform customers with whom they have concluded service contracts and other contractual relations in the context of the TAA that such customers are entitled, if they so wish, to renegotiate the terms of those contracts or to terminate them forthwith’.25 These contracts were not in themselves contrary to Article 101, but the concern was that the cartel had affected the terms of these contracts, so renegotiation would indeed serve to restore the status quo ante. The General Court ruled that, since this remedy was not normally imposed, a detailed explanation had to be given by the C ommission. Moreover, in determining the proportionality of this remedy, the Court intimated that account should be had of the private law rights resulting from an infringement of Article 101. As a result, the remedy was quashed for failure to state reasons.26 I am not sure why that was the case because the Court makes a comparison with Astra, but for me the level of reasoning in Astra is not so different or more elaborate.27 One interesting point that was made by the General Court in the TAA case was that a remedy that seeks renegotiation of contracts is something that the parties to the agreement can do themselves. This concerns the question of whether the restoring of competition is the job of a competition authority or of private parties. It appears that market players who suffer loss as a result of a competition infringement should secure compensation themselves: private law restores parties (and thus, often, m arkets) to the status quo ante, often insofar as money can do so. However, one might consider whether competition authorities should themselves perform the compensatory/restorative function, thereby widening and improving their remedial capacity. Given the present state of play, leaving restoration to private law seems desirable. After all, if private parties have been injured, then they have an opportunity to have their loss restored through private enforcement. Having said that, there was a report from the UK of a few years ago— the Macrory Report—which reviewed all forms of regulatory agencies and DD

Atlantic Container Line, above n 23, para 414. Ibid, paras 412–416. 27 Commission Decision of 23 December 1993 in Case IV/32.745—Astra. 25 26

66 Part I: Effective Enforcement of Competition Law remedies.28 There is a suggestion in that report that the compensation of the harm suffered is something that should be handled by regulatory agencies and that this may be more economical because the regulatory agency is in a good position to assess the damages. The Macrory Report has been implemented by all national regulators in the UK except for the Competition Commission. It might open up new perspectives to suggest that restoring competition and private enforcement may be done more efficiently by public enforcers. Let me turn to the issue of prevention. There are three points to make here. First, some conduct is ambiguous. Take rebates and the question when is it lawful or unlawful. If a decision tells the parties ‘do not offer a rebate’, this does not provide very useful guidance. If you look at some of the C ommission decisions on the abuse of a dominant position you will find that they are a mixture of an Article 7 prohibition that is either followed or preceded by a commitment decision. In Michelin II, you find a statement according to which, between 1998 and 1999, Michelin had been negotiating with the C ommission on what sort of rebate scheme would be allowed. Therefore, after the decision was issued, this new form of rebate scheme was allowed.29 Here I find a problematic contrast. Although Article 9 can be criticised, there is a degree of transparency in these decisions that people can see. With Article 7, however, there is a prohibition that is followed by a discussion as to how to comply with the decision. That is less transparent and can be worrying in the situation that the subsequent conduct of the parties is not market-tested before a formal decision is issued. It is particularly worrying if the discussion takes place before the formal decision is issued. The second point about prevention is that one should be able to monitor compliance in some way. It is not enough to simply say: ‘Here is what you should do’. It is possible to monitor compliance under Article 7 versus Article 9. One of the good things about Article 9 commitments is that, because they are voluntary, one can ask the party submitting the commitments to voluntarily set up a compliance scheme as well. Now, after the General Court’s decision in Microsoft, such compliance under Article 7 may be a little more difficult to carry out because of the General Court’s more restrictive approach as to what sort of powers can be subcontracted to third parties in that sort of case.30 So, again, in this context, Article 9 cases have a comparable advantage in terms of allowing greater flexibility in designing monitoring schemes for the remedies that you set forth. When it comes to monitoring, I want to mention the pharmaceuticals sector inquiry.31 One of the issues that

28 Richard Macrory, Regulatory Justice: Making Sanctions Effective (November 2006), http://www.berr.gov.uk/files/file44593.pdf. 29 Commission Decision of 20 June 2001 in Case COMP/36.041/PO—Michelin, Recital 350. 30 Case T-201/04 Microsoft Corp v Commission [2007] ECR II-3601, paras 1251–1279. 31 See http://ec.europa.eu/competition/sectors/pharmaceuticals/inquiry/index.html.

Part I: Effective Enforcement of Competition Law 67 the inquiry looked at was that of patent settlements. The Commission asked the parties, having identified the anticompetitive risks that arise, to notify such patent s ettlements to it. And we have now three annual reports that indicate the degree to which the parties have complied. So here, again, you have a situation where the probability of getting caught is 100 per cent because you must notify, and some breaches have been found. In the first year, 10 per cent of the settlements raised issues, in the second year it was 3 per cent and then it went back up to 11 per cent in the final year. The final spike upward is explained by some special cases. This raises questions about whether this is a method of enforcement, the message being: ‘You are being looked at, and we would like you to tell us what you are doing’. It also raised the question of whether 10 per cent non-compliance is a good number. Does this tell us how much compliance we want? So I think the experiment here leaves a number of questions open, but it is clearly an imaginative way to monitor compliance by simply asking the parties to report what they do. The third point about prevention is that remedies are a one-off. You impose them and then there is a risk of over- or under-enforcement with the remedy. It has been said before that the risk of over- or under-enforcement lies at the theory of harm stage; but with remedies it is even more pernicious, because you cannot correct. Should we build in a review so it can be checked whether the remedy has worked? This is what happens in Article 9 cases, and Microsoft, the browser case, is the primary example of this—after two years of implementation of the unbundling remedy, it can be asked whether that remedy has been efficient.32 Can you do so in Article 7 and is this advantageous? I know this is costly, but it makes sense to do. It strikes me that if you have a review clause it is likely to be to the advantage of the competition authority. That is to say that it is easier for the competition authority to claim that the remedy is not working than it is for the parties to say that the remedy has been excessive. This suggests that a review clause specifically protecting the undertaking might be warranted. My final point is about the remedy of a positive injunction. This is a duty to deal. It strikes me that some of the remedies imposed in some of these cases appear to be a bit excessive. This goes through the history of Article 102 cases. Going back to Commercial Solvents, the AG probably rightly held that the quantities that the defendant was asked to supply were excessive, given the nature of the problem.33 More problematic is an injunction, as in Microsoft, where the obligation is to deal with anyone. The aim of remedies is to terminate the infringement. The infringement in duty to deal cases is that it reduces

32 Commission Decision of 16 December 2009 in Case COMP/C-3/39.530—Microsoft (tying), Annex I, paras 20–21. 33 Cases 6 and 7–73 Istituto Chemioterapico Italiano SpA and Commercial Solvents Corporation v Commission [1974] ECR 223, para 48.

68 Part I: Effective Enforcement of Competition Law market access in downstream markets and therefore reduces workable competition in the downstream market. Now suppose a dominant firm has supplied six players in the downstream market so that there is effective competition on the downstream market since you have seven actors. Now number 8 comes along and says I too would like to have access to the relevant information in order to compete on the downstream market. Surely at that stage it should be right that the defendant can deny supplying access since the market is effective downstream so there is no competitive harm that can be fixed? An openended duty to deal with anyone therefore goes too far and goes beyond the rationale for the remedy, which is simply to terminate the infringement and not to regulate competition in the market to this degree. I’ll stop here. Frank Maier-Rigaud: I hope I will not be disappointing, because although I’m an economist I will talk about a purely legal question in front of a group of well-versed legal professors and lawyers, and it seems to be really the crème de la crème that is assembled in this room. I want to talk about proportionality, necessity and effectiveness. I will speak mainly about effectiveness, the subtopic in this group today. I will talk about Regulation 1/2003 and Article 7(1)—and, to be more precise, on the third sentence of this provision. Obviously, the words proportionality, necessity and effectiveness (or actually ‘effectively’ and ‘equally effective’) appear in Article 7(1). To bring an infringement (equally) effectively to an end, what does that really mean? ‘Equally effective’ implies that effectiveness is a gradual concept, not a concept that a particular remedy is either effective or not effective. Of course, there are such remedies, but Article 7(1) does not adopt a binary type of concept. The English version and the German version are not very clear on this. The French version of Article 7(1) is somewhat clearer, since there the word effectivement in the first part is used in the sense of actually terminating the infringement. This is a particular way of implementing this part of Article 7(1) also in the choice of words, and then in the second part, in contrast to the German and English versions, the word used is efficace; that is to say, ‘reaching the goal’. So effectiveness is not binary; it is a gradual concept. This then raises the question of which remedy is the most effective one. What I said is that I want to talk about effectiveness, but I still need to spend a second on proportionality and necessity. Here my argument would be that these are general filters, so it is not an ultimate selection criterion but is more of a filter that determines whether the remedy in the necessity part is capable of bringing the infringement to an end and does not go beyond this particular goal or address aspects that do not belong to the procedure. This is in stark contrast in some sense to the much more developed German law, where proportionality would automatically say something about effectiveness and vice versa. My claim would be that, under EU law, a more effective remedy is not automatically a less proportionate one and is not necessarily of milder DD

Part I: Effective Enforcement of Competition Law 69 means. So what does the test look like? The Commission may impose on the infringing firm any behavioural or structural remedies. Then two filters apply. In the second sentence it is specified that any such measure needs to be necessary and proportionate to the infringement. Therefore, obviously necessity and proportionality are criteria that both types of remedies (behavioural and structural) need to fulfil. The second type of filter, that is the third sentence, then basically says: choose from all the possible remedies the most effective one. This is conditional on what sentence 3 in Article 7(1) says, and that is that structural remedies can only be imposed if either there is no equally effective behavioural remedy or where any equally effective behavioural remedy would be more burdensome for the undertaking concerned than the structural remedy. This formulation has probably created the impression that structural remedies are subsidiary to behavioural remedies under EU law. But I argue this is not the case. So what does follow from this? There are two possibilities: either there is only a single remedy that comes out of this and that is proportionate and necessary, and then this is the remedy that you would want to choose; or, if there are many solutions that are proportionate and necessary but only a single one is the most effective, then this is the one to choose. In these situations, one would not care too much about the third sentence. However, if there are several remedies that are tied in the first place in terms of effectiveness, then one could distinguish two more situations. First, it may be that all the remedies are either all behavioural or all structural. Then, according to the case law,34 the undertaking should be allowed to choose; or, if the remedies include both structural and behavioural remedies, then the least burdensome solution has to be chosen. This may actually be, if one thinks, for example, of the E.ON case, the structural remedy. In conclusion: Regulation 1/2003 does not imply the subsidiarity of structural remedies. There is no preference for behavioural remedies, and the structural or behavioural nature of a remedy is immaterial from the perspective of Article 7 so long as they are not equally effective. Only in this highly theoretical case, when indeed they cannot be distinguished by the very criteria set forth in Article 7, does the Article suggest a preference for behavioural remedies. There is thus only a residual preference for behavioural remedies. The final point is a quote of the OECD Council on the role of structural remedies. As the text reads, Behavioural policies, unlike structural policies, do not eliminate the incentive of the regulated firm to restrict competition … despite the best efforts of regulators, regulatory controls of a behavioural nature, which are intended to control the ability of an integrated regulated firm to restrict competition, may result in less competition

Case T-24/90 Automec Srl [1992] ECR II-2223, paras 51 et seq.

34

70 Part I: Effective Enforcement of Competition Law than would be the case if the regulated firm did not have the incentive to restrict competition.35

This statement concerns regulated industries, but there is some truth to it for competition matters in more general terms since I spoke only about a very specific aspect of effective remedies. I thought it would be useful to put this more economic argument on one of my slides in order not to have talked only about the legal interpretation of Article 7. Monti: OK, so we’ve had 10 very rich and very diverse papers. Now we have about an hour for discussion, so what I suggest I do is that I will collect a few comments and suggestions at a time and go back to the panel and ask for their responses—so if you do want to comment, you just need to raise your flag like this so I know you want to say something, and I will try and pick you in as fair a manner as possible. So I have Jochen Burrichter first. DD

Jochen Burrichter: Yes, thank you very much. I think I have to explain that the German system is basically a criminal system and not an administrative system: if the BKA has imposed a fine, it is regarded as an indictment. So to some extent you can compare the system we have in Germany with the system in the US. The court has complete power to control the decision both from the factual and from the legal side. This implies that the court has to investigate the case as far as necessary, and if I look at the amount of fines that can be imposed under German law and also the fines that can be imposed on individuals, I think the criminal system is better equipped to deal with these aspects of the cases. In the case that lasted for three years, where we had to appear in more than 130 court hearings, the Vice-President of the BKA admitted that it had learned its lesson. The second point is that the court handled the case and the investigation in this way due to the deficiencies of the BKA, and it handled the evaluation of evidence in a way that was not completely appropriate for the handling of the case. But the system as such, in my view, is better than the EU system. In the EU, you appeal and then you have only 50 pages to make your case and you have 50 minutes to plead, and then the court is limited to manifest errors, giving broad discretion to the Commission and then confirming fines that can threaten a company’s very existence. That is not an appropriate system, and under German law if an EU-type system were adopted in Germany this would meet very substantial constitutional objections. We will have to see whether our Constitutional DD

35 OECD, Recommendation of the Council concerning Structural Separation in egulated Industries, C(2001)78/FINAL. See also OECD, Recent Experiences with R Structural Separation, DAF/COMP(2011)12.

Part I: Effective Enforcement of Competition Law 71 Court, which is actually dealing with a complaint against a fine of the EU, will confirm whether the fining system under Article 23 of Regulation 1/2003 is constitutional. Ian Forrester: The senior members around this table have seen, in their lifetime, fines growing from thousands to millions to hundreds of millions of euros. This phenomenon occurred in a period when judicial review in Luxembourg was in theory unlimited but in actual practice rather deferential. We have arrived in a situation where it is interesting that a fine of 500 million is imposed but the Court’s examination of such very heavy fines has not really been fundamentally rigorous. My suggestion is that the appropriate test for a fine should be at the level which achieves the legitimate amount of pursuing the goal of the administration. In other words, the right test is: ‘What level of fines should be achieved for the legitimate administrative objectives of deterrence and punishment’? Now, in Brussels, successive Commissioners are judged, at least by the press, and maybe by those who speak to the press, on the basis of the severity of the fines they have imposed. Therefore, Mr Monti was ‘better’ than Mr Van Miert because he imposed so many more millions in fines. That plainly equates heavy fines with good compliance, and that clearly is not the right way. So, penalties have moved from being what they should be—individualised and targeted on the particularly bad things that the individual did—to being arithmetical: they are the fruit of a mechanistic calculation. So they have evolved from alchemy, which was bad, but they have turned into a kind of taxation, which brings its own damage. There is a further problem, and that is that the fines go into the pocket of the European Commission, or at least they achieve the Commission’s achievement of its own budget, and that’s unacceptable. In any event, what I got from this morning’s discussion, which was intriguing, was the following proposition: in Germany, the competition rules are taken seriously and people worry about the BKA, and people worry and get angry when they’re investigated and when penalties are imposed as a sanction. Yet those penalties are very much lower than the penalties imposed in Brussels. Now I think it is a legitimate subject of inquiry whether the BKA deters more effectively or less effectively by reference to the levels of fines it imposes and the punishment it imposes than is the case with the Commission in Brussels. In short, I think that fines have got way out of control and that a public agency—at a time of concern regarding its legitimacy—doesn’t have the right to impose fines of that enormity without very careful reflection as to its entitlement to do so as a matter of justice, fairness, gravity and other individualised considerations. I think it would be a good thing at the end of this Workshop to reflect on what it is that other enforcement agencies in Europe do differently in terms of the setting of penalties and the success of those agencies in comparison to the success of the Commission in setting appropriate penalties. DD

72 Part I: Effective Enforcement of Competition Law Renato Nazzini: In terms of fines and sanctions again, the problem we have is that all the discussions about fines, in most of the papers, and certainly in defining the policy of most agencies in Europe, is about deterrence and is about what Ian Forrester was calling general deterrence, which comes again to the point that Simon Bishop was making well, namely: do we really know what the detection rate is—and we really cannot know this. Since the topic today is legitimate and effective enforcement, particularly legitimate enforcement, one question we need to ask ourselves is why punishment does not really feature highly in the pursuit of the goals of the administration and in the setting and calculation of these fines. Punishment as a factor would require looking at all sorts of elements, such as whether the infringement was intentional or negligent, the magnitude of the harm done to the economy, the duration of the infringement and so forth. The task at hand is to bring these factors to appropriate, commensurate, proportionate levels of punishment— and not general deterrence. This, I think, means a completely different framework in terms of thinking about fines, and it could address a number of issues that have been raised. In Article 102 cases, for example: what is the required degree of culpability—should it be intentional conduct only? What kind of abuse should be punished? Is it, what is termed in the US, a ‘naked’ abuse— something that is deliberate, intentional behaviour that cannot have any possible business justification other than excluding rivals, or is it something much more difficult and complicated, like loyalty rebates? Has the company excluded the only competitor that ever could have emerged in that market, or was the company just reacting to competition which was becoming more and more effective and stronger—as in the British Airways case,36 which cannot fail to be mentioned in this context? DD

Lowe: I came here this morning thinking that Alexander Italianer would have to deal with all the cases mentioned this morning, but all the cases mentioned were in the period when the Director-General for Competition was me. [Laughter] Apart from Michelin II and British Airways, which thankfully we’ve moved on from, I wanted to start off with legitimacy and say that, beyond the public recognition that something is harmful to society, surely an authority charged with policing that concern must demonstrate that it is active in pursuing it. Now, of course, it is the duty of the authority to pursue the objective of preserving competition, and to prevent and sanction abuses. From that point of view, if we don’t act proportionately and with necessity in relation to an abuse which manifestly gives legitimacy in our action, then that is the other side of the coin from what I heard about over-enforcement: there is under-enforcement, which certainly leads to a loss of legitimacy, a DD

36 Case T-219/99 British Airways plc v Commission [2003] ECR II-5917; appeal dismissed: Case C-95/04 P, [2007] ECR I-2331.

Part I: Effective Enforcement of Competition Law 73 loss of credibility for an institution if it is not willing to look at cases, hear evidence and investigate the facts—while respecting proportionality, necessity and due process. So there is a clear balance there—and a danger on both sides. Of course, in areas where the facts are more difficult to interpret and the theories of harm are more complex, the burden on the administration is much greater to get the answer right. From the point of view of the authority, in looking through the case, right to the first appeal and the second appeal, it is absolutely essential in the initial assessment of the case to consider whether you are going to harm or enhance the authority of the agency or the administration. So legitimacy is connected with not just law but the credibility of the action of the institution, which works both ways—over-enforcement and under-enforcement. Then you get to the issue of what is effective enforcement. It seems, before we get to the issue of what is the appropriate sanction or remedy, that the first question is: are we acting proportionately in relation to the supposed infringement—proportionately and in a necessary way either to get a cease-and-desist order or to achieve a situation on the market where things are at least not worse than they were before, and maybe even better but without any legal obligation to make them better? In that sense, there is still some fundamental logic in the law as far as Article 7 and Article 9 are concerned. The default position is that if there is a proven infringement then Article 7 applies and it is for the Commission or the national authority to judge. Alternatively, the parties come forward and say: ‘we have a solution to your problem, here is the remedy’. Now, there is absolutely nothing wrong—in fact, it enhances the due process and enhances the transparency of the operation—that there should be discussion with parties on possible remedies, notwithstanding the fact that if the parties are not willing to offer a remedy which is sufficient, the default option is an Article 7 decision. Now of course there are situations where, in order to preserve the legitimacy of an institution in relation to, for example, persistent abuses of market power of different kinds, maybe there isn’t an immediate remedy offered and therefore possibly the right thing from the point of view of the legitimacy of the action of the administration is to go for a cease-and-desist order. I just want to mention the case which Ian and I got heavily involved in, Microsoft, where the key issues at the beginning were the duty to deal and interoperability on the one hand, and the browser on the other. On the duty to deal, the situation was relatively clear and the defendant, the company concerned, was willing to offer what was needed. On the other issue, that of bundling, the situation was unclear, both from the point of view of the effectiveness of the remedy proposed by the parties but also because we had no particular answer to it, to be quite honest. So, in the initial decision, what we felt was proportionate and legitimate was a sanction on the conduct of the company on both issues, but with the prospect of being able to come back to the issue of bundling later if, indeed, there was a solution to it, and that’s how it happened. The second level, of course, is when you’ve

74 Part I: Effective Enforcement of Competition Law got a decision out there and you don’t enforce it, either through the action of the company or because you haven’t put in place the appropriate monitoring arrangements to make it happen. I think there would be a degree of selfcriticism on both sides there that that didn’t happen. Then the third issue is if someone has a precise responsibility to do a particular thing and doesn’t do it, then that surely, and I come now to fines, is certainly the worst thing that we should be envisioning in terms of sanctions. That is to say, someone deliberately either ignores or flouts the law, notwithstanding the need for due process in reaching the actual original decision substantively and procedurally, and all those situations occurred in the Microsoft case. On the Alrosa case,37 actually, I’m not sure that the problem was really an issue of the institutional processes itself but more the fact that there was genuine disagreement on what was a proportionate remedy. The clear view of one faction was that the proportionate remedy was a gradual reduction of the sales of Alrosa to De Beers. On the other hand, the other view, backed by a certain number of protagonists in the market test, was that you had to eliminate this supply entirely. I can tell you where I stood on that matter: I was in the first camp. Why? Because in any market, particularly for a market like diamonds, you have such a mixture of types of diamonds that it is inevitable that there is a minimum level of trade even between competitors. But the decision was taken the other way, and there I think that the issue of proportionality affects our legitimacy and the issue of due process because Alrosa itself was not fully involved. This was a good reason for the judgment of the General Court, which was overturned, maybe on the basis of higher principles, by the Court of Justice. There are a number of issues there that needed to be addressed, and hopefully they have been addressed subsequently, because they were really important. On the fines, I don’t disagree with Ian on the need for looking more closely at how courts and the authorities interact on the level of fines. Nevertheless, when you are in a situation of an infringement, which I think was described earlier as any collusive agreement of any kind which should be in principle per se sanctioned, then the question is: what is the proportionate level of the fine which actually deters not just the defendant but anyone else in the sector? That is a question of calculation and judgement as to what the damage has been in the past; I’m not sure it has much relevance to what happens in the future because in the future, apart from simply deterring, the issue is whether there is any alternative in that area of abuse which is more effective than a sanction or a fine that has to be proportionate and at the necessary level, and that is going to be debated over and over again. However, on the issue of other types of remedies, where you’re looking for something that is not just correcting the error of the past, then it must be significant improvements in market conditions Alrosa, above nn 19 and 20.

37

Part I: Effective Enforcement of Competition Law 75 which were justified by the authorities saying: irrespective of the infringement of the past, I am prepared to envisage this deal because it ultimately meets what the public expectation is—that the markets will work better, thanks not just to the removal of abuse, but to the change of behaviour of the companies concerned, and that is something that can be achieved. Well, in the last stage of our Microsoft browser discussion, where we actually did go for a settlement, the fine that came later wasn’t related to the substance of the settlement but to an incapacity to make sure that every updated Windows version was accompanied by the appropriate screen. But actually on the subject itself, the substantive discussion was seriously about how the market could work better between the parties and the authority. Those situations are naturally going to occur because they achieve results straightaway while avoiding a long period of contestation or alternatively, as Simon Bishop was suggesting, a long period of uncertainty due to private enforcement. I am not sure whether Alexander would agree with me, but I would love for all these plaintiffs to come to me and say: ‘Don’t worry, we’re going to get after this company; we don’t need you at all’. In every Article 102 case, every single company comes and says you’ve got to get after them, this is important for the industry, this is absolutely essential, you will achieve a fantastic result by public enforcement, which is much more effective and efficient than dragging the company and ourselves through court processes that could last 5–10 years. So effectiveness and efficiency tend, in the view of the plaintiffs in an Article 102 case, to support doing something through public enforcement. But we don’t necessarily need to oblige, and that’s why I thought it was very important that we produce guidance on Article 102, which we did. Monti: OK, I have nine contributors left waiting to speak. So if everyone could try and keep their contributions to two minutes at the most, that would be helpful. I’ll take Rafael and Mario, and then I’ll pause and ask the panel to respond if they wish to, and then I’ll take the remainder of the queue. DD

Rafael Allendesalazar: Thanks, two comments: one on Simon ishop’s intervention. I agree and support your idea, I don’t know about not B imposing fines on Article 102, but certainly on rethinking the idea of fining in Article 102 cases. Just to add an idea: deterrence is not only fines. There are also reputational issues, and we also have to take into account damages actions. Follow-on actions are much easier here than in cartel cases, so that should also be taken into account if and when setting fines in Article 102 cases. My second comment refers to an idea that Fred mentioned on the existence of differences compared to cartels, and I also agree with this. When we think of cartels and cartelists, we are always thinking of situations in which the company is the worst of the worst and therefore deserves any penalty it receives. We think of cartels as something very easy to distinguish, it’s an DD

76 Part I: Effective Enforcement of Competition Law animal everyone recognises when we get into the room, and I disagree with that. When you have to advise companies which, for instance, have a big part of their business related with an administration and the administration often insists on common proposals, or when you have network industries in which cooperation between companies (such as in telecoms) is necessary to provide a better service, advising them on what is legal and not legal is not easy. What is legal and not legal will easily become a cartel in the definition of the Commission. Furthermore, the Leniency Notice also blurs and tends to enlarge the distinction, because companies have an interest in obtaining a fine reduction, they have to confess that they are part of a cartel and therefore have an interest in defining a practice which may be in a grey area as a cartel. So my conclusion is, I think that we should not exclude cartels from the analysis of whether the sanction is proportionate or not, and that the issue of overenforcement may also occur in cartels. Mario Siragusa: I don’t think that the high level of fines is the major problem that we have in terms of legitimacy in our system. I think that the real problems are two: one, the never-ending problem of the limited jurisdiction of review, that’s the major defect in our legal system, and of course the combination of very high fines with a very limited judicial review becomes a very big problem for legitimacy. The second one is the use of fines in the commitment procedure, the threat of fines by an authority, which has the power to judge on the merit of the case, to impose huge fines and to induce the company to give commitments in cases where there is no serious investigation of the infringement. The combination of all those elements is a serious problem of our system. I think that these are the two major problems which destabilise the system and make it problematic from a legitimacy and constitutionality perspective. The commitment decisions also have a number of other very negative effects: they deprive us of jurisprudence and they don’t give clarity as to the law; they contain very opaque findings, which are not guidelines to the practitioners in the field; and they create obscurity in the law. The more we process it that way, the more our system is going to become absolutely unmanageable. So, for me, those are the two real problems of the legitimacy in our system. I find that fines are acceptable if they are subject to serious judicial review, because a judge who is going to look at the matter completely on the merits is going to assess whether the fine equitable or not, and I trust that the judge will do that. Unfortunately, in many countries, and especially at the European level, we don’t have judges who do this because of the limitations in our Treaty. They do not perform that kind of vetting of the Commission decisions. Then very quickly on some of the other issues: I fully agree that we need a Regulation No 2. I think this is going to become increasingly important because, frankly, most of the implementation and enforcement of the law is done by the national authorities. The Commission likes to do leniency and to pursue battles with the giants of high tech, but it leaves all DD

Part I: Effective Enforcement of Competition Law 77 the rest to the national authorities. National authorities are today the main enforcers of competition law, and there is a dangerous divergence of national laws on procedure and right of defence. It is important to harmonise that and bring it to a more level field. Finally, just a comment on the very interesting intervention of Frank Maier-Rigaud on the interpretation of Article 7: I congratulate you, because you have done a perfect lawyer job. That is, you took a provision that is very clear and demonstrated that it means the opposite of what it says. [Laughter] That is what lawyers do all the time. But that provision nevertheless has a big impact because, before that provision was adopted, behavioural remedies were not often discussed. Before that, it was generally thought that structural remedies were the priority type of remedies that enforcers should look for, and the existence of the provision has had an influence because at the national level you often discuss behavioural remedies. It has a practical impact. Monti: Thank you. Do any of the speakers want to respond to any of the points? DD

Lianos: I agree with what Mario has just said. I don’t think the fines are the major problem we have, as that can be fixed quite easily, either by limiting the 10 per cent turnover cap or by enhancing the judicial review of sanction decisions. We have already seen some steps of the Court in that direction.38 The most difficult part is related to commitment decisions and what I will call ‘regulatory antitrust’. Fuller wrote a paper back in the 1970s called the ‘Limits of Adjudication’.39 Basically, he was saying that if you bring in polycentric disputes, meaning disputes that relate not only to the two parties but also to other actors in the system—consumers, the broader public, etc— these kinds of disputes might prove difficult to adjudicate. Therefore, you can see an evolution of the system towards more regulation, or to a form of contractual or negotiated type of approach. This is a little bit what you have seen in EU competition law as you move to a more effects-based approach. It is less market access and more consumer welfare, or even the ‘public at large’. That is a very polycentric issue, and we have seen an evolution towards regulatory remedies or negotiated types of disputes. This resonates with what Philip Lowe said before in the sense that the Commission is even trying to improve the market equilibrium that prevailed before the infringement, which DD

38 See Case C-386/10 P Chalkor AE Epexergasias Metallon v Commission [2011] ECR I-13085, para 62. See also Case C-389/10 P KME Germany AG, KME France SAS and KME Italy SpA v Commission [2011] ECR I-13125. 39 Lon Fuller and KI Winston, ‘The Forms and Limits of Adjudication’, 92 Harvard Law Review 353, 397–398 (1978) (noting that ‘concealed polycentric elements are probably present in almost all problems resolved by adjudication’).

78 Part I: Effective Enforcement of Competition Law from a private law perspective is slightly strange. The other effect is to see an increasing number of commitment decisions. There I just want to say that the effects-based approach also leads to a focus on theories of harms related to narratives that have no clear principles in the sense of legal obligations. This increases the remedial discretion of the Commission. One of the major sideeffects of the Alrosa case is that if there is a complex economic assessment, the judicial review of the Commission is limited. But if you have an effects-based approach and a theory of harm which is an economic assessment and you try to devise remedies on the basis of this particular assessment that means that you also need to design remedies that forecast on the future, there the court is basically relieved of its obligation to do a review. In that sense, you can see that the effects-based approach, this polycentricism, leads to wider discretion for the Commission, and this is the major problem we have. The rise of commitment decisions is the most problematic development. Ost: On the issue of German procedural law, I want to restate that I think the system is fundamentally flawed. If you have 10–15 days reading out pleadings in the courtroom due to the principle of oral proceedings, it shows something is wrong in the system. I understand that you like the system as a defence lawyer. The undertaking infringing competition law likes to have an enforcement system which is too difficult to enforce on the basis of procedural means. I’m not advocating a one-to-one blueprint of procedural rules for the German system. There are ways in between. Some features can be taken over: the requests for information, the need to attack a decision quite precisely. Then you can streamline the procedure and allow for efficient judicial control. We need to see if it is really necessary to have the individual within this procedure because there we have a higher standard of fundamental rights to be applied. I wonder whether we should keep having the individual in our proceedings. Concerning the level of fines: to respond to Ian Forrester, we were obliged by our former law to calculate the effects, and the fines were pretty high then as well, but it was very burdensome to arrive there. In our cement cartel case, we had to inflict a 400 million euro fine on the wrongdoers, and in the liquid gas cartel case it was 240 million. But with the additional burden it is very difficult to take more cases. On the procedural issue, and on the call of Fred and Simon to do more on effects, I am very sceptical about whether we should really calculate the harm. We have to look at the impact, but more on general terms. It is not so much the effect of conduct that constitutes the wrongdoing as the conspiracy that intends to produce that effect. DD

Louis: You seem to be saying that the reason this is administrative is because it is an economic crime, but we don’t have to show the impact of the crime on the economy. You cannot have both. That is what the enforcer community has been trying to do for 10 years. I agree with Mario: the level DD

Part I: Effective Enforcement of Competition Law 79 of the fines in itself is not a problem, except for the fact that you cannot have the current level of fines and say: ‘But I can impose that in a way that (with the possible exception of Germany) is much easier than it is for somebody to fine me 25 euros because I went five kilometres per hour above the speed limit’. It is easier to impose a fine of 500 million euros for a cartel than it is to say ‘I have a picture of your car showing that you went too fast’. The argument that says ‘Don’t worry, it is corporations not individuals’ is flawed. We don’t have to be Mitt Romney to say that when you impose a 500 million euro fine on a company you are impairing its ability to invest, and this has a real impact on real people. You cannot say, for administrative convenience: ‘It is too hard to go for individuals’. This is a huge legitimacy issue because your aim is not compliance but to go for the quick and headline-grabbing fine. Refusing to pursue individuals means that you will never eliminate cartels from the world because the best compliance programme doesn’t impede an individual from doing something. You cannot have, at all times, one auditor per manager. It is impossible. And the problem is that the entire leniency system means that often we find ourselves in discussion with regulators where they are saying: ‘You cannot fire this person—we need them as witness’. Your policies sabotage internal compliance systems. You have to know that. We have to negotiate promotion plans over a 5- to 10-year period to get cooperative witnesses. What does that say about compliance? So you also have to look at individuals, and not only at astronomical fines. Maier-Rigaud: I’d like to thank Mario for complimenting me as being a good lawyer—although, as an economist, I had naïvely thought that lawyers normally tend to stay close to the true meaning of the law. I see I was wrong! My comment is more on the issue of fines, because it seems to be a hot topic here. I thought it would be useful to throw in that at least most of the economic studies I have that look at the deterrent effect of fines indicate that they are not sufficiently high and not sufficiently deterrent to actually match the illegal gains if the cartel. I think in this context I just wanted to throw in the deterrence trap that Christine Parker highlighted before. This is more of an issue, I think, than the notion that fines will over-deter. This is perhaps an empirical issue. DD

Monti: I’m not sure that we have enough time for all the speakers to rebut, but we’ll see. Alexander? DD

Alexander Italianer: Thanks. I have some quick comments. First, in response to Ian, fines go into the EU budget, but Member States then pay less, so there is no net effect on the budget. Second, I would like to put the size of the fines in perspective. Commitment decisions can have a much bigger monetary effect on the profits of a firm than a fine. Big numbers can have a

DD

80 Part I: Effective Enforcement of Competition Law small impact on a company’s bottom line. Some of the biggest fines that we have imposed amount to less than 1 per cent of the annual turnover of the companies involved. And finally, we do differentiate in fines where this is justified. For example, in the Telefónica/Portugal Telecom case, outright market partitioning was done not in secret, but publicly.40 We took that into account when setting the fine. We also impose a lower fine when there are novel issues. Turning to Fred’s point about us bringing in the parents, we interpret the 10 per cent cap not in the German way but in relation to a firm’s ‘capacity to pay’. It is very easy to set up a construction whereby the company that is being fined can duck the fine, and that is why we need to go after the parents. Finally, Fred makes a correct point about the 10 per cent cap being reached not in all cases but in cases where you have so-called mono-product companies. That may then lead to a problem of differentiation of fines imposed on the company, and the Court has actually expressed an obiter dictum with respect to that issue. But we have been responsive to that. It may not have been noticed yet, but if you look at the fines that were calculated in the Mountings case,41 which involved many mono-product companies, we had fines below the 10 per cent range, introducing more differentiation. I’ll leave it at that. Alberto Heimler: I have two points. One concerns what Frank just said, because we wrote a paper showing that fines were indeed too high and that there are ways to take into account issues of the market that would tend to reduce the fine. Above all, it can be relevant to look at elasticity of demand, which is often neglected. That could be something that can be assessed during the investigation to link the fine to the overcharge. Regarding the estimates that Simon was mentioning, where John Connor referred to a 25–30 per cent overcharge, that does not take into account the fact that what people look at is not the overcharge but the increased profits originating from the overcharge. That is quite important, because every overcharge brings with it a reduction in quantities that should be taken into account in the setting of fines. The s econd point is on effective remedies, and here I would just like to draw in the discussion of what should be meant by ‘effective’ in an antitrust context because ‘effective’ should mean bringing the infringement to an end. In abuse cases, in particular, this has not been the case, and very often we have introduced remedies going beyond that. And I would just like to refer to the Microsoft case again—sorry, Philip. In the Microsoft case, the effective remedy was a reasonable price that was imposed on the company for interface information, and this is quite clear both from the decision and in DD

Commission Decision of 23 January 2013, Telefónica/Portugal Telecom, above n 11. Commission Decision of 28 March 2012 in Case COMP/39452—Mountings for indows and window-doors. w 40 41

Part I: Effective Enforcement of Competition Law 81 articular from the judgment of the General Court. I’m wondering whether p such remedies are appropriate in an antitrust context, and whether ‘reasonable’ is the standard we should refer to, or whether indeed more appropriately we should refer to something like an exclusionary standard. It would be much more in line with antitrust thinking, and otherwise there would be a risk of the antitrust rules blurring with regulation. After all, the objective of regulation is to introduce reasonable pricing and remedies. Savvas Papasavvas: I need to underline at the start that opinions expressed by myself here are purely personal and do not reflect any official opinion of the General Court. Is the 10 per cent turnover rule excessive? I don’t think it is, and I hope to explain tomorrow why. With respect to the question of administrative or criminal sanctions, the General Court is, of course, part of an administrative system or procedure which is fully justified by the fact that, despite what Strasbourg might say, the Commission fines are administrative sanctions and not criminal or quasi-criminal. At least, that is my position. Do we get to limit the excessive fines? Last year we had E.ON/GDF, where there was a 1.1 billion euro combined fine, and we reduced it to 640 million.42 So we do get to reduce fines a lot, I know that quite well. Last but not least, with regard to the 50 pages, even though we do allow parties to have more than that, 50 pages are more than sufficient for a competent lawyer to point out the flaws of any kind of Commission decision. You need to focus on the one, two or three pleas before the Court, and you will be able to make your argument. There is no need for drowning the Court with more than that. I will save the rest for tomorrow. DD

Stephen Calkins: I do think it is harmful to have people suggesting that fines are high because there is a bidding war and that people want to outdo their predecessor and that fines are high to ensure that agencies are adequately funded. It is harmful to have those opinions out there. That is point one. Point two: I personally don’t think that we have sufficient deterrence, looking at the major cartels that indeed include some of the major corporations that end up price fixing in a way that is sort of shocking after all the publicity of the competition law enforcement. Obviously, something is not working. I would agree that taking huge amounts of money out of companies’ pockets does do harm to the companies sometimes. For me, the problem is that it is harming the wrong people because it is often harming successors or shareholders. It is not really harming the people that are responsible for the infringement. Which is why I get out of bed in the morning with the point that Christine Parker made in her paper. At least in theory, criminal enforcement is something that makes DD

42 Joined Cases T-360/09 E.ON Ruhrgas and E.ON AG, EU:T:2012:332; Case T-370/09 GDF Suez SA v Commission, EU:T:2012:333.

82 Part I: Effective Enforcement of Competition Law a great deal of sense—both in sending a message and in seriously deterring the individuals. I can’t think how many times I have seen corporate law firms giving advice, saying: ‘You really shouldn’t do this because you are risking criminal exposure but now let’s talk about all those other things where it’s going to be a judgement call and it’s a cost–benefit kind of thing’. When you look at the people that are trying to avoid the US as a matter of geography to avoid a criminal sanction, I really think that, if you can make it work, criminal sanctions against individuals are a terribly effective way to communicate a message and increase deterrence. Where that is not an option, focusing on the individual would solve the perception problems, and would perhaps be more deterrent than increasing the fines. John Davies: I just quickly wanted to comment on this question of fines versus remedies. I have had the privilege of being an enforcer in a system where fines were not available for abuse of dominance. That was a law that was heavily influenced by a couple of British economists (neither Simon nor me). A remedy-only system for abuse of dominance was great, and I still think it is great—with one exception, and that is repeat offences and essentially the same thing being done by other companies. Because what you see is that you remedy something and someone else comes along and does the same thing, but they are not quite the same products and not quite the same behaviour. In those circumstances you do need some sort of very narrow fine. Otherwise you are going around, trying to have an opinion on every single contract in the economy, and that’s not feasible. I would not take that too broadly, because you come back to a position of per se prohibitions established by precedent. But at the very minimum, the competition authority needs to have something like that to effectively give wider force to its decisions. DD

Bill Kovacic: I want to suggest that there are two corrective mechanisms that will influence a competition agency as sanctions increase. One of these is that, as sanctions go up, courts will exercise more intrusive oversight. There will come a point at which courts will begin to do two things. One is that the margin of discretion will disappear. The second thing is that they will start to reshape doctrine. They will begin to reformulate the legal standard without doing obvious harm to the existing framework. I can’t define for you the exact threshold at which that will happen, but the threshold does exist and the competition agency has to be attentive to that as it seeks additional monetary penalties—especially where the monetary penalties are sought for more ambiguous behaviour rather than so-called clear-cut offences. Again, it is hard to predict what that boundary is, but it will happen at some point, and perhaps even in an unconscious way. Judges will begin to respond to arguments about a lack of proportionality. The second thing that an agency invites over time is more intrusive political oversight. Again, there is no sign DD

Part I: Effective Enforcement of Competition Law 83 at the frontier that says you are approaching the boundary of more political oversight, but I think that boundary exists and that it will take several forms. One is that you will be called more often before parliamentary groups to explain what you are doing. There is the hidden threat in the background that they will change the law and take away your power. These are somewhat indistinct sources of constraint, but I think that, for a competition agency, day in, day out, as you climb the staircase for more powerful sanctions, you have to realise that it becomes a bit more dangerous up there. It requires greater efforts to provide justifications, foundations, a sound basis and facts that demonstrate action is warranted.

84

Joshua D Wright* The Federal Trade Commission and Monetary Remedies

I’m delighted to participate in this Workshop at the European University Institute, and I’d like to thank Philip Lowe, Mel Marquis and Giorgio Monti for inviting me to join you here in Florence and to share my views.

1. Introduction Today I would like to talk about the approach taken by the Federal Trade Commission (FTC) in pursuing monetary remedies against defendants in competition cases. Unlike the Department of Justice, which has the broad legal authority to pursue all sorts of remedies against antitrust defendants, the FTC’s authority is limited to pursuing so-called ‘equitable’ remedies.1 Equitable remedies typically involve court orders directing a defendant or defendants to engage or not engage in certain behavior, ie an injunction. In a merger case, the FTC will often ask a court to direct the parties divest a business in a particular geographic market. Or, in a case alleging anticompetitive conduct, the FTC might ask a court to issue an injunction preventing the defendants from engaging in the conduct the FTC alleges to be anticompetitive. Although monetary remedies are typically remedies awarded ‘at law’ rather than ‘in equity’—this distinction is a function of the twin court systems of law and equity in Anglo-American law—a court exercising its equitable powers can order a defendant to pay monies in certain circumstances. One broad category of equitable remedies that involves a defendant paying monies is ‘disgorgement’, which is a remedy requiring the defendant to in effect pay back his ill-gotten gains.2 * At the time writing, Commissioner, US Federal Trade Commission; currently: Professor of Law, George Mason University School of Law. The views stated here are my own and do not necessarily reflect the views of the FTC or of other Commissioners. For their invaluable assistance when I prepared this chapter I am grateful to my advisor Derek Moore and to my intern Brady Cummins. 1 15 USC § 53(b). 2 See United States v Grinnell Corp, 384 US 563, 577 (1966) (stating that adequate relief in a monopolization case can include ‘depriv[ing] the defendants of any of the benefits of the illegal conduct’); 2A Phillip Areeda and Herbert Hovenkamp, Antitrust Law: An A nalysis of Antitrust Principles and Their Application, 3rd ed, Aspen Law and Business, 2006, ¶ 325a (‘[E]quity relief may include… the disgorgement of improperly obtained gains’).

86 Joshua D Wright The FTC did not pursue disgorgement against antitrust defendants until 1999, when it sought $120 million from a pharmaceutical company the Commission alleged had overcharged consumers.3 In 2003, after accepting comments from interested parties, the FTC adopted a Policy Statement on Monetary Remedies in Competition Cases that was designed to provide guidance to the business community regarding when and under what conditions the FTC would seek monetary remedies against antitrust defendants.4 In 2012, before I joined the Commission, the Statement was clandestinely withdrawn—that is, without the opportunity for public comment from interested parties—over the dissent of Commissioner Ohlhausen.5 The Withdrawal of the Statement is concerning for several reasons. First, the Withdrawal was predicated upon a belief—that the Statement ‘chilled the pursuit of monetary remedies’6 while it was in effect—that has no empirical support. Second, the Commission, in withdrawing the Statement, incorrectly suggests that, because recent decisions by the Supreme Court of the United States have made it more difficult for a private plaintiff to bring a successful antitrust case, antitrust enforcement agencies in the US ought to ‘pick up the slack’ by pursuing monetary remedies with more frequency. In other words, the Commission suggests a perceived reduction in private enforcement of the antitrust laws requires greater public enforcement efforts to achieve optimal deterrence. This proposition makes sense only if there is some e vidence that the current state of the law results in under-deterrence of anticompetitive conduct. I am aware of no such evidence and, in fact, current economic thinking suggests that, at least in the context of single-firm conduct and vertical restraints, pursuing monetary penalties by seeking disgorgement against defendants with more frequency would almost certainly deter efficient behavior and harm consumers. There can be no doubt that disgorgement is a useful tool for the antitrust enforcer. In many cases, disgorgement of profits may be superior to divestiture or other behavioral remedies that require costly oversight by regulators to ensure compliance. Disgorgement in such contexts may sufficiently deter potential wrongdoing and reduce the cost of administering the antitrust system. In my view, however, the Commission ought not to pursue disgorgement to remedy against conduct that has plausible efficiency justifications. This is because, in the context of conduct that can be efficient and b enefit

FTC v Mylan Labs, Inc, 62 F Supp 2d 25 (DDC 1999). Policy Statement on Monetary Equitable Remedies in Competition Cases, 68 Fed Reg 45820 (4 August 2003) (hereinafter Policy Statement, or Statement). 5 Statement of the Commission, Effecting the Withdrawal of the Commission’s Policy Statement on Monetary Equitable Remedies in Competition Cases, 77 Fed Reg 47070 (31 July 2012) (hereinafter Withdrawal), http://www.ftc.gov/os/2012/07/120731commissio nstatement.pdf. 6 Ibid. 3 4

The Federal Trade Commission and Monetary Remedies 87 consumers in some contexts and harm competition and consumers in others—here I am referring to vertical distribution restraints imposed by a firm with market power—antitrust enforcers should be more worried about over-deterrence than under-deterrence. In any event, the Commission’s pursuit of certain remedies affects how businesses structure their dealings. Indeed, shaping the incentives of all m arket participants—and not punishment—is the point of pursuing remedies in many antitrust cases. For this reason, I am disappointed the Commission decided to withdraw the Policy Statement without a substantive discussion about the proper role of monetary remedies in the public antitrust enforcement system. Further, the Withdrawal is troubling because the absence of guidance creates uncertainty within the business community, which will undoubtedly affect firms’ behavior in ways that are unpredictable and will unnecessarily run the risk of harming the consumers we are charged with protecting.

2. Summary of the 2003 policy statement In the late 1990s and early 2000s, the FTC sought disgorgement in two antitrust cases.7 Though the Commission had long sought disgorgement of profits in consumer protection cases, seeking disgorgement in antitrust cases was breaking new ground.8 The Commission’s efforts to pursue disgorgement in antitrust cases led to critical commentary,9 and the Commission ultimately requested comments from the public about the conditions under which it would be appropriate for the Commission to pursue disgorgement and other monetary remedies in antitrust cases.10 After reviewing those comments, the Commission issued a Policy Statement in July 2003 with bipartisan support from all five commissioners.11 The Statement outlined three factors the Commission would consider: (i) whether the violation of law is ‘clear’; (ii) whether there is a reasonable basis upon which to calculate the disgorgement payment; and (iii) whether remedies in other litigations are likely to fail to accomplish fully the purposes of the antitrust laws.12 The Policy

See Mylan Labs, above n 3; FTC v The Hearst Trust, No 1:01CV000734 (DDC 2001). See Alden Abbott, ‘FTC Monetary Remedies Policy and the Limits of Antitrust’, Antitrust Source, December 2012, Article 2, 1; David Park and Richard Wolfram, ‘The FTC’s Use of Disgorgement in Antitrust Actions Threatens to Undermine the Efficient Enforcement of Federal Antitrust Law’, Antitrust Source, September 2002, Article 5, 1. 9 See Park and Wolfram, ibid. 10 Remedial Use of Disgorgement, 66 Fed Reg 67254 (28 December 2001). 11 Policy Statement, above n 4. 12 Ibid. 7 8

88 Joshua D Wright tatement was later endorsed without dissent by the Antitrust Modernization S Commission.13

3. The FTC’s 2012 Withdrawal Without asking for public comments from the business community, the Commission decided to withdraw the Statement in 2012 and gave several reasons that, in my view, are less than satisfying.14 The Commission offered both legal and policy reasons for withdrawing the statement. First, the Commission criticized the legal basis for the three factors laid out in the Guidelines. On the Statement’s guidance that monetary remedies will be pursued only when the violation is ‘clear’, the Commission observed that rarity or clarity of the violation is not an element considered by courts in disgorgement requests … Whether conduct is common or novel, clearly a violation or never before considered, has little to do with whether the conduct is anticompetitive; some novel conduct can violate the antitrust laws and can be even more egregious than clear violations.15

Next, the Commission criticized the requirement in the Policy Statement that the Commission consider whether alternative plaintiffs may seek monetary relief, which could potentially obviate the need for the Commission to seek disgorgement. The Commission reasoned that the presence of alternative plaintiffs seeking monetary recovery ‘is relevant in this context, but it is not dispositive. It is only one of several questions that might usefully be asked in deciding whether a Commission imposed monetary remedy is appropriate and necessary.’16 Finally, notwithstanding that the factors set forth in the Statement were not legal requirements, the Commission, in withdrawing the Statement, noted that it was ‘concerned that parties could mistakenly argue that the factors laid out in the Policy Statement are binding on the Commission, thus creating an unnecessary side issue in litigation’.17 The Commission also put forward two policy reasons for withdrawing the Statement. First, it argued that, in its experience, ‘the Policy Statement has chilled the pursuit of monetary remedies in the years [since it was issued]’.18 Second, and most interestingly from my perspective, the Commission also

Antitrust Modernization Commission, Report and Recommendations 288 (2007). Withdrawal, above n 5. 15 Ibid. 16 Ibid. 17 Ibid, 1, footnote 2. 18 Ibid, 2. 13 14

The Federal Trade Commission and Monetary Remedies 89 suggested that some recent decisions by the Supreme Court justify the Commission’s increased use of disgorgement as a remedy in antitrust cases: At a time when Supreme Court jurisprudence has increased burdens on plaintiffs, and legal thinking has begun to encourage greater seeking of disgorgement, the FTC has sought monetary equitable remedies in only two competition cases since we issued the Policy Statement.19

4. The FTC’s reasons for withdrawing the statement are not persuasive The Commission’s Withdrawal is troubling in many respects. First, unlike its decision to adopt the Policy Statement in 2003, the Commission’s decision to withdraw the Statement in 2012 was not unanimous. My colleague Commissioner Ohlhausen dissented from the Commission’s withdrawal and, in doing so, raised several interesting and important issues.20 Indeed, the fact that the Statement was withdrawn with minimal deliberation by the Commissioners, without a request for public comment and over the dissent of a Commissioner, is troubling by itself.21 Moreover, as Commissioner Ohlhausen points out, withdrawing the Statement obviates exactly what the Statement attempted to do in the first place: provide the business community with guidance regarding the Commission’s pursuit of monetary remedies in antitrust cases. Indeed, the Policy Statement was just that—a statement of policy regarding how the Commission would exercise its legal enforcement authority. If ‘rarity or clarity of the violation’ was an element considered by courts in deciding whether to award disgorgement, then pursing disgorgement only in such cases could not be a matter of Commission policy; under the law, the Commission would be forbidden from seeking disgorgement in cases in which the violation is somehow less than clear. This is simply another way of stating that a policy statement has no utility when it merely restates the law. Accordingly, it is of no moment to criticize a policy statement when it does something more than simply restate the law by arguing that the policy statement imposes

19 Ibid (citing Einer Elhauge, ‘Disgorgement as an Antitrust Remedy’, 76 Antitrust Law Journal 79 (2009)). 20 Statement of Commissioner Maureen K Ohlhausen, Dissenting from the Commission’s Decision to Withdraw its Policy Statement on Monetary Equitable Remedies in Competition Cases, 77 Fed Reg 47071 (31 July 2012) (hereinafter Ohlhausen Dissent), http://www.ftc.gov/os/2012/07/120731ohlhausenstatement.pdf. 21 Ibid.

90 Joshua D Wright a requirement that the law does not impose. As Commissioner Ohlhausen explains, a statement that the Commission will rely upon existing law could be used to justify virtually any decision by the Commission.22 More fundamentally, the Commission suggests in its Withdrawal that conduct cannot be ‘novel’ and constitute a ‘clear’ violation of the law. I see no reason why this observation must be true; indeed, it is not difficult to come up with examples of novel yet clear violations. Clarity does not come only from past experience. Novel conduct can certainly constitute a clear violation of the antitrust laws if it harms competition by restricting output and raising price and is without any efficiency justification. For example, there are myriad forms of deception, not all of which have been challenged in court as antitrust violations and, to the extent that an egregious example of deception by a firm with market power results in that firm increasing or maintaining monopoly power, that deception could constitute a clear violation of the antitrust laws that would justify disgorgement of profits even if it is not closely similar to a past successful antitrust case. The same can be said of rival firms concocting novel means of naked price-fixing. Novel or not in the form of their implementation, naked price-fixing schemes clearly violate Section 1 of the Sherman Act. Another reason the Commission gave for withdrawing the Statement is that the Statement ‘chilled’ the Commission from pursuing monetary remedies in antitrust cases.23 Yet the Commission cited no data or other evidence to support this assertion. Indeed, Commissioner Ohlhausen stated, ‘I have not been presented with any evidence that the Policy Statement has inappropriately constrained the Commission in the nine years it has been in effect’.24 Lack of data supporting this view is yet another troubling aspect of the Withdrawal. A review of Commission cases supports Commissioner Ohlhausen’s position. Indeed, the Commission sought disgorgement in two cases during the nine-year period before the Policy Statement was in effect,25 and also sought disgorgement in two cases during the nine-year period while the Policy Statement was in effect.26 These data could support the conclusion that the Commission does not pursue disgorgement very often in antitrust cases— only four times in 18 years—but they do not provide support for the conclusion that the Policy Statement has had any effect on the Commission’s pursuit

Ibid. Withdrawal, above n 5, 2. Ohlhausen Dissent, above n 20. 25 The Hearst Trust, above n 7 (settlement resulting in defendant paying $19 million in disgorgement); Mylan Labs, above n 3 (holding that the FTC could sue for monetary damages as well as injunctive relief, resulting in a $100 million settlement). 26 FTC v Lundbeck Inc, 650 F3d 1236 (8th Cir 2011) (affirming the district court’s ruling denying any relief on antitrust grounds); FTC v Perrigo Co, No 1:04CV01397 (RMC) (DDC 2006) (resulting in a settlement that included a $6.25 million disgorgement). 22 23 24

The Federal Trade Commission and Monetary Remedies 91 of monetary remedies, much less that the Statement has chilled Commission efforts to pursue disgorgement. It is possible, however, to expand the relevant universe of cases to consider remedies that are similar to disgorgement but do not actually require a monetary payment from the defendant. The Commission has, on a few occasions in recent years, sought a remedy whereby the owner of standardessentialpatents either cannot enforce the patents or must license the patents on a royalty-free basis.27 Requiring a patent holder to grant a royalty-free license to certain patents is similar to disgorgement in that the remedy effectively takes profits away from the defendant. In the typical context, a wouldbe licensee of the defendant’s patents complains that the defendant is causing ‘hold up’ by engaging in conduct that does not comport with the typical promise SEP-holders make to license their patents on fair, reasonable and non-discriminatory (FRAND) terms. This conduct can take the form of seeking an injunction against a willing licensee or demanding ‘supra-FRAND’ royalty rates. In this context, when the Commission requires the patentholder to license the patents royalty free, it is in effect ‘disgorging’ any profit the defendant would have made had it been able to reach licensing agreements with the complaining licensees. Indeed, Unocal, a defendant in one of the cases, argued that the royalty-free licensing requirement in that case amounted to ‘confiscation and disgorgement’.28 Adding the royalty-free licensing cases to the count of FTC actions further weakens the Commission’s claim that the Statement on monetary remedies has ‘chilled’ the Commission’s disgorgement efforts. Because only one of the four royalty-free licensing cases occurred before the Statement was adopted, adding these cases to the mix would mean that the Commission had in fact pursued disgorgement with greater frequency after the Statement was adopted. Of course, the sample size remains very small, so it would be a mistake to make too much of the numbers other than to point out that they do not appear to support the Commission’s concern that the Policy Statement chilled pursuit of disgorgement. Still, these cases raise a number of interesting issues regarding remedies. The Bosch case, the most recent decision by the Commission requiring royalty-free licensing, was followed six weeks later by the Commission’s consent agreement with Google relating to Google’s

27 Union Oil Co of Cal, 140 FTC 123 (2005); Dell Computer Corp, 121 FTC 616 (1996); Robert Bosch GmbH, No C-4377 (FTC 21 November 2012); Opinion of the Commission on Remedy, Rambus, Inc, No 9302 (FTC 5 February 2007), http://www.ftc.gov/os/adjpro/ d9302/070205opinion.pdf. 28 Respondent’s Trial Brief at 189, Union Oil Co of Cal, ibid, http://www.ftc.gov/os/ adjpro/d9305/041008unocalstrialbrief.pdf (‘Even though framed as a “cease and desist” remedy, there is no question that the essence of the relief sought in this action is a confiscation and disgorgement of Unocal‘s patent rights in California’).

92 Joshua D Wright failure to license standard-essential patents.29 In the agreement with Google, the Commission did not require royalty-free licensing to remedy the alleged law violation, instead requiring that Google go through a dispute resolution process before seeking injunctive relief against a willing licensee.30 The Commission failed to explain why it chose two different remedies in two cases involving similar conduct. Perhaps if the Policy Statement were in effect, the Commission would at least have had to grapple with this issue.

5. Would a reduction in private antitrust enforcement justify additional efforts by public enforcers to pursue monetary remedies? I would like to spend the balance of my time discussing what I find to be the most interesting aspect of the Commission’s decision to withdraw the Policy Statement: the idea that the FTC needs to pursue monetary remedies with more frequency because recent Supreme Court cases have made it more difficult for private plaintiffs to win antitrust cases. In its Complaint against Intel, former Chairman Leibowitz and former Commissioner Rosch issued a joint statement explaining why the Commission sued Intel under Section 5 of the FTC Act rather than under Section 2 of the Sherman Act: [C]oncern over class actions, treble damages awards, and costly jury trials have caused many courts in recent decades to limit the reach of antitrust. The result has been that some conduct harmful to consumers may be given a ‘free pass’ under antitrust jurisprudence, not because the conduct is benign but out of a fear that the harm might be outweighed by the collateral consequences created by private enforcement. For this reason, we have seen an increasing amount of potentially anticompetitive conduct that is not easily reached under the antitrust laws, and it is more important than ever that the Commission actively consider whether it may be appropriate to exercise its full Congressional authority under Section 5.31

Though the argument is framed in the context of the FTC’s increased use of its power to prosecute ‘unfair methods of competition’ under the FTC Act, the argument could just as easily apply to the FTC’s efforts to pursue

29 The Commission reached settlements with both defendants before I became Commissioner. I did not participate as Commissioner in considering the substance of either agreement. 30 Motorola Mobility LLC, No 121-0120, 2013 WL 124100 (FTC 3 January 2013). 31 Statement of Chairman Leibowitz and Commissioner Rosch in Intel Corp, No 9341, 2010 WL 4542454 (FTC 2 November 2010), http://www.ftc.gov/os/adjpro/ d9341/091216intelchairstatement.pdf.

The Federal Trade Commission and Monetary Remedies 93 disgorgement or other monetary relief from antitrust defendants. If certain defendants are given a ‘free pass’ under the current antitrust laws, then the FTC ought to pursue harsher remedies against those defendants it does sue to make up for the cases that would have been brought but for the Supreme Court’s decisions. For the former Commissioners’ position to make sense, two testable hypotheses would need to be confirmed: (i) that private antitrust lawsuits have become more difficult to win; and (ii) that anticompetitive conduct is currently under-deterred by the antitrust laws. I discuss both in turn below.

5.1 Is it now more difficult for a private plaintiff to win an antitrust case? The most commonly cited cases for the proposition that it has become more difficult for plaintiffs to win antitrust cases are a series of recent decisions by the Supreme Court beginning with Trinko in 2004.32 In that case, the Court rejected a monopolization claim brought under Section 2 of the Sherman Act alleging that an incumbent local telephone exchange carrier illegally monopolized the local telephone market by refusing to offer exchange service to its competitors.33 The Court held that any exception to the general proposition that a firm has the choice to deal or not deal with whomever it chooses was narrow and that, because the defendant had not discontinued a prior profitable course of dealing with competitors, the exception did not apply to the plaintiff’s case.34 Another case is Twombly, in which the Court altered the pleading standard a plaintiff must meet in alleging an illegal conspiracy in violation of Section 1 of the Sherman Act.35 The Court held that alleging an illegal conspiracy requires a complaint with enough factual matter (taken as true) to suggest that an agreement was made. Asking for plausible grounds to infer an agreement … simply calls for enough fact to raise a reasonable expectation that discovery will reveal evidence of illegal agreement.36

Another case is linkLine, in which the plaintiffs alleged that a vertically integrated defendant violated Section 2 by ‘squeezing’ the profit margins of downstream competitors that also purchase inputs from the defendant by increasing input prices and simultaneously decreasing downstream prices.37

32 Verizon Communications, Inc v Law Offices of Curtis V Trinko, LLP, 540 US 398 (2004). 33 Ibid, 404–405. 34 Ibid, 409. 35 Bell Atl Corp v Twombly, 550 US 544 (2007). 36 Ibid, 556. 37 Pac Bell Tel Co v linkLine Communications, Inc, 555 US 438 (2009).

94 Joshua D Wright The Court held that for a plaintiff to state a viable claim of ‘price squeezing’, the plaintiff must allege that the defendant charge a price below some relevant measure of cost and thereby satisfy the same test the Court applies to ordinary predatory pricing claims.38 Yet another case is Credit Suisse, in which the Court held that certain antitrust claims reached conduct governed by federal securities law and, because the antitrust claims were ‘repugnant’ to federal securities law, they could not proceed.39 I think it is worth unpacking exactly what it means to suggest that it has become ‘more’ difficult for private litigants to win antitrust cases. Saying the law has changed necessarily raises the question of identifying the appropriate benchmark by which to evaluate the change, not to mention asking whether the chosen benchmark is helpful. A soccer team might have won its last five matches, but observing that the team ‘always wins’ based upon that information without the proper context can be misleading about the team’s overall results. Perhaps the team lost five consecutive matches immediately before its five-match winning streak. Or perhaps it had lost the prior 15 matches. A recent winning streak may suggest that the tide is turning, but we must be careful not to overstate the evidence. And so it is with antitrust decisions made by the US Supreme Court. Judge Douglas Ginsburg and Leah Brannon examined the Supreme Court’s decisions in antitrust cases, looking at the last 45 years rather than at just the most recent 10.40 They found that in each of the four decades, beginning with 1967 to 1976 and ending with 1997 to 2006, defendants won antitrust cases with increasing frequency. Whereas defendants won all 13 of the antitrust cases decided by the Supreme Court from 1997 to 2006, defendants won only 16 of 44 cases decided by the Supreme Court from 1967 to 1976. The authors found several other telling trends. The number of decisions signed on to by a ‘supermajority’ of six or more of nine Justices increases over time, with 11 of the 13 pro-defendant decisions of the decade from 1997 to 2006 being decided by a supermajority. Also, the authors found that briefs submitted by the United States as amici in a private antitrust dispute pending before the Supreme Court also favored the defendant’s position with increasing frequency as time went on from 1967 to 2007, a period that includes presidential administrations of both political parties. Ginsburg and Brannon also argue convincingly that the Court’s move to decide cases in favor of defendants is a result of economic analysis becoming more engrained in antitrust law, rather than a result of bias. As the authors

Ibid, 451–452. Credit Suisse Sec (USA) LLC v Billing, 551 US 264 (2007). 40 Leah Brannon and Douglas Ginsburg, ‘Antitrust Decisions of the US Supreme Court, 1967 to 2007’, 3(2) CPI Journal 1, 3 (2007). 38 39

The Federal Trade Commission and Monetary Remedies 95 note, ‘the Court, far from indulging in a pro-defendant or anti-antitrust bias, is [instead] methodically re-working antitrust doctrine to bring it into alignment with modern economic understanding’.41 In a separate article, Ginsburg opines that The Court’s reliance upon modern economic analysis reflects the near consensus among academics on proper antitrust analysis. There is now broad and nonpartisan agreement in academia, the bar, and the courts regarding the importance of sound economic analysis in antitrust decision making. Such analysis has utterly transformed the dialogue within the Supreme Court.42

Putting aside the issue of whether the more recent pro-defendant decisions by the Supreme Court reflect a keener understanding of economics—and I most certainly agree that they do when compared to the antitrust decisions by the Supreme Court in the late 1960s and early 1970s—the data compiled by G insburg and Brannon are clear evidence of a different claim. The data show that, even if the Supreme Court has been deciding cases in favor of defendants in recent years, that trend simply reverses a prior trend in which the Supreme Court tended to decide cases in favor of plaintiffs. The claim that the FTC needs to pursue monetary remedies against antitrust defendants to replace the deterrent value of private lawsuits that would have been successful but for the Supreme Court’s recent jurisprudence simply assumes the law provided the proper incentives prior to the recent string of victories by defendants. Before addressing that question—whether the pre-Trinko state of antitrust law provided the appropriate level of deterrence against anticompetitive conduct—I would first like to examine some data on private antitrust lawsuits in the United States. It is difficult to evaluate empirically whether recent decisions by the Supreme Court have made it more difficult for plaintiffs to win cases because so many lawsuits result in settlement. Looking only at cases that proceed to a certain phase and result in a victory for plaintiffs would obscure plaintiffs’ true success rates. Case filings by private plaintiffs are a useful proxy, however, because a potential antitrust plaintiff will incorporate a higher or lower probability of success based upon a favorable or unfavorable decision by the Supreme Court into his decision regarding whether to file a lawsuit in the first place. Accordingly, one testable implication of the hypothesis that recent decisions have made life more difficult for antitrust plaintiffs is whether recent case filings have decreased.

Ibid, 4. Douglas Ginsburg, ‘Originalism and Economic Analysis: Two Case Studies of Consistency and Coherence in Supreme Court Decision Making’, 33 Harvard Journal of Law and Public Policy 217, 222 (2010). 41 42

96 Joshua D Wright The data suggest otherwise. Paul Godek has examined antitrust case filings in the United States, comparing the number of private suits to the number of suits brought by the federal agencies through 2008.43 According to Godek, in recent years private suits have outnumbered suits brought by the federal agencies by more than twenty to one.44 With regard to private suits, the data show that they have increased significantly between 2004 when Trinko was decided and 2008.45 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 0

1949 1953 1957 1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001 2005 Private Antitrust (Electrical Equipment)

Private Antitrust (All Other)

Government (Civil) Antitrust

Figure 1: Private and Government Civil Antitrust Cases Filed in Federal Courts Source: Godek (note 43).

The data on private suits between 2008 and 2012 tell a slightly different story.46 These data show a lull in private enforcement after 2008, perhaps explained by the economic collapse, but with a surge in 2012.47 It would be easy to make too much of this data; case filings are an imperfect proxy for win rates. Nevertheless, the data do indicate that the pro-defendant antitrust decisions starting with Trinko in 2004 have not resulted in a decrease in antitrust case filings by private plaintiffs.

43 Paul Godek, ‘Does the Tail Wag the Dog? Sixty Years of Government and Private Antitrust in the Federal Courts’, Antitrust Source, December 2009, Article 5, 1. 44 Ibid, 2. 45 Ibid. 46 Jonathan Randles, ‘Reversing Trend, Antitrust Suits Spiked in 2012’, Law360, 14 January 2013, http://www.law360.com/articles/406948/reversing-trend-antitrust-suitsspiked-in-2012. 47 Ibid.

The Federal Trade Commission and Monetary Remedies 97

5.2 Even if it is more difficult for a private plaintiff to win an antitrust case, does it follow that anticompetitive conduct is under-deterred? Even if there were evidence to support the claim that private antitrust lawsuits are declining, it would not follow that anticompetitive conduct is being under-deterred. Former Chairman Leibowitz and former Commissioner Rosch argue that the recent string of defendant victories in antitrust cases before the Supreme Court is because the Court is concerned about ‘costly’ private litigation—expensive discovery, class actions, treble damages—and not because the Court has substantive views about the competitive merits of the conduct alleged to be harmful in those cases.48 A close examination of the decisions in question reveals quite clearly that the Court is motivated by substance, ie the competitive merits of the conduct in question, and not solely by a concern about the cost of private litigation. This distinction is important because if the Court is motivated purely by a concern about designing antitrust rules to reduce the cost of private litigation, then those same concerns may not be present in antitrust cases brought by the government.49 To be sure, the Supreme Court has expressed concerns about the cost and uncertainty of private litigation.50 However, a much stronger theme in the Court’s jurisprudence is a concern about any court’s ability to separate anticompetitive conduct, which harms competition and consumers and ought to be condemned, from conduct that is benign or pro-competitive.51 The Court further recognizes that a court’s ability to distinguish between pro- and anticompetitive conduct is especially important in antitrust cases because it is a more severe error for a court incorrectly to condemn pro-competitive

Statement of Chairman Leibowitz and Commissioner Rosch, above n 31. See Statement of Howard Shelanski, Deputy Director for Antitrust (FTC), ‘Is There Life After Trinko and Credit Suisse? The Role of Antitrust in Regulated Industries: Before the Subcommittee on Courts and Competition Policy of the House Committee on the Judiciary’, 111th Congress (2010). 50 See Twombly, above n 35, 557–560 (explaining that discovery can be costly in antitrust cases and that a plaintiff with a ‘groundless’ claim may use the threat of discovery to improve positioning in settlement negotiations); Credit Suisse, above n 39, 281–282 (stating that ‘antitrust plaintiffs may bring lawsuits throughout the Nation in dozens of different courts with different nonexpert judges and different nonexpert juries’). 51 linkLine, above n 37, 451 (‘To avoid chilling aggressive price competition, we have carefully limited the circumstances under which plaintiffs can state a Sherman Act claim by alleging that prices are too low’); Credit Suisse, above n 39, 283 (‘[W]here the threat of antitrust lawsuits, through error and disincentive, could seriously alter underwriter conduct in undesirable ways, to allow an antitrust lawsuit would threaten serious harm to the efficient functioning of the securities markets’); Trinko, above n 32, 414 (‘Mistaken inferences and the resulting false condemnations are especially costly, because they chill the very conduct the antitrust laws are designed to protect’ (internal quotations omitted)). 48 49

98 Joshua D Wright behavior. This is because, as Judge Frank Easterbrook first explained, ‘the economic system corrects monopoly more readily than it corrects judicial errors’.52 Accordingly, the correct conclusion is that the Supreme Court has adopted rules designed to minimize the social costs of the antitrust system, which includes the cost of false negatives and false positives, but has explicitly recognized that the cost of false positives are greater and more intractable, and therefore are of greater concern in the antitrust system. The cost of discovery in suits brought by private plaintiffs is relevant to the Court’s approach within this framework, but it is certainly not the most important factor, much less the sole motivating factor. A close examination of the Court’s recent decisions makes clear that the Court is designing legal rules to reflect its concern about over-deterring conduct that is pro-competitive or benign.53 But potentially a more relevant question is what the economics literature says about the conditions under which certain conduct may harm the competitive process and reduce consumer welfare as a result. Perhaps the Supreme Court got it wrong and judicial modesty about a court’s ability correctly to identify anticompetitive conduct is resulting in would-be antitrust defendants walking away unpunished. If this proposition were true, then the FTC would be on firmer ground in pursuing monetary remedies against the defendants it does sue. But there is no empirical support for this proposition. In the realm of potentially exclusionary conduct by a dominant firm—a category that in my opinion includes vertical restraints, tying/bundling, discounting and exclusive dealing, among other forms of conduct—there is a rich theoretical literature modeling conditions under which certain conduct by a monopolist may have a harmful impact on consumer welfare.54 The empirical analysis of these restraints tells a different story. One study, authored by a group of economists from the FTC and the Department of Justice in 2005, concludes that, although ‘some studies find evidence consistent with both proand anticompetitive effects … virtually no studies can claim to have identified

Frank Easterbrook, ‘The Limits of Antitrust’, 63 Texas Law Review 1, 15 (1984). William Kovacic, ‘The Intellectual DNA of Modern U.S. Competition Law for Dominant Firm Conduct: The Chicago–Harvard Double Helix’, 2007 Columbia Business Law Review 1, 35–36 (2007) (‘Chicago School and Harvard School commentators tend to share the view that the social costs of enforcing antitrust rules involving dominant firm conduct too aggressively exceed the costs of enforcing them too weakly… [and that] antitrust rules and decision-making tasks must be administrable for the central participants in the antitrust system (courts, enforcement agencies, the private bar, and business managers)’). 54 See Alden Abbott and Joshua Wright, ‘Antitrust Analysis of Tying Arrangements and Exclusive Dealing’, in Keith Hylton, ed, Antitrust Law and Economics, Edward Elgar, 2010, chapter 8 (highlighting literature); Chiara Fumagalli and Massimo Motta, ‘Exclusive Dealing and Entry, When Buyers Compete’, 96 American Economic Review 785 (2006); Eric Rasmusen, Mark Ramseyer and John Wiley, Jr, ‘Naked Exclusion’, 81 American Economic Review 1137 (1991); Ilya Segal and Michael Whinston, ‘Naked Exclusion: Comment’, 90 American Economic Review 296 (2000). 52 53

The Federal Trade Commission and Monetary Remedies 99 instances where vertical practices were likely to have harmed competition’.55 Another study from 2009 concludes that, ‘it appears that when manufacturers choose to impose restraints, not only do they make themselves better off but they also typically allow consumers to benefit from higher quality products and better service provision’.56 Finally, in a paper considering newer empirical evidence concerning the competitive effects of vertical restraints, FTC economist Dan O’Brien concludes that, ‘with few exceptions, the literature does not support the view that [vertical restraints] are used for anticompetitive reasons’ and that vertical restraints ‘are unlikely to be anticompetitive in most cases’.57 The overall body of evidence supports a fairly strong conclusion that vertical restraints very rarely result in anticompetitive effects and most often benefit consumers. There is no reason to alter the design of legal rules and provide remedies to deter firms from engaging in conduct that rarely if ever harms consumers, even if there were reason to believe that suits challenging such conduct as illegal were decreasing over time. Moreover, in the United States at least, treble damages are available to private litigants in all antitrust actions arising under the two main US antitrust statutes, the Sherman Act and the Clayton Act.58 Economic theory teaches that penalties should be set at a level sufficient to induce offenders to internalize the full social cost of their crimes.59 In a world with imperfect detection and punishment, profit-maximizing market participants will need to face a potential damage award calibrated such that the gains from engaging in the prohibited conduct—the profits that accrue as a result of the anticompetitive behavior—are less than or equal to the expected penalty at the time the firm decides to engage in the challenged conduct. The expected penalty is, of course, a function of the probability of punishment and the magnitude of the penalty. Using this simple model, a damage multiplier can be justified on deterrence grounds if there is reason to believe plaintiffs—including both regulators and private litigants—do not successfully challenge every example of anticompetitive conduct in court. If, every time a firm engaged in anticompetitive conduct, it were sued in court and ordered to pay damages approximating the social harm caused by its conduct, all firms would

55 James Cooper et al, ‘Vertical Antitrust Policy as a Problem of Inference’, 23 International Journal of Industrial Organization 639, 658 (2005). 56 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, chapter 10. 57 Daniel O’Brien, ‘The Antitrust Treatment of Vertical Restraint: Beyond the Possibility Theorems’, in Konkurrensverket, The Pros And Cons of Vertical Restraints, SCA, 2008, chapter 3, 72–73 (2008). 58 15 USC § 15(a). 59 Gary Becker, ‘Crime and Punishment: An Economic Approach’, 76 Journal of Political Economy 169 (1968). See also William Landes, ‘Optimal Sanctions for Antitrust Violations’, 50 University of Chicago Law Review 652 (1983); Douglas Ginsburg and Joshua Wright, ‘Antitrust Sanctions’, 6(2) CPI Journal 3 (2010).

100 Joshua D Wright be properly deterred from engaging in such conduct in the future and there would be no need for a damage multiplier. This simple model hypothesizes that the probability of a lawsuit is 100 per cent and the expected magnitude of liability corresponds exactly to the amount of harm caused by the conduct. If the probability of a lawsuit falls significantly below 100 per cent, then to achieve optimal deterrence, the magnitude of liability ought to be adjusted accordingly.60 In the case of pricefixing cartels and other horizontal conspiracies, we can reasonably expect that regulators and private litigants do not ferret out and challenge every illegal conspiracy that exists because such conspiracies are clandestine by their very nature.61 On the other hand, most examples of potentially harmful single-firm conduct are open and notorious. A downstream distributor of a monopolist’s product will be keenly aware of any restraints on distribution put in place by the monopolist and, to the extent the distributor is harmed by the restraint, will have the appropriate incentive to challenge the conduct. Accordingly, to the extent there are successful private suits challenging vertical restraints by a firm with monopoly power that result in trebled damage awards, those suits are likely to over-deter such conduct, and therefore ought not to be supplemented by government enforcement actions seeking disgorgement.62 Indeed, in the context of single firm conduct, regulators and courts should be primarily concerned with over-deterrence and not under-deterrence. As I mentioned earlier, antitrust liability rules should be designed to minimize all social costs associated with enforcement.63 These costs include false convictions, false acquittals and the costs of administering the system. Not only is there strong support for the notion that the Supreme Court actually uses this framework in designing legal rules,64 economic analysis supports it as well. The reason to be more concerned with the costs of false positives than false negatives is that, if bona fide anticompetitive conduct goes unpunished, then market forces will in time almost certainly erode the monopolist’s

60 Landes, ibid, 657; Frank Easterbrook, ‘Detrebling Antitrust Damages’, 28 Journal of Law and Economics 445, 454 (1985). 61 Ginsburg and Wright, above n 59. 62 Easterbrook, above n 60, 460 (‘When the risk of false condemnation rises, and so the costs of engaging in beneficial conduct go up, the penalty should fall’); Bruce K obayashi and Joshua Wright, ‘Federalism, Substantive Preemption, and Limits on Antitrust: An Application to Patent Holdup’, 5 Journal of Competition Law and Economics 469, 509 (2009) (stating that the case for treble damages applied to ‘open and notorious’ conduct is ‘weak’); Donald Turner, ‘The Durability, Relevance, and Future of American Antitrust Policy’, 75 California Law Review 797, 798 (1987) (explaining that mandatory trebling ‘has adverse effects, not only encouraging baseless or trivial suits brought in hopes of coercing settlements, but also discouraging legitimate competitive behavior in the gray areas covered by the rule of reason’). 63 See Easterbrook, above n 52. See also Cooper et al, above n 55; Geoffrey Manne and Joshua Wright, ‘Innovation and the Limits of Antitrust’, 6 Journal of Competition Law and Economics 153 (2010). 64 See above nn 48–53 and accompanying text.

The Federal Trade Commission and Monetary Remedies 101 arket position. Anticompetitive vertical restraints, like other forms of prim vate monopoly power, are simply not strong enough to stem the tidal wave of creative destruction indefinitely.65 By contrast, improper condemnation of a practice that does not harm consumers will remove that practice from the toolkits of firms operating in other industries and, because the literature has established that vertical restraints usually benefit consumers, will prevent consumers in all areas of the economy from reaping those benefits. The only hopes, then, are to use a different and potentially less efficient and effective practice to achieve the same result, or to convince a court to reverse course.66 To summarize, it is my belief that recent pro-defendant decisions by the Supreme Court fail to justify increased efforts by the Commission to pursue monetary remedies against antitrust defendants for a number of reasons. First, there is no evidence to support the proposition that private antitrust lawsuits are decreasing as the result of recent Supreme Court decisions, which is a necessary but not sufficient condition to render plausible the Commission’s increased efforts seek disgorgement to ‘replace’ the foregone deterrent effect of fewer private lawsuits. Second, the ‘anti-plaintiff’ decisions that former Commissioners Leibowitz and Rosch point to simply reverse a prior trend of decisions by the Supreme Court that were favorable to plaintiffs. Moreover, the correct view is that these recent decisions are motivated by economic learning and an acceptance of the error-cost approach to designing liability rules in antitrust and not out of a concern related specifically to private antitrust suits. These factors together compel the conclusion, contrary to the Commission’s analysis supporting the withdrawal of its Statement, that recent decisions by the Supreme Court provide no basis for the FTC to increase systematically its efforts to pursue monetary remedies.

6. When is it appropriate for an enforcement agency to pursue disgorgement? I hope I have provided enough support to illustrate that the Commission’s reasons for withdrawing the Policy Statement do not pass muster. The next— and perhaps more relevant—issues to consider are the conditions under

65 Joseph Schumpeter, Capitalism, Socialism, and Democracy, 3rd edn, Harper & Brothers, 1950, 83. 66 Convincing the Supreme Court to reverse course is likely to take a very long time. Past experience suggests reversal could take at least 10 years. See United States v Arnold, Schwinn & Co, 388 US 365 (1967), reversed, Continental TV, Inc v GTE Sylvania, 433 US 36 (1977). In other instances it may take as many as 96 years. See Dr Miles Medical Co v John D Park & Sons Co, 220 US 373 (1911), reversed, Leegin Creative Leather Products v PSKS, Inc, 551 US 877 (2007).

102 Joshua D Wright which the agency ought to pursue disgorgement and other monetary remedies in the future. A look at the four cases in which the Commission sought disgorgement in the past may prove helpful. In its 1999 case against Mylan Laboratories, the FTC challenged an arrangement whereby Mylan and three other pharmaceutical companies conspired to make Mylan the monopoly producer of certain pharmaceuticals, which the FTC alleged resulted in $120 million in overcharges.67 In 2001, the FTC reached a consent agreement with Hearst after challenging a consummated merger in which the FTC alleged that the defendant did not timely disclose relevant documents in its pre-merger filings. The agreement required disgorgement of $19 million as well as certain divestitures.68 In Perrigo, the Commission again sought divestiture in a challenge to an agreement between drug manufacturers not to compete, which resulted in a price increase. In that case, the FTC reached a consent agreement in which the two companies agreed to pay a total of $6.25 million in disgorgement.69 Finally, Lundbeck involved an FTC challenge to a consummated acquisition by a pharmaceutical company of the rights to a drug that was allegedly the sole competitor to its own drug. The FTC sought disgorgement because prices had increased substantially. The United States District Court for the District of Minnesota denied the FTC’s request for relief on the ground that the two drugs did not compete in the same product market, and the district court’s decision was affirmed by the court of appeals.70 One principle to draw from these cases is that they all involve agreements or mergers between actual or would-be competitors, and not a single firm acting unilaterally. I would support a limitation on the FTC’s ability to pursue disgorgement only against naked price fixing agreements among competitors or, in the case of single-firm conduct, only if the monopolist’s conduct has no plausible efficiency justification. This latter category would include fraudulent or deceptive conduct, or tortious activity such as burning down a competitor’s plant. Declining to pursue disgorgement in cases involving vertical restraints has the virtue of taking the remedy off the table in cases that present the most difficulty in distinguishing between anticompetitive conduct that harms consumers and pro-competitive conduct that benefits them. Of course, the requirement in the withdrawn Policy Statement that the violation be ‘clear’ also would seem to exclude disgorgement in cases that include plausible efficiency defenses. Because I was not a Commissioner when the Statement was adopted or when it was withdrawn, and was not privy to all the commentary and discussion surrounding the relevant issues, it would be

Mylan Labs, above n 3. The Hearst Trust, above n 7. 69 FTC v Perrigo Co, No 1:04CV01397 (RMC) (DDC 2006). 70 Lundbeck, above n 26. 67 68

The Federal Trade Commission and Monetary Remedies 103 inappropriate for me to take a firm position on whether the Policy Statement itself supplies the correct guidelines for the Commission to follow in pursuing disgorgement in antitrust cases. I will say that the other two elements—that there be a reasonable basis for calculating the monetary relief and that the Commission consider the potential for monetary relief in related litigations— make sense at first blush. I share the same concern Commissioner Ohlhausen expressed in dissenting from the Commission’s Withdrawal of the Statement: the business community is now without any guidance whatsoever regarding the conditions under which the Commission will pursue disgorgement or other monetary remedies in antitrust cases. Disgorgement can certainly be a useful item in the antitrust enforcer’s toolkit—indeed, there may be some support for the proposition that disgorgement can be more practical and have the same deterrent effect as behavioral remedies or divestiture in certain circumstances.71 However, I fear that a lack of guidance from the Commission could cause much mischief. Risk-averse companies concerned about the financial and reputational effects associated with a disgorgement order from the FTC could respond to the lack of guidance by not engaging in conduct that could plausibly benefit consumers. And the threat of disgorgement could lead a company to settle a case even if it has a strong position on the merits.

See Elhauge, ‘Disgorgement’, above n 19, 88–89.

71

104

Ioannis Lianos* The Principle of Effectiveness, Competition Law Remedies and the Limits of Adjudication

1. Introduction The principle of effectiveness is closely related to the development of the emerging EU law on remedies. Its instrumental use has enabled the EU Courts to restrict the principle of national procedural autonomy when this is convenient in order to ensure the accomplishment of the aims set by EU competition law enforcement and to establish EU-granted remedies, the most notable one being the right to claim competition law damages. The principle of effectiveness may also influence the design of injunctive relief by the European Commission, which should be adequate to ensure not only that the results of the violation of competition law are reversed, but also that there is no risk that the aims of the competition law enforced will be jeopardised in the future (general deterrence, specific deterrence and prophylactic/preventive aims).1 The principle of effectiveness may therefore reinforce the remedial discretion of the Commission and national competition authorities when implementing EU competition law. Discretion to adopt effective remedies cannot, however, be unlimited. There are various forces that may operate in order to keep it in check. First, due process rights and fundamental rights jurisprudence, following the conferral of a binding effect to the EU Charter of Fundamental Rights, may balance the need for effective (read: expansive) remedies and therefore limit the risk of discretionary remedialism.2 Second, well-established conceptions

* Professor of Global Competition Law and Policy, UCL Faculty of Laws; Director, Centre for Law, Economics and Society, UCL Faculty of Laws; Chief Researcher, HSESkolkovo, Institute for Law and Development, National Research University, Higher School of Economics, Moscow; Alexander von Humboldt Research Fellow, WZB B erlin. The author would like to acknowledge the support of the Leverhulme Trust. He also thanks Claudio Lombardi and Florian Wagner-von Papp for helpful comments. Any errors or omissions are the sole responsibility of the author. 1 Ioannis Lianos, ‘Competition Law Remedies in Europe: Which Limits for Remedial Discretion?’, in Ioannis Lianos and Damien Geradin, eds, Handbook of EU Competition Law: Enforcement and Procedure, Edward Elgar, 2013, 362 et seq, 379–380. 2 ‘Discretionary remedialism’ is the view that courts and competition authorities have discretion to award the ‘appropriate’ or ‘optimal’ remedy in the circumstances of each

106 Ioannis Lianos in private and public law about the inherent limits of the remedial exercise may also impose constraints on the remedial discretion of the Commission and other competition authorities: the principle of proportionality, for public law, or the correlativity of private law disputes cannot accommodate fully the demands of optimal enforcement theory for deterrence. For instance, remedies should fit the competition law problem identified at the liability stage and the theory of harm the decision-maker has relied upon in order to find the defendant liable for the competition law infringement.3 This chapter explores a further limit to remedial discretion: that of the legitimacy of decision-makers. Legitimacy in this context does not relate only to the fact that public action should be legitimate from a legal perspective, that is, not illegal. Because of its potential for generating expansive remedies, the principle of effectiveness may render this purely legal limit ineffectual. Legitimacy, in this context, refers to the specific legitimacy-building mechanisms elaborated by the specific social order in order to guarantee that its action is politically acceptable by those upon whom a remedy will be imposed. As different social order systems dispose of different mechanisms to guarantee the legitimacy of their action, it is important to guarantee that the use of the powers conferred on each system corresponds to the specific functions exercised by it and do not expand to the realm/functions exercised by other forms of social ordering, which have their own legitimacygeneratingmechanisms. If such a trespass occurs, it may create a legitimacy crisis. Therefore, in the event of a trespass, the decision-maker should emulate the legitimacy-building tools of the social order whose boundaries were crossed in order to guarantee the legitimacy of its action. More concretely, it is argued in this chapter that remedial discretion is an instance of adjudication and, as such, is subject to the inherent limits of adjudication, as a separate form of social ordering from those of contracts/negotiation, managerial/ administrative discretion or legislation.4 If remedial discretion were to move beyond the limits of adjudication and cross over the ‘territory’ of regulation or contractual governance, the decision-maker should adapt by developing legitimacy-building tools that would emulate those used in the context of these other forms of social ordering. The participation in the remedial process of the interests affected constitutes an important source of legitimacy

individual case rather than being limited to specific (perhaps historically determined) remedies for each category of causative events. What is optimal in these cases is decided according to the interests and aims of the decision-maker. 3 For analysis of these limits to discretionary remedialism and the need for remedial iscretion to be compatible with established legal conceptions of remedies, see Lianos, d ‘Which Limits for Remedial Discretion?’, above n 1. 4 Lon Fuller, ‘The Forms and Limits of Adjudication’, 92 H arvard Law Review 353 (1978).

The Principle of Effectiveness 107 for managerial/administrative discretion.5 Consequently, if competition law remedies move closer to regulatory ones and therefore cross the limits of adjudication, they should give rise to increased participatory rights of the interests affected, including interests others than the parties to the dispute. I explore the interaction between the principle of effectiveness and remedial discretion before exploring how the limits of adjudication may be crossed over by the remedial action of competition authorities, in view of the polycentric dimension of competition law disputes. I then criticise in the following section the narrow view of adjudication and advance a distinction between the dispute resolution adjudication model and the structural adjudication model, the latter being closer to the managerial or administrative discretion form of social ordering and thus conducive to resolving polycentric disputes. The final part reflects on the emergence of the ‘consensual’ model of competition law enforcement, with the development of the practice of commitment decisions. I argue that, contrary to what has been thought so far, commitment decisions do not constitute a contractual form of social ordering but present characteristics similar to the structural adjudication model. In addition to being theoretically a more accurate description of the reality of such r emedial practice, the reconceptualisation of commitment decisions as a form of structural adjudication should enable the integration of legitimacybuilding mechanisms emulating those of the managerial or administrative discretion model, such as increased participation of all interests affected by the remedy in the decision-making process.

2. The principle of effectiveness and remedial discretion For a long time, the topic of remedies has been the focus of EU constitutional scholars researching national procedural autonomy and its limitations by the principles of effectiveness and equivalence. In particular, the principle of effectiveness has been expansively interpreted by the European judiciary to encompass all forms of remedial action for infringements of EU law, and was constitutionalised with the adoption of Article 47(1) of the Charter of Fundamental Rights in the EU (Charter) and Article 19(1) TEU.6 This has led to the development of the emerging field of EU remedies law and has had

5 James Henderson, ‘Comment: Settlement Class Actions and The Limits of Adjudication’, 80 Cornell Law Review 1011, 1016 (1995), noting that ‘[e]ach problem-solving process has its own source of legitimation’. 6 According to Article 19(1), remedies should be ‘sufficient to ensure effective legal protection in the fields covered by Union law’. Article 47 of the Charter of Fundamental Rights provides for a right to an effective remedy and effective judicial protection.

108 Ioannis Lianos important implications in the enforcement of EU competition law. Norbert Reich has described the development of the principle of effectiveness in the case law of the ECJ as following three distinctive approaches:7 —— The principle of effectiveness has initially been understood as an ‘ elimination rule’, enabling national courts to put aside or disapply national provisions making the implementation of EU rights ‘practically impossible’ or ‘excessively difficult’ or to interpret the national remedy in conformity with EU law. The jurisprudence of the European Courts requires Member States to establish a system of legal remedies and procedures which ensure respect for the right to effective protection of individuals’ rights under EU law, and hence put in place effective remedies for violations of EU law. When adjudicating the protection of EU rights in private legal disputes, the ECJ has mostly relied on its ‘effectiveness and equivalence’ test to assess national procedures, ‘effectiveness’ referring to the requirement that national procedural rules should not make the enforcement of EU rights impossible or excessively difficult. —— The second step in the development of the principle of effectiveness is its ‘use as a “hermeneutical” ie interpretative principle’. This ascribes some positive content to the principle of effectiveness, requiring Member States to provide for remedies ‘sufficient’ to ensure the protection of EU granted rights. This involves a two-step approach from national courts: first, to spell out the EU right which must be protected by an adequate remedy; second, the development of a minimum core of what constitutes an effective or ‘sufficient’ remedy to ensure the protection of the EUgranted right. For instance, in Manfredi, the ECJ held that: the full effectiveness of Article 81 EC (now Article 101 TFEU) and, in particular, the practical effect of the prohibition laid down in Article 81(1) EC would be put at risk if it were not open to any individual to claim damages for loss caused to him by a contract or by conduct liable to restrict or distort competition … As to the award of damages and the possibility of an award of punitive damages … it is for the domestic legal system of each Member State to set the criteria for determining the extent of the damages, provided that the principles of equivalence and effectiveness are observed … it follows from the principle of effectiveness and the right of any individual to seek compensation for loss caused by a contract or by conduct liable to restrict or distort competition that injured persons must be able to seek compensation not only for actual loss (damnum emergens) but also for loss of profit (lucrum cessans) plus interest.8

7 Norbert Reich, ‘The Principle of Effectiveness and EU Private Law’, in Ulf Bernitz, Xavier Groussot and Felix Schulyok, eds, General Principles of EU Law and European Private Law, Kluwer Law International, 2013, 301 et seq. 8 Joined Cases C-295 to 298/04 Vincenzo Manfredi v Lloyd Adriatico Assicurazioni SpA et al [2006] ECR I-6619 paras 60–62 and 95.

The Principle of Effectiveness 109 As it transpires from the above excerpt, the ECJ first refers to the principle of effectiveness (effet utile) in order to recognise an EU-granted right for any individual harmed by a competition law infringement to claim damages for the loss caused by the infringement (the EU right of compensation), before elaborating on the corresponding remedy for that right, the award of damages, which should at least ensure (the minimum core) that the injured person may receive compensation for the above-mentioned heads of damages. In essence, the procedural dimension, which, in theory, is left to the national level, is subject to the interpretation of the EU-law-granted right and its scope, as this is determined by the ECJ at the EU level. There have been attempts to enlarge this minimum core with additional requirements to be imposed to the national level as to the development of adequate remedies. For instance, in his Opinion in Donau Chemie, Advocate General Jääskinen examined the scope of Article 19(1) TFEU, advancing the view that, in the light of that Treaty provision, the standard of effective judicial protection for EU based rights seems to be more demanding than the classical formula [of the ‘effectiveness’ principle] referring to practical impossibility or excessive difficulty. In my opinion, this means that national remedies must be accessible, prompt, and reasonably cost effective.9

The ECJ did not follow the AG’s approach on this issue but retained its previous definition of the principle of effectiveness. In VEBIC and in Kone, the ECJ employed the principle of effectiveness (or full effectiveness) in order to derive from the interpretation of Article 101 TFEU an obligation for Member States to entitle national competition authorities to participate, as a defendant or respondent, in proceedings before a national court which challenge a decision that the authority itself has taken (in VEBIC) or the right for victims of umbrella pricing to claim compensation for harm suffered by an anticompetitive conduct in case there is a causal link between that conduct and the harm they suffered (in Kone). What is interesting in both cases is the insistence that the aims followed by Article 101 TFEU should not be put ‘at risk’.10 According to the Court, it is clear from the case-law of the Court … that national legislation must ensure that European Union competition law is fully effective … Those rules must therefore specifically take into account the objective pursued by Article 101 TFEU, which

9 Case C-536/11 Bundeswettbewerbsbehörde v Donau Chemie AG and Others, EU:C: 2013:366, para 47. 10 Case C-439/08 Vlaamse federatie van verenigingen von Brood—en Banketbakkers, Ijsbereiders en chocoladebewerkers (VEBIC) VZW [2010] ECR I-12471, para 58 (‘the very existence of such a risk is likely to compromise the exercise of the specific obligation on national competition authorities under the Regulation to ensure the effective application of Articles 101 TFEU and 102 TFEU’); Case C-557/12 Kone AG and Others v ÖBBInfrastruktur AG, EU:C:2014:1317, para 33.

110 Ioannis Lianos aims to guarantee effective and undistorted competition in the internal market, and accordingly, prices set on the basis of free competition.11

The consideration of the objective pursued by Article 101 (and Article 102) TFEU therefore seems to be a passage obligé for the hermeneutical endeavours of the ECJ with regard to the principle of effectiveness. National remedial provisions should be interpreted accordingly so as to avoid any possible risk as to the full realisation of the objectives of the above-mentioned p rovisions of the Treaty. One may note the almost unlimited hermeneutical potential of this construction for intervention by the ECJ in curtailing national procedural autonomy. —— The third level of development of the effectiveness principle follows naturally from the previous one and, according to Reich, consists in the development of ‘hybrid’ EU remedies, the function of which is to ‘upgrade’ national remedies. Reich explains that Article 19(1), sentence 2 TEU establishes a clear link between ‘effective protection’ and ‘sufficient remedies’, thus positing a unidirectional relationship between EU granted rights and national remedies: the recognition of EU rights should lead to the development of ‘sufficient’ national remedies to fully ensure their respective scope of protection. The absence of ‘sufficient’ national remedial tools cannot nevertheless put in question the scope of the EU granted right. National law should be interpreted creatively so as to ensure the minimum core of effective enforcement for the EU granted right. This requires, according to Reich, a three-step approach: (i) ‘the first step is concerned with finding appropriate national remedies in case of violations of EU granted rights’; (ii) the ‘national remedy has to be measured against the yardstick of the negative, eliminatory EU principles of effectiveness and equivalence and if the result is unsatisfactory from an EU law point of view as interpreted as providing “insufficient remedies”, it has to be “upgraded” to meet EU standards’; and (iii) ‘the remedy thus found is a “hybrid” insofar as it takes up elements of national law in the limits of the principle of procedural autonomy, as well as effectiveness in its different functions’.12 Precisely because of this ‘hybrid’ character, the definition of what constitutes an ‘effective’ remedy is a highly contextual issue, which depends on the perceived adequacy of the national remedial tools at the disposal of EU law and of the scope of the rights and corresponding duties generated by the provision of EU law interpreted. For instance, the scope of the secondary right (remedy) to claim damages for infringements of EU competition law, recognised by the EU Courts even if national law is silent on this issue, may

Kone AG, ibid, para 32. Reich, ‘The Principle of Effectiveness’, above n 7, 309.

11 12

The Principle of Effectiveness 111 have a different scope, depending on the interpretation one may have of the primary right preserved by this remedy.13 Although the EU Courts do not explicitly posit the link between this secondary right (remedy) and a primary right, this is essentially required, the first step in their reasoning consisting in interpreting the substantive provisions of the Treaty in competition law (in particular Article 101 TFEU) so as to derive a primary right ‘of any individual to seek compensation for loss caused by a contract or by conduct liable to restrict competition’. By doing so, EU law goes as far as defining the scope of the secondary right, which should presumably reflect the scope of the primary right it aims to preserve: the ability of the injured person to seek compensation for actual loss, loss of profit plus interest. One may advance that the scope of the secondary right would have been different—presumably wider—if it reflected a different primary right, for instance that of deterring future antitrust violations by the specific defendant (specific deterrence) or that of deterring future antitrust violations by any competition law infringer (general deterrence).14 In Manfredi, the Court did not exclude such an interpretation of Article 101 TFEU of the Treaty, recognising that national legal systems can provide for the possibility to award punitive damages.15 Yet it did not interpret Article 101 TFEU as requiring the inclusion of punitive damages in the minimum core of the ‘hybrid’ EU remedy of claiming damages for competition law infringements, as this did not relate to the scope of the primary right recognised by the Court (that of receiving compensation).16 In other words, the main function of the principle of effectiveness is to maximise the ‘sufficient’ attainment of the ends pursued by the primary right by providing the ‘adequate’ content to the secondary right of claiming antitrust damages, and not to create the primary right. The largely hermeneutic function of the principle of effectiveness has not escaped the attention of legal commentators. Some, for instance, argue that the principle has been used without much consistency, as a ‘kind of

13 On the distinction between primary and secondary (remedial) rights in the context of competition law enforcement, see Ioannis Lianos, ‘Competition Law Remedies: in Search of a Theory’, in Ioannis Lianos and D Daniel Sokol, eds, The Global Limits of Competition Law, Stanford University Press, 2012, 177 et seq; Lianos, ‘Which Limits for Remedial Discretion?’, above n 1. 14 The ECJ recognises the deterrent effect of the right to claim damages, but this is not perceived independently from the primary right to claim compensation. See, for instance, Kone AG, above n 10, para 23 (‘[t]he right of any individual to claim compensation for such a loss actually strengthens the working of the European Union competition rules, since it discourages agreements or practices, frequently covert, which are liable to restrict or distort competition, thereby making a significant contribution to the maintenance of effective competition in the European Union’). 15 Manfredi, above n 8, paras 92–93. 16 With regard to punitive damages, the ECJ relied in this case on the principle of equivalence rather than the principle of effectiveness. See ibid, para 99.

112 Ioannis Lianos jack-in-the-box instrument’ allowing the ECJ to justify almost every result.17 The ‘principle’ of effectiveness was derided as being essentially ‘an empirical concept’, requiring the EU Courts to proceed to a comparative examination of the national remedies available and also providing the authorities with the discretion to establish a low or a high level of enforcement, the principle being an instrument to achieve undefined policy objectives.18 Although there is some grain of truth to this characterisation of the ‘principle of effectiveness’ and the criticisms over an inconsistent application of that principle may be spot on, the fact remains that the principle does not have a substantive content of its own, but reflects the substantive (primary) right to which it is associated. The principle of effectiveness cannot be seen functioning outside the context of the adjudication of the primary right violated by the competition law infringer. Claims for damages are usually perceived as an archetypical private law dispute, by which I mean a dispute between two parties whose rights and duties are correlative to each other. For instance, the infringer’s wrong consists in the violation of the primary right of the victim. The secondary right of the victim for compensation correlates to the wrong committed by the infringer. Yet not all competition law disputes fit this mould. Some remedies may have a considerable impact on economic actors other than the parties to the dispute, in a direct or an indirect way. Damages claims may also emanate from complex causation chains. One may think of claims introduced by indirect purchasers, counterfactual customers or umbrella customers. It may not always be easy to interpret the substantive provisions of EU competition law as giving rise to specific primary rights, like that of compensation for losses incurred because of the competition law violation, giving rise to secondary rights, such as the right to claim damages or obtain injunctive relief. It may, for instance, be quite complex to transpose the infringement of Article 102 TFEU by means of a loyalty rebate and a corresponding remedy of mandatory injunction into (primary or secondary) ‘rights’ or ‘duties’ talk. How would the effectiveness of the remedy be judged in this context? Should the primary right in this case be that of the equal opportunity of the dominant firm’s competitors to access the market? If so, what size of market would be considered ‘sufficient’ to preserve the essence of that ‘primary’ right? Would the primary right be that of the consumers of the dominant undertaking who may be charged higher prices in the future, should the competitors be excluded from the market? In which case, the primary right would need to reflect the economic theory

17 Hans-Wolfgang Micklitz, ‘The ECJ between the Individual Citizen and the Member States—A Plea for a Judge Made European Law on Remedies’, in Hans-Wolfgang Micklitz and Bruno de Witte, eds, The ECJ and the Autonomy of Member States, Intersentia, 2012, 349 et seq, 397. 18 See ibid.

The Principle of Effectiveness 113 of consumer harm advanced—a quite difficult task. Or is the primary right that of the ‘public at large’, which has been deprived of a moral right to the protection of ‘competition as such’?19 This will inevitably raise the issue of the delimitation of this primary moral right. The conceptual plasticity of the substantive provisions of EU competition law may lead, if combined with an instrumental view of the principle of effectiveness, to a quite broad remedial discretion for decision-makers. Similarly, the secondary right (remedy) may be interpreted broadly so as to enable the adoption of competition law remedies that have a prophylactic (preventive) aim.20 These remedies seek to ensure that there remain no practices likely to result in distortions of competition and infringements in the future. Prophylactic remedies may be distinguished from specific deterrence as they affect the ability (and not the incentive) of the infringers to commit equivalent anticompetitive practices in the future by focusing on specific facilitators of potential infringements. These are not illegal practices in themselves but, in the specific circumstances of the case, they may facilitate illegal conduct. By prohibiting these practices, the decision-maker’s objective is not to deter the potential infringers from adopting such conduct, as this is not illegal, but to reduce their ability to commit illegal practices. In addition, in an economically oriented competition law, the definition of what is an ‘optimal’, ‘sufficient’ or ‘appropriate’ remedy may also be influenced by the view economists take on optimal deterrence (the optimal deterrence model) and on how the market equilibrium existing prior to the competition wrong may be improved by subsequent remedial action. The remedy may thus offer the opportunity to design a new market equilibrium, more competitive than the one following the infringement but also, in some circumstances, more competitive than the one existing prior to the infringement in order to m aximise welfare. Linking the principle of effectiveness with the concept of efficiency, perceived as wealth maximisation, may nevertheless open a Pandora’s box with regard to the remedial discretion enjoyed by competition authorities and the courts. These may adopt far-reaching remedies which would not correlate with the wrong committed and thus the scope of the primary right that was violated. Yet ‘discretionary remedialism’ cannot be the rule of the game.21 This is clear in view of the importance the principle of proportionality has in

19 See, eg Case C-8/08 T-Mobile Netherlands BV and Others [2009] ECR I-4529, para 38. In this case the Court of Justice accepted that Article 101, ‘like the other competition rules of the Treaty, 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 c ompetition as such’. 20 On prophylactic remedies in competition law, see Lianos, ‘Which Limits for Remedial Discretion?’, above n 1. 21 See Lianos, ‘Competition Law Remedies’, above n 13.

114 Ioannis Lianos elimiting the remedial discretion of competition authorities and courts in d EU competition law, if one takes a public law perspective on remedies. The principle of proportionality requires that measures adopted by Community institutions do not exceed the limits of what is appropriate and necessary in order to attain the legitimate objectives pursued by the legislation in question; when there is a choice between several appropriate measures, recourse must be had to the least onerous one, and the disadvantages caused must not be disproportionate to the aims pursued. This trade-off exercise involves, in addition to considering whether the means chosen are indeed a rational means to a purported end (step 1 of the test), some assessment of the possible excessive costs of the remedial action in relation to its benefits (step 2) and whether the means chosen are the least restrictive alternative (step 3). The first step of the proportionality principle is of particular interest for our purposes. The importance of remedial fit is often stressed by antitrust law literature.22 It is also indirectly linked with the existence of a causal relationship between the undertaking’s conduct and the theory of harm advanced. The scope of the obligations imposed on the undertakings concerned in order to bring the infringements to an end identified should be implemented according to the nature of the infringement declared, and the obligations imposed ‘must not exceed what is appropriate and necessary to attain the objective sought, namely re-establishment of compliance with the rules infringed’.23 This relates to the obligation of the Commission to give the undertakings concerned the opportunity of being heard on the matters to which it has taken objection. For example, the Commission is not entitled to impose a fine on an undertaking without having previously informed it in the statement of objections that it intended to do so, a requirement which also applies, by analogy, to remedies.24 The existence of a competition law violation (even if there is no explicit finding of an infringement) ‘constitutes the basis of the obligation of the parties to terminate the infringement’, hence the reason for imposing remedies.25 However, while the principle of proportionality requires a close fit between the harm and the remedy, determining the nature of that fit (functional, instrumental) remains an open question. The flexibility of the concept of the theory of harm, in particular its linkage to economic theory, enables the See Lianos, ‘Which Limits for Remedial Discretion?’, above n 1, with references. Case T-76/89 Independent Television Publications Ltd v Commission [1991] ECR II-575, para 93; Case C-279/95 P Langnese-Iglo v Commission [1998] ECR I-5609, para 74. 24 Case T-395/94 Atlantic Container Line AB [2002] ECR II-875, para 417; Case 322/81 NV Nederlandsche Banden Industrie Michelin v Commission [1983] ECR 3461, para 20. 25 Joined Cases 56 and 58/64 Établissements Consten SARL and Grundig-VerkaufsGmbH v Commission [1966] ECR 299, 338; Joined Cases T-125 and 127/97 The Coca-Cola Company [2000] ECR II-1733, para 80 (observing that ‘where an undertaking is in a dominant position it is obliged, where appropriate, to modify its conduct accordingly so as not to impair effective competition on the market regardless of whether the Commission has adopted a decision to that effect’). 22 23

The Principle of Effectiveness 115 ommission or the competition authorities to enjoy a wide remedial discreC tion, in particular when adopting prophylactic remedies. While in phase with the principle of effectiveness, prophylactic remedies have nevertheless been considered in certain circumstances as extending the scope of the remedy unreasonably, with regard to the scope of the p rimary right violated and thus the wrong committed. For instance, in a number of cases, the Commission required the undertakings concerned under its infringement decision to refrain in the future from any conduct which may have the same or a similar effect to those covered by the infringement decision, with the aim of preventing the undertakings from repeating the behaviour found to be unlawful.26 In Cartonboard, the Commission prohibited any future exchange of commercial information by which the participants directly or indirectly obtained commercial information on competitors, even if this was not by its nature unlawful under Article 101 TFEU as the information related also to certain aggregated statistical data.27 The General Court found that such prohibition exceeded what was necessary in order to bring the conduct in question into line with what was lawful because it was seeking to prevent the exchange of purely statistical information which was not in, or capable of being put into, the form of individual information which could thus be used for anticompetitive purposes. Indeed, the Commission had not considered the exchange of statistical data to be in itself an infringement of Article 101(1) TFEU. According to the Court, the mere fact that a system for the exchange of statistical information might be used for anti-competitive purposes does not make it contrary to Article [101(1) TFEU], since in such circumstances it is necessary to establish its actual anti-competitive effect.28

In Atlantic Container, the ECJ annulled part of the Commission’s d ecision for having imposed on the parties to the infringement an obligation to renegotiate or terminate the service contracts concluded between the s hippers and the maritime conference, which were not found to be illegal under Article 101 TFEU (as only the provisions of the TAA relating to price-fixing and capacity were found by the Commission to infringe this provision).29

26 Case T-310/94 Gruber + Weber GmbH & Co KG v Commission [1998] ECR II-1043, para 167; Case T-83/91 Tetra Pak International SA v Commission [1994] ECR II-755, para 220 (‘its aim is to put an end to all the practices found unlawful in the Decision and to preclude any similar practice’). 27 Gruber & Weber, ibid, para 177. 28 Ibid, para 178. 29 Atlantic Container Line, above n 24, para 398 et seq. Service contracts are agreements under which the shipper undertakes to ship a minimum volume or value of cargo during the period of the contract and, in return, the carrier undertakes to provide to the shipper specific service guarantees, such as a capacity guarantee, and negotiates a price lower than that which is normally applicable.

116 Ioannis Lianos The Commission had adopted this requirement of renegotiation or termination as a prophylactic remedy in order to prevent the members of a cartel from continuing to apply unlawfully fixed prices simply because these prices were incorporated in long-term contracts. In other words, the requirement to renegotiate or terminate the service contracts was justified by the fact that the effects of the infringements identified in the decision would have continued to exist if the addressees of that decision were able to continue to enjoy the economic advantages secured by ongoing contracts entered into on the basis of the horizontal agreement to fix prices and limit supply found illegal by the Commission.30 As this section of the decision formed part of the order bringing the infringement to an end, the Commission alleged that there was no need to give specific reasons or to draw it to the attention of the parties concerned in the statement of objections. The Court did not agree with this view, noting that the likelihood of private actions for damages should be a sufficient disincentive for the defendants to continue behaviour that would have maintained or facilitated the effects of the core infringement to Article 101 TFEU, and in any case, the Commission had the obligation to ‘explain its reasoning’ as to how the prophylactic measures suggested were ‘obviously necessary’ to bring the main infringement to an end—something that the Commission’s decision had not done.31 The Court even observed that, in view of the rights of defence of the defendants and the requirement that they should be offered a proper opportunity to make known their view,32 the statement of objections should in any event have set out, even briefly, but in sufficiently clear terms, the measures which the Commission intended to take in order to bring an end to the infringements and should have given the applicants all the information necessary in order to enable them properly to defend themselves before the Commission adopted a final decision on that point.

The instrumental use of the principle of effectiveness in order to expand the remedial discretion of the competition authorities is closely interrelated to the question of the legitimacy of competition authorities and the judiciary in competition law cases to adopt far-reaching remedies that may restrict individual rights.33 Both public and private lawyers recognise the existence of inherent limits to the remedial action of competition authorities and the judiciary, even if they may take different perspectives in this context.

Ibid, para 406. Ibid, paras 414–415. 32 Ibid, paras 418–419. 33 Arianna Andreangeli, ‘Between Economic Freedom and Effective Competition Enforcement: The Impact of the Antitrust Remedies Provided by the Modernisation Regulation on Investigated Parties Freedom to Contract and to Enjoy Property’, 6(2) Competition Law Review 225 (2010). 30 31

The Principle of Effectiveness 117 The private law account of remedial action emphasises correlativity as an essential feature of private law disputes.34 Correlativity perceives the parties’ relationship as a ‘bipolar unit in which each party’s normative position is intelligible only in the light of the other’s’.35 The duty of one party is the mirror image of the other party’s right. The structure of the remedy should thus ‘reflect the structure of the injustice, retracing and reversing the movement between the parties’. Courts may effectively exercise their adjudicatory function to determine, based on the evidence heard on the structure of the pre-existing relationship between the claimant and the defendant, the appropriate relational structure post-infringement, with the understanding that this should be equivalent to that prior to the infringement. The court’s or other decision-maker’s discretion is thus bound in the context of private law resolution of disputes by the fact that remedies should not go beyond the structure of the pre-existing relationship between the correlative entitlements and obligations of the parties to the dispute. An important difference between the private and public law accounts of the remedial process is that, while the former is based on the idea of correlativity, the latter is by essence polycentric, because of the variety of interests to be considered by the public authority charged with the protection of the general interest—by essence, a polycentric concept. In a public law context, because the judge or authority is seeking to implement generally articulated, aspirational norms in highly differentiated contexts, liability norms do not dictate the content of the remedy. Liability norms only provide the goals and boundaries for the remedial decision. In this context, the linkage between remedies and rights (or wrongs) is instrumental, as the liability stage indirectly constrains the targets of the remedial process, the aim of which is to provide a solution to a specifically determined (at the liability stage) problem.36 The normative parameters that have been set at the liability stage in the form of problems that the remedy must address operate simultaneously as a guide and as a constraint to the exercise of remedial discretion in a public law context. Some trade-off devices are thus developed in order to mediate between the different interests represented in these polycentric disputes, the principle of proportionality being one of them. It follows from the above that, despite the instrumental use of the principle of effectiveness in order to elaborate a minimum core of ‘hybrid’ EU/national remedies with the aim of ensuring the effective enforcement of competition law, legitimacy concerns may limit the remedial action of competition authorities and courts. The following section attempts to address the question

Earnest Weinrib, Corrective Justice, Oxford University Press, 2012, 2. Ibid. 36 See Lianos, ‘Which Limits for Remedial Discretion?’, above n 1. 34 35

118 Ioannis Lianos of legitimacy in the context of remedial action, by exploring the hypothesis that this is related to the limits of adjudication, as a form of social ordering distinct from, for instance, regulation or contract.

3. Legitimate remedies and the limits of adjudication In exploring the limits of adjudication, with the aim of setting clear limits to the legitimacy of judicial action, Fuller referred to the principle of polycentricity, which he defined as ‘situation(s) of interacting points of influence’ which ‘involve many affected parties and a somewhat fluid state of affairs’.37 Intervention in this context may have ‘complex repercussions’.38 Fuller observes that one may: visualize this kind of situation by thinking of a spider’s web. A pull on one strand will distribute tensions after a complicated pattern throughout the web as a whole … This is a ‘polycentric’ situation because it is ‘many centered’—each crossing of strands is a distinct center for distributing tensions.39

According to Fuller, the adjudication of polycentric disputes is problematic because the complexity of the dispute and the range of those affected, which it is sometimes difficult to foresee, render it quite difficult to organise the participation of all affected parties in the dispute so that they may represent their various positions. Informed only by the litigating parties and confined by limited competences, the decision-maker is ill-equipped to determine the impact of the decision reached on the different interests affected, with the consequence that the decision reached may negatively affect societal welfare. What becomes important is thus the ‘question of knowing when the polycentric elements have become so significant and predominant that the proper limits of adjudication have been reached’.40 Problems that are sufficiently polycentric may be unsuited for adjudication and may be resolved through managerial direction, through negotiation and contract,41 or left to the forces of the market. Some problems, such as the allocation of economic resources, may indeed ‘present too strong a polycentric aspect to be suitable for adjudication’.42 Fuller explains what happens when an attempt is made to deal by adjudicative forms with a problem that is essentially polycentric: three things can happen, sometimes all at once. First, the adjudicative solution may fail. Unexpected repercussions make the decision unworkable; it is ignored,

Fuller, ‘The Forms and Limits’, above n 4, 395 and 397. Ibid, 395. 39 Ibid, 395. 40 Ibid, 398. 41 Ibid. 42 Ibid, 400. 37 38

The Principle of Effectiveness 119 ithdrawn, or modified, sometimes repeatedly. Second, the purported arbiter ignores w judicial proprieties—he ‘tries out’ various solutions in post-hearing conferences, consults parties not represented at the hearings, guesses at facts not proved and not properly matters for anything like judicial notice. Third, instead of accommodating his procedures to the nature of the problem he confronts, he may reformulate the problem so as to make it amenable to solution through adjudicative procedures.43

This does not necessarily mean that contracts and managerial administrative discretion do not have problems of their own when dealing with polycentric disputes.44 Contracts are generally ill-suited to inequalities of bargaining power, and managerial and administrative discretion may raise important problems of unlimited discretion. However, these methods may integrate more fully the under-represented interests to the adjudicatory procedure and take into account evidence and arguments emanating from third parties rather than the initial parties to the dispute. Fuller’s reference to polycentric disputes is highly relevant for our discussion of competition law remedies, the remedial process constituting an instance of adjudication. The enforcement of competition law requires the balancing of different interests, even though recourse to balancing (or the rule of reason as it is called in US antitrust law) is in some circumstances limited by categorical thinking and the establishment of presumptions ( per se categories in US antitrust law or object restrictions in EU competition law). In fine, the formation of specific categories, such as specific types of abuse, subject to particular standards or tests (ie margin squeeze, refusal to deal), depends on some pre-balancing of the interests usually affected, together with considerations of procedural economy and error cost analysis.45 Hence, categorisation may be considered as a form of dynamic balancing of principles and policies with the aim of promoting the effectiveness of competition law enforcement: in this case, the reduction of administrative costs of enforcement while avoiding the risk of substantive errors. An illustration of this dynamic approach in establishing legal categories constitutes the evolution of the object box in EU competition law, which has been expanded to cover certain forms of information exchange,46 the prohibition of internet sales in the context of a vertical agreement,47 or even vertical practices that are particularly suspicious in view

Ibid, 401. JWF Allison, ‘Fuller’s Analysis of Polycentric Disputes and the Limits of Adjudication’, 53 Cambridge Law Journal 367 (1994). 45 See Arndt Christiansen and Wolfgang Kerber, ‘Competition Policy with Optimally Differentiated Rules’ Instead of “Per se Rules vs Rule of Reason”’, 2 Journal of Competition Law and Economics 215 (2006); Ioannis Lianos, ‘Categorical Thinking in Competition Law and the “Effects-based” Approach in Article 82 EC’, in Ariel Ezrachi, ed, Article 82 EC—Reflections on its Recent Evolution, Hart Publishing, 2012, chapter 2. 46 T-Mobile Netherlands, above n 19. 47 Case C-439/09 Pierre Fabre Dermo-Cosmétique SAS v Président de l’Autorité de la concurrence [2011] ECR I-9419. 43 44

120 Ioannis Lianos of the structure of the specific market,48 the latter being considered by some as an excursion outside what has been considered so far as the traditional domain of the object box, which does not require the consideration of information relating to the structure of the relevant market affected. The move towards a more effects-based approach has accentuated the polycentric dimension of competition law disputes. Final and intermediate consumers active in the specific relevant market were added to the list of the participants affected by the adjudicated transaction, their interest(s) being given the most important weight in the decision-making process (because of the emphasis on consumer harm). Because of the variety of interests that must be taken into account, the number of affected participants and the complexity of repercussions are particularly noteworthy in competition law disputes involving market leaders controlling industry standards, thus affecting an array of interrelated economic sectors that are not necessarily directly linked to the relevant market of the specific competition law dispute. For instance, cases like Google and Microsoft have profound implications on innovation and the development of competition in the broader IT sector globally. A similar conclusion may be reached with regard to competition law disputes involving key inputs to economic production, ie in energy and telecommunications, which may have an important impact on the economy overall. The remedies adopted in the context of such disputes have often been characterised as an instance of ‘regulatory antitrust’, the perception being that competition law has moved away from its archetype of adjudication towards a different form of social ordering, such as regulation.49 There seems to be a fine line between competition law remedies and regulatory remedies, which in some cases EU competition law has seemingly transgressed, thus leading to the accusation of ‘regulatory antitrust’, the latter perception being perceived as an oxymoron. For instance, Ibañez Colomo argues that this transgression may take different forms: first, competition may become regulatory in nature ‘when its application on a proscriptive basis (rather than prescriptive basis) would not be possible given the features of the relevant market’.50 Second, the

Case C-32/11 Allianz Hungária Biztosító Zrt and Others, EU:C:2013:160. Giorgio Monti, ‘Managing the Intersection of Utilities Regulation and EC Competition Law’, 4(2) Competition Law Review 123 (2008), noting that competition law has been applied, ex ante, in a regulatory fashion in the utilities sector and arguing that competition law should give way to regulation, in certain circumstances, ‘either because suspending competition law can yield greater efficiency or because it can allow the pursuit of other non-economic policy objectives’; Pablo Ibañez Colomo, ‘On the Application of Competition Law as Regulation: Elements for a Theory’, 29 Yearbook of European Law 261 (2010). 50 Ibañez Colomo, ‘Elements for a Theory’, ibid, 264. 48 49

The Principle of Effectiveness 121 ‘expected standards of intervention in a competition law case can be defined as the composite of (i) the gravity of the infringement identified and (ii) the remedies (if any) required to bring an end to it’, the relationship between the two being presumed to be a ‘linear one, that is, the intrusiveness of a given remedy increases in direct proportion with the gravity of the infringement’.51 However, when competition law is ‘instrumentalised’, and this expression is left without definition by the author, ‘the remedies imposed in a given case may exceed what would be necessary to bring an end to the infringement identified by the authority’.52 The correlative relationship between remedies and rights/wrongs is thus perceived as the defining boundary between competition law remedies and regulatory remedies. The two ‘do not follow the same logic’, as the former are generally concerned ‘with preserving the existing market structure from being further deteriorated’, thus leading to remedies that reflect the wrong committed, and not with redesigning the current market structure, by ‘sharing (or neutralising competitive advantages)’, which may give rise to remedies that go beyond the wrong committed.53 In an economically oriented competition law, this boundary is more easily described than practised. Theories of consumer harm may not only relate to the structure of the supply side but may also be generated by the specific characteristics of the demand side. Behavioural economics may provide insights into how some behavioural biases of consumers may be exploited by incumbents in order to raise prices. If the practices of the incumbents are caught by competition law, the remedy will need to address these behavioural biases in order to be effective. However, providing for remedies dealing with existing behavioural biases will not just restore the competitive process, but will generate a very different one than the one prior to the identified ‘competition law’ infringement. Would that remedy be considered as having a regulatory nature and hence as being outside the normal scope of competition law remedies, given that it aims to shape the conditions of competition in the market? Yet the polycentric character of complex competition law disputes is much more limited than that of pure regulatory law disputes. First, competition law remedies54 relate to the exercise of an adjudicative function, either of a competition authority or of a court, defined as the adversarial presentation

Ibid, 277 (emphasis in original). Ibid, 279. Ibid, 283. 54 With the notable exception of remedies imposed in the context of market investigation references in the UK, following Part 5 of the Enterprise Act 2002. In the EU context, sector inquiries do not carry the possibility for the Commission to impose remedies, but may instead lead to the initiation of competition law proceedings under Articles 101 and 102 TFEU. Consequently, the mandate of the Commission in exercising its competition law competence is exclusively adjudicatory. 51 52 53

122 Ioannis Lianos of evidence and reasoned arguments to a decision-maker,55 and should thus be distinguished from remedies adopted in the context of a rule-making activity, as is often the case in regulation. Second, even when regulators exercise an adjudicative function in enforcing regulation, the polycentric nature of the regulatory dispute is more pronounced than in the context of competition law, the decision explicitly taking into account the economic conditions of an entire sector of activities, rather than the competitive conditions of a specific relevant market on which the parties to the dispute are active—by definition, a narrower exercise. Third, the interests that are usually considered in a regulatory dispute are in principle more diverse than those taken into account in competition law disputes, the regulators assuming various responsibilities, such as the protection of the environment, universal service and security of supply, while the competition authorities’ role is primarily confined to the protection of the competitive process. As a result of the variety of regulatory goals, there is more extensive participation in the decision-making process of actors representing diverse interests, often not directly related to the dispute. Focusing, for illustration purposes, on merger control, which is the closest a competition law procedure can come to a regulatory law one, the EU Merger Implementing Regulation provides for the participation in the process of ‘third parties’, a category which is narrowly defined as including those having a ‘sufficient interest’ in the Commission’s procedure, such as customers, suppliers, competitors, members of the administration or management organs of the undertakings concerned, or recognised workers’ representatives of those undertakings.56 Certainly, the Commission has appointed a consumer liaison officer, and might also invite the views of other interested third parties, including consumer organisations,57 but these parties do not have a right to be heard in the absence of an explicit invitation by the Commission. In any case, the third parties are expected to comment only on the competition implications of the merger, rather than on broader issues, such as the protection of employment and the environment.58 This contrasts with the wide participation of various interests in the context of regulatory decision-making, often with the involvement of intermediary, although not elected, bodies representing a supposed public interest, and, less frequently, by direct intervention from the interested public. Consequently, despite the polycentric dimensions of most competition law decisions,59 remedies are precisely confined to address the

See Allison, ‘Fuller’s Analysis of Polycentric Disputes’, above n 44. On the role of third parties, see Article 18(4) of Regulation 139/2004 and Articles 16(1) and (2) of the Implementing Regulation. 57 Article 16(3) of the Implementing Regulation. 58 See DG Competition, ‘Best Practices on the Conduct of EC Merger Control Proceedings’, para 37, referring only to ‘competition concerns’. 59 Decisions are polycentric in that they affect not only the parties to the dispute but also other competitors, consumers, customers, shareholders and employees. 55 56

The Principle of Effectiveness 123 specific situation under adjudication. Indeed, the boundaries of remedial discretion are delimited by the interplay of the specific goals entrusted to competition authorities, the principle of remedial proportionality and the control of legality for misuse of powers.60 One may, however, criticise Fuller’s narrow view of adjudication as essentially involving adversarial presentation of evidence and arguments between the parties to the litigation. First, ‘concealed polycentric elements’ exist in all problems solved by adjudication.61 Second, dispute resolution does not constitute the only model of adjudication. Commenting on the development of constitutional adjudication, Fiss argued that a new model of adjudication, which he called ‘structural reform’, has emerged that challenges the narrow conception of the function of adjudication.62 While the dispute resolution model is a ‘sociologically impoverished universe’, ‘one that does not account for social groups and bureaucratic institutions’, its world being composed exclusively of individuals (the litigants), the ‘structural’ model is characterised by a multiplicity of parties and an array of competing interests and perspectives.63 Although dispute resolution ‘implies a unity of functions in party structure’, ie ‘the plaintiff is simultaneously the victim, the beneficiary of the remedy, and also the spokesperson’, the structural model is characterised by a fragmentation of the roles precisely because of the introduction of sociological entities, such as, in the context of competition law disputes, competitors or consumers. For instance, the possibility for the Commission, acting on its own initiative pursuant to Article 15(3) of Regulation 1/2003, to submit written observations (amicus curiae observations) to courts of the Member States where the coherent application of Article 101 or 102 TFEU so requires provides the opportunity to the national courts to hear evidence and arguments that may not have been proffered by the litigants and which relate to the broader EU interest to ensure a coherent enforcement of EU competition law throughout the Union. The function of remedies is also different in the context of the ‘structural’ model. Its main role consists in eliminating threats to the values (effective competition) protected by the law (a prophylactic aim), eventually restructuring the organisation that committed the infringement—‘a complex and difficult task wholly alien to the dispute resolution model’.64 Consequently, while the remedial process in the dispute resolution model is based on the assumption that the aim of adjudication is to assess the ‘abnormal’ event that has caused the dispute and to restore the 60 This ground of judicial review refers to cases where an authority uses its powers to take a measure with a purpose other than that for which the powers in question were conferred upon it. 61 Allison, ‘Fuller’s Analysis of Polycentric Disputes’, above n 44, 371. 62 Owen Fiss, ‘The Social and Political Foundations of Adjudication’, 6 Law and Human Behavior 121 (1982). 63 Ibid, 123. 64 Ibid.

124 Ioannis Lianos arties to their rightful position, that is, the position that they would have p occupied absent that specific abnormal event (in other words, the violation of EU competition law), the ‘structural’ model ‘reflects doubt on whether the status quo is efficient’, eventually promoting the establishment of a new status quo.65 One should therefore come to grips with the fact that competition law enforcement constitutes an instance of ‘structural’ adjudication, and not of the ‘dispute resolution’ model. This renders the distinction with the administrative/managerial model particularly difficult at times, at a practical level, although there is some value in thinking of the two (adjudication and administrative managerial model) as ideal types forming a continuum with regard to the ‘appropriate’ degree of discretion and consequently the legitimacy of the action of the institutions in charge, which is closely related to the participation of the interests affected. Put simply, the more EU competition law moves towards the regulatory/managerial model, and ‘structural’ adjudication comes close to that, the more it should integrate the legitimacy-building mechanisms of that model, with the enhanced participation of the entities subject to the remedies as well as of all those whose interests may be affected (ie consumers, competitors in related markets and interests vicariously represented by organisations and citizen’s groups, ie environmental associations). The triadic model of dispute resolution, limited to the parties and the adjudicator, needs to give way in circumstances of significantly polycentric disputes to the more inclusive model of ‘structural’ adjudication in order to preserve the legitimacy of competition law enforcement and its continued relevance and appeal to society at large.

4. Remedial discretion and commitment decisions: exploring the limits of adjudication The delimitation of the fine boundary between adjudication and the managerial or administrative model is not the only challenge to which mature competition law enforcement systems are confronted in devising effective remedies. In recent years, we have witnessed a significant displacement of the adjudicative EU competition law enforcement model by a presumably ‘consensual’ model of competition law enforcement, with the increasing use of commitment decisions under Article 9 of Regulation 1/2003.66 From 2004 Ibid, 124. Niamh Dunne, ‘Commitment Decisions in EU Competition Law’, 10 Journal of Competition Law and Economics 399 (2014). But see Wouter Wils, ‘Ten Years of Commitment Decisions Under Article 9 of Regulation 1/2003: Too Much of a Good Thing?’ (2015), ssrn.com/abstract=2617580, p 9 (‘[I]t cannot be concluded from the statistics that there is or has been an increasing trend to use commitment decisions.’). 65 66

The Principle of Effectiveness 125 to 2014 there were 35 commitment decisions, as opposed to only 21 infringement non-cartel decisions adopted under Article 7 of Regulation 1/2003. National competition authorities have also turned with increasing zeal to a sustained use of commitment decisions in order to resolve competition law disputes.67 Theoretically justified by the need to ensure an effective use of the scarce resources of competition authorities and procedural economy benefits,68 the use of the procedure of commitment decisions constitutes in reality another illustration of the limits of the classic model of adjudication in competition law enforcement, and the need to adopt procedures that maximise the aims of competition law enforcers, while keeping an eye on the legitimacy of their action. One may validly ask if commitment decisions fit the adjudicative model or may be considered as closer to the model of negotiation and contract or that of the regulatory model, which form distinct categories of social ordering if we follow Fuller’s perspective. One could explore the similarities and differences between commitments decisions in antitrust and in merger control in order to characterise such decisions as fitting in one of the models of decision-making advanced by Fuller. In merger control, remedies take the form of commitments offered by the parties to the merger, either at Phase I or Phase II of the merger procedure, which are eventually accepted by the Commission if they address its ‘serious doubts’ over the legality of the merger or the ‘competition concerns’ identified. This leads to the publication of a decision under Article 6(2) or Article 8(2) of the EC Merger Control Regulation, which makes binding the commitments offered by the parties. In the context of the ex post competition law enforcement of Articles 101 and 102 TFEU, Article 9 of R egulation 1/2003 enables the Commission to make commitments offered by parties to its proceedings binding on them, instead of issuing a regular prohibition decision, when those commitments address the concerns expressed in the Commission’s preliminary assessment. Such a decision may be adopted for a specified period and shall conclude that there are no longer grounds for action by the Commission. Technically, commitment decisions offered under Article 9 of

67 For an excellent analysis of national practices, see Heike Schweitzer, ‘Commitment Decisions in the EU and in the Member States: Functions and Risks of a New Instrument of Competition Law Enforcement within a Federal Enforcement Regime’, e-Competitions Bulletin, Special Issue on Commitment Decisions, 2 August 2012, http://ssrn.com/ abstract=2101630. 68 There has, however, been some doubt over this justification for the increasing use of commitment decisions. See Mario Mariniello, ‘Commitments or Prohibition? The EU Antitrust Dilemma’, Bruegel Policy Brief, Issue 2014/01 (January 2014). One may also further expect that negotiations between Commission officials and the legal/economic teams of global giants such as Microsoft, Google, Coca Cola, to cite randomly some major corporations having offered commitments to the Commission, are particularly time consuming and human resource-devouring.

126 Ioannis Lianos Regulation 1/2003 are not remedies as they do not aim to put the infringement to an end, as is the case for Article 7 of Regulation 1/2003 and Phase II merger control decisions, and do not make any finding as to whether there has been an infringement.69 Their only legal effect is to close the Commission’s proceedings. Essentially, it is considered a measure of ‘procedural economy’.70 In addition, as noted by AG Kokott in Alrosa, ‘unlike Article 7, Article 9 of Regulation 1/2003 is not an instrument for establishing infringements of competition law, but merely gives the Commission the possibility of effectively addressing concerns over competition for the future’.71 Contrary to Article 7 infringement decisions, Article 9 decisions cannot be used as conclusive evidence of the existence of an infringement of EU competition rules in followon private actions for damages.72 Yet, from a functional perspective, they can be qualified as ‘remedies’, as they aim to redress the situation of the victims of the competition law violation to that prior to the infringement. As both Article 9 of Regulation 1/2003 commitment decisions and decisions adopted in the context of merger control are formally suggested by the parties to the transaction, they have been distinguished from other competition law remedies and sanctions, which are imposed unilaterally by the Commission and are not the product of a ‘voluntary’ agreement between the Commission and the parties to the dispute (coercive remedies).73 In a similar vein, commentators as well as the ECJ consider that commitment decisions form part of what has been characterised as a ‘consensual competition law enforcement’ or a culture of ‘settlement’, hence accentuating the opposition between the voluntary nature of commitment decisions and the coercive nature of Article 7 of Regulation 1/2003 decisions imposing injunctions on the parties.74 69 John Temple Lang, ‘Commitment Decisions under Regulation 1/2003’, in Charles Gheur and Nicolas Petit, eds, Alternative Enforcement Techniques in EC Competition Law, Bruylant, 2009, 121 et seq, 122. 70 Case C-441/07 P Commission v Alrosa [2010] ECR I-5949, para 35. For a critical perspective on this justification for commitment decisions see Mariniello, above n 68. 71 Opinion of Advocate General Kokott in Alrosa, ibid, para 50. 72 Firat Cengiz, ‘Judicial Review and the Rule of Law in the EU Competition law Regime after Alrosa’, 7 European Competition Journal 127, 130 (2011). But see recent national case law hinting to the contrary—described in Anna Duron, ‘Private Damages Actions in the Wake of a Commitment Decision: New Risks after the Judgment of the Paris Commercial Court in Eco-Emballages?’, forthcoming. 73 Opinion of AG Kokott in Alrosa, above n 70, paras 51 and 55, noting the ‘voluntary’ character of commitments. According to the Advocate General, procedural economy requires that the Commission should not be asked to conduct a detailed assessment of the appropriateness of commitments to address the Commission’s concerns, the only requirement being that such appropriateness should be clear or manifest (para 53). Hence, procedural economy may be considered a facet of ‘effectiveness’ if this concept is given a broad definition as including anything that reduces the costs of enforcement while maintaining its corrective and/or deterrent effect. 74 See Alrosa, above n 70, para 48 (noting that undertakings ‘consciously’ accept concessions in the context of a commitment procedure under Article 9 of Regulation 1/2003). See also Florian Wagner von Papp, ‘Best and Even Better Practices in the European Commitment Procedure after Alrosa: The Dangers of Abandoning the Struggle for Competition Law’, 49 Common Market Law Review 929, 932–933 (2012) (noting the

The Principle of Effectiveness 127 The EU Courts have relied on the classification of remedies as voluntary or coercive when dealing with the question of the degree of the remedial discretion competition authorities may exercise in EU antitrust and merger proceedings. Competition authorities are subject to restrictions in the use of voluntary remedies, at least in antitrust proceedings. Recital 13 of R egulation 1/2003 warns that commitment decisions under Article 9 are not appropriate in cases where the Commission intends to impose a fine. Hardcore cartel cases, which are generally subject to fines, cannot be closed by a commitment decision.75 The principle of proportionality may also limit the remedial discretion of competition authorities in both merger and antitrust proceedings, thus reinforcing the legitimacy of the remedial action of the competition authorities. In Alrosa, however, the ECJ struck down the judgment of the General Court for having applied the proportionality control to a similar extent in Article 9 and Article 7 decisions.76 The ECJ noted that ‘the obligation on the Commission to ensure that the principle of proportionality is observed has a different extent and content, depending on whether it is considered in relation to the former or the latter article’.77 It further explained that: application of the principle of proportionality by the Commission in the context of Article 9 of Regulation No 1/2003 is confined to verifying that the commitments in question address the concerns it expressed to the undertakings concerned and that they have not offered less onerous commitments that also address those concerns adequately.78

Although both Article 7 and 9 decisions are subject to the principle of proportionality, the application of that principle differs according to which of those provisions is concerned. Hence, according to the ECJ: [T]here is therefore no reason why the measure which could possibly be imposed in the context of Article 7 of Regulation No 1/2003 should have to serve as a reference for the purpose of assessing the extent of the commitments accepted under Article 9

‘hybrid character of commitment decisions’, distinguishing the ‘public law paradigm’ of ‘an authoritative, unilateral top-down hierarchical command by the State’ from the ‘contract law paradigm’ which relies on voluntary negotiations between the parties). For critical discussion, see, inter alia, Denis Waelbroeck, ‘The Development of a New “Settlement Culture” in Competition Cases’, in Gheur and Petit, eds, Alternative Enforcement Techniques, above n 69, 221 et seq; Claus-Dieter Ehlermann and Mel Marquis, eds, European Competition Law Annual 2008: Antitrust Settlements under EC Competition Law, Hart Publishing, 2010. 75 See, however, the possibility for settlements in cartel cases: Commission Regulation (EC) No 622/2008 amending Regulation (EC) No 773/2004, as regards the conduct of settlement procedures in cartel cases, [2008] OJ L171/3. 76 Alrosa, above n 70. 77 Ibid, paras 38 and 48 (noting the specific characteristics of the mechanisms provided for in Articles 7 and 9 of Regulation 1/2003 and the voluntary character of commitments under Article 9). 78 Ibid, para 41.

128 Ioannis Lianos of the regulation, or why anything going beyond that measure should automatically be regarded as disproportionate … Undertakings which offer commitments on the basis of Article 9 of Regulation No 1/2003 consciously accept that the concessions they make may go beyond what the Commission could itself impose on them in a decision adopted under Article 7 of the regulation after a thorough examination.79

Following Alrosa, the distinction between voluntary and unilateral remedies leads to a different application of the proportionality principle, hence to a greater variation in the degree of judicial scrutiny of remedies and consequently the remedial discretion of the Commission. In addition to the differential implementation of the proportionality principle, in the context of judicial review, the perceived consensual nature of the commitment procedure may weaken the link between the wrong and the remedy adopted, presumably pushing the authority to use its bargaining power in order to achieve remedies that would not only attempt to reverse the situation to the status quo ante but would also aim to establish a new, allegedly more competitive, equilibrium, under the pretext that the parties subject to the remedy have consented to it. Indeed, the voluntary nature of commitments may enable them to go further in their scope than what could be the scope of an infringement decision, if the latter option were chosen.80 The distinction between infringement decisions under Article 7, allegedly inspired by a ‘public law’ paradigm, and commitment decisions relying on a ‘contract law paradigm’ does not, however, take sufficiently into account the important similarities between injunctions and commitments in EU competition law. The reference to the ‘public law paradigm’ as a separate pole to the ‘contract law’ one seems far-fetched, in view of the importance of ‘administrative contracts’ in continental administrative law, and also of the distinction between imperium merum (the power to coerce) and jurisdictio (the power to make legal decisions).81 Remedies do not form part of the imperium but do form part of the mixtum imperium, the power which a magistrate has for the purposes of administering the civil (not criminal) part of the law, which is incident to jurisdictio. If remedies were classified within the imperium, it would not have been possible, first, for arbitration clauses to be included in merger remedies, arbitration being in this case a forced ‘contract’, which is a

Ibid, paras 48–49 (emphasis added). Opinion of AG Kokott in Alrosa, above n 70, para 60; see also para 48 of the ECJ’s judgment. 81 Gerasimos Sofianatos, Injonctions et Engagements en Droit de la Concurrence, LGDJ, 2009, 3–5. 79 80

The Principle of Effectiveness 129 distinct possibility in EU merger control,82 and, second, for remedial injunctions to produce extraterritorial effects.83 More troubling is the opposition sometimes made between the passive role of the parties in Article 7 proceedings and their active role in commitment decisions, in merger control or in the context of Article 9 of Regulation 1/2003. Despite the ‘coercive’ appearance of an injunction, this is often the result of a prior (failed) negotiation between the Commission and the parties concerned, the Commission attempting at least to achieve some form of adhesion from the parties that will guarantee the proper execution of the remedy.84 The psychological pressure that an infringement decision might be adopted by the Commission in the absence of commitments offered by the parties largely belies the voluntary and consensual nature of the process, and enables the Commission to extract disproportionate remedies. To this one may add the possibilities offered to the Commission to make strategic use of commitment decisions in order to achieve regulatory objectives, such as the liberalisation of the energy sector, the regulation of p ayment systems or the opening up of systems competition in the high tech and car industry sectors, to cite but a few examples. As I have shown elsewhere, the parties ‘consent’ to these commitments, despite the obvious resource constraints of the Commission to follow all possible cases with an infringement decision, in view of the collective action problem they are facing, thus enabling the Commission and other competition authorities to leverage their investigative powers into substantial bargaining power, leading to commitments that may go beyond what would have been achieved had the C ommission chosen an Article 7 infringement decision.85 The competition authority may also adopt the strategy of investing first in some high-profile Article 7 infringement decisions for certain kinds of practices, which it can later capitalise on with a number of commitment decisions anchored to this quite favourable legal precedent of an infringement decision, pushing the parties to commit to far-reaching remedies. A closer look at the commitment decisions adopted by the European Commission in recent years indicates that these relied on contested theories of harm, which may have been risky for the Commission to rely upon in an Article 7 infringement

82 Christoph Liebscher, ‘Drafting Arbitration Clauses for EC Merger Control’, 21 Journal of International Arbitration 67 (2004); Gordon Blanke, ‘International Abritration in EC Merger Control: A Supranational Lesson to be Learned’, 27 European Competition Law Review 324 (2006); Laurence Idot, ‘Une innovation surprenante: L’Introduction de l’arbitrage dans le contrôle communautaire des concentrations’, 4 Revue d’arbitrage 591 (2000). 83 Sofianatos, Injonctions et Engagements, above n 81, 486. However, that does not guarantee the execution of remedial injunctions outside the EU. 84 Ibid, 188–191. 85 Lianos, ‘Which Limits for Remedial Discretion?’, above n 1, 451–454.

130 Ioannis Lianos decision: excessive prices,86 patent ambush,87 collective dominance,88 strategic underinvestment,89 anticompetitive use of court injunctions,90 to cite but a few. As negotiations occur at the highest levels between the Competition Commissioner and the CEOs of these global (in most cases) corporations, commitment decisions may offer more room for politics and the consideration of broader public interests and values than the customary narrow focus on competition concerns, hence highlighting the polycentric nature of most of these disputes. For instance, although issues of privacy were not as such touched upon in the Commission’s competition investigation against Google, it is inevitable that the Commission’s investigation and the final decision on Google’s commitments which should be taken by the college of commissioners will be influenced by the overall context of Google’s disputes with European regulators91 and, likely, by the European policy to form a single digital market.92 Broader competition policy concerns, such as market liberalisation, were particularly influential in the choice of the instrument of commitment decisions in various cases the Commission investigated in the gas and electricity markets.93 Similar concerns regarding the need for regulation may have played a role in the financial services sector.94 Finally, one should also factor in the considerable technical complexity of some markets investigated by the Commission, in particular the high tech sector. It is virtually impossible to develop sound and effective remedies without extensive and ongoing cooperation and sharing of information with the business entity that was found to infringe competition law and was asked to restructure its

86 See, eg Commission Decision of 9 December 2009 in Case COMP/38.636—Rambus; Commission Decision of 15 November 2011 in Case COMP/39.592—Standard & Poors; Commission Decision of 26 February 2014 in Case AT.39398—VISA MIF. 87 Rambus, ibid. 88 Commission Decision of 26 November 2008 in Case COMP 39.388—German Electricity Wholesale Market; Commission Decision of 26 November 2008 in Case COMP 39.389—German Electricity Balancing Market. 89 Commission Decision of 18 March 2009 in Case COMP/39.402—RWE gas foreclosure; Commission Decision of 3 December 2009 in Case COMP/39.316—Gaz de France. 90 Commission Decision of 27 April 2014 in Case AT.39939—Samsung—Enforcement of UMTS standard essential patents. 91 See, eg the references by former Commissioner Almunia to the various legal troubles Google is facing in Europe, including, among others, privacy disputes: Joaquín Almunia, ‘The Google Antitrust Case: What Is at Stake?’, SPEECH/13/768, Brussels, 1 October 2013. The Google case also seems to have divided the College of Commissioners, some of whom reportedly expressed concerns about the deal reached provisionally with the company. See ‘Google’s Almunia Deal Said to Be Criticized by EU Officials’, Bloomberg (12 February 2014), http://www.bloomberg.com/news/2014-02-12/google-deal-with-almunia-said-to-becriticized-by-eu-officials.html. 92 As the European Parliament’s Resolution of 27 November 2014 on supporting consumer rights in the digital single market (2014/2973(RSP)) reminded us. 93 For a discussion, see Dunne, ‘Commitment Decisions’, above n 66, 408–409. 94 Ibid, 410–411.

The Principle of Effectiveness 131 rganisation and/or modify its conduct.95 This may be best achieved in the o context of a commitment procedure, rather than an Article 7 infringement decision, as the atmosphere of open conflict that the latter may generate may be an impediment to cooperation. The above considerations, and not reasons of procedural economy, constitute the major drivers for the increasing use of commitment procedures in EU competition law. Remedies adopted in the context of commitment decisions often fit closely the model of ‘structural’ adjudication: being prospective by nature and aiming to eliminate any potential threat to the value of competition, by proceeding to the restructuring of the organisation involved in the violation of competition law and not just the issuance of a prohibition aimed at some specific act or conduct.96 For instance, in the Visa commitment decision, in addition to the various price caps on the setting of multilaterally agreed interchange fees to which Visa agreed, it also took the commitment to implement further transparency measures to reorganise its relationships with the merchants’ banks (acquirers).97 The process is also not solely at the hands of the competition authority and the infringer, as one would have expected to happen in view of the ‘contractual’ nature of the procedure, but involves the participation of interested third parties. These may contribute, according to Article 27(4) of Regulation 1/2003, to the ‘market test’ of the commitments, offering their view on the commitments offered by the defendant, before these are made binding by a decision of the Commission. In order to enhance the transparency of the process to the public, the Commission also publishes the full text of the commitments, as well as a press release setting out the key issues of the case. Interested parties may submit their observations within a fixed time limit, which cannot be less than one month. It is also mentioned in its best practices that the Commission may also ‘actively promote the market test’ by sending, for instance, the market test document to third parties ‘which can potentially

95 An illustration of the extensive cooperation between the Commission and the undertaking offering commitments needed in the design and implementation of remedies in high tech sectors may be offered by the Microsoft cases: Nicholas Economides and Ioannis Lianos, ‘A Critical Appraisal of Remedies in the EU Antitrust Microsoft Cases’, 10 Columbia Business Law Review 346 (2010). 96 For a similar conclusion, see Skyscanner Limited v CMA [2014] CAT 16, where, although the CAT cites approvingly the ECJ’s view that commitments are offered ‘voluntarily by the parties’ (para 40), it also noted that ‘[t]his is not an infringement finding and we agree with the CMA that this is more in the nature of a regulatory decision subject only to the normal control of judicial review’ (para 160, emphasis added). Although the CAT does not seem to note the distinction between the models of regulation and structural adjudication, the implications of this re-characterisation of commitment decisions with regard to participatory rights for third parties are obvious. 97 VISA MIF, above n 86.

132 Ioannis Lianos be concerned by the outcome of the case (eg consumer a ssociations)’ and by ‘informing in writing the complainant, inviting it to submit comments’.98 Yet, despite these efforts to enhance participation in order to address the legitimacy concerns raised by the extensive use of the commitment procedure, the Commission enjoys a significant degree of unfettered discretion in comparison to the one it can exercise when adopting an infringement decision. First, the Commission enjoys a wide discretion to make the commitments binding or to reject them, without that choice being framed by self-restraining guidelines, as is the case in some Member States.99 Second, there is no specific time-frame for accepting commitment decisions, which may also be agreed upon even after the statement of objections has been sent to the parties,100 whereas some national competition authorities explicitly limit the exercise of

98 DG Competition, ‘Best Practices on the Conduct of Proceedings Concerning rticles 101 and 102 TFEU’, http://ec.europa.eu/competition/consultations/2010_best_ A practices/best_practice_articles.pdf, para 116. 99 In comparison, Article 31D(1) of the UK Competition Act 1998 requires the publication of guidelines as to the circumstances in which it may be appropriate to accept commitments and provides that the Competition Authority ‘must have regard to the guidance for the time being in force’. More detailed guidance is provided in Part 4 and Annex A of the OFT’s Enforcement Guidance, still valid for the purposes of the Competition and Markets Authority (CMA), which, inter alia, notes that the ‘OFT will not accept binding commitments in circumstances where compliance with and the effectiveness of any binding commitments would be difficult to discern and/or where the OFT considers that not to complete its investigation and make a decision would undermine deterrence’. The Guidance of the CMA on its investigation procedures also provides that ‘10.15. [t]he CMA is likely to consider it appropriate to accept commitments only in cases where the competition concerns are readily identifiable, will be fully addressed by the commitments offered, and the proposed commitments can be implemented effectively and, if necessary, within a short period of time’. Further, ‘10.17. [t]he CMA is very unlikely to accept commitments in cases involving secret cartels between competitors or a serious abuse of a dominant position’. The concept of ‘serious abuse of a dominant position’ is further explained by the OFT’s Enforcement Guidance, where it is mentioned that ‘[w]hen making its assessment, the OFT will consider a number of factors, including the nature of the product, the structure of the market, the market share(s) of the undertaking(s) involved, entry conditions and the effect on competitors and third parties. The damage caused to consumers whether directly or indirectly will also be an important consideration’, this assessment being made on a case by case basis, ‘taking into account all the circumstances of a case’. Notwithstanding this individual assessment the OFT Enforcement Guidance recognises that, ‘[a]s a general rule, the OFT will regard predatory pricing as a serious abuse’. Furthermore, the UK Competition Appeal Tribunal (CAT) has set some broad standards governing the exercise of the discretion of the UK competition authorities to discuss binding commitments, requesting the consideration of the broader interests of deterrence and transparency, but also of those of third parties or consumers, commitments being explicitly excluded ‘in serious cases of infringement or where deterrence would be undermined’: Wanadoo UK Plc v Office of Communications, [2004] CAT 20, paras 128–129. Similar Notices have been published, for instance, by the French Competition Authority. See Notice on Competition Commitments, 2 March 2009, http://www.autoritedelaconcurrence.fr/doc/cpro_enga_2mars09_uk.pdf. 100 European Commission, Notice on best practices for the conduct of proceedings concerning Articles 101 and 102 TFEU [2011] OJ C308/6, para 109.

The Principle of Effectiveness 133 that discretion by imposing a specific time frame.101 Third, other interested parties and the complainant are not offered adequate opportunities to participate in the procedure and represent their own arguments and evidence on the anticompetitive conduct in question. Certainly, according to Article 27(4) of Regulation 1/2003, interested third parties may submit observations after the Commission has published a summary of the case, as well as the main content of the commitments or of the proposed course of action, in the Official Journal and the full text of the commitments has been published on the DG Competition website, within a time limit fixed by the Commission, which may not be less than one month from the date of publication. However, the interested parties do not have access to the full evidence collected by the Commission, in particular its preliminary assessment, which may affect their ability to form a more accurate view of the anticompetitive conduct in question and its impact on their interests.102 Fourth, interested third parties have limited access to the courts in order to challenge commitment decisions. Third parties constitute ‘non-privileged applicants’ and must satisfy the quite strict criteria of Article 263 TFEU. In particular they must demonstrate that they are individually concerned by the decision they challenge.103 It remains to be seen if these restrictive standing requirements will impede interested third parties from challenging in front of the ECJ commitment decisions that

101 For instance, in France, commitments may not be given once the statement of objections is issued. See Notice on Competition Commitments, above n 99 in fine, para 13. Compare with the CMA Guidance on its investigative procedures (2014), which provides that ‘10.18. (a) business under investigation can offer commitments at any time during the course of that investigation, until a decision on infringement is made’, although it is also made clear that the CMA is ‘unlikely to accept commitments at a very late stage in the investigation, in particular after the CMA has considered representations on the Statement of Objections’. 102 Furthermore, EU law establishes a distinction between the right of access to the public file conferred to the defendants or undertakings concerned (in the context of Article 9 commitment decisions), interested third parties involved in the proceedings (such as complainants and intervening parties) and third parties that are not involved in the competition authorities’ proceedings. The distinction between these three categories is deemed to be of particular importance. For instance, in Alrosa, the ECJ noted that even though Alrosa was previously a defendant in proceedings opened against it and De Beers under Article 101 TFEU, the company was only an interested third party as far as the Commission’s commitment decision in the investigation against De Beers under Article 102 TFEU was concerned. Consequently, it did not enjoy the same right of access as De Beers to the public enforcement file. See Alrosa, above n 70, paras 88–95. 103 With regard to the condition of individual concern, the relevant test (established in Case 25-62 Plaumann & Co v Commission [1963] ECR 95) is satisfied only if the challenged act affects them ‘by reason of certain attributes which are peculiar to them or by reason of circumstances in which they are differentiated from all other persons, and by virtue of these factors distinguishes them individually just as in the case of the person addressed’. For a more recent case supporting this restrictive interpretation of that condition, see Case C-583/11 P Inuit Tapiriit Kanatami and others v Parliament and Council, EU:C:2013:625.

134 Ioannis Lianos affect their own interests.104 Furthermore, as I have noted previously, in Alrosa the ECJ distinguished the proportionality ground of review for Article 9 of Regulation 1/2003 commitment decisions from that applying to Article 7 infringement decisions, thus transforming it into some form of reinforced control of the rationality (means–end test) of the remedy, even if the ECJ still keeps the proportionality label. The third step of the proportionality test, the assessment of the existence of a least restrictive alternative, which may usually exercise an important constraining effect upon the discretion of the Commission, finds itself significantly emptied of its content. Judicial scrutiny thus cannot constitute an effective check and balance mechanism of the remedial action of the Commission in the context of commitment decisions. Finally, the Commission is not formally obliged to take into account the views presented by third parties, but simply to acknowledge them without explaining the weight it did or did not apply to them in its final decision to accept the commitments ‘offered’ by the parties.105 The absence of extensive participatory rights for interested third parties in the context of commitment decisions, despite the polycentric nature of the disputes frequently addressed by them, constitutes one of the major weaknesses of that procedure. In view of the increasingly important role commitments play in the conduct of competition law enforcement by the Commission, expanding the participation of interested parties (including the complainant) is becoming an important area for procedural reform. As recognised in the ECN Recommendation on Commitment Procedures of 9 December 2013, ‘the views of other players in the market on the proposed commitments can play an important part in assessing their adequacy to meet the competition concerns and to allow such third parties the opportunity to submit their observations’.106 The courts may also play an important role in reminding competition authorities of the need to reinforce the position of third parties (and the complainant) in the commitment decisions procedure. For instance, in some of its judgments, the UK Competition Appeal Tribunal (CAT) has emphasised the need to provide the complainant with a fair and ‘structured’

104 Such litigation is emerging in the Member States. See, for instance, the appeal brought by Skyscanner against the decision of the OFT to accept binding commitments from online travel agencies (OTAs) such as Booking.com, Expedia and the Intercontinental Hotel Group. Skyscanner, one of the OTAs not subject to the OFT’s commitments, argued that the OFT acted ultra vires, given its powers under Section 31A of the Competition Act 1998, and that Skyscanner should not be subject to commitments to which it had not agreed but which impacted its business. See Skyscanner, above n 96. 105 For an interesting discussion of the nature of the obligation of the competition authorities to engage with the comments presented during the consultation period by the consulted third parties, see Skyscanner, above n 96, and discussed below. 106 http://ec.europa.eu/competition/ecn/ecn_recommendation_commitments_09122013_ en.pdf para 15.

The Principle of Effectiveness 135 opportunity to comment until informal commitments (‘undertakings’ in UK competition law jargon) have been accepted, the CAT noting that, although the commitment procedure does not envisage ‘an adversarial system’, ‘this does not preclude the need to afford the complainant an opportunity to defend its interests during the administrative proceedings’ and the need to ensure that he be ‘given an opportunity to be heard before the OFT closes its file’ and/or accepts the undertakings proposed.107 The judgment of the UK Competition Appeal Tribunal in Skyscanner v CMA may provide some food for thought.108 In this case, an appeal was brought by Skyscanner against a decision of the OFT (now replaced by the Competition and Markets Authority, or CMA), in which the OFT accepted commitments from some online travel agents and hotel groups. Skyscanner was a third party operating a price comparison website, also known as a metasearch site, assisting customers to compare pricing between various OTAs and hotels, with whom Skyscanner contracted for the inclusion of their offerings in its metasearch results. The main competition concern the commitments accepted by the OFT attempted to address concerned the restriction imposed on arrangements between the specific OTAs and the hotel groups to restrict each OTA’s ability to discount the rate at which room-only hotel accommodation bookings were offered to customers (so-called ‘price agreements’). The OFT also expressed concerns with regard to some most favoured nation clauses that ensured that the retail rates for hotel room bookings provided by hotels to OTAs were no less favourable than the lowest retail rate displayed by other distribution outlets (so-called ‘rate parity obligations’), which the OFT found were capable of reinforcing the restriction of competition resulting from the discounting restrictions, although it did not consider whether they could on their own be considered as a restriction of competition. The OFT accepted commitments proposed by the parties, who agreed to remove the complete prohibition on discounting room-only rates and to replace these by limited discounting to ‘closed groups’ of consumers—basically a membership group which consumers could actively choose to join, up to the level of the commission or margin of OTAs. Although, according to the commitments, the OTAs could not publicise information about the specific level or extent of discounts outside the ‘closed group’, they could publicise information regarding the availability of discounts on their own websites or on

107 Pernod Picard/Campbell Distillers [2004] CAT 10, paras 241–243 (emphasis in original). See also Wanadoo UK, above n 99, para 132, noting that, ‘in the course of any negotiations about binding commitments, the regulatory authority should bear in mind, among other appropriate matters, the position of the complainant. The weight to be given to a complainant’s interests may well vary from case to case … This may particularly be so where the complainant has a detailed knowledge of the market and/or may be closely affected by the outcome.’ 108 Skyscanner, above n 96.

136 Ioannis Lianos price-comparison and metasearch websites. Skyscanner alleged that the OFT failed to take into account properly or not at all its representations during the consultation (‘market-testing’) process on the impact that these commitments had on the metasearch sector in which Skyscanner was active. Skyscanner expressed concern that the commitments only focused on intra-brand competition by allowing OTAs to offer discounts, but that allowing hotels to prevent OTAs from publicising specific information about discounts to consumers outside the OTAs’ closed group, and hence on metasearch websites, had a negative effect on inter-brand competition (and price transparency for consumers), as consumers would be unable to use metasearch sites in order to compare the actual room prices and discounts offered by different hotels, the only information these metasearch websites would be able to provide being that discounts are available, but not the exact level of discounts. Of particular interest for our discussion is the allegation that the OFT violated its statutory duty to consult and to consider representations made in response to its consultation notice. Despite the fact that the OFT had acknowledged the concerns expressed by Skyscanner and had even met with Skyscanner representatives to discuss those concerns, and despite the fact that Skyscanner was a ‘third party respondent to the consultation’ and not a complainant in the OFT’s original investigation, the CAT found that the consultation process was ‘defective’ as it had ‘failed conscientiously to address the objections raised’ by Skyscanner and two other respondents to the consultation, which concerned the possible effect of the proposed commitments on price transparency and metasearch websites.109 The CAT noted that the fact that Skyscanner was a third party, and not a complainant, should not affect ‘the significance the OFT would attach to a material point raised by a respondent’,110 and that the demands made by the OFT to Skyscanner for further evidence as to the impact of the proposed commitments on the metasearch website sector or price transparency ‘was not acceptable’.111 The CAT noted that the fact that the OFT was pursuing its investigation on the basis that it had identified a restriction ‘by object’ may have deprived it ‘of the ability properly to appreciate the significance of the role of operators such as Skyscanner’, but this did not excuse the OFT from not investigating ‘a plausible point further’ before taking a decision, and, in any case, the OFT should not have insisted on more evidence or supporting material from Skyscanner itself on this point.112 This finding was accompanied by some derogatory remarks of the CAT, raising the possibility that ‘the OFT found

Ibid, para 74. Ibid, para 71. 111 Ibid, paras 86–93, especially para 92. 112 Ibid, paras 93 and 100. 109 110

The Principle of Effectiveness 137 the S kyscanner objection inconvenient because it threated to upset a carefully constructed edifice that the OFT believed’.113 Consequently, the CAT upheld Skyscanner’s appeal on this ground.114 The Skyscanner judgment, with the emphasis put by the CAT on the consultation process and the need to actively engage with the representations made by third parties, illustrate the polycentric nature of commitment decisions. This specific nature calls for the consideration of the views (and interests) of all those affected by the decision, beyond the parties to the dispute.115 One cannot, therefore, apply the same principles as in the adjudicatory model. While conceptually distinct from the model of administrative management/ regulation, the ‘structural model’ of adjudication offers a more appropriate description of the hybrid nature of commitment decisions than the model of contractual governance. This has implications for the role and conceptualisation of the consultation process, or of the intervention of the judiciary, among other issues, in order to guarantee the legitimacy of the decisionmaking process.

5. Conclusions Left unbounded, the principle of effectiveness may offer the opportunity for competition authorities to expand their remedial discretion and re-design market processes and outcomes in accordance with the dominant interpretation of their statutory objectives. This may appear to be a sound sacrifice of legal form in favour of ‘future-oriented’ remedies that ‘lay down rules about how markets should behave in the future’.116 The point made in this chapter is that, whatever one thinks of the appropriateness of an expansive view of remedial discretion, which in my view is not supported by the restrictive interpretation of the principle of effectiveness in EU law, remedial discretion is naturally limited by the specific function exercised by the remedial process

Ibid, para 97. The CAT also noted that, ‘if a consultation response raises an important and obvious point of principle, it is for the authority to examine it further. This is particularly so where the authority has not carried out an analysis of the economic effects of the practices which it proposes to address with its commitments decision and where that decision itself may generate its own economic effects within the market’ (ibid, para 90). 115 However, requirements of participation for third parties in commitment procedures should not be as expansive as requirements of participation in merger regimes involving ‘public interest’ considerations such as, for example, the South African regime. There is a difference of degree in polycentricity between ‘competition’ concerns and ‘public interest’ considerations. 116 Daniel Crane, ‘Optimizing Private Antitrust Enforcement’, 62 Vanderbilt Law Review 675, 708, 721 (2010) (discussing the need for private enforcement to be forward looking, insisting on the need to ensure an ‘effective’ enforcement). 113 114

138 Ioannis Lianos chosen or, more contentiously, imposed by the nature of the dispute. Drawing on Fuller’s account of the existence of various forms of social ordering, each of them emerging in specific circumstances/contexts and having their own principles and limitations, I offered some reflections on the possible limits that the essence of each ideal type of social ordering sets on the expansive interpretative potential of the principle of remedial effectiveness. For instance, the limits of adjudication are of a different sort than those of contracts/negotiation or managerial/regulatory discretion. For each of these ideal types, legitimacy concerns operate differently and generate dissimilar demands on process or substantive rules bounding discretion. The polycentric nature of competition law disputes calls for flexibility in the choice of the adequate form of social ordering aiming to achieve the objectives set by the legislator. Its specificity also breaks with the classic view of the adjudication model and hints at the prevalence, in a significant number of cases with a pronounced polycentric element, of what has been called the ‘structural adjudication’ model, still distinct from the model of regulatory governance. I discussed the nature of commitment decisions as an illustration of the difficulties of classification without a proper consideration of the functions and respective limits of each form of social ordering. Commitment decisions have erroneously been characterised as closer to forms of contractual governance or, taking a starkly different perspective, to regulation. In reality, these characterisation disputes offer little, and may be as confusing and unpurposeful as discussions over the gender of angels, if one does not develop an overall theory with regard to the meaning of each category and the practical implications of such categorisation. This study aimed to sketch such a theory by insisting on the need to keep an eye on the specific limitations of remedial discretion, of a procedural or substantive nature, that have emerged in each form of social ordering with the aim of guaranteeing the legitimacy of decision-making. It is not impossible for a competition authority to make the choice of one or the other form of social ordering, should it enjoy the power to make such a choice, if this advances its purpose. Yet the authority should also adopt the mechanisms that were put in place in order to respond to the calls for the legitimacy of decision-making in the context of the specific type of social ordering. New tools, such as commitment decisions, which do not fit into the existing categories, may require the development of new mechanisms or a different conceptualisation of existing mechanisms in order to ensure the legitimacy of decision-making. This creative process is presently ongoing in EU competition law.

Damien MB Gerard * Negotiated Remedies in the Modernisation Era: The Limits of Effectiveness

The entry into force of Regulation 1/2003, which marked the formal roll-out of a multi-faceted process known as ‘modernisation’, has given rise to a tension between effectiveness and legitimacy in the enforcement of EU competition law which has not yet been solved. Inasmuch as it revamped the procedures governing the enforcement of antitrust law in the European Union, Regulation 1/2003 is largely driven by effectiveness considerations. Notably, in pursuance of ‘the objective of enhancing administrative efficiency and effectiveness in dealing with competition concerns’,1 Regulation 1/2003 introduced a new form of proceedings whereby the Commission can close cases by making commitments binding on the ‘undertakings concerned’2 without reaching a finding of infringement. As such, commitment decisions constitute the closest EU equivalent to the consent decrees heavily relied upon to resolve civil antitrust complaints in the United States.3 Whereas commitment decisions were originally perceived as a peripheral tool that the C ommission would use on ‘unusual and rare’ occasions,4 they quickly developed into the

* University of Louvain (UCL, Belgium). Comments welcome at Damien.Gerard@ uclouvain.be. 1 Report on the functioning of Regulation 1/2003 (COM/2009/0206 final), paras 13 and 18 (note that ‘administrative [or procedural] efficiency’ and ‘effectiveness’ are often used as synonymous in policy documents, and so they are in this contribution). 2 Article 9 of Regulation 1/2003 on the implementation of the rules on competition laid down in Articles [101] and [102] of the Treaty [2003] OJ L1/1: ‘Where the Commission intends to adopt a decision requiring that an infringement be brought to an end and the undertakings concerned offer commitments to meet the concerns expressed to them by the Commission in its preliminary assessment, the Commission may by decision make those commitments binding on the undertakings.’ 3 Article 9 of Regulation 1/2003 was even directly ‘inspired by “consent decrees” in the Americal system’: Alexander Schaub, ‘The Commission’s Position within the Network’, in Claus-Dieter Ehlermann and Isabela Atanasiu, eds, European Competition Law Annual 2002: Constructing the EU Network of Competition Authorities, Hart Publishing, 2005, 237 et seq, 240. US consent decrees typically provide that they do ‘not constitute any admission by any party regarding any issue of fact or law’. See, eg United States v Microsoft, No. 98 Civ 1232 (CKK), 2002 (DDC 12 November 2002). 4 John Temple Lang, ‘Commitment Decisions and Settlements with Antitrust Authorities and Private Parties under European Antitrust Law’, in Barry Hawk, ed, International Antitrust Law and Policy: Fordham Corporate Law Institute, Juris Publishing, 2005, chapter 13, 270.

140 Damien MB Gerard enforcement mechanism of choice in the modernisation era, outside of the field of hard-core cartels.5 Hence, commitments have come to embody the advent of a ‘negotiated’ form of enforcement and of associated transformations in the EU antitrust enforcement system. These transformations, which have profoundly modified the respective incentives of enforcement authorities and firms, were not immediately perceived, let alone anticipated. Yet the revelation of these consequences gave rise to situations of acute cognitive dissonance, as is apparent from the infamous Alrosa case,6 and to serious legitimacy concerns. By growing in prominence, commitments have caused the focus of enforcement proceedings to shift towards the negotiation of ‘effective’ remedies to address concerns expressed succinctly by the Commission in what is known as a ‘preliminary assessment’.7 As such, commitments have embodied a shift in the focus of EU antitrust enforcement from the wrong to the remedy.8 Starting from that observation, this chapter questions whether effectiveness is the proper benchmark to govern the design of remedies in the EU ‘negotiated’ enforcement framework. To that end, after mapping the central position taken by negotiated remedies in the modernisation era (Section 1), it attempts to identify some of the limits of effectiveness and endeavours to set forth concrete policy proposals to address these limits (Section 2). Generally, it aims to explore persisting tensions in the ‘modern’ EU antitrust enforcement framework between negotiated procedures, effectiveness, due process and legitimacy.

1. Negotiated remedies in the modernisation era Over the past 10 years, the nature, scope and design of remedies have grown in importance in the enforcement of competition law in the European Union. That evolution can be viewed as one of the many consequences of a process known as the modernisation of EU competition law. Modernisation

5 Indeed, Recital 13 of Regulation 1/2003 provides that: ‘[c]ommitment decisions are not appropriate in cases where the Commission intends to impose a fine’, which would typically be the case for hard-core cartels. In that sense, see also the ECN Recommendation on Commitment Procedures, 9 December 2013. 6 Decision of the Commission of 22 February 2006 in Case COMP/38.381—De Beers; Case T-170/06 Alrosa Company Ltd v Commission [2007] ECR II-2601; Case C-441/07 P Commission v Alrosa Company Ltd [2010] ECR I-5949. 7 Article 9(1) of Regulation 1/2003. 8 A similar shift has been associated in the US, since the mid-1990s, with the growing reliance on consent decrees by enforcement authorities and the increased sophistication of e conomic tools. See, eg A Douglas Melamed, ‘Antitrust: The New Regulation’, 10 Antitrust 13, 15 (Fall 1995).

Negotiated Remedies in the Modernisation Era 141 consists in a comprehensive attempt to experiment with a utility-maximising approach in the regulation of economic competition. This is apparent in its three intrinsically interrelated dimensions. ‘Substantive modernisation’ entailed a shift in the substantive rationality of EU competition law from the static protection of the freedom to trade of competitors to the dynamic promotion of a competitive process conducive to efficiency gains, contributing, as a result, to the welfare of consumers. That evolution is often presented as a move from a form-based to an effects-based approach in EU competition law enforcement, ie to greater attention being paid to case-specific contexts apprehended by means of increasingly sophisticated economic tools.9 In parallel, ‘institutional modernisation’ entailed a move towards a form of network antitrust enforcement in the European Union by means of the so-called ‘decentralisation’ process that allowed for the full applicability of EU competition law by national competition authorities (NCAs) involved with the Commission in enforcement coordination and policy design within the E uropean Competition Network (ECN).10 The third related dimension is that of ‘procedural modernisation’.

1.1 Procedural modernisation and the rise of negotiated procedures In addition to the transition from a prior notification to an exception s ystem for the review of coordinated practices, ‘procedural modernisation’ was marked by the emergence of and increasing reliance on negotiated procedures in the enforcement of EU competition law, ie leniency, settlement and commitments proceedings, against the background of a significant increase in the amount of fines. While the three dimensions of the modernisation process have contributed to a transformation of the inner rationality of EU competition law enforcement, procedural modernisation itself has profoundly

9 Several sources discuss the process of ‘substantive modernisation’. In relation to unilateral practices, see, eg Giorgio Monti, ‘Article 82 EC: What Future for the EffectsBased Approach?’, 1 Journal of European Competition Law & Practice 2 (2010); Nicolas Petit, ‘From Formalism to Effects?—The Commission’s Guidance on Article 82 EC’, 32 World Competition 485 (2009). In relation to coordinated practices, see, eg Alison Jones, ‘The Journey toward an Effects-based Approach under Article 101 TFEU—the Case of Hardcore Constraints’, 55 Antitrust Bulletin 783 (2010); Damien MB Gerard, ‘The EffectsBased Approach under Article 101 TFEU and its Paradoxes: Modernisation at War with Itself ?’, in Jacques Bourgeois and Denis Waelbroeck, eds, Ten Years of Effects-Based Approach in EU Competition Law—State of Play and Perspective, Bruylant, 2013, 18 et seq. 10 See, eg Firat Cengiz, ‘Multi-level Governance in Competition Policy: the European Competition Network’, 35 European Law Review 660 (2010); Damien MB Gerard, ‘Public Enforcement: The ECN—Network Antitrust Enforcement in the European Union’, in Ioannis Lianos and Damien Geradin, eds, Handbook on European Competition Law— Enforcement and Procedure, Edward Elgar, 2013, chapter 3.

142 Damien MB Gerard odified the respective incentives of enforcement authorities and firms in m resolving antitrust issues. Hence, procedural modernisation raises important systemic issues, including as to the design of the procedural safeguards guaranteeing the legitimate exercise of public authority in the enforcement of EU antitrust rules. As noted, a prominent feature of procedural modernisation has been the formalisation by Regulation 1/2003 of the possibility of closing cases by means of commitments made binding by the Commission. In practice, when adopting a commitment decision, the Commission ‘conclude[s] that there are no longer grounds for action’,11 ie it closes the case without reaching a formal finding of dominance or infringement. Though the Commission did resolve cases by means of informal commitments in the past, the possibility of making commitments binding was timidly put forward by the 1999 Modernization White Paper as a ‘useful’ complement to existing enforcement tools.12 A few years later, however, in its first report on the functioning of Regulation 1/2003, the Commission acknowledged that the commitment procedure had in fact ‘added considerable value’ to the enforcement toolbox.13 Moreover, Article 5 of Regulation 1/2003 also empowered NCAs to accept commitments as a means of closing national cases involving the (joint) application of EU competition law; as a result, commitment procedures are now expressly provided for in the legal rules of almost all EU Member States,14 albeit with some procedural variations. The transformations induced by an increased reliance on negotiated procedures are therefore not limited to the EU level but have permeated antitrust enforcement throughout Europe. As further developed below, national enforcement systems have sometimes proven to be better equipped to accommodate and respond to these transformations and, generally, to ensure a sustainable balancing of effectiveness and legitimacy considerations.

1.1.1 Modernisation and the shift in enforcement rationality The three dimensions of the modernisation process have altered the inner rationality of EU competition law enforcement, which is to say they have altered its objectives and means of action. Interestingly, that phenomenon presents striking similarities with the move, associated with the growing reliance on consent decrees in the United States, from a law-enforcement to a regulatory approach in the application of antitrust rules. That novel approach

Article 9(1) of Regulation 1/2003. White Paper on modernisation of the rules implementing Articles 85 and 86 of the EC Treaty—Commission Programme No 99/027, 28 April 1999, para 90. 13 Report on the functioning of Regulation 1/2003, COM/2009/0206 final, para 18. 14 See ECN Recommendation on Commitment Procedures. 11 12

Negotiated Remedies in the Modernisation Era 143 was labelled ‘regulatory’ because of its forward-looking nature, its tendency to work by guidance and its consensual decision-making model.15 In turn, challenges raised by the move towards that regulatory model in the US appear equally similar to those prompted by the EU modernisation process,16 including in terms of legitimacy, even if the defining features of the US enforcement system arguably limit the magnitude of these challenges17 compared to the situation in the Union. The effectiveness paradigm From an enforcement point of view, the entire modernisation process has been driven by effectiveness considerations.18 This is particularly apparent when it comes to procedural modernisation. According to the European Court of Justice, the commitment procedure, in particular, ‘is a new mechanism introduced by Regulation No 1/2003 which is intended to ensure that the competition rules laid down in the EC Treaty are applied effectively’, inasmuch as it enables ‘a more rapid solution to the competition problems identified by the Commission’ and is inspired by ‘considerations of procedural economy’.19 In the procedural modernisation context, however, effectiveness is not a purely functional concept but has a normative ambition, namely to maximise the allocation of enforcement resources with a view to strengthening compliance on the part of economic actors and the societal impact of competition policy at a time when markets are becoming increasingly dynamic. Driven by effectiveness, the enforcement practice of the Commission appears increasingly inspired by cost/benefit considerations in the selection of cases and the conduct of proceedings.20 ‘When the EU competition authority decides

Harry First, ‘Is Antitrust ‘Law’?’, 10 Antitrust 9 (Fall 1995). Ibid, 11. See also Melamed, ‘The New Regulation’, above n 8, 13–14. 17 Relevant features include the much greater importance of private enforcement and the need for public enforcement agencies in many instances to bring infringement actions to court, as well as the judicial approval of DOJ consent decrees. 18 Anecdotally, the terms ‘effective[ly]’ and ‘effectiveness’ appear more than 10 times in the Recitals of Regulation 1/2003. 19 ECJ in Alrosa, above n 6, para 35. 20 On the link between the emergence of a culture of negotiated enforcement and antitrust agencies’ selection of cases, see, eg Douglas Ginsburg and Joshua Wright, ‘Antitrust Settlements: The Culture of Consent’, Concurrences, No 2-2013, Article No 51634, 56 et seq, 59 (‘Another consequence of an agency bringing cases primarily with an eye to settlement is to change the agency’s case selection criteria. In addition and to some extent in lieu of the criteria that would otherwise make a case attractive, such as the benefit to consumers from terminating an anticompetitive business practice, the probability and ease with which the case will settle become part of the mix. That is, instead of pursuing the cases that advance most the agency’s law enforcement mission, it will tend to pursue cases with the best prospect for settlement, cases that will consume few investigative resources, settle quickly, and are more likely to result in a consent decree that provides a continuing role for the agency’). 15 16

144 Damien MB Gerard to pursue an antitrust case’, such considerations determine for example the choice between ‘one of two main paths: a prohibition decision under article 7 of Regulation 1/2003 or a commitment decision under article 9’.21 In such an effectiveness-driven approach, the notion of risk becomes the main driver of the application of the law with, as a c orollary, a substantive alteration of notions of legal certainty and due process. Nowadays, antitrust compliance has very much become a matter of risk management: companies are invited to self-assess their commercial practices and self-comply with general guidance while managing the risk of complaints and proceedings, which, if and when these occur, they strive to contain or settle. Though it may have achieved results inasmuch as most companies today internalise the ‘antitrust risk’ when devising commercial strategies, that new paradigm has upset many, in particular counsel, whose traditional role (providing legal certainty) appears ill-suited to the new effectiveness p aradigm, which calls instead for acute damage control skills. Moreover, it is a widely shared perception that the move towards a negotiated form of antitrust enforcement has led to an increase in the Commission’s discretion to the detriment of due process, notably in the absence of a credible alternative.22 Generally, the focus on effectiveness and the promotion of negotiated procedures has had the effect of putting greater emphasis on sanctions and remedies, and on how they are tailored. In fact, negotiated procedures are largely driven by the nature and scope of remedies, rather than by an attempt to apply the law to the facts. Yet the impact of the effectiveness paradigm goes further: it also governs the design of remedies. As former Director-General Italianer emphasised, ‘the most important “innovation” in our remedies these last years has been a simplification push on our side for both mergers and antitrust, leading to greater effectiveness’.23 By hailing the virtues of ‘simplification’, which include ‘considerations of expediency’,24 DirectorGeneral Italianer seemed to mean that remedies ought to be designed in ways that make them more administrable, thereby maximising both the monitoring resources of the Commission and the social impact of the remedies. However legitimate, these considerations raise the question of the relationship between effectiveness and efficiency in the design of remedies and how to align them— in particular, in negotiated proceedings driven by procedural economy.

21 Joaquín Almunia, ‘Remedies, Commitments and Settlements in Antitrust’, SPEECH/ 13/210, Brussels, 8 March 2013. 22 See, eg Denis Waelbroeck, ‘Le développement en droit européen de la concurrence des solutions négociées (engagements, clémence, non-contestation des faits et transactions): que va-t-il rester aux juges?’, Global Competition Law Center Working Paper 1/08, http:// www.gclc.coleurope.eu. 23 Alexander Italianer, ‘Legal Certainty, Proportionality, Effectiveness: The Commission’s Practice on Remedies’, speech, Brussels, 5 December 2012. 24 Commission Staff Working Paper accompanying the Report on the Functioning of Regulation 1/2003, SEC(2009) 574 final, para 99.

Negotiated Remedies in the Modernisation Era 145 Effectiveness versus efficiency As noted, substantive modernisation has led to the emergence of a new consensus whereby competition law is viewed as a dynamic process conducive to efficiency gains.25 In turn, that process is subject to rules designed to tackle restrictions of competition which distort the incentives to achieve efficient outcomes. As a reflection of that consensus, the modernised approach in the EU entails that those practices which are liable to have an ‘appreciable adverse impact on the parameters of competition on the market, such as price, output, product quality, product variety and innovation’, will be caught by Article 101/102 TFEU.26 To make the efficiency consensus operational, it is submitted, enforcement must transcend effectiveness considerations to pursue welfare maximisation objectives, in particular in the design of remedies. In other words, it is not sufficient for an enforcement authority like the Commission to pursue the maximisation of its own administrative resources; it is bound to maximise the efficiency of the activity that it is regulating, ie competition. That proposition can be captured as a need to move from effective to optimal enforcement, in order to ensure the legitimacy of the Commission’s remedial practice. Hence, a key threshold question is how to reconcile the maximisation of enforcement resources with the promotion of competition law as a tool of welfare maximisation and, as a corollary, how to ensure the optimal character of negotiated remedies. Admittedly, achieving legitimate remedial outcomes involves complex trade-offs which, in the EU administrative enforcement framework, have historically been captured by concepts known as proportionality and due process (understood as ‘fair trial’). However, as further explained below, negotiated procedures, and commitment proceedings in particular, involve a relaxation of the substance of both concepts, thereby creating a legitimacy gap. Indeed, unilaterally favouring effective over optimal enforcement by loosening proportionality and due process requirements carries risks for the substantive legitimacy of the EU competition law enforcement process, and for the acceptance of its outcomes by those to whom it is addressed. The concern is heightened by the fact that negotiated decisions now account for the bulk of EU antitrust enforcement outcomes. In turn, filling that legitimacy gap requires a review of the sequencing of commitment proceedings and associated procedural rights, as well as the (re)development of a credible alternative to those proceedings,

25 That consensus is reflective of a broader-scale agreement (with sometimes important nuances in the way it is implemented) to the effect that ‘consumers benefit from competition through lower prices and better products and services’ and promotes reliance on a ‘sound analytical framework firmly grounded in economic principles’. ICN, ‘A Statement of Achievements’, April 2010, 6–8. 26 See, eg Guidelines on the application of Article 81(3) of the Treaty, [2004] OJ C101/08, para 16.

146 Damien MB Gerard in order to carve out space for fair dialectical exchanges conducive to efficient remedial outcomes and compatible with effectiveness concerns.

1.1.2 The inexorable attractiveness of negotiated procedures The utility-maximising approach driving modernisation (notably insofar as it has led to a dramatic increase in the amount of fines) combined with, first, the typical aversion of firms to uncertainty and, second, a persistently suboptimal judicial review system, have made the temptation to negotiate rather than litigate cases almost irresistible, to the point of becoming the norm in noncartel cases. This section attempts to document and explain the rise in negotiated procedures. Statistical overview At the EU level, commitment decisions were introduced by Regulation 1/2003, which entered into force on 1 May 2004. Whereas commitment decisions were originally relied upon to settle cases that had been ongoing for years prior to the entry into force of Regulation 1/2003, they subsequently emerged as the default antitrust enforcement tool in the modernisation era and are still today ‘becoming increasingly frequent’.27 For illustrative purposes, the following chart summarises the Commission’s antitrust decisional practice over the May 2004–July 2013 period, by type of decision and nature of infringement. 35

31

30 Article 9 decisions

25 16

20 15 7

10

9 4

5 0

6 0

Total

Article 102

Article 101 vertical

3

Infringement decisions

Article 101 horizontal

Figure 1 *excludes collusive practices that are not amenable to Article 9 decisions

27 Alexander Italianer, ‘To Commit or Not to Commit, That is the Question’, speech, Brussels, 11 December 2013.

Negotiated Remedies in the Modernisation Era 147 As is apparent from Figure 1, commitment decisions have outnumbered infringement decisions by a ratio of more than three to one. Moreover, commitment decisions have been adopted predominantly in the field of dominance, but also in relation to coordinated practices, including in horizontal cases.28 In recent years, the trend towards negotiated enforcement has become even more pronounced; between January 2008 and July 2013, only three non-cartel prohibition decisions were adopted,29 compared to 19 commitment decisions. The Commission, therefore, favoured a negotiated outcome in 85 per cent of non-cartel cases in that five-year period.30 Since the entry into force of Regulation 1/2003, a large number of EU Member States have replicated Article 9 (with some variations) in their domestic competition law enforcement system.31 While ECN statistics do not track the number of commitment decisions adopted at national level, selected national statistics reveal that the NCAs of large Member States have built a strong record of commitment decisions over the past few years,32 though the situation is not comparable to that at the EU level, especially when it comes to the proportion of commitment versus infringement decisions. Whereas the NCAs’ statistics do not often isolate cartel cases from other cases, Figure 2 suggests that NCAs continue to resort primarily—or at least equally—to infringement decisions to resolve antitrust cases, which is a key difference with the situation prevailing at the EU level, even if cartel cases are also accounted for.33 One may speculate at length about the possible reasons 28 Interestingly, in all but one of these horizontal cases, an ‘object restriction’ was alleged. This reflects the Commission’s willingness to test the limits of the scope of Article 9, which, as noted, ‘is not appropriate in cases where the Commission intends to impose a fine’. Recital 13 of Regulation 1/2003. 29 Commission Decision of 13 May 2009 in Case COMP/37.990—Intel; Commission Decision of 22 June 2011 in Case COMP/39.525—Telekomunikacja Polska; Commission Decision of 23 January 2013 in Case AT.39839—Telefónica/Portugal Telecom. 30 As a comparison, it is estimated that ‘over 60 per cent of antitrust disputes are resolved … by means of consent decrees’. Richard Epstein, Antitrust Consent Decrees in Theory and Practice, AEI Press, 2007, 1. Compare with the figure of ‘roughly 70%’ provided in 1995 by Michael Weiner, ‘Antitrust and the Rise of the Regulatory Consent Decree’, 10 Antitrust 4 (Fall 1995). 31 The practice existed in Spain prior to the entry into force of Regulation 1/2003. Some Member States also derived the power to make commitments binding directly from the direct applicability of Article 5 of Regulation 1/2003 empowering NCAs to apply Articles 101 and 102 TFEU in individual cases, including by ‘accepting commitments’. See the position taken by the Belgian NCA in Decision 2006-I/O-12 of 31 August 2006 in Case CONC-I/ O-00/0049, Banksys SA, 10. 32 The 2012 Annual Report of the Polish NCA also reports (at page 11) the adoption of 38 commitment decisions out of a total of 86, including only eight decisions pertaining to horizontal agreements (none of which was a commitment decision). With respect to smaller Member States, the situation appears to vary significantly from one NCA to another. For example, the Danish NCA appears to have adopted five commitment decisions in 2013 alone (compared to four infringement decisions), whereas the Belgian NCA has only adopted two commitment decisions since 2004. 33 In December 2013, Director-General Italianer indicated that, ‘[s]ince May 2004, the Commission has adopted thirty-two commitment decisions under Article 9 and seventeen

148 Damien MB Gerard 2007

2008

2009

2010

2011

2012

GERMANY Infringements

13

10

6

5

Commitments

7

19

3

5

FRANCE Infringements

24

15

15*

11*

6*

10*

Commitments

8

7

3

6

5

5

Infringements

11

5

6*

6*

9*

9*

Commitments

9

11

9

13

5

3

Infringements

12

19

16*

14*

33*

34*

Commitments

0

3

4

11

2

5

*

*

ITALY *

*

SPAIN *

*

Figure 2: Antitrust Enforcement Statistics of Selected NCAs (2007–2012) *

includes collusive practices not amenable to commitment decisions

for this difference in the ratio of commitment decisions at national compared to EU level. It is submitted that, among them, the design of the competition law enforcement framework and its impact on the sequencing of proceedings, the associated exposure to and magnitude of fines, and the strength of the judicial review system can all influence the incentive of firms to favour negotiation over litigation.34 Explaining the success of the commitment procedure As noted, explanations for the success of commitment decisions, particularly at the EU level, can be manifold. This section inquires successively into the role played by two types of factors: (i) the benefits directly associated with engaging in commitment procedures; and (ii) the growing financial exposure and risks associated with infringement decisions. It is submitted that, taken together, these factors go a long way to explaining the incentive of firms to forgo litigious strategies and favour negotiated solutions, in spite of the uncertainties and pitfalls that such negotiations entail. prohibition decisions under Article 7’. Italianer, ‘To Commit or Not to Commit’, above n 27. Thus, even including cartel cases, the ratio between committment and infringement decisions remains close to two to one at the EU level. 34 Empirical testing of that claim falls beyond the scope of this contribution, but is part of a broader research project conducted by the author on the occasion of the 10th anniversary of the entry into force of Regulation 1/2003.

Negotiated Remedies in the Modernisation Era 149 Engaging in commitment proceedings carries many benefits, both for enforcers and for firms. Judging by the success of commitment decisions, these benefits appear to outweigh by far the costs associated with such proceedings, at least at the EU level, whether in terms of wealth transfer or repressive utility. Former Director-General Italianer cited ‘procedural efficiencies’, ‘quick market impact’, ‘swift resolution of concerns’ and the involvement of ‘fewer resources’ as the main benefits of commitment proceedings from the Commission’s perspective.35 As noted, gains in terms of effectiveness constitute a powerful driver of enforcers’ preference: even if the monitoring of commitments can prove demanding,36 procedural economies can be significant in terms of human capital and duration of proceedings,37 as illustrated in Figure 3.38 31

35 30 21

25 20

14

15 10 5 0

Commitment proceedings

6

5

6 0

< 1 year

Infringement proceedings

0 1 to 2 years

2 to 4 years

> 4 years

Figure 3: Duration of Proceedings: Commitment vs Infringement

The reduced formalities of commitment proceedings also allow enforcement authorities to secure the effectiveness of their intervention in highly dynamic markets on the basis of a ‘preliminary assessment’ and with the comfort of being entitled to reopen formal proceedings in case of ‘material change’ in the market context in question.39 Conversely, free from the constraint of

Italianer, ‘To Commit or Not to Commit’, above n 27. Monitoring the commitments made binding in the Repsol case, for example, required the detailed review of a large number of contracts between Repsol and individual gas stations, notably to determine the conditions according to which the termination of such contracts was to take place. See Commission Decision of 12 April 2006 in Case COMP/38.348—REPSOL CPP SA—Distribution de Carburants et Combustibles. 37 In the US, widespread reliance on consent decrees was also associated with a willingness to ‘economize on enforcement and compliance resources’, as well as on litigation costs. See Melamed, ‘The New Regulation’, above n 8, 13. 38 Commitment proceedings that lasted more than four years relate to cases started under Regulation 17/62 and then concluded by means of commitments under Regulation 1/2003. 39 Articles 9(1) and 9(2) of Regulation 1/2003. 35 36

150 Damien MB Gerard e stablishing an infringement and, beforehand, of developing formal objections, enforcers can limit the articulation of their theory of harm and rely on a narrower set of evidence. Relaxed proportionality requirements also allow them to experiment with ‘innovative’ theories and remedies, including of a structural nature.40 The increased flexibility in the design of negotiated remedies has also enabled the Commission to pursue regulatory objectives under the shadow of commitments ‘proposed’ by private parties.41 Finally, by resolving cases through the use of commitments, enforcers de facto shield their decisions from judicial review. Thus, as has been pointed out, appeals brought against Article 9 decisions have in practice been quite rare.42 In the speech mentioned above, Italianer also contended that ‘[f]or companies under investigation, the decision whether to commit or not to commit seems easy’ in view of the possibilities to ‘avoid paying fines’, ‘escape entanglement in long drawn out legal procedures’, ‘limit the reputational damage that goes with the finding of an infringement’ and, more positively, ‘offer tailor made solutions to competition concerns’.43 These benefits are, indeed, very appealing: escaping a finding of infringement means avoiding a fine, adverse publicity and possible follow-on damages actions,44 but also increased scrutiny in the future and, as the case might be, a label of dominance. The ability to better control the solution of the case is also real, but should not

40 According to Article 7 of Regulation 1/2003 and Recital 12: ‘structural remedies can only be imposed either where there is no equally effective behavioural remedy or where any equally effective behavioural remedy would be more burdensome for the undertaking concerned than the structural remedy’. To date, structural remedies have not been imposed pursuant to Article 7, whereas they have been imposed in a number of commitment decisions. See Commission Decision of 26 November 2008 in Cases COMP/39.388 and 39.389—German Electricity Balancing Market and German Electricity Wholesale Market [2009] OJ C36/8; Commission Decision of 18 March 2009 in Case COMP/39.402—RWE Gas Foreclosure [2009] OJ C133/10; Commission Decision of 29 September 2010 in Case COMP/39.315—ENI [2010] OJ C352/8; and Commission Decision of 10 April 2013 in Case AT.39727—CEZ [2013] OJ C251/4. The German Electricity and RWE Gas Foreclosure decisions were considered by the Commission to be ‘unprecedented’ insofar as they were ‘expected to open two separate markets to competition’. Commission Staff Working Paper accompanying the Report on the Functioning of Regulation 1/2003 SEC(2009) 574 final, para 64. 41 Indeed, various commitments entered into by major actors of the energy sector went beyond what energy regulators could achieve by means of the tools available under the applicable EU energy framework. For commentary, see, eg Suzanne Rab, Daphne Monnoyeur and Anjali Sukhtankar, ‘Commitments in EU Competition Cases—Article 9 of Regulation 1/2003, its Application and the Challenges Ahead’, 1 Journal of European Competition Law & Practice 178 (2010). 42 Italianer, ‘To Commit or Not to Commit’, above n 27; ECN Recommendation on Commitment Procedures, para 24. 43 Ibid. 44 This is acknowledged by the Commission in its Manual of Procedure (2012), http:// ec.europa.eu/competition/antitrust/antitrust_manproc_3_2012_en.pdf. See chapter 16, para 12.

Negotiated Remedies in the Modernisation Era 151 be overstated, given the many other contingencies involved in the process which also affect the alleged voluntary character of commitments.45 In terms of legal certainty, a more direct benefit lies with the de facto validity of commitments across the EU.46 Generally, though, the decisive factor driving the incentive of firms to seek negotiated solutions lies in the growing financial exposure and risks associated with infringement proceedings and, as a result, the perception of a lack of alternative in order to achieve a satisfactory outcome, independently of the strengths and weaknesses of the case at hand. The increase in the amount of fines observable at the EU level since the publication of the Commission’s first fining guidelines in 1998 is a welldocumentedphenomenon, illustrated in Figure 4.47 At the time, and ever since, the underlying rationale was ‘to set the fine at a level which ensures that it has a sufficiently deterrent effect’.48 Though it has rarely reached the statutory 10 per cent worldwide turnover ceiling provided for by Regulation 1/2003,49 the Commission has continuously increased the amount of fines over the past decade to the point of exceeding the one-billion-euro threshold on three occasions since 2009.50

45 Consider in this regard the saga surrounding the Google case, in which Google’s second commitment proposal was again deemed ‘not acceptable’ by the Commission in D ecember 2013. See Frances Robinson, ‘EU’s Almunia: Google’s Antitrust Proposals Are “Not Acceptable”’, Wall Street Journal (Europe edition), 20 December 2013. 46 As the Commission indicated, it usually ‘strives in commitment cases to address the competition concerns in such a way that parallel enforcement action by the national competition authorities within the territorial coverage of the Commission decision should in principle not be needed’. Commission Staff Working Paper accompanying the Report on the Functioning of Regulation 1/2003 SEC(2009) 574 final, para 106. See, however, the decision of the Spanish NCA finding a breach of the commitments made binding in Repsol (above n 36), which was a central element supporting the judgment of the General Court in Case T-342/11 Confederación Española de Empresarios de Estaciones de Servicio (CEEES) and Asociación de Gestores de Estaciones de Servicio v Commission, EU:T:2014:60, especially para 66. 47 See, eg Damien Geradin and David Henry, ‘The EC Fining Policy for Violations of Competition Law: An Empirical Review of the Commission Decisional Practice and the Community Courts’ Judgments’, 1 European Competition Journal 401 (2005). For a more recent account, see John Connor, ‘Has the European Commission Become More Severe in Punishing Cartels? Effects of the 2006 Guidelines’, 32 European Competition Law Review 27 (2011). 48 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, as confirmed by the EU Court of Justice in Joined Cases C-189/02 P, C-202/02 P, C-205/02 P to C-208/02 P and C-213/02 P Dansk Rørindustri and Others v Commission [2005] ECR I-5425, paras 260 and 292. 49 Article 23(2) of Regulation 1/2003. 50 See Commission Press Releases IP/09/745 of 13 May 2009 (Intel—€1.06 billion), IP/12/1317 of 5 December 2012 (TV and computer monitor tubes—€1.47 billion) and IP/13/1208 of 4 December 2013 (interest rate derivatives—€1.71 billion).

152 Damien MB Gerard ( million) 9414

10000 8000

5358

6000

3463

4000 2000 0

540 1990–1994

293 1995–1999

2000–2004

2005–2009

2010–2012

Figure 4: Fines: Cartels

It is tempting to link together the evolution in the amount of fines and the rise of commitment decisions, as if the underlying deterrence rationale had expanded its reach to the point of deterring firms from litigating cases in the first place and of creating an insuperable incentive to engage in negotiated procedures. However, that expansion has also been fuelled by a growing perception of the limits of the EU judicial review system, notably as it submits complex economic and technical appraisals inherent in modern substantive antitrust assessments to a so-called ‘manifest error’ standard, ie to limited review. Limited review has been a trademark of dominance cases in particular—an area in which the EU Courts have not overturned any Commission decision (except on minor ancillary points) in close to 15 years.51 However, similar considerations can be found in horizontal cases, including recent ones.52 Even though some Advocates General have endorsed—more or less openly—the need to move beyond the limited review standard in cases where the Commission has imposed a fine, these calls have so far had limited effect on the case law of the Court of Justice.53 The perception of a lack of alternative to negotiated solutions is equally supported by the outcome of cases involving comparable practices resolved

51 A prime example is the Microsoft case, where the General Court started the substantive part of its reasoning by stating that ‘in so far as the Commission’s decision is the result of complex technical appraisals, those appraisals are in principle subject to only limited review by the Court, which means that the Community Courts cannot substitute their own assessment of matters of fact for [that of the Commission]’. Case T-201/04 Microsoft Corp v Commission [2007] ECR II-3601, paras 88, 379 and 482. 52 See, eg Case T-111/08 MasterCard, Inc and Others v Commission, EU:T:2012:260, paras 82 and 201; appeals dismissed: Case C-382/12 P, EU:C:2014:2201. 53 See, in particular, Opinion of Advocate General Sharpston in Case C-272/09 P KME Germany AG v Commission, EU:C:2011:63, paras 60–70; Opinion of Advocate General Wathelet in Case C-295/12 P Telefónica SA v Commission, EU:C:2013:619, paras 107–145.

Negotiated Remedies in the Modernisation Era 153 either by means of a commitment decision or of an infringement decision, as illustrated in a stylised way in Figure 5:54 Siemens/Areva (2012) (Commitment proceedings)

Telefónica/Portugal Telecom (2013) (Infringement proceedings)

—

on-compete obligation (‘NCO’) N part of a nuclear power plants joint venture

—

on-compete obligation (‘NCO’) N included in a stock purchase agreement entered into between Telefónica and Portugal Telecom, whereby Telefónica acquired sole control over the Brazilian mobile operator Vivo and sold its stake in Portugal Telecom

—

CO not only covered the lifetime — N of the JV but, in its original form, was to continue for a period of 8 to 11 years after Siemens’ loss of joint control of the JV

CO was to apply for 15 months, N but the parties in fact removed the NCO after 4 months

—

he Commission came to the T conclusion that the post-JV NCO raised concerns as to its compatibility with Article 101, due to its broad scope and duration

—

he Commission found that the T NCO constituted an infringement by object

—

he Commission closed its T investigation pursuant to Article 9 by accepting commitments from the parties to set aside the NCO

—

he Commission fined Telefónica T €67 million and Portugal Telecom €12 million, imposed fines for a total of €79 million

Figure 5

In view of the above, adopting a litigious approach before the Commission with the hope of containing the exposure to fines and relying on the EU Courts to possibly overturn a finding of liability has become a notoriously risky option. Conversely, the loss in repressive utility resulting from the Commission’s acceptance of commitments appears relative, notably since it is able to maintain a high level of general deterrence by adopting a tough stance in infringement cases.55 The above analysis may have been candidly— if perhaps unwittingly—captured by former Director-General Italianer in a 54 See, respectively, Commission Decision of 18 June 2012 in Case COMP/39736— Siemens/Areva and Commission Decision of 23 January 2013 in Case AT.39839—Telefónica/ Portugal Telecom. Example suggested by Frédéric de Bure, counsel with Cleary Gottlieb LLP (Paris). 55 On the notion of ‘general deterrence’, see Guidelines on the method of setting fines imposed pursuant to Article 23(2)(a) of Regulation 1/2003, [2006] OJ C210/2, para 4.

154 Damien MB Gerard comparison between the choice of firms pondering the option of engaging in commitment proceedings and that of Hamlet in Shakespeare’s famous play, where the young Danish prince faces an agonising question: ‘“To be, or not to be”, he asks himself, and carefully weighs the merits of life and death. He chooses life, but for negative reasons: only because he fears the alternative may be worse.’56 In closing, as yet another illustration of the interrelationship between the different dimensions of modernisation, one may also formulate the hypothesis that the increase in enforcement/justification costs resulting from the advent of an effects-based approach associated with substantive modernisation may have paradoxically increased the incentives of both enforcement actors and firms to resort to negotiated procedures instead of engaging in protracted inquiries as to the actual impact of the practices in question on competition and welfare.57 That hypothesis finds support in the link suggested between the growing sophistication of economic theories relied upon in the application of antitrust rules and the turn towards a regulatory approach to competition law enforcement in the US. Already in the mid-1990s, Melamed contended that ‘[t]he role of economics has changed’ and The power of economic insights appears subtly to have promoted a kind of thought process in which government antitrust officials sometimes ask, not whether the defendant violated a legal norm, but whether constraining the defendant’s use of its property might improve or ensure its contribution to consumer welfare.58

In turn, the prospective nature of that analysis was viewed as regulatory in nature and ‘fundamentally nonlegal’, with the consequence that ‘[i]t can thus erode the rule of law and leave government officials largely unconstrained’, ‘especially when prospective analysis is combined with consent decree process’, because that process is subject to a ‘more limited judicial oversight’.59 Assuredly, these considerations echo concerns associated in the EU with the transformations induced by procedural modernisation.

1.2 Negotiated procedures and the focus on remedies At the EU level today, ‘[r]emedies lie at the core of competition law enforcement’.60 That move of remedies from the back to the fore of antitrust

Italianer, ‘To Commit or Not to Commit’, above n 27. Compare, eg Commission Decision of 19 December 2007 in Case COMP/34.579— MasterCard (infringement decision with fine) and Commission Decision of 8 December 2010 in Case COMP/39.398—Visa MIF (commitment decision). 58 Melamed, ‘The New Regulation’, above n 8, 15. 59 Ibid. 60 Italianer, ‘Legal Certainty’, above n 23. 56 57

Negotiated Remedies in the Modernisation Era 155 enforcement largely mirrors the shift from adversarial to negotiated enforcement. By nature, commitment proceedings focus on devising remedies rather than on establishing an infringement. The definition of remedies necessarily implies the pre-existing determination of anticompetitive concerns, ie a ( prima facie) theory of harm. However, in negotiated procedures, that theory of harm largely emerges from the discussion of possible remedies, as the Commission’s preliminary assessment is typically issued after a first (thorough) discussion of the envisaged commitments and some informal testing with third parties. In turn, focusing on remedies can bias the definition of the possible harm because it may lead to a looser (ie less specific) outcome, because it may be more vulnerable to capture by third parties and because sometimes remedies may not really be ‘negotiated’. In the end, all these factors can contribute to threatening the optimal nature of remedies.

1.2.1 The exposure of negotiated remedies to suboptimal outcomes In negotiated procedures, ‘[r]emedy design is not an after-thought at the end of a process that is abstracted from substance’; rather, ‘[i]t is intimately connected to substantive analysis’.61 Though remedies are never abstracted from substance, discussions on remedies in negotiated procedures actually assist in shaping the substance of the underlying concerns, notably as they contribute to the testing of substantive theories and possibly expose their weaknesses. To play that role, however, discussions on remedies need to be preceded by an agreement on the relevant set of facts, and they need to be conducted in a relatively stable legal environment that allows for open dialectic exchanges between the enforcer and the party or parties subject to the proceedings.62 Moreover, ascertaining the possible remedy and the underlying harm in parallel creates an incentive to settle on some ‘rough’ substantive theory the main merit of which is to match the scope of remedies inasmuch as they are effective, ie ‘simple, workable and easy to implement’.63 That incentive is reinforced by the loss of the possibility to convince the Commission to drop its case altogether once a firm decides to take the commitment route.64 Hence, the risk of a suboptimal outcome arising from the design of negotiated remedies appears heightened by the lack of robust pre-existing theory of harm.

61 Thomas Graf, ‘The Google Commitments—Testing Substantive Theories through Remedy Discussion’, kluwercompetitionlawblog.com, 4 July 2013. Mr Graf is Google’s lead counsel in pending cases COMP/C-3/39.740, COMP/C-3/39.775 and COMP/C-3/39.768. 62 See ECN Recommendation on Commitment Procedures. 63 Italianer, ‘Legal Certainty’, above n 23. 64 Italianer, ‘To Commit or Not to Commit’, above n 27. Interestingly, the C ommission has nonetheless closed proceedings in two cases after a set of commitments was s ubmitted to market testing. See Case COMP/37.749—Austrian Airlines + SAS (file closed on 13 March 2008); Case COMP/37.984—SkyTeam (file closed on 23 January 2012).

156 Damien MB Gerard The design of remedies in general, but even more so in a negotiated context, is also highly dependent on the feedback received in response to the market testing of the remedies in question. As Italianer highlighted, ‘[m]arket testing is a key tool which allows us to tailor the remedies to the competition concerns’.65 Admittedly, seeking the views of third parties is an important step in balancing the information asymmetry between firms and enforcers and, as a result, preventing the capture of the negotiation process. Naturally, third parties investing resources in actively responding to market test exercises are rarely motivated by purely benevolent interests. Likewise, even though a ‘market test is not an opinion poll which determines the fate of the remedies’,66 it still opens up space for building constituencies promoting s pecific business interests.67 The design of negotiated remedies, it is submitted, may be particularly vulnerable to the influence of such vocal interests for at least three reasons. First, as noted, discussions about the nature and scope of remedies do not take place against the background of a robust theory of harm but, rather, against preliminary concerns that are still in a state of flux, which creates a risk of over-reliance on the outcome of the market test, at least if the contributions to the market test are not properly disclosed to the firm proposing commitments. Second, considerations of procedural economy may also result in greater incentives to (push to) accommodate even loose concerns expressed by third parties in the hope that they will refrain from challenging the substance of the decision.68 Third, once constituencies have been created, they can also form powerful obstacles to the adaptation of commitments over time, whereas the possibility to reflect changing market contexts constitutes a commanding reason for favouring negotiated solutions, in particular when ‘concerns about competitive harm are real but inchoate or dependent on conduct and market dynamics’.69 Thus, for these reasons as well, negotiated procedures appear to be particularly exposed to settling on suboptimal remedies.

Italianer, ‘Legal Certainty’, above n 23. Manual of Procedure, above n 44, chapter 16, para 66. The Google case, for example, has given rise to the development of highly sophisticated coalitions such as FairSearch.org, which includes the likes of Microsoft, Oracle, TripAdvisor and Foundem, and ICOMP, which includes a diverse set of companies and organisations involved in internet commerce such as Mediaset, Premier League and the American Society of Media Photographers, but also Microsoft and Foundem. 68 See, eg James Kanter, ‘Europe Warns Google It Could Face Further Concessions’, New York Times, 28 May 2013, in particular: ‘Thomas Vinje, the chief lawyer for FairSearch Europe, a group of Google competitors including Microsoft, Nokia and Oracle, said he would be ‘very surprised’ if some companies did not bring an appeal’. 69 Yane Svetiev, ‘The Tools of Experimentalist Enforcement: Competition Policy as a Learning Platform’, on file with the author, 17. 65 66 67

Negotiated Remedies in the Modernisation Era 157

1.2.2 A biased negotiation The fact that they ‘are offered by undertakings on a voluntary basis’ appears to serve as a justification for all concerns expressed about the conduct and outcome of commitment proceedings.70 However, that alleged ‘voluntary’ character is hardly a guarantee of efficient remedial outcome, notably as the choice to engage in commitment proceedings and possibly to agree on a final set of remedies results from a mixture of factors that are largely detached from optimum considerations.71 Moreover, the scope of the freedom of choice to engage in commitment proceedings is a function of the available alternative(s) which, when it comes to EU antitrust enforcement, is/are highly hazardous. Once that choice is made, the voluntary character of the outcome of the ‘negotiation’ is even more dubious since, as the Commission insists, discussions on remedies ‘are not a bargaining process’ and can be discontinued at any moment.72 Possibly, the ‘voluntary’ nature of commitments allows one, at best, to assume that a firm would not compromise the essential features of its business and of the strategic development of that business, which is far from ensuring the efficiency of the remedies in question. In fact, there is a fundamental ambivalence in the discourse about negotiated procedures. On the one hand, competition authorities emphasise the voluntary nature of the process to justify a departure from established proportionality and due process requirements, while on the other hand, they openly acknowledge that accepting (or not) commitments belongs to their sole discretion and that commitment decisions remain ‘a unilateral act of State as opposed to a contract’.73 As a result, the Commission expressly admits that it may ‘make proposals during [commitment] discussions on how to modify certain elements of the text, and may even provide concrete drafting proposals on specific issues’,74 as it has done on a number of occasions.75 Yet, with

Commission Best Practices, para 117. As Marsden put it: ‘A dynamic exists where the threat of years of investigation with the significant legal and commercial costs, distraction of senior management, ongoing negative publicity, uncertainty, and possibility of huge fines combine to make defendants particularly prone to offer commitments as a practical matter, no matter what the theory of harm may be or allegations they are facing’. Philip Marsden, ‘The Emperor’s Clothes Laid Bare: Commitments Creating the Appearance of Law, While Denying Access to Law’, 10(1) CPI Antitrust Chronicle (October 2013) 1, 3. 72 Italianer, ‘Legal Certainty’, above n 23. 73 See, eg Commission Best Practices, para 115; OFT Enforcement Guidelines, 11; Notice on Competition Commitments issued by the French NCA on 2 March 2009, para 42. 74 Manual of Procedure, above n 44, chapter 16, para 43. 75 In the De Beers/Alrosa case, for example, the Commission candidly acknowledged that it had ‘suggest[ed] amendments to the proposed commitments’ that would have resulted in the termination of all trading relations between Alrosa and De Beers. See De Beers, above n 6, Recital 42. 70 71

158 Damien MB Gerard the discretion to accept commitments reflecting its own requirements,76 it is submitted, comes the responsibility on the part of the Commission to ensure that the remedies made binding ‘unilaterally’ on the firm subject to the proceedings are indeed optimal. As further developed in Section 2 below, reaching that optimal outcome is fundamentally dependent on the design of the negotiation process. As the European Court of Justice crudely presented it in Alrosa, the alternative for firms may well be to either: (i) ‘consciously accept that the concessions they make may go beyond what the Commission could itself impose on them in a decision adopted under Article 7 of the regulation after a thorough examination’; or (ii) ‘a finding of an infringement of competition law and a possible fine’.77 The illusory character of that ‘alternative’ is magnified by the fact that, when the Commission judges that ‘cooperation with the companies is not satisfactory’, it may ‘always revert to the prohibition path’,78 which it is not shy to reassert publicly in relation to specific cases.79 Although that possibility aims to secure the incentive of firms to cooperate, it largely negates the voluntary and consensual nature of the process, especially if it leaves the Commission with the possibility to impose fines, and is therefore no guarantee of optimal remedial outcome. *

*

*

This first section sought to explain how the procedural modernisation process driven by effectiveness considerations contributed to moving remedies to the core of antitrust adjudication in the European Union as a result of a shift towards ‘negotiated’ enforcement largely resulting from the attractiveness of commitment proceedings. However, effectiveness is inapt as such to ensure the design of efficient remedies, it was submitted, as negotiated procedures are particularly vulnerable to suboptimal outcomes. N otably, they are not conducted on the basis of a robust pre-existing theory of harm; they are particularly dependent on third-party comments formulated in response to market tests; and they are de facto not voluntary. Building on these premises, the second section further inquires into the limits of effectiveness for delivering welfare-maximising remedies, and it reflects on ways to overcome these limits with a view to reconciling effectiveness and efficiency considerations in the design of remedial outcomes.

76 See ECJ in Alrosa, above n 6, para 94: ‘It follows from Article 9(1) of Regulation No 1/2003 that the Commission has a wide discretion to make a proposed commitment binding or to reject it’. 77 Ibid, para 48. 78 Almunia, ‘Remedies, Commitments and Settlements’, above n 21. 79 As reported by, eg Kanter, ‘Europe Warns Google’, above n 68.

Negotiated Remedies in the Modernisation Era 159

2. The limits of effectiveness In the modernisation context, as noted, effectiveness can be defined as an attempt to maximise the allocation of enforcement resources with a view to strengthening compliance on the part of economic actors and the societal impact of competition policy. As such, effectiveness has been a main driver of the move towards negotiated enforcement and the increasing reliance in the EU on commitment decisions to resolve antitrust cases. Indeed, commitments have been praised for their ability to bring about a ‘swift resolution of concerns’ and to ‘expedite market changes’, while involving ‘fewer resources’.80 These benefits are real, and so are the advantages that firms perceive by engaging on the negotiated route. Still, while they are not an a priori evil that ought to be repealed at all costs, negotiated procedures have profoundly modified the respective incentives of both enforcers and firms, to the point where they give rise to a ‘culture of consent’ in the EU.81 As in the US, but without the same safeguards built into the overall enforcement framework, that culture has many side effects. Notably, the culture of consent is not naturally geared towards the achievement of optimal solutions.82 This is because, in a nutshell, the maximisation by competition authorities of their own administrative resources by resorting to negotiated outcomes, whether in the selection of cases, or in the conduct of proceedings and the design of remedies, does not square with the maximisation of welfare. In turn, the culture of consent reveals a tension between procedural and substantive modernisation, which needs to be addressed. After elaborating on the consequences induced by the culture of consent on the role and design of remedies, including in terms of legitimacy, this second section explores ways to overcome the tension between procedural and substantive modernisation and to reconcile the maximisation of enforcement resources with the promotion of competition law as a tool of welfare maximisation, ie effectiveness with efficiency. To that end, it revisits the notions

80 Italianer, ‘To Commit or Not to Commit’, above n 27; Commission Staff Working Paper accompanying the Report on the Functioning of Regulation 1/2003 SEC (2009) 574 final, para 102. 81 This expression is borrowed from Ginsburg and Wright, ‘Antitrust Settlements’, above n 20, 56. 82 For a similar concern expressed in relation to commitment proceedings, see, eg F lorian Wagner-von Papp, ‘Critical Considerations on the Commission’s Commitment to the Commitment Procedure’, 3(3) CPI Antitrust Chronicle (March 2013) 1, 5 (‘Just as the negotiations cannot guarantee that the commitments are proportionate, they do not necessarily guarantee that the commitments are adequate. The Commission does not have the benefit of a full investigation into the facts, and may inadvertently even agree to commitments that do themselves have anticompetitive effects. The dynamics of commitment negotiations may make them more focused on achieving an agreement than on safeguarding competition’).

160 Damien MB Gerard of proportionality and due process, which have historically governed the achievement of legitimate enforcement outcomes, and questions their interpretation and relevance in a ‘modernised’ context. In turn, while both proportionality and due process appear appropriate to mediate the effectiveness/ efficiency conundrum, it focuses on due process as a possible procedural proxy for proportionality and puts forward policy proposals to adjust the sequence of commitment proceedings with a view to addressing its current ‘systemic’ shortcomings. Ultimately, reshaping the incentives of enforcers and firms so as to steer negotiated procedures towards the achievement of optimal solutions also requires considering the broader EU antitrust enforcement framework, including the incentives to resort to the normal infringement procedure and the conditions for ensuring that it remains a sustainable alternative to a negotiated procedure.

2.1 Effectiveness and the design of negotiated remedies As noted, effectiveness considerations driving modernisation have led to the development of a negotiated form of enforcement focusing on devising remedies rather than on establishing infringements. This is particularly the case for commitment proceedings, which aim to resolve cases by negotiating remedies to address concerns expressed succinctly by the Commission in a ‘preliminary assessment’. In view of their success, commitments embody a shift in the focus of EU antitrust enforcement from the wrong to the remedy. This section inquires into the significance of that shift for the nature and function of remedies at the EU level and, subsequently, for the legitimacy of EU antitrust law enforcement in general.

2.1.1 From corrective to regulatory remedies Historically, competition law remedies in the EU have been associated with the termination of infringements, ie with the Commission’s power to ‘require the undertakings or association of undertakings concerned to bring such infringement to an end’.83 Thus, as the Court of Justice held in Commercial Solvents, the Commission may, ‘in relation to the infringement which has been established’, order the infringing firm to ‘do certain acts or provide certain advantages which have been wrongfully withheld as well as prohibiting the continuation of certain action, practices or situations which are contrary to

83 Article 3(1) of Regulation 17/62: First Regulation implementing Articles 85 and 86 of the Treaty 1962 OJ 13/204.

Negotiated Remedies in the Modernisation Era 161 the Treaty’.84 Remedies have therefore been essentially conceived as means to ensure corrective justice, ie to allow for the ‘reestablishment of compliance with the rules infringed’85 or, stated otherwise, to ‘re-establish the situation which existed prior to the infringement’.86 On that basis, the EU Courts have sanctioned the imposition of remedies that were found to go beyond the restoration of the status quo ante, notably as they were designed to prevent the possible recurrence of the infringement in the future.87 Interestingly, Regulation 1/2003 follows the same approach by expressly granting the Commission the ‘power to impose any remedy, whether behavioural or structural, which is necessary to bring the infringement effectively to an end’.88 However, modernisation has modified the nature and function of r emedies. Substantive modernisation, first of all, by promoting a more contextual approach to antitrust enforcement driven by efficiency considerations, naturally calls for the design of remedies challenging the prevailing market structure and aiming to minimise the recurrence of the observed anticompetitive conduct.89 Moreover, as Lianos has suggested, in an economically oriented enforcement framework, infringements rely on theories of consumer harm that may ‘not only relate to the structure of the supply side but may also be generated by the specific characteristics of the demand side’,90 thereby requiring remedies that may be less strictly related to the wrongful conduct identified. More fundamentally, procedural modernisation has entailed a profound revamp of the EU approach to remedies because commitment decisions, in essence, do not ‘conclud[e] whether or not there has been or still is an infringement’,91 ie they do not establish a wrong. Rather, they aim at

84 Joined Cases 6 and 7-73 Istituto Chemioterapico Italiano SpA and Commercial Solvents Corporation v Commission [1974] ECR 225, para 45. 85 Joined Cases C-241 and 242/91 P Radio Telefis Eireann (RTE) and Independent Television Publications Ltd (ITP) v Commission (Magill) [1995] ECR I-808, para 93. For the US equivalent, see, eg Ford Motor Co v United States, 405 US 562, 577 (1972) (‘[a]ntitrust relief should unfetter a market from anticompetitive conduct’). 86 Case T-24/90 Automec Srl v Commission [1992] ECR II‑2223, paras 51–52. 87 See, eg Case T-395/94 Atlantic Container Line AB and Others v Commission [2002] ECR II-875, paras 398 et seq; Case T-310/94 Gruber + Weber GmbH & Co KG v Commission [1998] ECR II-1043, paras 177–178; Case T-7/93 Langnese Iglo GmbH v Commission [1995] ECR II-1533, para 205. 88 Regulation 1/2003, Recital 12 and Article 7. 89 As Mariniello puts it: ‘Despite the criticism of proactive remedies in the legal literature, they might be justifiable on economic grounds’ because proactive remedies may ‘increase competition and therefore potentially lead to higher welfare levels for consumers’. Mario Mariniello, ‘Commitments or Prohibition? The EU Antitrust Dilemma’, Bruegel Policy Brief No 2014/1, January 2014, 6. 90 Ioannis Lianos, ‘Competition Law Remedies in Europe—Which Limits for R emedial Discretion?’, CLES Research Paper Series 2/2013, 64. Published as ‘Competition Law Remedies in Europe: Which Limits for Remedial Discretion?’, in Ioannis Lianos and Damien Geradin, eds, Handbook of EU Competition Law: Enforcement and Procedure, Edward Elgar, 2013, 362 et seq. 91 Regulation 1/2003, Recital 13.

162 Damien MB Gerard addressing concerns and at altering competitive interactions with a view to preventing harm to competition. Hence, the purpose of commitments is less corrective than prescriptive in kind, and is indeed of a ‘prospective nature’.92 As noted, the increasing reliance on consent decrees in the US was also viewed as a move away from a law-enforcement model towards an approach of antitrust enforcement as a means of ‘prospective regulation’, whereby ‘antitrust enforcers no longer confine their inquiry to whether the defendant has violated the law’ and how to redress wrongs.93 Thus, the ‘widespread use of consent decrees’, as Melamed contended, ‘dispenses with the adjudication process and therefore sometimes entails both remedies for conduct that might not be unlawful and remedies that go beyond law enforcement purposes’.94 The surge in commitment decisions entailed in various ways a paradigm shift from a corrective towards a regulatory approach to the design of remedies in the EU, which also appears to extend beyond negotiated procedures.95 A first indication of the transformative nature of commitment decisions lies in the fact that, unconstrained by the limitations set forth by Article 7 of Regulation 1/2003 with respect to infringement decisions,96 they have enabled the Commission to impose structural remedies.97 These structural remedies cases invariably involved the energy sector and are widely perceived as being supported by regulatory motives, namely a willingness to design new equilibriums in the markets in question.98 Commenting on the obligation made binding on RWE to divest its transmission grid and on ENI to dispose of its stake in certain international pipelines, Director-General Italianer openly indicated that these remedies ‘contributed to effectively [opening up] energy markets to competition, in addition to the regulatory provisions fostering liberalisation’, as they ‘ensured that the abuse could never be repeated and created the conditions for undistorted competition downstream’.99

Opinion of Advocate General Kokott in Alrosa, above n 6, para 74. Melamed, ‘The New Regulation’, above n 8, 13. 94 A Douglas Melamed, ‘Antitrust as Law Enforcement’, 27 Antitrust 76, 77 (2012). 95 This is not to say that competition law remedies did not exhibit regulatory features in the past or, in general, to deny the application of competition law as regulation in certain economic contexts. For a compelling discussion not limited to the antitrust context, see Pablo Ibáñez Colomo, ‘On the Application of Competition Law as Regulation: Elements for a Theory’, 29 Yearbook of European Law 261 (2010). 96 Compare Article 7 of Regulation 1/2003, which limits the imposition of structural remedies to situations ‘either where there is no equally effective behavioural remedy or where any equally effective behavioural remedy would be more burden some for the undertaking concerned than the structural remedy’. 97 See, in particular, German Electricity Markets, RWE Gas Foreclosure, ENI and CEZ, all above n 40. 98 For a thorough discussion, see, eg François Marty, ‘Une régulation du secteur de l’énergie au travers des procedures d’engagements? Réflexions sur le contentieux concurrentiel européen’, 26–27 économie publique 93 (2011). 99 Italianer, ‘Legal Certainty’, above n 23. 92 93

Negotiated Remedies in the Modernisation Era 163 Secondly, the Commission has relied on commitments to stretch the boundaries of prevailing antitrust legal standards and, in effect, to tackle practices that could not be comfortably addressed under these standards. Examples include the notoriously complex question of excessive prices, whether in the form of ‘abusive’ royalty rates or ‘unfairly high’ licensing fees for standards or identifiers codes, which commitment decisions endeavoured to set at a c ertain level.100 Similarly, by challenging agency agreements and most favoured nation clauses, which had historically been considered positively under antitrust standards, the Commission sought a ‘competitive reset’ of the launch of e-book retail sales in the European Union.101 The ongoing Google case also challenges the boundaries of antitrust standards, notably with respect to socalled ‘scraping’ practices, ie the ‘use by Google without consent of original content from third party websites in its own vertical Web search services’.102 A willingness to shape the future development of online search interfaces also transpires from the successive rejection of commitments aimed at addressing vertical search concerns;103 Commissioner Almunia subsequently acknowledged that ‘a decision making commitments binding on Google will be similar in effects to a regulation of Google’s activities’.104 All these cases, and others,105 display a tendency to go beyond established antitrust standards and

100 Commission Decision of 9 December 2009 in Case COMP/38.636—Rambus [2010] OJ C30/17; Commission Decision of 15 November 2011 in Case COMP/39.592—Standard & Poor’s [2011] OJ C31/8. 101 Commission Decision of 25 July 2013 in Case AT.39847—E-Books, paras 108 and 117. 102 See Communication from the Commission published pursuant to Article 27(4) of Council Regulation 1/2003 in Case AT.39740—Google [2013] OJ C120/22, para 3. 103 See, eg Joaquín Almunia, ‘The Google Antitrust Case: What is at Stake?’, SPEECH/13/768, Brussels, 1 October 2013: ‘Many respondents during the market test said that in this Google proposal the links to rivals that would be displayed for certain categories of specialised search services were not visible enough. In my opinion, the new proposal makes these links significantly more visible. A larger space of the Google search result page is dedicated to them. Rivals have the possibility to display their logo next to the link, and there will be a dynamic text associated to each rival link to better inform the user of its content.’ 104 Joaquín Almunia, ‘Concurrence et croissance pour l’après-crise’, SPEECH/14/150, 5ème Conférence, New Frontiers of Antitrust, Paris, 21 February 2014 (free translation of ‘une décision avec des engagements contraignants aura des effets similaires à ceux d’une régulation des activités de Google’). 105 Ibáñez Colomo observed similar features in cases such as Premier League (Commission Decision of 22 March 2006 in Case COMP/38.173, [2008] OJ C7/18) and the 2009 Microsoft (Tying) decision (Commission Decision of 16 December 2009 in Case COMP/39.530, [2013] OJ C120/15). In the former, the Commission ‘did not explain why the “single buyer rule” was necessary to bring the joint selling of football rights in line with Article 101 TFEU’ and ‘why the case was different from similar joint selling cases decided in previous years and in which the remedy was not required’. Ibáñez Colomo, ‘Elements for a Theory’, above n 95, 306. With respect to the latter, the corrective measures obtained in 2009 went further than the ones imposed in the 2004 prohibition decision despite the fact that the context appeared less serious. See ibid.

164 Damien MB Gerard an ambition to recast market equilibriums in accordance with certain policy objectives. Thirdly, the regulatory nature of the remedial approach adopted in commitment decisions is further illustrated by the acknowledged convergence with the approach adopted in merger control cases. While merger control is casually referred to as ‘the [closest] a competition law procedure can come to a regulatory law one’,106 the Commission has revealed its practice of applying ‘the same guiding principles’ to the design of remedies in both merger control and commitment cases, the latter being presented as reflective of the Commission’s general practice in the antitrust field.107 Although informative, the assimilation of merger control and antitrust proceedings is controversial, as the nature of what is to be remedied is essentially different. Merger control involves the ex ante assessment of the welfare effects of a concentration, while antitrust enforcement is supposed to prohibit ex post commercial practices qualified as restrictive according to principles derived from the o pen-textured nature of Articles 101 and 102 TFEU. This difference translates into the different nature of these two kinds of proceedings, of the standards applied and of the risks involved. In turn, the stated convergence between the merger control and antitrust areas reveals the emergence of a regulatory approach beyond the design of remedies, in the interpretation of antitrust rules centred on ‘whether the defendant’s conduct maximises consumer welfare or otherwise serves the public interest’.108 The convergence of merger control and commitment proceedings towards a regulatory approach to the design of remedies may also be apparent from the focus on ‘remedies that are simple, workable and easy to implement’,109 ie primarily effective. This is understandable because the credibility of the Commission’s intervention in both types of cases is conditioned on the successful implementation of remedies.110 Yet the increased remedial discretion accompanying that claim for effective remedies

Lianos, ‘Which Limits’, above n 90, 66. Italianer, ‘Legal Certainty’, above n 23. Melamed, ‘The New Regulation’, above n 8, 13. 109 Italianer, ‘Legal Certainty’, above n 23. 110 With respect to commitments, see the Manual of Procedure, above n 44, chapter 16, para 82. See also the €561 million fine imposed on Microsoft for failure to comply with its commitment to give Windows users a choice between several web browsers (Commission Decision of 16 December 2009 in Microsoft (Tying), above n 105), which was justified by Commissioner Almunia by the ‘duty to preserve the integrity and credibility of the system’ (see Joaquín Almunia, ‘Remedies, Commitments and Settlements’, above n 21). For national cases sanctioning the breach of commitments, see: (i) Decision of 23 January 2013 of the Spanish NCA in Case S/0020/07 TRIO PLUS (fine of €188,646 imposed on Prisa Television and Telefonica for failing to respect commitments to terminate joint distribution agreements for pay-TV and electronic communication services); and (ii) Decision No 11-D-10 of 6 July 2011 of the French NCA in Case 08-D-34, Pompes funèbres Marseille (fine of €60,000 imposed on the city of Marseille for breach of a commitment relating to the exchange of sensitive information between funeral service operators). 106 107 108

Negotiated Remedies in the Modernisation Era 165 in commitment proceedings is bound to be more controversial, as commitments imply that ‘even if only on a preliminary basis … there may be or may have been an infringement of EU competition law which should be addressed’ and thus to which remedies must directly relate.111

2.1.2 The effectiveness conundrum and the legitimacy gap The shift from a corrective to a regulatory approach in the design of remedies at the EU level, as induced by the widespread recourse to commitment decisions, has far-reaching consequences for the interpretation of antitrust principles and for the legitimacy of antitrust enforcement. Indeed, while becoming the default mechanism to enforce antitrust principles in the EU, commitment decisions have become a primary source of guidance on the interpretation of these very principles. As noted, however, commitment decisions tend to stretch the boundaries of antitrust legal standards with the paradoxical effect of affecting the overall predictability of the scope of Articles 101 and 102 TFEU. While premised on a search for greater enforcement effectiveness, negotiated procedures therefore also affect the substantive effectiveness of antitrust principles, ie their ability to be intelligently grasped and accurately internalised by firms. In turn, inasmuch as it leads to suboptimal outcomes, that effectiveness conundrum affects the (output) legitimacy of antitrust law as a means geared towards the achievement of efficiency gains and the maximisation of consumer welfare. This section aims to substantiate that conundrum by discussing the precedential value of commitment decisions and then to articulate the paradox arising therefrom, before exploring ways to address the legitimacy gap that it creates. In theory, even though ‘the Commission must have at its disposal sufficient facts to make an informed and sound assessment of the relevant competition concerns’, commitment decisions ‘are not based on full investigations and do not reach definitive conclusions on the facts of a case or the application of the law’.112 Their precedential value is therefore ‘different … c ompared to final decisions under Article 7’ of Regulation 1/2003, ie infringement decisions, and yet they do provide illustrative guidance, as ‘[o]ther market participants can learn from the decision and the commitments what was considered by the Commission sufficient to remove the competition concerns’; in this sense, they ‘serve to clarify the Commission’s competition policy’.113

Manual of Procedure, above n 44, chapter 16, para 18. Ibid, chapter 16, paras 7 and 18. Interestingly, according to Bruegel estimates, ‘[f]or decisions published since May 2004, the average length of commitment decisions is 21 pages compared to 160 for prohibition decisions’. Mariniello, ‘Commitments or Prohibition?’, above n 89, 7. 113 Manual of Procedure, above n 44, paras 6 and 11. 111 112

166 Damien MB Gerard Thus, while not entailing ‘definitive conclusions’ on the application of antitrust principles, commitment decisions at least inform firms of ‘sufficient’, but not necessarily optimal, ways to comply with these principles. In practice, however, the Commission’s staff has publicly acknowledged that a commitment decision can also ‘serve as a model for addressing similar situations’.114 The Distrigaz decision, for example, is said to have ‘set out guidance on the Commission’s approach towards foreclosure by long-term contracts’.115 Moreover, after adopting the Premier League decision, the Commission apparently ‘considered that further action in this field should not be a priority for it as the existing commitment decisions appeared to provide sufficient orientation to operators and national competition authorities to deal with domestic media markets effectively and consistently’.116 This is particularly telling because the Premier League commitments were considered to include more stringent requirements than previous decisions adopted pursuant to Article 101 TFEU on the joint selling of broadcasting rights, notably by providing for a so-called ‘single buyer rule’.117 The precedential value of commitment decisions also extends to proceedings before NCAs; the four commitment decisions adopted in relation to the disclosure of technical information by automobile manufacturers,118 for instance, directly guided the commitments offered by Citroën in a French procedure.119 The precedential value of commitment decisions must also be assessed against the paucity of infringement decisions adopted by the Commission in recent years, irrespective of the fact that ‘prohibition decisions set a stronger precedent for future cases’.120 Commentators have been particularly concerned by the deficit of ‘carefully investigated and reasoned decisions’, in particular in very dynamic sectors of the economy.121 ‘When commitments decisions espouse novel theories of harm in fast-moving markets,’ it has been suggested, ‘they create important precedents, considered relevant by the industry as a whole who otherwise have little direct relevant case law or Commission

114 Commission Staff Working Paper accompanying the Report on the Functioning of Regulation 1/2003 SEC(2009) 574 final, para 109. 115 Ibid. See also Commission Decision of 11 October 2007 in Case COMP/37.966— Distrigaz [2008] OJ C9/8. 116 Commission Staff Working Paper, above n 114, para 109. See also Premier League, above n 105. 117 According to the ‘Single Buyer Rule’, no single bidder can be awarded all live coverage packages on an exclusive basis. 118 See Commission Decisions of 14 September 2007 in Case COMP/39.140— Daimler Chrysler [2007] OJ C317/76; Case COMP/39.141—Fiat [2007] OJ L332/77; Case COMP/39.142—Toyota [2007] OJ L329/52; Case COMP/39.143—Opel [2007] OJ L330/44. 119 See Decision of the Conseil de la concurrence No 07-D-31—Citroën, 9 October 2007. 120 Almunia, ‘Remedies, Commitments and Settlements’, above n 21. 121 Ken Daly, ‘A Plea for Plea Bargaining—Closing the Gaps between the EU’s Leniency, Settlement, and Commitments Procedures’, 3(3) CPI Antitrust Chronicle (March 2013) 1, 6.

Negotiated Remedies in the Modernisation Era 167 guidance.’122 Hence, ‘rules can end up being set for an industry based only on case-specific facts and the interactions of a case team, a defendant, and at most some self-interested third parties’.123 Conversely, others have suggested that the commitment procedure is precisely ‘adapted to situations where there exists some uncertainty about the law or the character of the conduct and its effects’ and therefore about ‘the existence or seriousness of a violation’; by refraining from finding an infringement, the Commission expressly refrains from ‘dressing up a decision with too much certainty’ but, rather, engages in a collaborative learning process designed to minimise competition concerns in the future and allowing for adjustments over time.124 While this is indeed a convincing rationale in favour of commitment proceedings, the overwhelming reliance upon them, including in cases that do not fit that pattern or do not allow for adjustments, results in a deficit of certainty that is only partly compensated for by the multiplication of guidance documents, which typically do not address hard issues. Moreover, commitment decisions are rarely subject to appeal, which means that the decisional practice of the Commission is generally not tested in court anymore. This results in a lack of guidance by the EU Courts both for companies and the Commission (and NCAs).125 Eventually, an effectiveness paradox emerges. The promotion of negotiated procedures as part of a utility-maximising approach to competition law enforcement was designed to increase the effectiveness of enforcement. However, insofar as it leads to negotiated procedures becoming the default enforcement mechanism, that approach has the reverse effect of blurring the contours of the law and reducing the predictability of antitrust principles, thereby leading to a loss in (substantive) effectiveness. The paradox comes full circle once one considers that modernisation as a whole was premised on the stabilisation of competition law principles. So a question of principle emerges: can the (over-)promotion of (enforcement) effectiveness entail a destruction of the law itself ? Conversely, how can one ensure that increased (enforcement) effectiveness is sustainable over time? These questions are all the more acute given that the effectiveness conundrum so identified also raises questions about the (output) legitimacy of antitrust, and specifically about its ability to deliver welfare-maximising outcomes. Indeed, while b lurring the boundaries of antitrust principles, commitment decisions create a p arallel

Philip Marsden, ‘The Emperor’s Clothes’, above n 71, 4. Ibid. 124 Svetiev, above n 69, 5 and 14–17. 125 For a similar concern, see also Daly, ‘A Plea for Plea Bargaining’, above n 121, 6. For Mariniello, the lack of judicial review also results in ‘less of an incentive for accuracy in the [substantive antitrust] assessment’. Mariniello, ‘Commitments or Prohibition?’, above n 89, 4). 122 123

168 Damien MB Gerard ‘suboptimal’ case law126 based on ‘preliminary concerns’ and case-specific ‘sufficient’ remedies, which turns out to be a main source of guidance available to firms (and counsel). Yet the continued acceptance of negotiated outcomes and thus the sustainability of commitments as an effective enforcement mechanism requires the ‘negotiation’ to take place in the shadow of the law, ie of sufficiently established and robust principles. Conversely, closing the legitimacy gap would also address the effectiveness conundrum, it is submitted, because reconciling effectiveness with efficiency would also improve the precedential value of commitment decisions and therefore the predictability of antitrust principles.

2.2 Conditions for legitimate (and efficient) negotiated remedies In the EU, the legitimacy of antitrust enforcement and remedial action has historically been guaranteed by concepts known as proportionality and due process. On the one hand, the principle of proportionality entails that action by the Commission may not go beyond what is appropriate and necessary to achieve the objective sought.127 On the other hand, due process materialises in antitrust proceedings by the recognition and exercise of the right to be heard in response to objections (or concerns) stated in writing by the Commission and by the associated right of access to file in the preparation of that response, as well as of the right to the judicial review by the EU Courts of any adverse decision.128 These two concepts are complementary because due process requirements essentially enable the Commission to identify the relevant issues and to properly tailor its analysis,129 while proportionality ensures that the substance of the adversarial process is translated into the outcome of the case. However, negotiated procedures, and commitment proceedings in particular, entail a relaxation of the substance of both concepts, thereby limiting their ability to ensure the legitimacy of such procedures and to address the effectiveness conundrum identified above. This section discusses the application of the proportionality and due process concepts in commitment

126 In this sense, see also the contributions in Claus-Dieter Ehlermann and Mel Marquis, eds, European Competition Law Annual 2008: Antitrust Settlements, Hart Publishing, 2010. 127 Proportionality is recognised by the EU Courts as a general principle of EU law and applies to the exercise of the Union’s competences pursuant to Article 5 TEU. Generally, Article 49(3) of the EU Charter of Fundamental Rights provides that ‘[t]he severity of penalties must not be disproportionate to the criminal offence’. 128 Due process requirements derive from Article 6 of the European Convention of Human Rights and Fundamental Freedoms and Articles 41(2) and 47 of the EU Charter of Fundamental Rights, as interpreted by the Court of Justice and implemented by the EU legislator in, eg Article 27 of Regulation 1/2003. 129 See, eg Case T-15/02 BASF AG v Commission [2006] ECR II-497, paras 44 et seq.

Negotiated Remedies in the Modernisation Era 169 roceedings and explores ways to address their shortcomings. Specifically, it p suggests that a solution to the current state of affairs lies in the strengthening of the input legitimacy of the commitment procedure and therefore largely depends on a review of the applicable procedural safeguards.

2.2.1 Proportionality inquiry It is settled case law that, in infringement proceedings, the principle of proportionality requires that the burdens [ie remedies] imposed on undertakings in order to bring an infringement to an end do not exceed what is appropriate and necessary to attain the objective sought, namely re-establishment of compliance with the rules infringed.130

Appropriateness requires remedies to be capable of curing the infringement and restoring compliance, while the necessity condition implies that, when facing a choice between several potentially appropriate measures, the Commission must opt for the least onerous one.131 Hence, p roportionality has historically been an important constraint on the Commission’s remedial discretion. As noted, the EU Courts have in the past annulled (part of) remedies that were found to go beyond—ie were not necessary for—the restoration of the status quo ante.132 One should not overstate, though, the level of scrutiny exercised by the EU Courts when reviewing the Commission’s proportionality assessments, notably in view of the applicable judicial review standard.133 Moreover, aside from the fact that proportionality is not easily justiciable in the absence of full appellate jurisdiction on the part of the EU Courts, it is inherently linked to a corrective approach to remedies because it implies the existence of a clear benchmark against which to define whether a particular remedial measure is appropriate and necessary, ie typically an infringement. In contrast, proportionality is less suited to a utility-maximising approach to enforcement and, specifically, to effectiveness-driven negotiated procedures that exclude a finding of infringement and involve instead the remediation of ‘concerns’, most of the time briefly articulated. The application of the proportionality principle to remedies made binding by means of commitment decisions was central to the Alrosa case, which

130 Microsoft, above n 51, para 1276. See also, eg Joined Cases T-25, 26, 30 to 32, 34 to 39, 42 to 46, 48, 50 to 65, 65, 68 to 71, 87, 88, 103 and 104/95, Cimenteries CBR and Others v Commission [2000] ECR II-491, para 4705. 131 RTE and ITP v Commission, above n 85, para 93. 132 See above note 87. 133 The EU Courts exercise a control of legality over Commission antitrust decisions and apply a ‘manifest error’ review standard to so-called ‘complex technical appraisals’. See above note 51.

170 Damien MB Gerard involved the termination of all trading relations—direct and indirect, and for an indefinite duration—between Alrosa and De Beers, the two largest producers of rough diamonds in the world, against concerns that previous supply agreements between the two firms had contributed to entrench De Beers’ dominant position on the rough diamonds market and its ‘marketmaker role’.134 The General Court annulled the Commission Decision because the commitments made binding on De Beers were disproportionate,135 but the Court of Justice reversed that annulment.136 In essence, the General Court stuck to a corrective approach and held that, when making commitments binding on firms, the Commission had to opt for the least onerous appropriate measure known to it and ‘[could] not lawfully propose to the parties that they should offer it commitments which go further than a decision which it could have adopted under Article 7(1) of Regulation No 1/2003’.137 Moreover, by denying the necessity of the remedy, the General Court indirectly questioned the theory of harm put forward by the Commission, as the mere holding of a dominant position does not infringe Article 102 TFEU if there is no abuse.138 The Court of Justice clearly took a different approach, as it validated the Commission’s position and thereby stretched significantly the boundaries of Article 102 TFEU while adopting a regulatory approach to remedies. The starting point of the Court of Justice’s reasoning was that proportionality applies to ‘any act of the institutions of the Union’, including commitment decisions, but that ‘the precise extent and limits of the obligations which flow from the observance of that principle’ vary according to the nature of the proceedings, ie whether in infringement (Article 7) or commitment (Article 9) proceedings.139 In turn, the Court found that Application of the principle of proportionality by the Commission in the context of Article 9 of Regulation No 1/2003 is confined to verifying that the commitments in question address the concerns it expressed to the undertakings concerned and that they have not offered less onerous commitments that also address those concerns adequately.140

As a result, not only is the necessity prong of the proportionality test virtually inapplicable, the control of the appropriateness of the remedies is also reduced

De Beers, above n 6, Recital 28. See the General Court in Alrosa, above n 6. 136 See the ECJ in Alrosa, above n 6. 137 As mentioned earlier (note 75), the Commission acknowledged that it had ‘suggest[ed] amendments to the proposed commitments’ that would have brought all trading relations between Alrosa and De Beers to an end. See De Beers, above n 6, Recital 42. 138 The case should arguably have been dealt with under Article 101 TFEU, but Alrosa was not ready to provide the commitment sought by the Commission. 139 Alrosa, above n 6, paras 36–37. 140 Ibid, para 41. 134 135

Negotiated Remedies in the Modernisation Era 171 to a minimum because, the Court added, ‘[j]udicial review for its part relates solely to whether the Commission’s assessment is manifestly incorrect’.141 In other words, the principle of proportionality would be breached only if ‘the Commission’s conclusion [were] obviously unfounded, having regard to the facts established by it’.142 The justifications for doing away with that key constraint on ‘discretionary remedialism’ are twofold:143 (i) effectiveness, ie the fact that the commitment procedure ‘is based on considerations of procedural economy’, so that,144 as suggested by Advocate General Kokott, ‘the general interest in finding an optimum solution from the point of view of speed and procedural economy justifies restricting the choice of possible measures in the context of Article 9 of Regulation 1/2003’;145 and (ii) the voluntary nature of commitments, which implies that firms ‘consciously accept that the concessions they make may go beyond what the Commission could itself impose on them in a decision adopted under Article 7 of the regulation after a thorough examination’.146 Ultimately, the Court of J ustice upheld the unlimited prohibition of all trading relations between Alrosa and De Beers as an adequate remedy to a situation of an essentially contractual nature which cannot easily be characterised as abusive under Article 102 TFEU. Reducing the necessity prong of the proportionality principle to the least onerous option among the ‘adequate’ ones proposed by a firm is conceptually understandable in view of the fact that commitments are indeed, at least formally, proposed by the firm(s) in question. However, it becomes problematic when considered against the lenient review standard applicable and therefore the wide discretion left to the Commission in determining whether a set of commitments is adequate or not. Moreover, given the particularly unattractive alternative of long and costly infringement proceedings,147 the allegedly voluntary character of commitments is at most a safeguard against the possible ‘obviously unfounded’ nature of the Commission’s concerns. Hence, the reduced version of the proportionality principle applicable in commitment proceedings, as construed by the Court of Justice, incentivises—rather than limits the risk of—firms to settle on some suboptimal remedies, in particular if the deal reached with the Commission involves ancillary sweeteners.148

Ibid, para 42. Ibid, para 63. 143 For a general discussion on the limits to ‘discretionary remedialism’, see Lianos, ‘Which Limits’, above n 90. 144 Alrosa, above n 6, para 35. 145 Opinion of Advocate General Kokott in Alrosa, above n 6, para 60. 146 Alrosa, above n 6, para 49. 147 Interestingly, Commission official Wouter Wils voiced a similar, though softened, concern in ‘Settlements of EU Antitrust Investigations: Commitment Decisions under Article 9 of Regulation No 1/2003’, 29 World Competition 352 (2006). 148 In the Alrosa case, for example, De Beers was also faced with a parallel investigation of a new selective distribution system called ‘Suppliers of Choice’, which was closed a few 141 142

172 Damien MB Gerard In other words, it is unable to reconcile effectiveness and efficiency considerations, and it leaves the effectiveness conundrum unaddressed. Arguably, though, the parallel drawn by the General Court between the infringement and commitment decisions on the basis of the adverse effects that they both entail for firms irrespective of their specific features was unwarranted. Commitment and infringement decisions differ inasmuch as the former do not entail a finding of infringement but aim to address concerns at the end of an ideally collaborative learning process. Even if the commitment procedure requires adjustments, that possibility should be preserved, it is submitted, because it responds to a concrete need to guarantee the effectiveness of antitrust enforcement in highly (or increasingly) dynamic markets. To the extent that due process safeguards are properly implemented, as discussed below, it is therefore not per se unacceptable for the Commission to benefit from reasonable flexibility in the application of the proportionality principle in commitment cases. In the US, as Ginsburg and Wright have observed, consent decrees also allow enforcers to impose ‘conditions that extend beyond what [they] would likely be able to obtain after successful litigation’.149 Likewise, while the judicial review of DOJ consent decrees is mandatory,150 the public interest standard of review applied by US federal courts pursuant the Tunney Act does not fundamentally differ from the ‘manifestly incorrect’ standard set by the EU Court of Justice for the assessment of proportionality in commitment cases,151 notably as it has also historically been deferential in

weeks after the adoption of the commitment decision of 22 February 2006, as revealed by the General Court in a judgment of 11 July 2013 in Joined Cases T-107 and 354/08 Diamanthandel A Spira BVBA v Commission, EU:T:2013:367. See especially para 16: ‘[o]n 29 March 2006, the Commission sent Spira a case-orientation letter informing it of its initial view, namely that there was insufficient Community interest to investigate the complaint further and inviting it to consider withdrawing its complaint’ against the Suppliers of Choice system. Ginsburg and Wright, ‘Antitrust Settlements’, above n 20, 56. The mandatory character of the judicial review of DOJ consent decrees can be viewed as a major difference with the situation at the EU level, where such review is very exceptional. However, as Frankel observes, ‘[b]ecause it would be awkward for settling defendants to object to entry of a decree they have agreed to, rarely is there a challenge alleging that the decree is “too strong” or unnecessary. Instead, challenges typically come from competitors who complain that the remedy does not impose sufficient obligations on the defendants.’ Lawrence Frankel, ‘Rethinking the Tunney Act: a Model for Judicial Review of Antitrust Consent Decrees’, 75 Antitrust Law Journal 549, 609, footnote 225 (2008). Still, ‘[b]y forcing the DOJ to reveal the rationale behind its settlement, and subjecting the settlement to public comment and judicial review, the Act gives the DOJ an incentive to make sure its proposed remedy is sound’. Ibid, 595. 151 The Tunney Act provides that, before entering any consent judgment proposed by the DOJ, the court must consider: (A) the competitive impact of such judgment, including termination of alleged violations, provisions for enforcement and modification, duration of relief sought, anticipated effects of alternative remedies actually considered, whether its terms are ambiguous and any other competitive considerations bearing upon the adequacy of such judgment necessary to a determination of whether the consent judgment is in the public interest; and (B) the impact of entry of such judgment upon competition in 149 150

Negotiated Remedies in the Modernisation Era 173 order to ‘preserve consent decrees as an effective enforcement mechanism’.152 Rather, the main systemic feature that constrains the US authorities’ remedial discretion lies in the fact that they have to take a case through the many hurdles of a trial in order to establish an infringement, ie there is a real alternative to negotiated procedures capable of preserving the consensual character of proposed remedies and of supporting their presumably proportionate nature.153 In contrast, the combination of the Alrosa case law, the absence of mandatory judicial review over the entry of commitment decisions and the lack of a credible alternative to a negotiated procedure has created the wrong set of incentives at the EU level. In the end, compliance with the proportionality principle in the post-Alrosa era has largely become a matter of self-discipline on the part of the Commission, with all the structural limitations that it implies, in spite of some noticeable efforts. Concretely, the Commission’s Manual of Procedure summarises the Alrosa proportionality standard but also underlines that each commitment decision should ‘explain why the commitments resolve the identified competition concerns in a proportionate manner’.154 In effect, the relevant market or markets, upon the public generally and individuals alleging specific injury from the violations set forth in the complaint including consideration of the public benefit, if any, to be derived from a determination of the issues at trial (15 USC § 16(e)(1)). In practice, while the review standard applied by district courts has varied, many courts historically approved decrees without much consideration in the understanding that ‘the statute’s requirement that a decree be in the public interest did not mean that the decree had to embody the relief that would serve the public best’. Frankel, ‘Rethinking the T unney Act’, ibid, 556. Further to amendments passed in 2004, Frankel summarises the public interest review carried out under the Tunney Act as follows: ‘(a) whether the consent decree may have resulted from improper influence or other improprieties; (b) whether it is ambiguous, readily enforceable, and not unduly harmful to third parties; and (c) whether it is likely to be effective and adequate’. Ibid, 606. In turn, the efficacy and adequacy test entails that the court ‘needs to review the decree to ensure that the proposed remedy has a logical connection to, and a reasonable probability of resolving, at least some of the alleged harm’. Ibid, 609–610. Hence, ‘[a]lthough a court ought to check that the facts alleged in the complaint are at least plausible and set out a claim that is arguably in violation of the antitrust laws (to, among other things, prevent blatant DOJ overreaching), as long as the claims made in the complaint are not obviously false or illogical, a judge ought not to delve in depth into the veracity of the facts, the illegality of the challenged practice, or the validity of possible defenses, since to do so would lead to the sort of litigation the parties attempted to avoid by settlement’. Ibid, 610, footnote 225. 152 Lloyd Anderson, ‘Mocking the Public Interest: Congress Restores Meaningful Judicial Review of Government Antitrust Consent Decrees’, 31 Vermont Law Review 594 (2007). For Anderson, the proper standard of judicial review under Tunney Act, as amended in 2004, is a ‘within the reaches of the public interest’ standard whereby courts should enter consent decrees only if ‘they are firmly convinced, after serious consideration of the enumerated factors, that they are reasonably calculated to protect competition’. Ibid, 612. 153 As Anderson explains, the Tunney Act was originally introduced to avoid ‘sell-outs and sweetheart deals’ between defendants and the government (ibid, 594, 596 and 614), rather than to constrain the remedial discretion of the DOJ. 154 Manual of Procedure, above n 44, chapter 16, paras 46 and 70. Guidance issued by certain NCAs in relation to commitment proceedings adopts a strict stance on the issue of proportionality. The French NCA, for example, indicates that ‘[t]he Autorité does not

174 Damien MB Gerard all commitment decisions nowadays undertake to justify the nature and scope of the remedies made binding on the firm(s) proposing commitments in light of the proportionality principle,155 and those involving structural remedies do so with particular care. Likewise, former Commissioner Almunia was adamant that the Commission ‘make[s] sure that the commitments are tailored to the competition problem at hand at all times’.156 Remarkably, the Commission has also indicated that it has been guided by proportionality considerations when assessing comments received in response to the market test and submissions filed by complainants, thereby hinting that it resisted requests for more far-reaching remedies on proportionality grounds.157 In the French long-term electricity supply case, for example, the Commission refused to limit the duration of EDF’s new supply contracts to a three-year period as it would have been ‘disproportionate in the context of the proposed commitments’, instead settling in favour of a five-year duration.158 The Commission has also endeavoured to ensure the proportionality of negotiated remedies by limiting commitment decisions in time or by including a review clause in them. While Article 9 of Regulation 1/2003 merely states that commitment decisions ‘may be adopted for a specified period’,159 the Manual of Procedure refers generally to the factoring of duration into the Commission’s proportionality assessment and favours deadlines over review clauses.160 In practice, about two-thirds of all commitment decisions adopted so far have included limitations in time, and that proportion reaches beyond four-fifths if one discounts for those obvious cases where remedies were of a one-off nature, such as structural ones. The duration of commitments has ranged between two years and nine months and 10 years, with an

accept binding commitments going beyond the resolution of competition concerns even though it can, when necessary, acknowledge additional remedies proposed by the undertaking concerned, for example, in order to facilitate the implementation of commitments that have been accepted’. Notice on Competition Commitments of 2 March 2009, above n 73. 155 Decisions typically use the formulation of the proportionality principle that can be found in EU Courts’ precedents involving infringement cases and then include a caveat referring to the Alrosa judgment as guiding the application of proportionality in commitment cases. 156 Almunia, ‘Remedies, Commitments and Settlements’, above n 21. 157 See, eg Microsoft (Tying), above n 105, recital 97; Commission Decision of 17 March 2010 in Case COMP/39.386—Long-term contracts France, paras 54 and 64. 158 Long-term contracts France, ibid, para 80. See also paras 83 and 85 in relation to secondary suppliers. 159 In Alrosa, the General Court refused to consider the indefinite period of time for which the commitments were made binding on De Beers as disproportionate, pointing to the fact that Article 9 of Regulation 1/2003 does not ‘require’ the Commission to limit the duration of commitment decisions, as was allegedly the intention in the original Commission proposal that preceded the adoption of the Regulation. See Alrosa, above n 6, para 91. The Court of Justice affirmed this point. 160 Manual of Procedure, above n 44, chapter 16, paras 51–52.

Negotiated Remedies in the Modernisation Era 175 average of (around) five years.161 Such limitations in time preserve the adjustable character of commitments and their experimental nature, while relieving firms proposing commitments of the need to petition the Commission prior to adapting their conduct to market changes and of the associated risk of (again) having to face constituencies of adverse interests favouring the maintenance of the restrictions in place.162 In theory, limiting commitment decisions in time can usefully contribute to mitigating the effects of suboptimal remedies. In practice, though, much depends on their actual duration, and on whether that duration fits the commercial and technical cycles in the relevant market contexts; commitments made binding for a period of five years, for example, may severely impact firms in many industries. As a result, the negotiation of the duration of commitments can sometimes be highly contentious. *

*

*

Overall, the traditional role of the proportionality principle as a constraint to the Commission’s remedial discretion appears significantly weakened in negotiated procedures. As noted, beyond the inherent limitations to the justiciability of proportionality, compliance with that principle has very much become a matter of self-discipline on the part of the Commission in the aftermath of the Alrosa judgment of the Court of Justice. Hence, proportionality appears to be of limited help in ensuring the lasting legitimacy of antitrust enforcement and remedial action, and in addressing the effectiveness conundrum arising from the move towards a regulatory paradigm. This is notably because the voluntary character of commitments can hardly be construed as a guarantee of proportionality of the proposed remedies, let alone of their optimal character. As a result, it was suggested that a solution to reconcile effectiveness and efficiency is rather to be found in the strengthening of the input legitimacy of the commitment procedure, ie in the other constraint historically put on the Commission’s remedial action, namely due process. This is also because appropriate due process standards can contribute to restoring the voluntary nature of commitment as a proportionality safeguard and therefore can act as a proxy for a proportionality standard compatible with the specific features of negotiated enforcement. In theory, indeed, firms do not have an incentive to consent to remedies that are unrelated to antitrust concerns identified in the course an open dialogical process with the C ommission

161 Although they are said to ‘often last for too long a period’ (Epstein, above n 30, 6), some prominent consent decrees in the US have also speficied an original period of five years. See, eg United States v Microsoft, above n 3. 162 Article 9(2) of Regulation 1/2003 generally allows the Commission to ‘upon request or on its own initiative, reopen the proceedings’, notably ‘where there has been a material change in any of the facts on which the decision was based’.

176 Damien MB Gerard and third parties. However, as explained below, EU-negotiated procedures do not guarantee the openness of that process at this stage, thereby further deepening the legitimacy gap previously identified.

2.2.2 Due process requirements The Court of Justice gradually set due process requirements in infringement proceedings at a time when antitrust enforcement was still governed by Regulation 17/62.163 These requirements mainly entail the exercise of the right to be heard in response to objections stated in writing by the Commission, and the associated right of access to file in the preparation of that response. In addition, the judicial review of adverse decisions has been construed by the Court of Justice as an additional guarantee against the potential arbitrariness of the Commission and, generally, of the legitimacy of the EU antitrust enforcement framework as a whole.164 The procedural safeguards associated with each step in infringement proceedings were then largely codified in Reg ulation 1/2003 and Regulation 773/2004.165 The picture is radically different when it comes to negotiated proceedings, which nowadays form the bulk of antitrust enforcement in the European Union: Regulation 773/2004 does not contain a single reference to commitment proceedings and Regulation 1/2003 only refers to the issuance by the Commission of a preliminary assessment in response to which commitments can be offered (Article 9(1)), and to the obligation to market test commitments before making them binding (Article 27(4)). Hence, irrespective of all its benefits, choosing the commitment route has so far been a leap of faith, as the due process safeguards built into the process are virtually non-existent. Interestingly, the application of due process requirements in commitment proceedings was the second issue at the heart of the Alrosa case, and the General Court and the Court of Justice again differed squarely in their analysis. For the General Court, the right to be heard was a ‘fundamental principle of [EU] law’ that ‘must be guaranteed even in the absence of any rules governing the proceedings in question’ and which requires a right ‘of access to the Commission’s file’ in order to be exercised effectively.166 In the end, it concluded that the Commission had breached the right to be heard by failing to provide

163 Among the most significant cases in that respect, see, eg Case T-30/91 Solvay SA v Commission [1995] ECR II-1775; Case T-36/91 Imperial Chemical Industries plc v Commission [1995] ECR II-1847. 164 See, eg Case T-38/02 Groupe Danone v Commission [2005] ECR II-4407, para 51; Case T-54/03 Lafarge SA v Commission [2008] ECR II-120*, paras 42 et seq. 165 Regulation 773/2004 relating to the conduct of proceedings by the Commission pursuant to Articles 81 and 82 of the EC Treaty, [2004] OJ L123/18, in particular chapters V and VI. 166 General Court in Alrosa, above n 6, paras 191 and 197.

Negotiated Remedies in the Modernisation Era 177 Alrosa with a copy of the comments filed by third parties in the framework of the market test of its commitments, whereas The third-party observations were of particular importance in the proceedings, in so far as the Commission took them into account in concluding that the market testing was negative and that only the cessation of all trading relations with effect from 2009 constituted an acceptable solution.167

Moreover, the General Court found that the Commission was under a ‘duty to hear the parties on those observations, and on the other factual elements justifying its new conclusion’.168 The Court of Justice reversed, finding that ‘the proposition that the Commission was obliged to provide Alrosa with a reasoned explanation of why the observations of the third parties had changed its position’ was ‘incorrect’,169 and that Alrosa was not entitled to be heard or to have access to the file in relation to the rejection of its proposed commitments as a result of the market test.170 As construed by the Court of Justice, commitment proceedings are therefore affected by a due process deficit. In effect, while it is bound to formalise its concerns in writing, the Commission is freed from any specific requirements in handling proposed remedies tabled in response to its preliminary assessment. Generally, the Article 9 procedure is not subject to a clear sequence with specific procedural rights associated with each step in the process leading to the adoption of the decision, and with clearly defined implications for the firms making the commitments. Combined with the relaxation of the proportionality principle, the lack of formalisation of the interactions between firms, the Commission and third parties means that no guarantee is built into that process to ensure that remedies found ‘appropriate’ by the Commission to address its concerns are optimal from a welfare point of view. Put otherwise, the increased discretion benefiting the Commission as a result of the relaxation of proportionality requirements appears further widened by the lack of procedural safeguards, thereby creating an asymmetry in the respective incentives of firms and the Commission, which threatens the lasting legitimacy of negotiated enforcement. In the end, doing away with both proportionality and due process requirements allows effectiveness considerations to go unchecked, to the point of possibly affecting the selection and assessment of the evidence, and deprives commitment proceedings of any guarantee as to the (more) efficient character of new market equilibriums sought by the Commission. Thus, beyond concerns about ‘[t]he current lack of adequate

Ibid, para 202. Ibid, para 194. 169 ECJ in Alrosa, above n 6, para 92. 170 Ibid, paras 88–89. Specifically, the Court of Justice held that Alrosa could not ‘claim the procedural rights reserved to the parties to the proceedings’ because, in the end, the decision made binding a (revised) set of commitments offered unilateraly by De Beers. 167 168

178 Damien MB Gerard checks and balances on the commitments procedure’,171 what is at stake is the ability of the Commission to achieve the objectives of substantive modernisation through means other than its own self-discipline. Conscious, no doubt, of the risks inherent in the broad discretion it was given, the Commission endeavoured to frame the exercise of that discretion in three ways. First, it allowed firms proposing commitments to call upon the Hearing Officer at any time during the proceedings in order to facilitate the exercise of their procedural rights.172 That possibility has been available since the very first commitment case, and all subsequent decisions have been accompanied by the publication of a report of the Hearing Officer. Interestingly, while Hearing Officers did, at times, grant the request of firms offering commitments to obtain access to specific documents,173 their reports also revealed the Commission’s early practice of subjecting the adoption of a commitment decision to the signature of a declaration whereby firms acknowledged that they had received sufficient access to the information necessary to propose commitments meeting the Commission’s concerns.174 That practice appears to have been discontinued in late 2010,175 without any clear explanation.176 Secondly, in the Best Practices Notice, the Commission indicated a willingness to ‘take into consideration the interests of third parties’ when carrying out its proportionality assessment and,177 when commitments cannot be implemented without the agreement of third parties, to request evidence of the agreement of the relevant third parties prior to making the commitments binding.178 These guarantees reflect a formal condition imposed by the Court of Justice in light of the principle of proportionality in commitment procedures and at the same time reflect a willingness to prevent a repetition of the Alrosa saga, inasmuch as it was rooted in a decision making binding

Marsden, ‘The Emperor’s Clothes’, above n 71, 5. Manual of Procedure, above n 44, chapter 13, para 68. See also Article 15(1) of Decision 2011/695 of the President of the European Commission of 13 October 2011 on the function and terms of reference of the hearing officer in certain competition proceedings, [2011] OJ L275/29. 173 See, eg Final report of the Hearing Officer in Case COMP/38.636—Rambus [2010] OJ C30/15. 174 See, eg Final report of the Hearing Officer in Case COMP/39.530—Microsoft (Tying) [2010] OJ C36/5. 175 The report of the Hearing Officer in the ENI case appears to be the last one to refer to such a declaration having been filed with the Commission. See Final report of the Hearing Officer in Case COMP/39.315—ENI [2010] OJ C352/6. 176 Note, however, that the discontinuance of that practice follows by a few months the ECJ’s judgment in Alrosa, dated 29 June 2010. 177 Commission notice on best practices for the conduct of proceedings concerning Articles 101 and 102 TFEU, [2011] OJ C308/06, para 115 (Best Practices Notice). 178 Ibid, para 128 and Manual of Procedure, above n 44, chapter 16, para 47. See also the French NCA’s Notice on Competition Commitments of 2 March 2009, above n 73, para 38. 171 172

Negotiated Remedies in the Modernisation Era 179 c ommitments of a contractual nature offered by De Beers unilaterally without the consent of Alrosa.179 Generally, the willingness to give voice to third parties also supplements Article 27(4) of Regulation 1/2003 and confirms the importance of the Commission market testing the proposed commitments ‘with all stakeholders before taking a final decision to make sure that it is a good outcome for the market and that no issue is overlooked’.180 Thirdly, in spite of the lack of obligation in this respect, the Commission has endeavoured to inform firms proposing commitments of the outcome of the market test,181 ie of the observations received in response to the publication of the ‘concise summary of the case and the main content of the commitments’ known as the Article 27(4) notice. However, because of the flexibility allowed by the Alrosa judgment, most of the time the Commission simply ‘inform[s] the parties orally or in writing of the substance of the replies’ by means of a summary of the observations submitted by third parties.182 In at least two cases, though, for reasons that are not public, firms making commitments were ‘provided with non-confidential versions of the responses received in response to the market test’, and not just a summary of them.183 This variety of situations and treatment reveals the unsustainable character of the Court of Justice’s position in Alrosa, the wide discretion of the Commission and the uncertain character of the procedural rights available in commitment proceedings. Overall, these self-limitations do not go far enough, it is submitted, in putting in place the due process requirements necessary to ensure that commitment proceedings allow for an open dialogical exchange capable of reconciling effectiveness and efficiency and of closing the legitimacy gap previously identified. More stable and ambitious steps are thus required to strengthen the input legitimacy of commitment proceedings and turn them into the kind of collaborative learning process capable of achieving optimal remedial outcomes.184 Short of deep institutional transformations that would bring the

179 On the first point, see in particular Alrosa, above n 6, para 41. Of course, the second guarantee sounds like an acknowledgement by the Commission that it mishandled the Alrosa case, even though it managed to escape the censure of the Court of Justice for formalistic reasons. 180 Almunia, ‘Remedies, Commitments and Settlements’, above n 21. 181 See, eg Final report of the Hearing Officer in Case COMP/39.351—Swedish Interconnectors [2010] OJ C142/27; Final report of the Hearing Officer in Case COMP/39.398—VISA MIF [2011] OJ C79/6; or Final report of the Hearing Officer in Case COMP/39.736— Siemens/Areva [2012] OJ C280/7. 182 Best Practices Notice, above n 177, para 132; Manual of Procedure, above n 44, chapter 16, para 64. 183 Final report of the Hearing Officer in Case COMP/B-1/37.966—Distrigaz [2008] OJ C9/6; Final report in Rambus, above n 173. 184 On the importance of ensuring sufficient interactions between firms proposing commitments and the relevant competition authority, see ECN Recommendation on Commitment Procedures, para 10 and recommendation 4.

180 Damien MB Gerard EU antitrust enforcement framework more in line with the US system, these steps should involve, on the one hand, departing from the Alrosa case law with respect to due process requirements and, on the other hand, rebalancing the respective incentives of firms and the Commission. To ensure that they achieve the objective sought, these steps should be clarified in an official communication that creates (enforceable) legitimate expectations. Departing from the Alrosa precedent essentially involves: (i) consistently providing firms proposing commitments with a full non-confidential version of the observations submitted by third parties in response to the market test, which is already the case in certain national systems,185 notably as the Commission frequently requires ‘improvements’ to the commitments as a result of the market test; and (ii) the (even succinct) motivation of the Commission’s decision to reject market-tested commitments, inasmuch as it is liable to fundamentally affect the interests of the undertakings concerned.186 As noted, the Commission has already communicated a non-confidential version of the responses to the market test in some past cases. Moreover, it has developed a practice of sending letters to executives of the firms proposing commitments to explain its assessment of those commitments in light of third parties’ comments.187 Hence, while they would more clearly frame the sequence of commitment proceedings and clarify the procedural rights of firms offering commitments, these steps should not raise serious effectiveness issues. The second set of steps to undertake in order to strengthen the input legitimacy of negotiated enforcement aims to rebalance the incentives to engage in commitment proceedings on the part of the Commission and firms, respectively. The overall objective is to salvage the voluntary character of commitments as a guarantee of proportionality and to turn commitment proceedings into a collaborative process capable of delivering optimal outcomes from both an effectiveness and an efficiency point of view. The first step is simple, radical and consistent with the underlying rationale of Article 9: engaging in commitment proceedings should immunise firms from the risk of fines even

185 See the French NCA’s Notice on Competition Commitments of 2 March 2009, above n 73, para 27: ‘The applicant and the undertaking concerned … have access to the documents used by the R apporteur to establish the preliminary assessment and to those used by the Autorité to decide on the commitments, that is to say at least the preliminary assessment and the third parties’ comments resulting from the market test’. 186 As indicated by the Manual of Procedure, above n 44, ‘[t]he obligation to carry out a market test should not be misunderstood as requiring the approval of the market for the commitments’. Chapter 16, para 66. 187 See, eg Joaquín Almunia, ‘Statement of VP Almunia on the Google Antitrust Investigation’, SPEECH/12/372, Brussels, 21 May 2012; Frances Robinson, ‘EU Almunia Tells Google to Improve Antitrust Search Proposal Offer’, Wall Street Journal, 17 July 2013, http://online.wsj.com/article/BT-CO-20130717-703287.html.

Negotiated Remedies in the Modernisation Era 181 if the Commission ultimately finds an infringement. In essence, that proposition amounts to taking Recital 13 of Regulation 1/2003 seriously, inasmuch as it provides that ‘[c]ommitment decisions are not appropriate in cases where the Commission intends to impose a fine’.188 Practically, the Commission should be deemed to have forfeited the right to impose fines upon the issuance of a ‘preliminary assessment’ within the meaning of Article 9, since the Commission is at that point ‘convinced of the undertakings’ genuine willingness to propose commitments which will effectively address the competition concerns’.189 If a statement of objections has already been issued, the same consequence should attach to the market testing of proposed commitments, ie to the publication of an Article 27(4) notice.190 Abiding by such a principle would radically contribute to closing the legitimacy gap identified e arlier because it would force the Commission to ‘weigh carefully’191 the optimal character of negotiated solutions and would create a more balanced framework for the negotiation of (the final set of) commitments. Interestingly, in the only case so far where the Commission did return to the infringement procedure after an inconclusive market test, it refrained from imposing fines.192 The second step is more systemic in nature and aims to restore infringement proceedings as a credible alternative to the negotiated route for firms willing to litigate the Commission’s concerns or objections, and as a source of law capable of providing stable guidance. In turn, it should also contribute to restoring the voluntary nature of the choice to engage in commitments proceedings and ensure that the negotiated procedure takes place in the shadow of the law. In the US, the fact that antitrust enforcement authorities have to take a case through the many hurdles of a trial in order to establish an

188 The same principle was repeated in the Commission’s Best Practices Notice, above n 177, para 116. It contrasts, however, with the ECN Recommendation on Commitment Procedures, which states that ‘[a]n Authority can at any stage continue proceedings with a view to adopting a prohibition decision to bring to an end an agreement or conduct that is found to infringe the competition rules and may provide for the imposition of remedies and/or fines’ (para 4). 189 Commission Notice on best practices for the conduct of proceedings concerning Articles 101 and 102 TFEU, [2011] OJ C308/06, para 116; Manual of Procedure, above n 44, chapter 16, para 23. In an Article 9 procedure, the Commission is also supposed to have positively ‘ascertain[ed] the willingness of the parties to settle … and that the commitments appear sufficiently likely to resolve the identified competition problems’. Manual of Procedure, chapter 16, para 10. 190 As the Commission acknowledges, ‘[i]n cases, in which the Commission sent a Statement of Objections announcing its intention to impose a fine it would be preferable for the commitment decision to explain why a fine was no longer necessary’. Manual of Procedure, chapter 16, para 14. In addition, the Commission would have already informally ‘market tested’ or discussed the offered commitments with specific third parties prior to the publication of the Article 27(4) notice. See Manual of Procedure, chapter 16, para 54. 191 According to the Manual of Procedure, ‘[t]he advantages and disadvantages of a commitment decision have to be weighed carefully in each individual case’. Chapter 16, para 5. 192 Commission Decision of 16 July 2008 in Case COMP/C2/38.698—CISAC.

182 Damien MB Gerard infringement means that there is a real alternative to the consent decree procedure capable of preserving the consensual character of negotiated remedies and of supporting their presumably proportionate nature. The EU antitrust enforcement context is different, and deep institutional transformations bringing it more in line with the US system are unlikely to take place in the near future. Similarly, introducing a systematic public interest review of proposed commitment decisions would not fit well into the EU framework and, judging by the limitations of that review in the US, it would be unlikely to deliver the expected result. Rather, a more promising approach to rebalancing the EU antitrust enforcement framework lies, it is submitted, in strengthening the role of the EU Courts in adjudicating appeals brought against the Commission’s infringement decisions.193 Indeed, empowering the EU Courts to operate as courts of full appellate jurisdiction, at least when reviewing ‘[antitrust] decisions whereby the C ommission has fixed a fine or periodic penalty payment’,194 would likely change the dynamics of the EU antitrust enforcement system because it would force the Commission to conduct proceedings in the shadow of full review. Hence, it carries the potential of freeing up space for more open interactions over the substance of cases, thereby strengthening infringement proceedings as a more effective alternative to negotiated solutions. Moreover, such an evolution would not require any major structural changes to the EU competition enforcement and institutional framework but could be directly implemented by the EU Courts on the basis of Article 31 of Regulation 1/2003. In fact, it would also bring the EU model closer to that of certain Member States, as the decisions of many national competition authorities are already subject to full appellate jurisdiction today. *

*

*

Reconciling effectiveness and efficiency in the design of negotiated remedies requires strengthening the input legitimacy of commitment proceedings. This, in turn, requires a strengthening of the other constraint historically imposed on the Commission’s remedial action, namely due process. This is notably because appropriate due process standards can contribute to restoring the voluntary nature of commitments as a proportionality safeguard and therefore act as a proxy for a proportionality standard reflecting the specific features of negotiated enforcement, including effectiveness concerns. However, in the post-Alrosa era, commitment proceedings are blemished

193 For an in-depth discussion of that proposition, see Damien MB Gerard, ‘Breaking the EU Antitrust Enforcement Deadlock: Re-empowering the Courts?’, 36 European Law Review 457 (2011). 194 Article 31 of Regulation 1/2003.

Negotiated Remedies in the Modernisation Era 183 by a clear due process deficit, as they are not subject to a clear sequence articulating identifiable procedural rights. Moreover, the self-limitations with which the Commission has endeavoured to comply are insufficient to turn commitment proceedings into a collaborative learning experience. This chapter therefore proposed to depart from the Alrosa precedent by systematically providing firms proposing commitments with a non-confidential version of the observations submitted by third parties in response to the market test and, as the case may be, by giving them a reasoned statement explaining why the scope of the proposed commitments is deemed insufficient. In addition, it was submitted that, in order to rebalance the respective incentives of firms and the Commission in the design of negotiated remedies and to address the shortcomings associated with the current situation, it is necessary to immunise firms engaging in commitment proceedings from the risk of fines and to empower the EU Courts to operate as courts of full appellate jurisdiction in reviewing infringement decisions, thus restoring infringement proceedings as a credible alternative to the negotiated route.

3. Conclusion Over the past 10 years, the nature, scope and design of remedies have grown in importance in the enforcement of competition law in the European Union. That evolution can be viewed as one of the many consequences of the process known as the modernisation of EU competition law, understood as a comprehensive attempt to experiment with a utility-maximising approach to the regulation of economic competition, with substantive, procedural and institutional dimensions. In effect, modernisation has notably entailed a revamp of enforcement strategies driven by effectiveness considerations. In turn, that effectiveness paradigm has led to a shift towards a ‘negotiated’ form of enforcement by means of tools such as leniency, settlements and commitment proceedings. Commitments, in particular, have developed into the default mechanism for enforcing the antitrust provisions of the EU Treaties outside the field of cartels, where the two other instruments are steadily relied upon. The shift towards negotiated enforcement has moved remedies to the core of antitrust adjudication. However, the effectiveness rationality driving negotiated procedures is as such inapt to ensure the design of efficient remedies, it was submitted, ie effectiveness has limits when it comes to defining optimal remedies, which in turn affects the legitimacy of those negotiated remedies. In the EU context, these limits are magnified by the relaxation in commitment proceedings of the safeguards that have historically limited the Commission’s remedial discretion, as captured by the concepts of proportionality and due process. While there is a case for relaxing proportionality

184 Damien MB Gerard requirements in commitment cases, doing away with both concepts creates a gap that threatens the legitimacy of competition law itself. In turn, filling that gap requires a review of the sequencing of commitment proceedings and associated procedural rights, as well as the (re)development of a credible alternative to those proceedings, in order to salvage the voluntary character of commitments as a guarantee of proportionality and to turn commitment proceedings into a collaborative process capable of delivering optimal outcomes both from an effectiveness and an efficiency point of view.

Giorgio Monti * Behavioural Remedies for Antitrust Infringements—Opportunities and Limitations

1. Introduction In this chapter the focus is on the approach of the European Commission and the EU Courts in imposing remedies for infringements of Articles 101 and 102 on the basis of Article 7(1) of Regulation 1/2003. This provides that, when the Commission finds an infringement, it ‘may by decision require the undertakings and associations of undertakings concerned to bring such infringement to an end’.1 The Court of Justice has expanded on this phrase, indicating that it allows the Commission to prohibit ‘the continuation of certain action, practices or situations which are contrary to the Treaty’.2 However, the Court has also recalled that remedies must satisfy the requirement of proportionality: they may ‘not exceed what is appropriate and necessary to attain the objective sought, namely re-establishment of compliance with the rules infringed’.3 These judge-made limitations made in cases pursuant to Regulation 17/62 have now been codified in Article 7 of Regulation 1/2003, which provides that the remedies selected must be ‘proportionate to the infringement committed and necessary to bring the infringement effectively to an end’. There is nothing in the wording of Regulation 1/2003 that indicates that the role of commitments accepted by the Commission should be any different: pursuant to Article 9, commitment decisions may be issued when the Commission ‘intends to adopt a decision requiring that an infringement be brought to an end’, meaning that the commitments offered by the parties must serve the same functions as the remedies that would have been imposed

Professor of Law, European University Institute, Florence 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, [2004] OJ L1/1. This power was already provided for in Article 3(1) of Regulation 17/62, First Regulation Implementing Articles 85 and 86 of the Treaty. OJ, English special edition: Series I Chapter 1959–1962, 87. 2 Joined Cases 6 and 7-73 Istituto Chemioterapico Italiano SpA and Commercial Solvents Corp v Commission [1974] ECR 223, para 45. 3 Joined Cases C-241 and 242/91 P Radio Telefis Eireann (RTE) and Independent Television Publications Ltd (ITP) v Commission (Magill) [1995] ECR I-743, para 93. * 1

186 Giorgio Monti by the Commission. However, commentators have suggested that remedies in Article 9 proceedings have gone further than this.4 Indeed, that this may be so is acknowledged by the ECJ in the Alrosa judgment, where it suggested that parties may be tempted to offer far-reaching commitments in exchange for a quicker and less formal resolution that avoids fines.5 As commitment decisions have been the subject of ample commentary, they are outside the scope of this chapter, which focuses on remedies imposed in infringement decisions. A further delimitation of this chapter is that it focuses on remedies, and not sanctions. This division is also found in Regulation 1/2003: Article 7 empowers the Commission to impose remedies, while Article 23 provides that fines may be imposed by the Commission in cases of infringement. Moreover, the chapter excludes claims for damages and/or injunctive relief by persons harmed by wrongful conduct, even if, in some legal traditions, these too count as remedies, and, in some, these ‘remedies’ may be awarded by administrative agencies.6 Only behavioural remedies have been imposed under Article 7, and while these remedies may at times have structural effects, the Commission has not yet used its powers to impose a structural remedy in an infringement decision at the time of writing. Remedies have seldom received attention, yet they can be a useful, and sometimes essential, mechanism by which to secure compliance.7 While, in cases of price-fixing cartels, the undertakings will generally have been aware of what they have done to infringe Article 101 and so are able to self-assess how to avoid a subsequent finding and are also able to cease the infringement unaided, in instances where the scope of the antitrust infringement is less clear it may be beneficial for the undertaking to be instructed as to how to behave to avoid further penalties. Likewise, the competition authority can specify how best to comply so as to ensure that the infringement really ceases. Accordingly, deployment of remedies can be an enforcement strategy that might be seen as a complement (or an alternative) to the more conventional deterrence-based approach based on fines or incarceration.8 Instead of 4 See, eg Malgorzata Sadowska, ‘Energy Liberalization in an Antitrust Straitjacket: A Plant Too Far?’ (2011) 34 World Competition 449; Yves Botteman and Agapi Patsa, ‘Towards a More Sustainable Use of Commitment Decisions in Article 102 TFEU Cases’, 1 Journal of Antitrust Enforcement 347 (2013). 5 Case C-441/07 P Commission v Alrosa Company Ltd [2010] ECR I-5949, para 48. 6 Bernard Schwartz, Administrative Law, 2nd ed, Little, Brown & Co, 1984, 74–89. More generally, and providing a critique of terminologies, see Ioannis Lianos, ‘Competition Law Remedies in Europe: Which Limits for Remedial Discretion?’, in Ioannis Lianos and Damien Geradin, eds, Handbook of EU Competition Law: Enforcement and Procedure, Edward Elgar, 2013, 362 et seq. 7 An important exception is the study by Per Hellström, Frank Maier-Rigaud and Friedrich Wenzel Bulst, ‘Remedies in European Antitrust Law’, 76 Antitrust Law Journal 43 (2009). 8 For a helpful discussion of the differences between these two types of enforcement models, see Daniel Crane, The Institutional Structure of Antitrust Enforcement, Oxford University Press, 2011, 103–107.

Behavioural Remedies for Antitrust Infringements 187 a ssuming compliance by increasing the level of the penalties or the probabilities of getting caught, remedies serve to direct the firm to act in a manner that can remove the harmful effects in the future. Indeed, given some criticisms that the deterrence model may fail to secure compliance completely, having an alternative enforcement tool that is well defined and functional helps strengthen the authority’s tasks.9 While remedies are useful, it is also imperative to bear in mind certain limitations that exist. This chapter considers two aspects: first, Section 2 looks at the role remedies might play and how this role has been shaped by the Commission and the Courts. Then Section 3 examines certain procedural issues pertaining to the imposition of remedies. In this discussion, I use examples from Commission practice to show the variety of remedies that have been selected. The chapter closes with an assessment of the role of r emedies in EU competition law in light of the issues considered here.

2. The purposes of remedies I start from the premise that remedies are effective if they serve the three functions identified in the withdrawn US Department of Justice Report on unilateral conduct: to terminate the defendant’s wrongful conduct, to prevent its recurrence and to re-establish the opportunity for competition in the market.10 The General Court has also suggested that bringing the infringement to an end also includes bringing to an end the effects of the anticompetitive conduct.11 I first consider termination and prevention, and then explore the policy of re-establishing (the opportunities for) competition.

9 For discussion of the deterrent effect of fines, see, eg Cento Veljanovski, ‘Cartel Fines in Europe: Law, Practice and Deterrence’, 29 World Competition 1 (2007); Mario M ariniello, ‘Do European Union Fines Deter Price-Fixing?’, Bruegel Policy Brief No 2013/4, May 2013. 10 US Department of Justice ‘Competition and Monopoly: Single Firm Conduct under Section 2 of the Sherman Act’ (2008), chapter 9, http://www.justice.gov/sites/default/files/atr/ legacy/2009/05/11/236681.pdf. While the report was later withdrawn by the DOJ, it remains one of the few comprehensive surveys of the role of remedies by an antitrust authority. The same goals were identified for the European Commission. See Philip Lowe and Frank Maier-Rigaud, ‘Quo Vadis Antitrust Remedies’, in Barry Hawk, ed, International Law & Policy: Fordham Competition Law Institute 2007, Juris Publishing, 2008, chapter 20, 599. 11 Case C-119/97 P Union française de l’express (Ufex), formerly Syndicat français de l’express international (SFEI), DHL International and Service CRIE v Commission [1999] ECR I-1341, paras 93–94. Here the Court speaks of the power to ‘eliminate or neutralise’ anticompetitive effects.

188 Giorgio Monti

2.1 Termination and prevention The first two functions are closely related. In principle, it is possible to design a remedy that terminates the conduct and prevents its recurrence. A representative example is the remedy imposed in the Water Management Products decision. Article 1 details the nature of the infringement (a continuing agreement and/or concerted practice in the sector for water management services). Article 3 then provides that the undertakings shall bring the infringement to an end and ‘shall refrain from repeating any act or conduct described in Article 1, and from any act or conduct having the same or similar object or effect’.12 One might perhaps query whether the quoted passage is sufficiently precise, not least because, pursuant to Article 24 of Regulation 1/2003, fines may be applied for failure to bring the infringement to an end. However, in a case with broadly the same remedy, the General Court has taken the view that the undertaking should be able to understand what is p rohibited by reference to the decision.13 In some cases, the Commission is more prescriptive—here a valuable illustration is the order in Hilti, which guides the firm not only in explaining what kinds of discounts are forbidden, but also under which circumstances the dominant player may offer discounts.14 However, being too prescriptive may lead to the remedy being quashed. In Cartonboard, the Commission foresaw that the parties to this cartel would continue to exchange information, so it specified the kinds of information that may be exchanged and those that may not. It noted that the parties had already begun to modify their scheme for information exchanges and the Commission took steps to request further delimitations on the kinds of communication that might be passed among the undertakings.15 However, on appeal, some of the requirements established by the Commission were disallowed. The General Court noted that the Commission sought to prohibit the exchange of certain aggregated statistical data, and it took the view that this was not necessary to bring the infringement to an end because it

12 Article 3, Commission Decision of 27 June 2012 in Case COMP/39611—Water Management Products. This was a cartel settlement decision, but rightly the same remedy as in ordinary decisions was issued. For ordinary decisions see, eg Article 3, Commission Decision of 17 February 1992 in Cases IV/31.370 and 31.446—UK Agricultural Tractor Registration Exchange [1992] OJ L68/19: ‘The AEA and the eight members of the Exchange shall immediately put an end to the infringement established in Article 1, in so far as they have not already ended the infringement, and shall in future refrain from entering into any agreement or concerted practice that may have an identical or similar object or effect’. 13 Case T-34/92 Fiatagri UK Ltd and New Holland Ford Ltd v Commission [1994] ECR II-905, para 39; appeal dismissed: Case C-8/95 P, [1998] ECR I-3175. 14 Commission Decision of 22 December 1987 in Cases IV/30.787 and 31.488— Eurofix-Bauco v Hilti [1988] OJ L65/19, Annex, para 1(b). 15 Commission Decision of 13 July 1994 in Case IV/C/33.833—Cartonboard [1994] OJ L243/1, paras 165–166 and Article 2.

Behavioural Remedies for Antitrust Infringements 189 was not information that could be used for anticompetitive purposes. If the Commission wished to prohibit the exchange of such information, it would have to explain how these anticompetitive effects might materialise.16 These examples suggest that there are two policies at play that may conflict: on the one hand, the Commission may guide the undertakings to redesign their commercial conduct, and the undertakings may wish to secure this guidance to avoid fines for non-compliance. On the other hand, the Commission cannot overplay this role by seeking to eliminate anticompetitive effects that may arise in ways not specifically foreseen and condemned by the decision itself. To take a very simple but illustrative example: if what is penalised is cooperation among competitors, the decision cannot forbid the undertakings from entering into vertical agreements with distributors that may facilitate tacit collusion among those competitors. There are two legal principles that underpin this limitation: the first is that an extensive remedy infringes the rights of the defence—in Cartonboard, therefore, unless the Commission in the decision were to explain how it considered the exchange of aggregated statistical data as facilitating collusion, the parties would have no means of responding to this. The second is that there must be a link between the remedy and the violation: according to consistent case law, it is the infringement, not the anticompetitive effects, that has to be tamed.17 It follows from the above that when one speaks of remedies being preventative, this should be read to mean that they prevent the specific antitrust offence that has been discovered from recurring; they do not n ecessarily prevent the anticompetitive effect if the parties can find different ways of behaving that have not been forbidden but lead to similar anticompetitive effects. In this sense, the preventative effect of a behavioural remedy is different from that of antitrust fines. The Commission’s fining policy allows a 100 per cent increase in the fine for recidivism even in situations where the undertaking has committed the same or a similar infringement.18 Accordingly, if a party continues to cause anticompetitive effects by means not specifically prohibited by the first decision, it will draw little comfort that it is not in breach of the obligations in that decision, for its penalty will be much more severe for its choice to circumvent the original decision. By contrast, in a behavioural remedy, the Commission may only prohibit the undertaking from continuing the specific infringement. Nevertheless, preventing a recurrence of the infringement is valuable because the case law shows that the undertaking will normally wish to

16 Case T-310/94 Gruber + Weber GmbH & Co KG v Commission [1998] ECR II-1043, paras 177–178. 17 See the cases cited in nn 2 and 3 above. 18 Commission Guidelines on the method of setting fines imposed pursuant to Article 23(2)(a) of Regulation 1/2003, [2006] OJ C210/2, para 28(a).

190 Giorgio Monti c ontinue the kind of behaviour it has engaged in, so establishing parameters for legality is essential. This also justifies the Commission imposing a fine and a remedy at the same time: the two elements provide for different goals. The fine is designed to provide for general deterrence, meaning that it acts as a signal to everyone that a competition infringement does not pay, so the fine is not part of the toolbox that brings the infringement to an end.19 However, a counter-argument might be that the fine is also designed to achieve specific deterrence. The Guidelines on fines indicate that the size of the penalty should deliver specific and general deterrence,20 but if the remedy under Article 7 (recalling that breach of the remedy makes the undertaking liable to pay periodic penalties) delivers specific deterrence, then one might legitimately ask for the decision-maker to deduct from the fine any amount which is covered, as it were, by the behavioural remedy. A reply to this might be that the Commission will already take this into consideration as a mitigating factor when it considers that the undertaking has ceased the infringement and is cooperating with the Commission21—one aspect of such cooperation may be in aiding the Commission in designing an appropriate remedy. A good example of this is the approach taken in Hilti: after the Commission secured an interim injunction, the dominant firm undertook to cease the practices the Commission objected to and agreed a certain mode of conduct to be followed pending the final decision, and this modality of behaviour was then restated as the remedy in that case. The result of this cooperation was a reduction in the fine.22 This approach justifies mixing the deterrence-based approach with a more prescriptive enforcement style from a legal perspective. Mixing up these two enforcement styles is also explained by the limited scope of the behavioural remedy—it merely seeks to avoid continuation of the infringement, and nothing more.

2.2 Re-establishing competition The third objective of remedies is a little vague, so an example will help elucidate it. Suppose you sanction a cartel and impose the sort of cease-and-desist orders described above: is the opportunity for competition re-established?

Hellström et al, ‘Remedies’, above n 7, 45. Fining Guidelines, above n 18, para 4. Ibid, para 29. 22 See Eurofix-Bauco v Hilti, above n 14, Recital 103; affirmed: Case T-30/89 Hilti AG v Commission [1991] ECR II-1439; Case C-53/92 P Hilti AG v Commission [1994] ECR I-667. The remedy was not challenged on appeal; the fine was upheld. In Case T-83/91 Tetra Pak International SA v Commission [1994] ECR II-755, para 222, the General Court stated that cooperation in designing remedies during the administrative procedure may lead to a reduction in fines. 19 20 21

Behavioural Remedies for Antitrust Infringements 191 On the one hand, the answer is yes, because henceforth the parties will compete and not collude. On the other hand, the parties’ financial position, their market shares and the rate of innovation may all have been altered by the cartel, so the conditions of competition are not precisely the same as they were before the cartel. Accordingly, it is unlikely that any remedy will re-establish competition completely, so it is better to speak of a remedy as restoring the possibility for competition. In instances where the conduct is exclusionary, the interest to restore competition more radically has led to some more aggressive remedies. In a situation where a dominant firm acts to exclude its major competitor, it is understandable that the Commission will seek to protect that one firm, since it is unlikely that there will be further new entrants to challenge the dominant company. This worry is particularly prominent in the IMS Health case: The Commission considers that it is justified to make an order which will as far as possible ensure that the applicant is not put out of business, and that no intolerable damage to the public interest occurs, pending the final outcome of the administrative procedure by means of which the applicant complains. In the present case the most urgent need is to maintain the status quo by permitting the other undertakings which, according to the Commission’s investigation, are currently present in the relevant market to continue to compete on this market.23

The decision was quashed, but it is illustrative of the Commission’s policy of protecting the nearest competitors from the suspected infringement. Here again, however, what gets restored is the capacity of the competitors to compete with the dominant firm, and some of the advantages the dominant undertaking obtained during the period of the abuse are not removed by the remedy. For this, the best one can do is seek damages afterwards for loss sustained during the period of the abuse. Similar protective considerations may have played a hand in the r emedies imposed in Commercial Solvents (a case concerning a dominant player’s refusal to continue to supply its rival), where the dominant firm was requested to continue to deal with its only rival (Zoja) and the Commission set out the quantities of chemicals that should be delivered to the rival firm. The A dvocate General had not been convinced by the quantities that the Commission had asked the dominant player to supply. He noted that Zoja exported a number of products outside the then EEC, so that by supplying so many the C ommission was possibly going beyond what was necessary to bring the infringement, the effects of which were felt in the EEC, to an end. However, the ECJ accepted the choice made by the Commission, deeming that it was reasonable for it to

23 Commission Decision of 3 July 2001 in Case COMP D3/38.044—NDC Health/IMS Health: Interim measures, [2002] OJ L59/18, Recital 214; decision subsequently suspended: Case T-184/01 R IMS Health v Commission [2001] ECR II-3193.

192 Giorgio Monti consider that maintaining Zoja’s sales worldwide was necessary to ensure it remained a viable competitor.24 The lesson here is that, at least in the early days, the Court deferred to the Commission in its design of remedies. In more recent case law, the Court may be seen to be somewhat more circumspect with the remedies imposed. This occurred, for example, when the Court reviewed the remedies the Commission imposed in ECS/AKZO (a predatory pricing case where the dominant firm used unfair pricing tactics to take away some of its rival’s customers), where the Court was wise to restrict their impact. The Commission’s decision framed the remedy (in part) in these terms: that AKZO shall refrain from offering or applying prices or other conditions of sale for flour additives in the EEC which would result in customers in respect of whose business it competes with ECS paying to AKZO Chemie BV prices which are dissimilar from those being offered by AKZO Chemie BV to comparable customers.25

AKZO complained that this went too far because it appeared to forbid it from any sort of price differentiation—after all, ECS could contact any of AKZO’s clients. The Court softened this remedy by insisting that it was merely meant to afford the option for ECS to regain its customers, and it did not affect AKZO’s policy vis-à-vis other clients: In that regard it must be pointed out that the measure in question is intended to prevent repetition of the infringement and to eliminate its consequences. That is the light in which that provision must be seen. Firstly, it prohibits AKZO from approaching ECS’s customers again by quoting to them advantageous prices without extending those prices to its own customers. Secondly, if ECS endeavours to win back from them the customers whom it has taken from ECS unlawfully, it prevents AKZO from aligning itself on ECS’s prices without giving its customers the benefit of this adjustment. On the other hand, the situation evoked by AKZO in support of its allegation is not covered by that provision. In fact, the provision does not prohibit them from making defensive adjustments, even aligning itself on ECS’s prices, in order to keep the customers which were originally its own.26

This approach sits well with the role that the rights of the defence and proportionality play in setting the limits on remedies: the Commission cannot do more than address the specific abuse that was committed; it cannot guide the dominant firm’s pricing policy more generally. At the same time, what the

24 Commercial Solvents, above n 2, para 48. See also the discussion in Lianos, ‘Which Limits?’, above n 6, 6–7. 25 Commission Decision of 14 December 1985 in Case IV/30.698—ECS/AKZO [1985] OJ L374/1, Article 3. 26 Case C-62/86 AKZO Chemie BV v Commission [1991] ECR I-3359, paras 155–156.

Behavioural Remedies for Antitrust Infringements 193 remedy seeks to achieve here is to allow ECS to try to regain those customers that had been taken away from it by AKZO. The impact on the rival here is the same as in Commercial Solvents: in both of these cases the remedy tries to put the competitor in the same place, economically, as before the infringement.27 This narrow function of remedies and the Court’s close scrutiny is confirmed by the approach the General Court took in the Transatlantic Agreement (TAA) case. In its decision, the Commission had imposed an obligation whereby the undertakings who had infringed Article 101 must inform customers with whom they have concluded service contracts and other contractual relations in the context of the TAA that such customers are entitled, if they so wish, to renegotiate the terms of those contracts or to terminate them forthwith.28

These contracts were not in themselves contrary to Article 101, but the concern was that the cartel had affected the terms of these contracts, so renegotiation would indeed serve to restore the status quo ante. The G eneral Court ruled that as this remedy was not normally imposed, a detailed explanation had to be given by the Commission; moreover, in determining the proportionality of this remedy, the Court intimated that account should be taken of the private law rights resulting from an infringement of Article 101. As a result, the remedy was quashed for failure to state reasons.29 Conversely, the General Court noted that in the one decision where a similar remedy had been set (Astra I ) the Commission had given more explanation. However, this seems to give too much credit to what was said in that case and too little to what was said in TAA. In Astra I, the Commission merely noted that the illegal joint venture necessarily had an effect on the terms of the downstream contracts, so that ‘the restrictive effects which these contracts perpetuate can only be eliminated when the customers have been given the right of r eadjustment’.30 However, in TAA, the Commission had also noted, when gauging the effect of the illegal cartel, that it had an effect on downstream contracts.31 The depth of reasoning in the two instances does not differ greatly. This suggests that the General Court is trying to raise the bar for the Commission with a view to ensuring that the rights of the defence are safeguarded when prescriptive

27 It is thus not surprising that some US commentators believe that the Europeans protect competitors and not competition. However, what this line of critique seem to miss is that in these two instances there was only one competitor, and if it had not been saved the dominant firm would likely have succeeded in eliminating competition completely. 28 Case T-395/94 Atlantic Container Line AB and Others v Commission [2002] ECR II-875, para 414. 29 Ibid, paras 412–416. 30 Commission Decision of 23 December 1992 in Case IV/32.745—Astra [1993] OJ L20/23, Recital 33. 31 Commission Decision of 19 October 1994 in Case IV/34.446—Trans-Atlantic Agreement [1994] OJ L376/1, paras 289 et seq.

194 Giorgio Monti remedies are imposed that try to regulate contracts that are not in themselves in breach of Article 101. The General Court was also concerned about the proper role of public and private enforcement, considering that it would be for private parties to safeguard their rights under EU law in national courts— a matter discussed below. The instances I have identified so far, and the Court’s evolving jurisprudence, suggest that when remedies try to restore competition, they do so within a narrow compass: in Commercial Solvents and AKZO the Commission could just rescue the firm that was being eliminated, which was necessary to afford the possibility of some competition against the incumbent, and in TAA the Court denied a wider reading of the scope of remedies that might correct market failures that are more remote than that. A situation which perhaps goes a little further is found in the Microsoft decision. Here, in the context of Microsoft’s refusal to supply, the remedy imposed was for the dominant firm to release information to ‘any undertaking having an interest in developing and distributing work group server operating system products’.32 The difference with the other cases is that the beneficiary was not a named competitor (although the identity of some of the beneficiaries was probably known) but rather any person wishing to enter the relevant market. Is this a remedy that goes too far? Suppose a dominant player has afforded access to five others, so that now the downstream market is workably competitive. Can the dominant player refuse to supply a sixth entrant? From an economic perspective, the answer might be yes: if competition is workable, the infringement has by now been put to an end, so there is no need for more remedies because the opportunities for competition are now established. Obviously the sixth could be that brilliant new entrant whose market presence revolutionises the downstream market. But competition law is not about regulating markets to optimise welfare, but removing a market failure caused by the dominant player. However, this consideration does not have much bite. In Magill, for example, the Commission took the view that simply requiring the three broadcasters to supply TV listings inter se, and thus allowing for three comprehensive TV guides, would not do: the three dominant undertakings would also have to supply Magill. But the reasoning has nothing to do with considering if this remedy would ensure the availability of competitively priced guides; rather, it considered that this would discriminate against third parties in a manner incompatible with Article 102.33 Likewise, in

32 Commission Decision of 24 March 2004 in Case COMP/C-3/37.792—Microsoft, Article 5(a), http://ec.europa.eu/competition/antitrust/cases/dec_docs/37792/37792_4177_1. pdf. 33 Commission Decision of 21 December 1988 of Case IV/31.851—Magill TV Guide/ ITP, BBC and RTE [1989] OJ L78/43, para 27 and Article 2.

Behavioural Remedies for Antitrust Infringements 195 Microsoft the Commission insisted on the need to provide the relevant information to everyone because, if access were to be granted on a discriminatory basis, this would lead to other restrictions of competition.34 Accordingly, the wide access rights provided in Microsoft fit within the legal framework for remedies. One might argue that there is a tension between the principle of proportionality (making sure the anticompetitive effects are removed) and the principle of equal treatment (giving anyone access on equal terms). However, in the Court’s view, granting access to some but not others would in itself be an abuse of dominance.35 Some authors would like there to be a wider role for remedies, such that in appropriate cases remedies could tackle the anticompetitive effects and restore competition to what it would have been like but for the infringement.36 For example, it was suggested that in a case of illegal tying the dominant player should take steps to assist competitors who had been excluded unlawfully to gain a foothold in the market (eg by imposing an advertising ban on the dominant player or even requesting that it actively promote the goods of the victims of its abuse). The authors obviously had in mind the Microsoft commitment decision, where it was felt that the undertaking had tied the web browser to the operating system illegally. The commitments here were designed to ensure that Microsoft facilitated the entry of alternative browsers by both ‘untying’ the products in question, thereby allowing original equipment manufacturers (OEMs) not to install Internet Explorer, and not retaliating against any OEM making this choice (which would apply to new buyers), and also by offering existing users of Windows OS the option to choose an alternative browser (via a ‘choice screen’).37 Some might well quibble with this example in that proportionality has quite a different sound when it is performed under Article 9 as opposed to Article 7.38 In fact, this remedy was not imposed by the Commission, but offered by Microsoft. Moreover, what we do not see in the Commission’s assessment of proportionality is a reason why protecting both new and old users is necessary: would a mere prospective commitment to unbundle henceforth not suffice? In the decision fining Microsoft for failing to implement the ‘choice screen’ commitment fully, it was estimated that this had affected around 15 million users.39 This sounds

Commission Decision in Microsoft, above n 32, para 1006. When the dominant player is active both upstream and downstream, then secondaryline injury is a credible antitrust problem. See, eg William Baumol and Daniel Swanson, ‘The New Economy and Ubiquitous Competitive Price Discrimination: Identifying Defensible Criteria of Market Power’, 70 Antitrust Law Journal 661 (2003). 36 This suggestion is found in Hellström et al, above n 7, especially at 58. 37 Commission Decision of 16 December 2009 in Case COMP/39.530—Microsoft (Tying). 38 See Alrosa, above n 5. 39 Commission Decision of 6 March 2013 in Case AT.39530—Microsoft—Tying, para 46 (down from the Commission’s original estimate of 30 million users). 34 35

196 Giorgio Monti impressively high, but, absent any knowledge of how other browser suppliers had fared notwithstanding this infringement, we remain none the wiser as to the necessity to protect established users. This is particularly so because the theory of harm in the decision is foreclosure of rivals, not harm to c onsumers. Nowhere is there a clear statement that an as-efficient competitor would not have managed to enter the market successfully with a lesser commitment. Accordingly, this wide remedy would likely not have survived judicial scrutiny had it been imposed under Article 7. At the same time, one can see the policy attraction for widening the role of remedies in this manner: rather than just prohibiting a specific form of conduct, or protecting a discrete market player, it would be desirable for a competition authority to use remedies to stimulate the market players to act in a manner that maximises welfare. However, as the law stands, this kind of enforcement strategy seems one reserved for commitment decisions under Article 9, not for formal decisions under Article 7. This might explain the surge in the use of commitment decisions: after all, if the deterrent effect of fines is poor, and extensive behavioural commitments can serve to restore competition in the affected market more effectively, then the attraction of using formal decisions is reduced, since the scope for remedies is much less, having regard to the limiting principles set out by the case law.40

3. Practical considerations 3.1 The procedure by which remedies are designed As noted above, the remedies prescribed can be fairly detailed because the Commission seeks to guide the undertaking.41 This makes the following statement by the General Court in Microsoft somewhat unhelpful: When the Commission finds in a decision that an undertaking has infringed Article 82 EC, that undertaking is required to take, without delay, all the measures

40 Query whether, for example, this was a consideration in going for a commitment decision in the E-Books case (Commission Decision of 25 July 2013 in Case AT.39847— E-Books [2013] OJ C378/25), where fines were likely the appropriate remedy but where the Commission might have considered that it enjoyed more latitude in imposing a remedy with a commitment decision. The US Department of Justice, it will be recalled, prosecuted the same infringement, securing settlements from the publisher. The US Court of Appeals for the Second Circuit endorsed this approach. See United States v Apple Inc, No 13-3741 (2d Cir 2015). 41 See also, for example, Article 4 in the Microsoft decision of 24 March 2004, above n 32. The first paragraph of Article 4 of the contested decision provides that Microsoft is to bring to an end the abusive conduct established in Article 2, in accordance with the procedures

Behavioural Remedies for Antitrust Infringements 197 necessary to comply with that provision, even in the absence of specific measures prescribed by the Commission in that decision.42

This statement is the mirror image of the well-known ‘special responsibility’ that dominant players have not to infringe Article 102. Knowing this, however, does not necessarily help one in working out how to act in all circumstances. This is why sometimes the most effective remedy to prevent the reoccurrence of the anticompetitive effects is to guide the parties.43 Securing guidance requires some cooperation between the undertaking and the Commission. In commitment decisions, the procedure by which this takes place is well settled, in that one can see the toing and froing between the parties and the Commission, and the analysis of the market test. The process is set out in DG Competition’s Antitrust Manual of Procedure,44 even if we should not be surprised if in these decisions the Commission probably leans on the parties somewhat.45 What is the corresponding procedure when remedies are imposed in formal decisions? In Commercial Solvents, the ECJ had ruled that the Commission has the power to ‘require the undertaking concerned to submit to it proposals with a view to bringing the situation in conformity with the requirements of the Treaty’.46 This passage is unsurprising when we refer to Article 3(3) of Regulation 17/62, which made express provision for the Commission, prior to making a remedies decision, to address to the undertaking recommendations for termination of the infringement. However, Article 7 of Regulation 1/2003 no longer includes this power. Irrespective of this, as Advocate G eneral Warner noted in Commercial Solvents, the absence of specific guidance would (given periodic penalty payments for non-compliance) place the undertaking in an ‘intolerable position’.47 Procedurally, therefore, there should be a place

laid down in Articles 5 and 6 of that decision. Microsoft is also required to refrain from repeating any act or conduct having the same or equivalent object or effect as that abusive conduct (second paragraph of Article 4). Case T-201/04 Microsoft Corp v Commission [2007] ECR II-3601, para 1256. Indeed, it may even be necessary. There is a tension between the sweeping statement in the Microsoft judgment and the position taken by the same court in Case T-128/98 Aéroports de Paris v Commission [2000] ECR II-3929, paras 81–83, where the applicant complained that the Commission decision did not indicate what was required to comply, and the General Court considered whether this was true and found that a clear obligation had been put in place. This suggests that designing the right form of words for the remedy is vital. 44 See chapter 16. The Manual of Procedure is available at http://ec.europa.eu/ competition/antitrust/antitrust_manproc_3_2012_en.pdf. 45 See the Opinion of Advocate General Capotorti in Case 730/79 Philip Morris Holland BV v Commission [1980] ECR 2671 in the context of settlement procedures under the old regime prior to Regulation 1/2003. 46 Commercial Solvents, above n 2, para 45. 47 Opinion of Advocate General Warner in Commercial Solvents, above n 2, 272. 42 43

198 Giorgio Monti for discussion between the parties and the Commission on how best to design the remedy. Indeed, it is normally the case that the statement of objections also describes possible remedies to counter the anticompetitive conduct. From what I can gather in the reported materials, discussions on remedies are carried out in parallel with the infringement procedure. This is reported to have taken place in Deutsche Post II, where the dominant firm was found to have abused its dominant position, inter alia, by imposing unlawful surcharges, and during the course of the proceedings Deutsche Post had given an undertaking that it would no longer impose surcharges, once three months passed from the issuing of the decision.48 This means that, from the time the Statement of Objections is issued, the defendants are pursuing parallel agendas: one is trying to convince the Commission that the infringement has not taken place, while the other is exploring how the remedy should be designed, should the Commission conclude that there is an infringement. In other words, there is an adversarial stage and a more conciliatory stage that take place simultaneously.49 Sometimes compliance with the remedy may be discussed even after the decision. An example is the Pierre Fabre decision, where the French NCA instructed the undertaking to submit to it its revised contracts to test if they complied with the prohibition.50 In some cases the parties are given time after the decision in order to decide how to comply—so, for example, in Magill the parties had two months to submit proposals of the terms they would introduce to comply—which the Commission would then approve.51 However, even in these cases, the gist of the remedy had been determined at the time of the decision. Sometimes negotiations take place in a different setting from the d ecision: for instance, in the Michelin II rebates decision of 2001, there had been meetings and correspondence between Michelin and the Commission as early as 1998 discussing what commitments the former ‘might enter into in order

48 Commission Decision of 25 July 2001 in Case COMP/36.915—Deutsche Post AG (Deutsche Post II—Interception of cross-border mail) [2001] OJ L331/40, para 78. It appears that these satisfied the Commission, although there is no discussion in the decision save the obvious finding that the abuse was ongoing at the time of the decision (see para 188)— clearly so, since the parties undertook to implement the relevant measures after the decision. 49 This can lead to some awkwardness, as the undertaking may not wish to be seen to admit liability merely by agreeing remedies. See the apparent dilemma faced by Tetra Pak in Commission Decision of 24 July 1991 in Case IV/31043—Tetra Pak II [1992] OJ L72/1, Recitals 178–180, where the firm at times appeared to agree to the Commission’s approach and at others contested it. 50 All the documents relating to this case may be obtained at http://www.autoritedela concurrence.fr/user/avisdec.php?numero=08-D-25. 51 Magill TV Guide, above n 33, Recital 27.

Behavioural Remedies for Antitrust Infringements 199 to bring the infringement to an end’. And by 4 February 1999, Michelin had submitted agreements which in its view would bring the abuse to an end.52 One striking consideration is how relatively more transparent Article 9 procedures are in this phase of proceedings. One might wonder whether initiatives like market tests could also be beneficial in Article 7 decisions, and whether it would not serve some purpose for the final commitments to be made public. This would be particularly important if, as shown above, the remedies may become more wide ranging.53 Regulation 1/2003 does provide that the complainant and those able to show a sufficient interest may have a role in the procedure, but one may query whether the kind of transparency afforded in market tests under Article 9 might be imported.54

3.2 Ex post monitoring In the first Microsoft decision (2004), the remedy included a system for monitoring the firm’s compliance, which was in part annulled by the General Court on two grounds, both of which weaken the Commission’s ability to ensure that there is compliance with Article 7. First, the Commission had delegated too much power to the trustee (he could intervene on the basis of third party complaints and had access to a lot of information held by Microsoft) and the Court considered that these powers could only be exercised by the Commission and not delegated; second, the undertaking should not have to pay for the monitoring mechanism.55 It is unfortunate that this remedy was disallowed, because the monitoring trustees had actually found a violation for which Microsoft was sanctioned.56 This monitoring scheme was substituted by the appointment of technical consultants and by reliance on private parties who could exercise their rights in national courts. It is ‘explained’ in part by the changed nature of the market which made the original set-up unnecessary, but it seems it entails the loss of a major procedural device for checking compliance. Absent this, the Commission may only rely on its investigations or third party complaints to address any non-compliance. One way of facilitating monitoring is to impose a remedy by which the undertaking reveals to its

52 Commission Decision of 31 May 2002 in Case COMP/36.041/PO—Michelin [2002] OJ L143/1, Recital 350. 53 For discussion of the need to tailor procedures in public law in instances where the decision-maker is doing more than imposing liability, see Susan Strum, ‘A Normative Theory of Public Law Remedies’, 79 Georgetown Law Journal 1355 (1991). 54 As noted above (n 44), commitment decisions and the relevant procedure are discussed in chapter 16 of the Antitrust Manual of Procedure. 55 Microsoft, above n 42, paras 1251–1279. 56 See http://ec.europa.eu/competition/sectors/ICT/microsoft/implementation.html.

200 Giorgio Monti c lients that their contracts may now be revised as a result of the Commission’s decision. In JCB, for example, the parties had to inform their distributors that they were free to make passive sales and amend contracts accordingly, but importantly also had to send the Commission copies of these revised agreements.57 A similar approach was taken in Hasselblad, another case dealing with vertical restraints. Here, however, the Commission required that the dealers should be informed using a form of words that the Commission would approve beforehand. In addition to dealers, the undertaking also had to inform the public that aftercare services would be available to all products and not only those purchased in the Member State.58 This kind of approach shifts the implementation costs to the parties, and ensures the effectiveness of the remedy: if the contracts are approved by the Commission, the parties can rely on them in national courts; conversely, if the Commission considers the contracts inadequate to bring the infringement to an end, revisions may be secured. Perhaps even more intrusively, in Tetra Pak II, in addition to listing the practices that the dominant firm must cease and the kinds of conduct that is allowed, the Commission also requested that: [d]uring the period of five years beginning 1 January 1992, Tetra Pak shall, within the first six months of each year, give the Commission a report allowing it to establish if the actions taken by Tetra Pak pursuant to this Decision have indeed brought the infringements detailed in Article 1 to an end.59

This requirement was probably imposed because the Commission considered the abuse particularly egregious. What stands out is that the remedy is time-limited, which is unusual when compared to other cases; that there is no real consequence spelled out explicitly, so its sole purpose seems to be as a monitoring device to ensure compliance with the decision set; and it is not a reporting obligation that may afford the Commission the option of imposing additional remedies if they consider that compliance has not yielded the results sought for. This leads one to consider how far remedies must be final or may be revised.

3.3 Trial-and-error remedies As is well known, damages for a tort are generally awarded once and for all. If the victim needs more or less later, no adjustment is allowed. It seems that

57 Commission Decision of 21 December 2000 in Case COMP/35.918—JCB [2002] OJ L69/1, Article 3. 58 Commission Decision of 2 December 1981 in Case IV/25.757—Hasselblad [1981] OJ L161/18, Article 6. 59 Commission, Tetra Pak II, above n 49, Article 4.

Behavioural Remedies for Antitrust Infringements 201 this is also the case for antitrust remedies imposed under Article 7. The fi nality of the remedy is the reason why a premium is placed on getting the theory of harm right, so that at least Type 1 and 2 errors are avoided at the stage where liability is assessed, even if they are likely inevitable at the remedy stage. In contrast, trial and error might be attempted in commitment decisions, and Microsoft (2009) again furnishes an example: the commitment is to last for five years, but any time after the second year Microsoft or the C ommission may request a review either because market conditions have changed or because the choice screen remedy has not functioned.60 It is obvious that in this review clause the dice are loaded in favour of the review asking for more commitments, not less. It is not easy to see why an authority would be relatively happy to be given the option to change commitments in case of a Type 1 error. This leads one to question whether an undertaking might demand a review of its commitments. First, must the review be expressly provided for in the decision? And second, what grounds might be pleaded? The answer to the first question is that Article 9(2) provides for a set of grounds upon which a decision may be reopened; for current purposes, the relevant one is that there has been a material change in the facts on which the decision was based.61 (Incidentally, this makes the first ground for review in the Microsoft decision above unnecessary.) Whether a more specific review clause is required depends upon the breadth of the express provision in Article 9(2). Obviously if the market conditions change so that the theory of harm is no longer viable, so that the Commission’s concerns do not exist, this is a material change. What if the change is in the costs to the undertaking? Consider a situation like that in Commercial Solvents, where the dominant firm was asked to supply specific numbers of medicines: what if, sometime after the decision, the cost of supplying these medicines increases? Is a greater economic burden on the dominant firm a good ground for seeking a review of the commitment? It seems not, because the proportionality self-assessment carried out by the Commission does not include the burden on the undertaking making the commitment. This suggests that a review clause protecting the undertaking might be warranted. This review clause becomes even more important in Article 7 remedies because there is no statutory review clause in Regulation 1/2003. Not all decisions contain this clause and some harmonisation may be desirable. In Hilti, for example, it was decided that the remedy would last until the undertaking ceased to hold a dominant position and that Hilti would inform the Commission when it proposed to cease abiding by the remedy set.62

Microsoft (Tying), above n 37, Annex I, paras 20–21. Article 9(2)(a) of Regulation 1/2003. 62 Eurofix-Bauco v Hilti, above n 14, Annex, para 3. 60 61

202 Giorgio Monti In contrast, there is no duration in Microsoft, although it would be obvious that if technology changes, the duty to continue to supply would lapse.

3.4 The relationship between remedies and private enforcement In a good number of the remedies discussed above, contracts are rewritten or new contractual relations are formed. This raises the question as to the appropriate relationship between Article 7 and private enforcement. The issue has not been discussed much in the case law. For instance, once exceptional circumstances are identified that make a refusal to supply an abuse, the remedy imposing a duty to deal is imposed as a matter of course: in Microsoft, the Commission said that this would be the ‘natural remedy’.63 In both Commercial Solvents and Magill, the parties with whom the dominant players had to deal were identified and contractual relations were entered into under a Commission order. It is also notable that the duty to supply in some cases (eg Microsoft) is set in terms such that anyone with an interest in developing software may have access to it, even if they are not identified in the decision. However, when it comes to mandatory injunctions under Article 101, the dominant view is different. This view is expressed clearly by Richard Whish: an undertaking infringing Article 101 cannot be ordered to supply a distributor or a customer.64 However, the General Court’s judgment that is said to stand for this proposition, Automec II, is slightly more circumspect. The question had arisen in the context of a complaint by a dealer whose contract to distribute BWM cars had been terminated. One of the arguments made by the Commission to decline the case was that it was not empowered to impose the remedy sought by the complainant. The Court agreed, and explored the nature of the Commission’s remedial powers under Article 101, finding that the Commission did not have the power to order a party to enter into contracts because it must respect freedom of contract; the best it may do is to find that an infringement exists and to order the parties concerned to bring it to an end, but it is not for the Commission to impose upon the parties its own choice from

Microsoft, above n 32, para 998. See Richard Whish and David Bailey, Competition Law, 7th edn, Oxford University Press, 2012, 253. Ivo van Bael took a view closer to that adopted here in the text. Ivo Van Bael, Due Process in EU Competition Proceedings, Wolters Kluwer 2011, 209. Recall that the Automec judgment (Case T-24/90 Automec Srl v Commission (Automec II) [1992] ECR II-2223) was principally about setting out a standard by which to judge whether the Commission had rightly refused to follow up a complaint, and the court explained that the Commission could prioritise cases—it follows that it is free to prioritise an Article 101 case in situations where it wishes to issue a remedy akin to a mandatory injunction. 63 64

Behavioural Remedies for Antitrust Infringements 203 among all the various potential courses of action which are in conformity with the Treaty.65

However, while the Court appears quite categorical in this passage, in the one immediately following it seems to leave the door open for positive remedies in other types of case.66 In sum, the judgment is quite equivocal: some passages suggest the remedy is unavailable for Article 101 infringements, while others suggest it is available but only in certain, undefined, circumstances. The better view is the latter: there may be instances where the Commission may consider it appropriate to require the infringer to deal. The remedy prescribed in Eco System/Peugeot comes near to this: Eco System served as an intermediary for customers based in France who wished to buy a Peugeot car from abroad (where prices are lower). Peugeot sent a circular to three of its dealers not to handle orders from Eco System. On a complaint by the latter, the Commission issued interim measures, followed by a final decision requiring that Peugeot rescind the circular and that ‘the dealers must in future refrain from any behaviour that would perpetuate the effects of the circular complained of’.67 The impact of this remedy was to afford Eco System the possibility of placing orders and continue its business. Obviously, the specifics of the contracts are not supervised by the Commission, but its orders may require contracts to be negotiated in conformity with the antitrust rules. Therefore, it seems that the Commission has the power to impose behavioural remedies by which the infringer confers a benefit on its victims for infringements of both Articles 101 and 102. However, given that these victims would be able to launch follow-on actions to safeguard their position, is the imposition of remedies that benefit specific individuals unnecessary, and so disproportionate? It may be argued that it is more economical for the competition authority to provide private-law-type remedies than to have the entire issue re-litigated. In the UK, the Macrory Report on regulatory sanctions indicated that it may be appropriate to subsume compensation under the heading of sanctions available to regulators, which blurs the elegant line that some like to draw between remedies and sanctions but which appears too formalistic, while a

Automec, ibid, para 52. Ibid, para 53: ‘Accordingly, in the circumstances of the case, it must be held that the Commission was not empowered to issue specific orders requiring BMW to supply the applicant and to allow it to use BMW trademarks’ (emphasis added). 67 Commission Decision of 4 December 1991 in Case IV/33.157—Eco System/Peugeot [1992] OJ L66/1, Article 2; affirmed: Case T-9/92 Automobiles Peugeot SA v Commission [1993] ECR II-493 and Case C-322/93 P Automobiles Peugeot v Commission [1994] ECR I-2727. See also the discussion in Antoine Winckler, ‘Remedies Available in French Law in the Application of EC Competition Rules’, in Claus-Dieter Ehlermann and Isabela Atanasiu, eds, European Competition Law Annual 2001: Effective Private Enforcement of EC Competition Law, Hart Publishing, 2003, 119 et seq (discussing the follow-on claim brought in the French courts). 65 66

204 Giorgio Monti focus on effective enforcement in general is more desirable as a platform for discussing the remedial powers to be handed to competition authorities.68 Furthermore, in the context of many of the cases discussed above, the victims might favour aggressive behavioural commitments to damages, and it may also be the case that national courts would not award the sort of injunctive relief that the Commission was able to extract.69 For example, the British courts stipulate that specific performance in breach of contract cases would only be awarded if no less onerous remedy is available.70 One should add to this the time lapse between the Commission decision and the initiation of procedures at national level, which might be fatal to a plaintiff. For these reasons, it seems appropriate for the Commission to take steps to protect individuals directly through its decisions. Accordingly, there may be good reasons to support the Commission’s use of remedies that confer benefits on given market players.

4. Conclusion As I have noted above, behavioural remedies make infringement decisions a somewhat hybrid affair: they are neither solely premised upon deterrence nor are they solely concerned with resolving competition concerns. Accordingly, they belong to neither the ‘crime/tort model’ of antitrust enforcement nor the ‘administration model’ of enforcement, but fit partly into both sets of enforcement styles.71 Moreover, while they are remedies based in ‘public law’, they can have important effects in safeguarding individual victims of anticompetitive conduct, thus blurring the line between public and private enforcement as well. This chapter has revealed that the scope of behavioural remedies under Article 7 is quite narrow. There are four legal principles that guide the Commission. The first is that the remedy must be directly linked to the infringement: the remedy prevents the recurrence of the infringement, not necessarily of the anticompetitive effects. This is bolstered by a second legal principle, the right of the defence: for the Commission to impose a remedy, it must first build a theory of harm in the infringement decision to explain

68 Richard Macrory, ‘Regulatory Justice: Making Sanctions Effective’ (November 2006), available at http://webarchive.nationalarchives.gov.uk/20090609003228/http://www.berr. gov.uk/files/file44593.pdf. 69 This hints at the restorative justice potential of agreed remedies. See Christine Parker, ‘Restorative Justice in Business Regulation? The Australian Competition and Consumer Commission’s Use of Enforceable Undertakings’, 67 Modern Law Review 209 (2004). 70 Co-operative Insurance Society Ltd v Argyll Stores Ltd [1998] AC 1. The test is also strict for securing interim injunctions. See American Cyanamid v Ethicon [1975] 2 WLR 316. 71 This is the distinction made in Crane, Institutional Structure, above n 8.

Behavioural Remedies for Antitrust Infringements 205 how the remedy cures an infringement. If the Commission fails to do so, no remedy may be imposed for possible, hypothetical violations or for conduct that might have similar anticompetitive effects that have not yet been identified. Thirdly, proportionality also ensures that the remedy does not go beyond what is needed to restore the possibility for competition. A fourth legal principle, that of equal treatment, might at times afford a more intrusive role for remedies: as discussed in the context of the Microsoft case, it may be that the duty to deal with several rivals is more than what is required to restore competition. However, as I have suggested, the role that equal treatment plays in the context of refusals to deal is to avoid secondary line injury. These four principles serve to remove any risk of ‘discretionary remedialism’:72 that is, of the Commission using its power to impose aggressive remedies. They also allow the parties to be clear as to their obligations in the aftermath of an infringement decision. A consequence of these principles is that the Commission may prefer using Article 9 in situations where the competition concerns may require more wide-ranging remedies. Conversely, it may be argued that similar limiting principles could be deployed in curbing the current approach in commitment decisions. Procedurally, remedies are discussed between the Commission and the parties at the same time that the two are also contesting the anticompetitive practice. This might be queried as setting wrong incentives—a party scared of losing may opt for a cooperative attitude in the hope of securing a reduction in the fine, but may then end up with a more intrusive remedy than is needed. This mixing up of the prosecutorial and conciliatory roles of the Commission seems to be a weakness which is remedied by increasingly careful judicial review of the remedies imposed. Nevertheless, one wonders if remedies might be best discussed in separate procedures, which are also more transparent and afford wider participation, as well as market tests. However, this would increase the costs and the duration of procedures, which are already lengthy. The Commission may not pass the monitoring burden entirely away from itself, but in its decisions it has utilised a number of techniques to create incentives for third parties to report infringements as well as requiring the firms themselves to report to the Commission. What might be improved in the context of the remedy order is a clear stipulation as to the duration of the injunctive relief, and it may also be desirable to include review clauses to check if more or less aggressive remedies are required. However, this would add a further layer of costs onto the Commission, which might find it preferable to invest its limited resources elsewhere. Given the hybrid nature of infringement decisions where behavioural remedies are imposed, their close link to the infringement proceedings and

72 This concern animates the paper by Lianos, above n 6, which considers all sorts of remedies.

206 Giorgio Monti the procedural framework within which these remedies are fixed, it seems appropriate not to expand the role of remedies in ways that might allow the Commission to redesign the market to optimise performance. Remedies at best can restore the opportunity for competition by guiding undertakings to avoid conduct that repeats the antitrust offence. While this might appear to be a relatively modest role for remedies in Article 7, this is their scope as explained by the European Courts; and as the examples in this chapter have shown, the Commission has used this power well to try to secure compliance.

Frank P Maier-Rigaud * Behavioural versus Structural Remedies in EU Competition Law

… and it will fall out as in a complication of diseases, that by applying a remedy to one sore, you will provoke another; and that which removes the one ill symptom produces others … (Thomas More, Utopia, p 40)

1. Introduction The effectiveness of competition law in general and any competition authorities’ enforcement efforts in particular depends as much on the actual implementation of the adopted decisions via remedies as on the investigation of the infringement and the finding of liability. The purpose of this chapter is to discuss the role of structural remedies in EU competition law in a non-merger context. The focus is on the distinction between behavioural and structural remedies, the general role of structural remedies and the scope for implementing structural remedies from a legal and economic perspective.1 The motivation for this discussion stems from the curious asymmetry between the relatively frequent use of structural remedies in merger cases on the one hand and their sparse use in antitrust and in particular abuse of dominance cases on the other hand.

* Director and Head of NERA’s European Competition Economics Group; Professor, Department of Economics and Quantitative Methods at the IÉSEG School of Management Paris and the Université Catholique de Lille. The author would like to thank Marc Pirrung and Lars Kjølbye for highly instructive discussions on the topic of remedies throughout the years. Sources referenced in the text and in the following footnotes are listed in full at the end of this chapter. 1 General introductions to remedies under Article 101 and 102 TFEU can be found in Lowe and Maier-Rigaud (2008) as well as Hellström et al (2009). For the merger context, some of the relevant papers are the ex post evaluation exercise carried out by DG COMP (European Commission, 2005) and summarised in Kopke (2005) and the notice on merger remedies (European Commission, 2008). See also Lévêque and Shelanski (2003) and Davies and Lyons (2007). There seems to be a trend towards behavioural remedies also in merger control, at least in the US. See Kwoka and Moss (2011).

208 Frank P Maier-Rigaud This asymmetry seems mainly due to the legal perception that there exists a ‘strong presumption that structural remedies are disproportionate’2 in a nonmerger context and that in implicit recognition of their highly intrusive nature, Article 7(1) of Regulation 1/2003 makes clear that structural remedies are only to be employed in exceptional circumstances … From this, it follows that three cumulative conditions must be satisfied before structural remedies may be imposed …: (1) structural remedies are a remedy of last resort, ie behavioural remedies would be insufficient; (2) structural remedies must be effective; and (3) structural remedies must be proportionate.3

As stated by Wind (2005: 659), ‘Priority is given to behavioural remedies, as structural remedies are used only if it is not possible to use the first’. As will be argued below, this legal interpretation is doubtful, but from a narrow economic point of view it is surprising to find that a suspected (!) substantial lessening of competition that may not even amount to dominance is presumably treated ‘more fiercely’ and ‘rigorously’ than the abuse of an already existing dominant position. As long as the abuse of a dominant position has a more significant anticompetitive impact than the projected lessening of competition of a therefore prohibited merger (or the avoided effects due to a partial divestiture in the context of an approval), such an asymmetry may be incompatible with an economic effects-based approach to competition policy. As a result, it would seem reasonable to make use of the full spectrum of remedies available to the Commission in resolving competition problems in antitrust cases which may benefit not only the consumers and companies harmed by these infringements but ultimately even the perpetrators. This chapter sets out to contribute to the clarification of the role of structural remedies from a competition policy perspective, but also from the perspective of the concerned companies on which such measures would be made binding. In addition to the economic arguments that would speak in favour of rehabilitating structural remedies in dominance cases, it is argued that the Commission clearly has the legal means of following such a more economic approach in the choice of remedies to be imposed. By clarifying this latter point, the chapter contributes to lifting the myth of a primacy of behavioural over structural remedies that does not withstand scrutiny of a detailed analysis of the relevant wording of Regulation 1/2003.4

O’Donoghue and Padilla (2006: 683) and identically in (2013: 887). O’Donoghue and Padilla (2006: 735). The test proposed on these and earlier pages has already been criticised in Adam and Maier-Rigaud (2009: 142f). Unfortunately, these parts have not been updated and are reprinted identically in O’Donoghue and Padilla (2013: 946). 4 See Maier-Rigaud (2012), where some of the key arguments made here were already presented. 2 3

Behavioural versus Structural Remedies in EU Competition Law 209

2. Behavioural versus structural remedies There is no generally accepted definition of structural remedies in the literature, and the discussion of structural remedies has typically focused only on a subset of cases where such remedies could be imposed, namely in merger control. Davies and Lyons (2007) propose a ‘clean break principle’ in order to distinguish between structural and non-structural remedies. According to their definition, a structural remedy is one that neither requires ongoing monitoring by the enforcement authority nor establishes ongoing links between firms. A structural remedy is thus characterised as a one-off measure, as opposed to a measure creating an ongoing relationship between a firm and a regulator or with other firms. This follows the ICN Merger Remedies Review Project Group (ICN, 2005) that characterises structural remedies as ‘one-off remedies that intend to restore the competitive structure of the market’. Another approach, focussing on property rights, is formulated by Motta et al (2003). The approach is based on the notion that ‘structural remedies modify the allocation of property rights and create new firms: they include divestiture of an entire ongoing business, or partial divestiture’.5 This is contrasted to non-structural remedies, which set constraints on the merged firms’ property rights: they might consist of engagements by the merging parties not to abuse of certain assets available to them. They might also consist of contractual arrangements such as compulsory licensing or access to intellectual property.6

Lévêque (2000) finally considers the dichotomy of structural versus behavioural remedies as ‘over-simplifying and confusing’, and proposes instead to characterise remedies along two different dimensions: first, as to their target, such as a firm’s environment, output or property rights; and second, as to whether they change the incentives of the addressee to behave in a certain way or whether they prescribe certain behaviour. In his terminology, economic instruments lead to a change in incentives that make an infringement of antitrust rules a non-profitable strategy, whereas regulatory instruments do not change these incentives and therefore need monitoring and enforcement (command-and-control regime). Council Regulation 1/2003 proposes the terms ‘structural’ and ‘behavioural’ remedy without clear definition. In the present context, a structural remedy is defined as a measure that effectively changes the structure of the firm by a transfer of property rights regarding tangible or intangible assets, including the transfer of an entire business unit, that does not lead to any ongoing

Motta et al (2003: 108). Ibid.

5 6

210 Frank P Maier-Rigaud r elationships between the former and the future owner. After its implementation, a structural remedy does not require any further monitoring. This definition implies that a structural remedy removes the incentive or the means of a firm to repeat the infringement of the antitrust laws that was at the source of the administrative procedure.7 One of the reasons why there is a need for monitoring behavioural remedies stems precisely from the fact that a company may still be facing incentives to circumvent or simply not implement the behavioural remedy. In other words, absent monitoring and enforcement, firm compliance is a dominated strategy. Although a structural remedy also needs some monitoring and—where necessary—enforcement until a divestiture is completed, this concerns only a limited amount of time (usually several months). After implementation, the firm no longer has an incentive to infringe competition law and no further monitoring or enforcement is necessary.8 As stated by the OECD (2001), behavioural policies, unlike structural policies, do not eliminate the incentive of the regulated firm to restrict competition … despite the best efforts of regulators, regulatory controls of a behavioural nature, which are intended to control the ability of an integrated regulated firm to restrict competition, may result in less competition than would be the case if the regulated firm did not have the incentive to restrict competition.9

A behavioural remedy, on the other hand, requires permanent monitoring and enforcement.10 The incentive to circumvent a behavioural remedy also has implications at the design stage. Requiring certain behaviour with respect to one dimension of the strategy space (quantity, price, quality, choice of contractual partners, etc) can easily lead to circumvention by changing the behaviour with respect to another dimension. As a result, behavioural remedies that are designed to prevent circumvention tend to be very detailed and complex. As a consequence, such remedies resemble firm-specific regulation that removes a lot of the flexibility that firms need in order to operate

7 A mandatory, exclusive and non-limited as well as irrevocable licence that is paid up front is in effect equivalent to a structural remedy. 8 There exist structural remedies that may require a behavioural flanking in form of a line-of-business restriction. Without such a restriction, divestitures may only lead to temporary relief if the divested business can easily be replicated or repurchased in the future. 9 See also OECD (2011a: 9, 119), reviewing the OECD Council Recommendation on Structural Separation in Regulated Industries. 10 Given the similarities between behavioural remedies and regulation, the scope of structural remedies extends to situations where competition problems persist despite sector-specific regulation. For an example of structural remedies in a regulated sector see, eg Maier-Rigaud et al (2011), discussing the ENI case, or Hellström et al (2009), discussing, among other aspects, the E.ON case. Noteworthy is also the recent Article 9 decision in Case AT.39727—CEZ on capacity hoarding. On the friction between the US notion of a regulated conduct defence and the approach in the EU, see OECD (2011b), and in particular the contribution by Richard Brunell therein.

Behavioural versus Structural Remedies in EU Competition Law 211 efficiently and compete effectively in a changing environment.11 In such an environment, behavioural remedies usually require ongoing revision and adaptation in order to avoid becoming ineffective and/or detrimental to the competitive process.12 The inflexibility in responding to future dynamic market developments, which cannot be properly antedated either by competition authorities or by market participants, is the key weakness of behavioural remedies.13 Structural remedies, on the other hand, are within the logic of the market system and allow an efficient adaptation to changing market conditions. They are directly compatible with and have a direct bearing on the incentive structure of market participants. Structural remedies do not affect the flexibility of management to make appropriate business decisions on a lasting basis.14 A fundamental property of structural remedies is that, from a spectator point of view, they cannot be distinguished from the usual workings of a market system in which mergers and divestitures are a characteristic feature of normal market developments. In that sense, they take full advantage of market allocation dynamics in resolving the competition problem in an efficient way and allow the concerned businesses to move on. While it is possible to look at remedies from a property rights perspective, ie expropriation versus restriction of use, and therefore consider a divestiture as a harsher remedy than a behavioural one, one may well come to a different conclusion when considering the underlying economics. The primary goal of competition policy is to guarantee the proper and efficient functioning of markets in order to achieve the highest possible degree of welfare for society.15 Structural remedies make use of the dynamics of markets in removing the incentives for committing similar infringements in the future, thereby eliminating competition problems. Behavioural remedies, on the other hand, do not make use of market dynamics but constrain market forces based on some strategy dimension of the firm,16 thereby distorting market

11 Only from a static point of view is the loss in flexibility the deliberate means to remedy the infringement. 12 Besides the necessity to monitor compliance continuously, the market development has to be monitored to ensure the continued appropriateness of the remedy itself, ie to ensure that the remedy is neither creating advantages nor becoming too restrictive for the firm. The same obviously applies with respect to the impact of the remedy on other m arket players, requiring a corresponding impact assessment and potential adjustments to the remedy. 13 The fundamental uncertainty about the future is the main reason for market-based allocation protected by competition policy to begin with and should not be confused with the rather static notion of asymmetric information that an antitrust authority may additionally face. 14 See Levinson et al (2000). 15 For present purposes, the question whether total or consumer welfare is the right measure is secondary. 16 In fact, behavioural remedies typically try to simulate the outcome that would be produced by markets in the absence of an infringement by either imposing such an outcome directly (so-called performance or outcome remedies) or by imposing constraints on the

212 Frank P Maier-Rigaud allocation. As a result, behavioural remedies have a more significant social cost than a property rights approach would suggest. Whenever the underlying competition problem is structural, behavioural remedies will hardly be ‘selfexecuting’. Of course, structural remedies are not self-executing either, since they are not in the interests of the company concerned.17 The major difference is that (appropriately designed) structural remedies do not require ongoing monitoring after their implementation. Finally, as structural remedies generate revenue from the sale of the relevant asset at current market value, a structural remedy is different from expropriation as the firm is remunerated.18 The restriction of the decisional scope of the firm under a behavioural remedy in turn is not compensated as it is considered to be exclusively related to the infringement. This may or may not already be the case at the time the behavioural remedy is imposed, and even behavioural remedies that are successful in a static sense are likely to restrain the firm beyond what is necessary once market conditions have changed. As a result, firms will often prefer to avoid restrictions on their business conduct, as behavioural remedies often come with additional trustee costs and ongoing supervision that bears the risk of detracting attention from the firm’s core business. A structural remedy therefore has the added advantage of allowing the firm to move on.

3. Conditions under Regulation 1/2003 Leaving mergers aside, structural remedies in European competition law occur either in the context of an infringement decision under Article 7 or in the context of a commitment decision under Article 9 of Regulation 1/2003. Under Article 7, an infringement is identified by decision and is terminated with remedies, whereas Article 9 renders commitments offered by the parties

behaviour of the firm in order to achieve this outcome indirectly (so-called process or conduct remedies). The presumption that market outcomes in a dynamic context can be rationally anticipated is the key weakness of behavioural remedies. From that perspective, it is surprising to see how widespread behavioural measures are in a competition law context as they are fundamentally at odds with the idea that a decentralised competitive process (constrained by competition law) alone allows the efficient allocation of resources and the highest welfare. 17 If the concept is taken literally, antitrust remedies are by definition never selfexecuting. Antitrust remedies are designed and adopted in order to end an infringement that occurred precisely because of existing incentives pushing the firm towards infringing competition law. 18 Of course, the strategic value of the asset, enabling the firm to partake in anticompetitive conduct with abnormal returns, is not priced-in. Problems may also arise if the number of buyers is low or if the market is already heavily concentrated—rendering the asset uninteresting for new entrants, for example.

Behavioural versus Structural Remedies in EU Competition Law 213 binding if those address and eliminate the competition concerns voiced by the Commission without formal finding of an abuse.19 The distinction between Article 7 and Article 9 would once not have been considered important for the purposes of a discussion of behavioural versus structural remedies, with the CFI20 arguing that the treatment under Article 7 is also the benchmark for Article 9 cases. This has, however, partially been reversed by the ECJ,21 which stated that, under Article 9, the Commission’s obligation in relation to the principle of proportionality ‘is confined to verifying that the commitments in question address the concerns it expressed to the undertakings concerned and that they have not offered less onerous commitments that also address those concerns adequately’.22 As a result, the proportionality requirements under Article 7 are stricter than those under Article 9, so that focusing on Article 7 is appropriate in exploring the possibilities for imposing structural remedies. Article 7(1) of Regulation 1/2003 provides the legal basis for finding and terminating an infringement of EU competition law. This provision, with emphasis added, reads as follows: Article 7(1) Finding and termination of infringement 1. Where the Commission, acting on a complaint or on its own initiative, finds that there is an infringement of Article 81 [101] or of Article 82 [102] of the Treaty, it may by decision require the undertakings and associations of undertakings concerned to bring such infringement to an end. 2. For this purpose, it may impose on them any behavioural or structural remedies which are proportionate to the infringement committed and necessary to bring the infringement effectively to an end. 3. Structural remedies can only be imposed either where there is no equally effective behavioural remedy or where any equally effective behavioural remedy would be more burdensome for the undertaking concerned than the structural remedy. 4. If the Commission has a legitimate interest in doing so, it may also find that an infringement has been committed in the past.

It is preceded by Recital 12, which largely reiterates the Article in stating: This Regulation should make explicit provision for the Commission’s power to impose any remedy, whether behavioural or structural, which is necessary to bring the infringement effectively to an end, having regard to the principle of proportionality. Structural remedies should only be imposed either where there is no equally effective behavioural remedy or where any equally effective behavioural remedy would be more burdensome for the undertaking concerned than the structural

19 A general introduction to remedies under EU law including a discussion of Articles 7, 8 and 9 can be found in Lowe and Maier-Rigaud (2008). 20 See Case T-170/06 Alrosa Company Ltd v Commission [2007] ECR II-2601, para 101. 21 See Case C-441/07 P Commission v Alrosa Company Ltd [2010] ECR I-5949, para 45. 22 Ibid, para 41.

214 Frank P Maier-Rigaud r emedy. Changes to the structure of an undertaking as it existed before the infringement was committed would only be proportionate where there is a substantial risk of a lasting or repeated infringement that derives from the very structure of the undertaking.

As will be discussed later, the last sentence is merely an example of conditions that need to be met under the proportionality screen, ie that structural remedies may not be proportionate where they would alter the structure of the undertaking as it existed prior to the infringement (in contrast, for example, to vertical agreements, where the infringement modifies the structure) unless there is a risk of a lasting or repeated infringement in the future. In the following, the meaning of the first three sentences that make up Article 7(1) will be analysed in turn to provide an economically sound interpretation of Article 7(1) that, as will be subsequently argued, is also in line with the original intent of the drafters of these provisions.

3.1 Article 7(1)—the first sentence Article 7(1)1 is an introduction to the Commission’s powers of requiring offenders to put an end to antitrust infringements. It specifies the domain of application and authorises the Commission (‘the Commission may’) to force the undertakings or associations of undertakings to put an end to the competition infringement identified. To ‘bring such an infringement to an end’ must, however, be interpreted broadly, as it is perfectly conceivable to envision (in particular, behavioural) measures that ‘bring the infringement to an end’ in the short run but cannot guarantee that effect in the longer term. Such a situation may in particular arise if incentives remain to commit competition law violations in the future.23 As a result, and as is further specified in Article 7(1)2, what must be chosen is not any possible measure that brings the infringement to an end, but only those that effectively end the infringement. This emerges further from

23 The importance of incentives in the context of structural remedies has in particular been recognised in energy cases. See, eg Maier-Rigaud et al (2011) for a description of the ENI case. The fact that incentives for infringing competition law need to be removed to effectively bring an infringement to an end is, for instance, also acknowledged by Kjølbye and Gauer (2014: 1631f) in their discussion of the RWE and ENI cases: ‘In these cases the infringement stemmed from the very structure of the undertaking, namely the vertical integration. The undertakings had an incentive to maximize the profit of the group as a whole and to leverage their control of the network to their benefit on the downstream market. Any remedy that would not have removed this incentive would not have been equally efficient.’

Behavioural versus Structural Remedies in EU Competition Law 215 Recital 12 of Regulation 1/2003, which emphasises the risk of a lasting or repeated infringement in its final sentence: Changes to the structure of an undertaking as it existed before the infringement was committed would only be proportionate where there is a substantial risk of a lasting or repeated infringement that derives from the very structure of the undertaking.

In addition, the Ufex case law has been interpreted as suggesting that ‘bringing an infringement effectively to an end’ is not only a matter of effectively ending the conduct, but also requires ending its effects, ie its repercussions on the market.24 Based on Ufex, ‘bringing an infringement effectively to an end’ therefore encompasses terminating the consequences and repercussions (ongoing effects) of the infringing conduct that may continue completely decoupled from the fact that the conduct that originally caused the effects has ceased.25 It is not the subject of this chapter to discuss the inherent risks associated with any approach that would aim to eliminate all consequences and repercussions of an infringement in the name of putting ‘an infringement effectively to an end’. Suffice it to say that this aspect is not evidently grounded in Regulation 1/2003 and that it opens up the major difficulty of establishing the status quo ante, a process that may (positively and negatively) affect a range of economic actors and consumers that were neither part of the infringement nor aware of its existence.26 Irrespective of whether one would like to follow this interpretation of Ufex or not, the question of the role of structural versus behavioural remedies remains irrespective of whether eliminating the effects caused by the infringement is considered a legitimate policy goal.

24 See Case C-119/97 P Union française de l’express (Ufex), formerly Syndicat français de l’express international (SFEI), DHL International and Service CRIE v Commission [1999] ECR I-1341, paras 93 et seq. 25 ‘If anti-competitive effects continue after the practices which caused them have ceased, the Commission thus remains competent’ (Ufex, para 94); ‘the Commission therefore cannot rely solely on the fact that practices alleged to be contrary to the Treaty have ceased, without having ascertained that anti-competitive effects no longer continue’ (Ufex, para 95). On Ufex, see Hellström et al (2009: 45f) or Dalheimer et al (2005: 56). See also the CEZ commitment decision (Commission Decision of 10 April 2013 in Case AT.39727— CEZ), in which it is succinctly stated in para 79 that ‘Given that the conduct which gave rise to the Commission’s concerns consisted of pre-emptive booking of transmission capacity which in turn effectively prevented new entry into the Czech electricity market, transfer of some of ČEZ’s generation capacity to a competitor represents a clear-cut solution to the identified competition concerns. Transfer of generation capacity is necessary in this case as no other type of remedy can address the effects of ČEZ’s conduct.’ This seems to be an example of structural remedies following the Ufex logic. It would have been preferable if the remedies in ČEZ had been motivated by the continuing incentive of ČEZ to engage in pre-emptive booking. The commitment of transferring generation capacity could thus have been motivated as a remedy eliminating the incentive to overbook in the future (thereby effectively ending the infringement), as opposed to a way of addressing the persisting effects of the conduct in an effort to turn back the clock. 26 As an indication of the difficulties involved, see the discussions of damages claims in the context of Article 102 TFEU cases in Maier-Rigaud and Schwalbe (2013a, 2013b).

216 Frank P Maier-Rigaud

3.2 Article 7(1)—the second sentence Article 7(1)2 explains in more detail what is meant under Article 7(1)1 and that the Commission does not have to stop at a cease-and-desist order but that positive remedies may be imposed. For the purpose of bringing antitrust infringements to an end, any necessary remedy that is proportionate to the infringement can be employed, ie any type of remedy, behavioural or structural, can be imposed if the general filters of proportionality and necessity are fulfilled.27 As this sentence does not yet differentiate between behavioural and structural remedies, the main qualification concerns measures that go beyond a simple cease-and-desist and the general filters. Proportionality is typically not an ultimate selection criterion for remedies, and the same applies to necessity. Necessity is a general filter that determines whether the remedy is generally capable of bringing the infringement to an end, does not go beyond this goal and does not concern aspects that are not part of the procedure. In contrast to German law, the European concept of proportionality says nothing about effectiveness, just as effectiveness says nothing about proportionality. A more effective remedy is therefore not automatically a less proportionate one in the sense of a milder means.28

3.3 Article 7(1)—the third sentence As already suggested above, the infringement not only needs to be brought to an end, but needs to be brought effectively to an end, possibly requiring the use of remedies that go beyond a cease-and-desist order. What does the deliberate qualification of ‘effectively’ in Article 7(1)2 or the use of ‘equally effective’, as employed in Article 7(1)3, suggest? If it is possible to bring an infringement effectively to an end, it must also be possible, as discussed above, that an infringement is brought to an end but that this is not effective, ie that it does not last and that the infringement, albeit brought to an end temporarily, is ineffectively brought to an end, for instance, from a more dynamic, longer term perspective. This may arise, for example, in the case of a cease-and-desist order or a behavioural remedy that does not effectively eliminate the underlying incentives to commit competition law violations.29

27 Supposedly the same criteria would apply concerning remedies aimed at restoring the status quo ante. 28 On the issue of structural versus behavioural remedies in general and the specifics of the proportionality test under German law in particular, see Maier-Rigaud (2012). 29 It has been argued that it is sufficient for conduct to cease as it is the role of fines to ensure deterrence. From an economic perspective, however, it may be preferable to e liminate

Behavioural versus Structural Remedies in EU Competition Law 217 The reference to ‘equally effective’ remedies suggests that effectiveness is a gradual concept, that it can be measured ranging, for example, from noneffective to less effective to highly effective. If that were not the case, the requirement in Article 7(1)3 would only be ‘effective’ in order to contrast it to a situation where the comparison is made between an effective and an ineffective remedy in a binary interpretation of the concept of effectiveness. The argument becomes clearer when considering the French version of the relevant sentences: A cette fin, elle peut leur imposer toute mesure corrective de nature structurelle ou comportementale, qui soit proportionnée à l’infraction commise et nécessaire pour faire cesser effectivement [in the sense of ‘to actually terminate’] l’infraction. Une mesure structurelle ne peut être imposée que s’il n’existe pas de mesure comportementale qui soit aussi efficace [in the sense of reaching the goal] ou si, à efficacité égale, cette dernière s’avérait plus contraignante pour l’entreprise concernée que la mesure structurelle.

Whether one considers the use of the term ‘effectively’ in sentence 2 to be different from the use of the term ‘equally effective’ in sentence 3, as suggested in the French version, is not that important. What the French version renders clear, however, is that the effectiveness of a remedy is not a binary measure according to which a whole range of measures may be classified as either effective or not effective, but a gradual concept. If that were not the case and the term ‘equally’ were intended to mean ‘as well’, all three official language versions could dispense with the qualifier, ie the ‘equally’ in front of ‘effective’, the ‘aussi’ in front of ‘efficace’ and the ‘gleicher’ in front of ‘Wirksamkeit’, as, indeed, it would only be a question of ‘effective’ or ‘ineffective’ where the latter remedies would automatically be disqualified. Additional support for this interpretation is found in De Smijter and K jølbye (2007: 120), who were both directly involved in the negotiations c oncerning Regulation 1/2003: ‘If one proportionate remedy is more effective than other available remedies, the Commission can impose the most effective one’.30 If the gradual nature of the concept of ‘effectiveness’ is accepted, it raises the obvious question of which remedy is to be considered the most effective.

the business rationale for committing infringements directly instead of putting the pressure on fines to achieve the same effect indirectly. While this paper is not about the appropriateness of fines in general or specific levels of fines, it is clear that the approach to remedies foreseen under Article 7 generates less pressure on fines. This may be relevant in the context of the current debate on fines as, clearly, every time a fine is paid, deterrence has failed. 30 In addition to clarifying the gradual nature of effectiveness, De Smijter and K jølbye (2007) also clarify that the effectiveness of the remedy relates not only to the capacity of the remedy to address a competition problem, but also to its enforceability. This is an important aspect, as the authors clearly point to the higher burden on the competition authority in the monitoring of behavioural remedies. They write: ‘For a remedy to be e ffective it must be

218 Frank P Maier-Rigaud

3.4 The test: how to choose among proportionate and necessary remedies Having discussed the individual sentences of Article 7(1) above, it is now possible to put the different elements together in order to reconstruct the implicit test that Regulation 1/2003 prescribes and that the Commission is applying. To determine the remedy to be imposed out of a large set of potential candidates, one can think of Article 7(1) as a filtering process, as depicted in Figure 1. Pool of all potential remedies necessity filter

Pool of necessary remedies proportionality filter

Pool of necessary and proportionate remedies

A. Only one remedy passes both filters

effectiveness filter

No distinction between structural and behavioural remedies

Pool of necessary, proportionate and equally effective remedies

B. Only one remedy passes all filters

Pool of homogenously structural or homogenously behavioural remedies

Pool of mixed behavioural/structural remedies

C. Firm is allowed to choose the remedy

D. Least burdensome remedy has to be chosen

No distinction between structural and behavioural remedies

Figure 1: The Filters of Article 7

Behavioural versus Structural Remedies in EU Competition Law 219 The first filter foreseen is the one of Article 7(1)2 that specifies that any measure that is necessary to bring an infringement to an end and that is proportionate to the infringement can be imposed. Necessity and proportionality are criteria that any remedy, irrespective of whether it is a structural or a behavioural remedy, needs to fulfil.31 This filter implies that, among all necessary and proportionate remedies, the most effective one is to be chosen conditional on Article 7(1) sentence 3,32 which specifies that: Structural remedies can only be imposed either where there is no equally effective behavioural remedy or where any equally effective behavioural remedy would be more burdensome for the undertaking concerned than the structural remedy.

Article 7(1)3 allows for exactly four possibilities. Either there is only a single remedy that is both proportionate and necessary (Outcome A in Figure 1)

possible to monitor compliance. Complex behavioural remedies are often ineffective. As the complexity of a remedy increases, the less likely it is that it will address the competition problem effectively. Behavioural remedies often suffer from a high degree of complexity relative to structural remedies.’ That the problem of complexity is not only one concerning the monitoring of the competition authority but can also negatively affect firms will be discussed below. Similarly, Kjølbye and Gauer (2014: 1631) state, in the context of their discussion of the incentives faced by RWE and ENI as vertically integrated companies, that ‘Any remedy that would not have removed this incentive would not have been equally efficient. In addition, the difficulties inherent in monitoring any conceivable behavioural remedy (eg, a commitment to market all unused capacity or to avoid a margin squeeze) would arguably have rendered it ineffective.’ 31 The last sentence of Recital 12 discusses proportionality of structural remedies, istinguishing the case where the infringement is based on conduct from the case where the d structure of the undertaking—as, for example, in the case of a joint venture agreement—is itself the infringement. It only specifies explicitly what would follow anyhow from the proportionality test, namely that a structural remedy modifying the structure of the undertaking as it existed prior to the infringement can only be considered proportional if there is ‘a substantial risk of a lasting or repeated infringement’. Recital 12 therefore only clarifies that structural remedies are an obvious remedy for example in vertical agreements, where the ‘structural agreement’ is the infringement, but also remain proportional in cases where the structure of the undertaking is driving the infringing conduct so that a structural remedy is required in order to eliminate the incentives to infringe competition law in the future. As long as the infringement directly derives from the structure of the undertaking, either as it was created by the infringement or as it existed prior to the infringement, and as long as the incentives provided by that structure continue to push for a repetition of the infringement unhampered, structural remedies will be proportionate. In that sense, Recital 12 does not add anything to the test except maybe to clarify Magill on the question of effectiveness and proportionality (see paras 89–92 in Joined Cases C-241/91 and 242/91 P Radio Telefis Eireann (RTE) and Independent Television Publications Ltd (ITP) v Commission (Magill) [1995] ECR I-743). It should be noted that Kjølbye and Gauer (2014: 1634), in their discussion of the CEZ case, interpret Recital 12 more narrowly as describing a situation of vertical break-up. They argue that ‘Recital 12 cannot be understood as limiting the imposition of any type of structural remedy to infringements deriving from the very structure of the undertakings’, drawing a line between the divestiture of some of the company’s assets and full vertical separation. Their interpretation is not identical to the one proposed here, but entails the same consequences. 32 Recital 12, second sentence is identical down to one word: the word ‘can’ is replaced by ‘should’.

220 Frank P Maier-Rigaud or there are many remedies that are proportionate and necessary but only a single of these remedies can be considered the most effective one (Outcome B in Figure 1). If either of these two cases arises, the third sentence is irrelevant and the distinction between structural and behavioural remedies is super fluous in terms of the test. If, however, there are several remedies that are all tied on first place in terms of effectiveness, two more situations can be distinguished (Outcomes C and D in Figure 1). If all tied remedies are either uniformly behavioural or uniformly structural, the concerned undertaking should be allowed to choose (Outcome C in Figure 1).33 If the tied remedies include both behavioural and structural measures, then the remedy that has to be chosen is the one that is the least burdensome for the undertaking (Outcome D in Figure 1). In either of these two additional cases, the behavioural or structural nature of the remedy plays only an indirect role in the sense that the composition of the pool of remaining remedies determines whether the concerned undertaking chooses the remedy or whether the least burdensome remedy is imposed. It is often assumed that this will typically be the behavioural remedy. This is, however, neither conclusive nor empirically to be expected.34 As a result, and despite a largely contrary literature, Council Regulation 1/2003 does not imply a preference for behavioural remedies (nor the subsidiarity of structural remedies). On the contrary, the structural or behavioural nature of a remedy is immaterial as long as such remedies are not equally effective (Outcomes A and B in Figure 1). Only in the highly theoretical case of when they cannot be distinguished by the very criteria set forth in Article 7 (Outcomes C and D in Figure 1) does Article 7 suggest drawing on the distinction between behavioural and structural remedies to determine whether the concerned undertaking should choose (Outcome C in Figure 1) or the least burdensome measure has to be adopted (Outcome D in Figure 1). Arguably, even that is not a real distinction, as the least burdensome measure is not unlikely to be the one that the concerned undertaking would be choosing.

33 This can be derived from the case law. See, eg Case T-24/90 Automec Srl v Commission [1992] ECR II-2223, paras 51 et seq. 34 In the E.ON case (Commission Decisions of 26 November 2008 in Case COMP/39.388—German Electricity Wholesale Market and Case COMP/39.389—German Electricity Balancing Market), this question was explicitly dealt with. After stating that there was no equally effective behavioural alternative, the question of whether a hypothetical equally effective behavioural remedy would have implied a lower burden for the firm is discussed. That is later found not to be the case, and not only because E.ON never proposed such a behavioural remedy. In any case, this is an empirical question and not one of legal interpretation. On the E.ON cases, see Chauve et al (2009) or Hellström et al (2009) for a discussion of the remedies.

Behavioural versus Structural Remedies in EU Competition Law 221

4. Conclusion In its proposal for what later became Council Regulation 1/2003,35 the Commission suggested the following wording for Article 7: Article 7(1) Finding and termination of infringement Where the Commission, acting on a complaint or on its own initiative, finds that there is an infringement of Article 81 or of Article 82 of the Treaty, it may by decision require the undertakings and associations of undertakings concerned to bring such infringement to an end. For this purpose, it may impose on them any obligations necessary, including remedies of a structural nature. If it has a legitimate interest in doing so, it may also find that an infringement has been committed in the past.

The proposal also contains a justification for that wording: Secondly, the Commission is empowered to impose all remedies necessary to bring the infringement to an end, including structural remedies. Structural remedies can be necessary in order to bring an infringement effectively to an end. This may in particular be the case with regard to cooperation agreements and abuses of a dominant position, where divestiture of certain assets may be necessary.

Even if this wording appears to be different from Article 7 as well as Recital 12 of Council Regulation 1/2003, the Commission, albeit at the price of an interpretable formulation, seems to have asserted itself despite the intense and controversial debates at the time. This becomes apparent if one changes the sentence structure of Article 7 slightly and thereby visibly turns the (wrong) impression that one may get from the current formulation on its head. The following content preserving reformulation of Article 7(1)3 demonstrates how misleading the current formulation is and that this ultimately contributed to the myth of a subsidiarity of structural measures. A logically equivalent formulation of the third sentence of Article 7(1), ie a formulation that prompts the same action as the original, assuming a non-binary interpretation of effectiveness, is the following sentence. In contrast to the original version, the wording now rather suggests (with emphasis added) a subsidiarity of behavioural measures: Behavioural remedies can only be imposed either where there is no more effective structural remedy or where any equally effective structural remedy would be equally or more burdensome for the undertaking concerned than the behavioural remedy.

35 See Proposal for a Council Regulation on the implementation of the rules on c ompetition laid down in Articles 81 and 82 of the Treaty …, COM(2000) 582 final of 27 September 2000, http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2000:0 582:FIN:EN:PDF.

222 Frank P Maier-Rigaud Even if this wording no longer suggests a priority for behavioural remedies, the (exclusively empirical) question of the relative effectiveness of necessary and proportionate structural and behavioural remedies remains. An additional indication that speaks against a priority of behavioural over structural remedies consists in Council Regulation 17/62. In the context of the negotiations of Council Regulation 1/2003, which were overshadowed by the discussions of structural remedies in the Microsoft case, Council Regulation 17/62 represented the status quo and therefore the fall-back position of the European Commission. As the Commission already had the powers to terminate an infringement with structural remedies prior to Council Regulation 1/2003, the Commission would have likely preferred to forgo any mention of structural remedies instead of implementing categorical limitations of its powers.36 The fact that the Commission did not simply drop the structural remedies from Article 7 reverting to the old formulation but ultimately accepted the version of Article 7 that appears in Regulation 1/2003 today can, from a negotiation point of view, only easily be reconciled with the interpretation offered here. Finally, the finding that there is an important role for structural remedies in competition law is hardly surprising. Efficiency considerations should of course also play a role in the remedy design stage. This follows directly from the justification for competition law enforcement, which has emphasised the role of efficiencies at least since the introduction of Regulation 1/2003. If, however, one of the reasons for having competition law enforcement to begin with is its efficiency increasing impact on the economy overall, it would be inconsistent to restrict such consideration to the identification of competition law violations and generally exclude efficiency considerations at the remedial stage, even if a trade-off between efficiency and increased interference with (fundamental) property rights is perceived. The fact that the European Commission clearly has the legal means to follow a more economic approach also in this often neglected but absolutely central part of competition law enforcement is reassuring, and will typically not be detrimental even to those firms subjected to structural remedies.

36 The thesis that Regulation 17/62 did not foresee structural remedies has often been repeated. While Regulation 17/62 did not mention structural remedies explicitly, the Commission had already imposed such remedies prior to Regulation 1/2003 outside the realm of the Merger Regulation, including, in particular, in the Continental Can and Magill cases. See, eg Gauer et al (2003) or Riley (2003: 607). Even in the hypothetical absence of structural remedies and these court cases, the absence of any explicit reference to structural remedies in Regulation 17/62 can clearly not be interpreted as implying that such measures could not have been imposed. To take another example from competition law, only the eighth version of the German GWB started to mention structural remedies explicitly, but such measures were already permitted prior to the changes in 2013.

Behavioural versus Structural Remedies in EU Competition Law 223

Bibliography Adam, Michael and Frank Maier-Rigaud (2009) ‘The Law and Economics of Article 82 EC and the Commission Guidance Paper on Exclusionary Conduct’ 1 Journal of Competition Law (Zeitschrift für Wettbewerbsrecht) 131. Chauve, Philippe; Godfried, Martin; Kovács, Kristoff; Langus, G; Nagy, Károly and Stefan Siebert (2009) ‘The E.ON electricity cases: an antitrust decision with structural remedies’, EU Competition Policy Newsletter 1, 51–54, http://ec.europa. eu/competition/publications/cpn/2009_1_13.pdf. Dalheimer, Dorothe; Feddersen, Christoph and Gerald Miersch (2005) EU-Kartellverfahrensordnung—Kommentar zu VO 1/2003 (CH Beck). Davies, Stephen and Bruce Lyons (2007) Mergers and Merger Remedies in the EU: Assessing the Consequences for Competition (Edward Elgar). De Smijter, Eddy and Lars Kjølbye (2007) ‘The Enforcement System Under Regulation 1/2003’, in J Faull and A Nikpay (eds), The EC Law of Competition, 2nd edn (Oxford University Press) 87 et seq. European Commission (2005) Merger Remedies Study, http://ec.europa.eu/ competition/mergers/legislation/remedies_study.pdf. European Commission (2008) Commission Notice on remedies acceptable under the Council Regulation (EC) No 139/2004 and under Commission Regulation (EC) No 802/2004, [2008] C267/1. Gauer, Céline; Dalheimer, Dorothe; Kjølbye, Lars and Eddy de Smijter (2003) ‘Regulation 1/2003: A Modernised Application of EC Competition Rules’, EC Competition Policy Newsletter 1, 3–8, http://ec.europa.eu/competition/publications/ cpn/2003_1_3.pdf. Hellström, Per; Maier-Rigaud, Frank and Friedrich Wenzel Bulst (2009) ‘Remedies in European Antitrust Law’ 76 Antitrust Law Journal 43. ICN (2005) Merger Remedies Review Project, International Competition Network Merger Working Group: Analytical Framework Subgroup, Bonn, June, http://www. internationalcompetitionnetwork.org/uploads/library/doc323.pdf. Kjølbye, Lars and Celine Gauer (2014) ‘Energy’, in J Faull and A Nikpay (eds), The EU Law of Competition, 3rd edn (Oxford University Press) ch 12. Koch, Oliver; Nagy, Károly; Pucinskaite-Kubik, Ingrida and Walter Tretton (2009) ‘The RWE Gas Foreclosure Case: Another Energy Network Divestiture to Address Foreclosure Concerns’, EC Competition Policy Newsletter 2, 32–34. Kopke, Alexander (2005) ‘Merger Remedies Study’, EU Competition Policy Newsletter 3, 3–5. Kwoka, John and Diana Moss (2011) ‘Behavioral Merger Remedies: Evaluation and Implications for Antitrust Enforcement’, working paper, http://ssrn.com/ abstract=1959588. Lévêque, François (2000) ‘The Conduct vs Structural Remedies Controversy: An Irrelevant Dichotomy of Antitrust Policy Instruments’, CERNA, Ecole Nationale Supérieure des Mines de Paris working paper, http://www.cerna.ensmp.fr/ Documents/FL-MSMay00.pdf. Lévêque, François and Howard Shelanski (eds) (2003) Merger Remedies in American and European Union Competition Law (Edward Elgar).

224 Frank P Maier-Rigaud Levinson, Robert; Romaine, Craig and Steven Salop (2000) ‘The Flawed Fragmentation Critique of Structural Remedies in the Microsoft Case’, Georgetown University Law Center, Business, Economics, Regulatory Law Working Paper No 204874. Lowe, Philip and Frank Maier-Rigaud (2008) ‘Quo Vadis Antitrust Remedies’, in Barry Hawk (ed), International Antitrust Law & Policy: Fordham Competition Law 2007 (Juris Publishing) ch 20, 597–611. Maier-Rigaud, Frank (2012) ‘Zur Idee der Subsidiarität struktureller Maßnahmen im europäischen Wettbewerbsrecht’ 5 Journal of German and European Competition Law (Wirtschaft und Wettbewerb) 487, http://ssrn.com/abstract=1906335. Maier-Rigaud, Frank and Ulrich Schwalbe (2013a) ‘Quantification of Antitrust Damages’, in David Ashton and David Henry (eds), Competition Damages Actions in the EU: Law and Practice (Edward Elgar), http://ssrn.com/abstract=2227627. Maier-Rigaud, Frank and Ulrich Schwalbe (2013b) ‘Private Enforcement under EU Law: Abuse of Dominance and the Quantification of Lucrum Cessans’ 11(2) CPI Antitrust Chronicle 1, http://ssrn.com/abstract=2359327. Maier-Rigaud, Frank; Manca, Federica and Ulrich von Koppenfels (2011) ‘Strategic Underinvestment and Gas Network Foreclosure—the ENI Case’, EC Competition Policy Newsletter 1, 18–23, http://ssrn.com/abstract=1903788. Motta, Massimo; Polo, Michele and Helder Vasconcelos (2003) ‘Merger Remedies in the European Union: An Overview’, in François Lévêque and Howard Shelanski (eds), Merger Remedies in American and European Union Competition Law (Edward Elgar). O’Donoghue, Robert and Jorge Padilla (2006) The Law and Economics of Article 82 EC (Hart Publishing). O’Donoghue, Robert and Jorge Padilla (2013) The Law and Economics of Article 102 TFEU, 2nd edn (Hart Publishing). OECD (2001) Recommendation of the Council on Structural Separation in Regulated Industries, http://www.oecd.org/dataoecd/24/49/25315195.pdf. OECD (2011a) Recent Experiences with Structural Separation: A Report to the Council on Implementation of the 2001 Recommendation Concerning Structural Separation in Regulated Industries, DAF/COMP(2011)12, http://www.oecd.org/ daf/competition/50056685.pdf. OECD (2011b) Regulated Conduct Defence, OECD Policy Roundtable, http://www. oecd.org/regreform/sectors/48606639.pdf. Riley, Alan (2003) ‘EC Antitrust Modernisation: The Commission Does Very Nicely— Thank You! Part One: Regulation 1 and the Notification Burden’ 24 European Competition Law Review 604. Wang, Wei (2011) ‘Structural Remedies in EU Antitrust and Merger Control’ 34 World Competition 571. Wind, Elena (2005) ‘Remedies and Sanctions in Article 82 of the EC Treaty’ 26 European Competition Law Review 659.

3 AGENCIES AS AMICI CURIAE Chair: Giorgio Monti Speaker: Stephen Calkins Calkins: The first point is that amicus participation in the US is really an amazingly big part of the US legal system. In the Supreme Court this last term there were 77 cases decided, 1046 amicus briefs for an average of thirteen and a half per case, with the leading cases attracting 97, 92, 82 and 80 amicus briefs.1 The leading competition case there, which was a straight antitrust case, was FTC v Actavis, with 27 amicus briefs.2 It’s a big part of what the agencies do: if you go to the FTC it has an amicus website, and if you go back to 1998 you’ll get around 90 FTC amicus briefs; and if you go to the Justice Department, its similar website reports filings of 70 amicus briefs. But do all these briefs make a difference? The answer is that there are lots of different ways that people look at things, and generally all perspectives give a sense that amicus participation makes a difference both in deciding what cases the US Supreme Court takes on board and then to some extent deciding the merits of different cases. There are a number of different things that people look at to draw those conclusions, and I am not going into that now. There are wonderful examples of cases where you can see that an amicus participation may have made a difference, the caveat being that you can never know for sure; all judges and justices could have done everything without any advice or suggestions from anybody, so you never know for sure. You do get surveys of former law clerks and interviews with judges that suggest that amicus briefs really do make a difference, but again you can never be sure. There DD

1 American Bar Association Preview of Supreme Court Cases, http://www.americanbar. org/publications/preview_home/alphabetical.html. The breakdown is as follows: 97 briefs in Hollingsworth v Perry, No 12-144; 92 briefs in Fisher v University of Texas at Austin, No 11-345; 82 briefs in Kiobel v Royal Dutch Petroleum, No 10-1491; and 80 briefs in United States v Windsor, No 12-307. 2 FTC v Actavis, Inc, No 12-416.

226 Part I: Effective Enforcement of Competition Law are a couple of cases I would highlight—one is Phoebe Putney.3 This is a state action case worth talking about. The issue arose because the defendant in a hospital merger case was basically trying to say that a rather general authorisation from the state showed that the state had contemplated anticompetitive behaviour, so therefore the merger should be immune to antitrust attack. The problem with this argument was in part revealed when Justice Kagan in oral argument said to him: You are arguing that these general prohibitions show a contemplation of anticompetitive behavior but I have here a brief from 20 states pleading with us not to agree with you or else they’ll be royally in trouble every time such legislation comes along.4

This reminded me of one of the wonderful amicus cases back in the Kodak tying case, when the defendants to this one were arguing that there’s no problem in linking a product to an aftermarket product in services and such because the purchasers can engage in life cycle costing, and therefore there will be immediate punishment for anybody trying to overcharge. In that case there was a wonderfully engaging amicus brief from a group of states that said: Trust us, your Honours, we are utterly incompetent to do life cycle costing. The people that buy the main product are in one building, the people that buy the after-market product are in another building, they hate each other, don’t talk to each other, and can’t possibly do this kind of math. So please don’t even think about going there.5 Another famous case was the Leegin case; this was the resale price maintenance case where five to four Justices of the Supreme Court embraced a rule of reason.6 The Court, in its majority opinion, specifically points out that both the Justice Department Antitrust Division and the Federal Trade Commission, the primary enforcers, were both in favour of a rule of reason approach. Both the majority opinion and the dissent each cite amicus briefs 10 times. The majority relies, interestingly enough, on a brief by one private company,

FTC v Phoebe Putney Health System, Inc, 568 US ___ (2013). See Transcript at 26–27: ‘JUSTICE KAGAN: Mr Waxman, we do have a brief from quite a number of states, and the brief basically says: We do this all the time; we set up these local authorities, and then we give them powers because they have to act in the world. We give them normal powers, like the ability to make contracts and the ability to buy property.’ ‘And when we do that, we don’t mean that they can do anything they want notwithstanding the antitrust laws. And to construe these very normal powers that we would give to a state entity in order to allow it to operate as a permission to violate the antitrust laws is not at all consistent with our own intentions.’ 5 Amicus briefs in Eastman Kodak Co v Image Technical Servs, Inc, 504 US 451 (1992); Texas Indus Inc v Radcliff Materials, Inc, 451 US 176 (1980); Monsanto Co v Spray-Rite Serv Corp, 465 US 752 (1984); Brown v Pro-Football, Inc, 518 US 231 (1996); Hartford Fire Ins Co v California, 509 US 764 (1993); Texaco Inc v Hasbrouck, 496 US 543 (1990). 6 Leegin Creative Leather Products, Inc v PSKS, Inc, 551 US 877 (2007). 3 4

Part I: Effective Enforcement of Competition Law 227 Ping, a high-end golf club company, as evidence that the way of doing business, the alternative of using so-called Colgate rights (ie taking advantage of broad rights to choose customers and refuse to deal with discounters), was inefficient and costly—and that was cited by the majority as an important part of their decision.7 Meanwhile, the amicus briefs on both sides were relied on by the authors of those opinions as warfare among economists as to what the majority view was. Briefs from economists were cited by both sides, and during oral argument Justice Breyer said: ‘So we’re supposed to count economists, is that how we decide?’ So amicus briefs clearly were an important part of the litigation. Finally, there is another case, linkLine Communications,8 which Josh Wright referenced. This was interesting in part because we only have a decision here because of an amicus. The plaintiff in that case abandoned its argument before the Supreme Court proceeded to hear the case and said: ‘We give up, we’re not going to pursue our argument anymore’, at which point one might thought the thing for the Court to do was to dismiss the case, and indeed that’s what four Justices thought the Court should’ve done, but the majority said wait, we have two amicus briefs that are still making the argument, so we’re going to appoint to one of them to make the argument orally, and that’s good enough for us to hear the case. So in linkLine we wouldn’t have a decision at all if it hadn’t been for the amicus participation. That was, interestingly enough, the participation of the American Antitrust Institute, a pro-enforcement institute which I’ve had some connection with and which has filed more than 60 amicus briefs in court cases, generally on the pro-enforcement side. And, indeed, you can find a couple of cases where they’ve made a difference.9 So, in the US, amicus briefs and amicus participation really are a big part of the competition law system. In Europe—it’s just not as big a deal here. There are many reasons for this. First, there is an Advocate General, so the role played by the Solicitor General in the US, who participates as an amicus, isn’t as necessary for getting a fair and impartial perspective. Another is that the system is set up more for participating as an intervener rather than as amicus curiae. But that has consequences, because intervention gives you more rights than you would have as an amicus, which means that the courts have to limit the number of interveners. It’s not physically possible to have 90 interveners; that just wouldn’t work. So you’ve got a difference there. You’ve got more courts with a sense that a real amicus ought to be neutral, whereas US courts never really believed that

551 US at 903 (citing Brief of PING, Inc as Amicus Curiae in Support of Petitioner). Cited above n 22. 9 See http://www.antitrustinstitute.org/content/activities-amicus-program. 7 8

228 Part I: Effective Enforcement of Competition Law and abandoned that idea several decades ago. So it’s much less common in Europe to have an amicus-like intervention, if you will, than in the US. It does happen, though, and certainly the Commission does this in a very important way pursuant to Regulation 1/2003, recognising that if Member State courts are going to be enforcing the competition rules of the Treaty there has to be a way for the Commission to be involved.10 Indeed, the Commission has made cooperation with national courts a priority.11 Its website reports 11 opinions responding to court requests for information or an opinion, where the national court gave permission for publication.12 That site also reports 10 ‘written observations’ made to national courts (again limited to where the court approved of publication).13 The Commission’s ‘observations’ have met with considerable success. So, as I wrap up a little, I can confidently predict that, in Europe, interventions and amicus work will increase. I have complete confidence in that because private enforcement is increasing, and private enforcement goes hand in hand with the interest of government regulators getting more involved—and that would apply to national enforcers as much as it does to the Commission. Realistically, the American Antitrust Institute would not exist—I’m not speaking for them, but they would not exist—if there were no vigorous private enforcement system, which creates among other things a group of lawyers that tend to be on the side of the plaintiff and who can help to write the briefs and help to do the funding in order to make it all work. With private enforcement comes more scope for amicus intervention work in Europe. Folks will be looking across the ocean and think about the US and draw some lessons, and there are lots of wonderful things about amicus work in the US. There are some concerns, burden does get to be a concern at some point: you know, 90 briefs is an awful lot for a law clerk to read. The people with more resources can do more to generate the briefs from third parties; we have a lot of briefs in the US now from scholars, almost always written by a lawyer who is a friend of the parties or who has done business with one of the parties. Not always, but I’m not saying it’s done nefariously—it’s usually done pro bono, for good purposes—but the people with more resources can coordinate it b etter and

10 See Emil Paulis, ‘Coherent Application of EC Competition Law in a System of arallel Competences’, in Claus-Dieter Ehlermann and Isabela Atanasiu, eds, European P Competition Law Annual 2000: The Modernization of EC Antitrust Policy, Hart Publishing, 2001. More recently, see Kathryn Wright, ‘European Commission Interventions as Amicus Curiae in National Competition Cases: The Preliminary Reference in X BV’, 30 European Competition Law Review 309 (2009). 11 See Commission Notice on the co-operation between the Commission and the courts of the EU Member States in the application of Articles 81 and 82 EC, [2004] OJ C101/54. 12 See http://ec.europa.eu/competition/court/antitrust_requests.html. As of 2015, 23 opinions had been made available. 13 See http://ec.europa.eu/competition/court/antitrust_amicus_curiae.html. As of 2015, 14 ‘written observations’ documents were available.

Part I: Effective Enforcement of Competition Law 229 do it on time, and so you have to worry about that a little bit. You worry about the courts over-relying on this to do their research. You go back to the Supreme Court and ask: is it the best way to decide things to look at Ping’s filings? To close, there are lots of good things about it and I just can’t resist mentioning that one of the amicus highlights this last year came on the question of whether isolated DNA should be patentable. The Supreme Court decided it should not, thereby agreeing with an amicus brief filed by James D Watson, Nobel Laureate and co-discoverer of the double helix.

Stephen Calkins* The Antitrust Conversation (Continued)

JUSTICE BREYER: … I have 32 briefs here that explain very clearly what you said [at length] in a sentence … And then I have some very nice dark green briefs that clearly say … there could be offsetting justifications.1

Although Justice Breyer exaggerated for effect (only 27 amicus curiae briefs were filed), he rhetorically used the outpouring of argument to make his point: applying the antitrust laws is not so challenging as to require creation of a new presumption. And he could do this because (a) all those briefs were filed, and (b) everyone involved in the argument knew that they had been filed. Those briefs were part of the conversation. Active participation by amici is the norm in US Supreme Court litigation, including antitrust litigation, and increasingly in lower court litigation as well. Amicus participation has soared in recent decades, stimulating a small industry of commentary on whether this is a blessing or a curse and whether it makes a difference and, if so, when and why. As a modest contributor to that commentary2 who now resides in Ireland, it seemed appropriate to revisit the subject from a foreign perspective. This paper comments on recent amici efforts in the US, first in general and then specifically in antitrust cases. It then briefly reviews the important but much more limited amicus activity outside the US. Assuming the efforts to stimulate private competition enforcement in Europe continue to succeed, there will be both more need and more potential for amicus activity on this side of the Atlantic.

* Professor of Law, Wayne State University Law School (on leave); Member and Director, Mergers Division, The Competition Authority of Ireland; and Adjunct Professor, University College Dublin. This chapter is written in a personal capacity and does not represent the views of the Competition Authority. The author thanks numerous friends for suggestions and observations, and Joan Wadsworth for research assistance. The usual disclaimer applies. 1 Transcript of oral argument, FTC v Actavis, Inc, No 12-416 (25 March 2013), 16. Briefs with dark green covers are amicus briefs supporting respondents (here the defendants). 2 See Stephen Calkins, ‘The Antitrust Conversation’, 68 Antitrust Law Journal 625 (2001).

232 Stephen Calkins

1. Introduction The US Supreme Court heard argument in 77 cases during 2012–13.3 It received 1046 merits amicus briefs for those cases—an average of 13.58 briefs per case, with a median of seven.4 Numbers ranged from zero briefs (two cases)5 to highs of 80,6 82,7 928 and 979 briefs in high-profile cases involving gay marriage (two cases), affirmative action in higher education and potential corporate tort immunity for violations of the law of nations (such as torture). By any measure, the 2012–13 term saw a remarkable number of amicus briefs. Fifteen cases had 20 or more merits amicus briefs: 97 briefs: Hollingsworth v Perry, No 12-144 92 briefs: Fisher v University of Texas at Austin, No 11-345 82 briefs: Kiobel v Royal Dutch Petroleum, No 10-149110 80 briefs: United States v Windsor, No 12-307 49 briefs: Association for Molecular Pathology v Myriad Genetics, Inc, No 12-398 48 briefs: Shelby County, Alabama v Holder, No 12-96 33 briefs: Adoptive Couple v Baby Girl, No 12-399 27 briefs: FTC v Actavis, Inc, No 12-416 24 briefs: Decker v Northwest Environmental Defense Center, No 11-338 24 briefs: Kirtsaeng v John Wiley & Sons, Inc, No 11-697 23 briefs: Bowman v Monsanto Co, No 11-796 23 briefs: Maryland v King, No 12-207 22 briefs: Los Angeles County Flood Control District v Natural Resources Defense Council, Inc, No 11-460 21 briefs: Arizona v Inter Tribal Council of Arizona, Inc, No 12-71 20 briefs: American Express Co v Italian Colors Restaurant, No 12-133 Yet the truly remarkable thing is that this was not an aberration, but part of a pattern.

3 American Bar Association Preview of Supreme Court Cases, http://www.americanbar. org/publications/preview_home/alphabetical.html. 4 Calculated from the Preview, ibid, adjusted by consulting Westlaw, where ‘Preview’ indicated its results were incomplete. 5 United States v Bormes, No 11-192; Kloeckner v Solis, No 11-184. 6 United States v Windsor, No 12-307 (defense of marriage act). 7 Kiobel v Royal Dutch Petroleum, No 10-1491 (potential corporate tort liability for violations of the law of nations). The number was inflated in that 21 of the briefs were supplemental briefs, filed because the case was scheduled for reargument this term. 8 Fisher v University of Texas at Austin, No 11-345 (affirmative action in higher education). 9 Hollingsworth v Perry, No 12-144 (California’s limiting marriage to a man and a woman). 10 As noted in n 7 above, 21 of the briefs were supplemental briefs.

The Antitrust Conversation (Continued) 233

2. US amicus practice in general Picture a scene—as in an old movie—with lawyers arguing before a judge off in some remote village. The judge is stumped by an obscure legal point. He notices a leading member of the bar watching the proceedings, and calls on him to share his view. This was the original American amicus curiae.11 This was the repeating of a scene familiar in common law history: ‘The function of the amicus curiae at common law was one of oral “Shepardizing,” the bringing up of cases not known to the judge’.12 Although the concept of amicus curiae had its origins in Roman law, it has been associated with common law legal systems and is a regular part of common law legal systems. It has achieved particular prominence in the US legal system. The original and most common American amici, typified by the scene portrayed above, were neutrals—genuine friends of the Court.13 Partisans were never wholly missing, however, and starting in the 1830s, with the shift to a more paper-oriented process of litigation, they came to represent the norm. Amici participation in Supreme Court proceedings had become a regular feature by the late 1960s, and in 1971 they appeared in a majority of cases.14 At the start of the Burger Court (1969–85), amici appeared in fewer than 40 per cent of Supreme Court cases; by the end of that Court, the number was 70 per cent; the Rehnquist Court (1986–2003) peaked in 2001, with participation in nearly 95 per cent of cases.15 The Court received more than 800 per cent more amicus briefs in the decade 1986–95 (when 85 per cent had briefs, with an average of 4.23 briefs per case) than the decade 1946–55 (when 23 per cent of argued cases had briefs, an average of 0.5 briefs per case).16 In the 1995 term, 76.9 per cent of all cases disposed of by the Court had amicus participation—92.1 per cent of cases disposed of with oral argument.17

11 Samuel Krislov, ‘The Amicus Curiae Brief: From Friendship to Advocacy’, 72 Yale Law Journal 694, 694 (1963). 12 Krislov, above n 11, 695. 13 Stuart Banner, ‘The Myth of the Neutral Amicus: American Courts and their Friends, 1790–1890’, 20 Constitutional Commentary 111, 113 (2003). 14 Ryan Owens and Lee Epstein, ‘Amici Curiae during the Rehnquist Years’, 89 Judicature 127, 128 (2005). 15 Ibid. 16 Joseph Kearney and Thomas Merrill, ‘The Influence of Amicus Curiae Briefs on the Supreme Court’, 148 University of Pennsylvania Law Review 743, 752 (2000). 17 Paul Collins and Lisa Solowiej, ‘Interest Group Participation, Competition, and Conflict in the US Supreme Court’, 32 Law and Social Inquiry 955, Table 1 (2007).

234 Stephen Calkins Another way to appreciate the change is to realize that landmark cases such as Brown v Board of Education18 and Baker v Carr19 had only six and seven amicus briefs, respectively.20 All of the cases that have attracted 20 or more briefs have apparently come since 1970.21 The change in frequency of filing is not explained by rule changes, since the relevant rules have remained relatively unchanged.22 What changed was the Court’s attitude. In the 1940s and early 1950s, the Court was not receptive to amicus briefs and regularly denied permission to file where the parties did not consent.23 This changed over time, such that in recent decades there has been no meaningful limitation on amicus participation. Amicus briefs are filed in lower court cases, but much less frequently. During the first part of the 1990s, they were filed in only 8 per cent of court of appeals cases.24 And while the total number of court of appeals cases with an amicus filing increased by 14.6 per cent between 1992 and 2002, the number of cases increased at a greater rate.25 Amicus filings in state courts have also increased, but from a low base.26

2.1 Influence of amici Does amicus participation make a difference? Proving influence is inevitably a challenge: who can say whether a litigant would have prevailed without good lawyering by retained counsel or amici? And how can one prove that Justices would not have found the same approach, uncovered the same facts or appreciated some special issue on their own? Commentators have examined citations and quotations, surveyed judges and their law clerks, performed statistical analyses and conducted in depth reviews of particular cases. The uniform answer is that amicus participation can make a difference. Amicus participation appears particularly likely to make a difference at the petition for certiorari stage, when the Court (including its law clerks) is sifting

347 US 483 (1954) (separate schools for minorities unconstitutional). 369 US 186 (1962) (one person, one vote). Kearney and Merrill, above n 16, footnote 28. 21 Ibid, 754–755. 22 Ibid, 761. 23 Ibid, 763. 24 Paul Collins and Wendy Martinek, ‘Who Participates as Amici Curiae in the US Courts of Appeals’, 94 Judicature 128, 130 (2010). 25 John Harrington, ‘Amici in the Federal Courts of Appeals: How Friendly Are They?’, 55 Case Western Reserve Law Review 667, 680 (2005’). 26 Matthew Laroche, ‘High Court Studies: Is the New York State Court of Appeals Still “Friendless?” An Empirical Study of Amicus Curiae Participation’, 72 Albany Law Review 701, 708–709 (2009). 18 19 20

The Antitrust Conversation (Continued) 235 through 10,000 petitions to choose the privileged 75–80 cases it will hear.27 Causation is hard to prove, but powerful amicus briefs, most especially by the Solicitor General (SG), seem to be important in helping a petition stand out.28 Perhaps not surprisingly, as this understanding has become more commonly shared, the number of amicus briefs at the petition stage has shot up, doubling between 1970 and 2010,29 and increasing by 35 per cent in the past five years.30 (On the other hand, with the increasing number of amicus briefs there has been a decline in the strength of association of grants of c ertiorari and a single amicus brief, or even two or three.)31 On the merits, too, amicus participation makes a difference, although not in a mindless number-counting way.32 The two most common theories explaining influence are the ‘affected groups’ theory (the very fact of filing shows that various groups care) and the ‘information’ theory (it is the content that matters), with the ‘information’ theory drawing much the greater support.33 The most prominent examples are provided by headline cases. In the University of Michigan affirmative action case, a then-record 107 amicus briefs were filed, and the brief filed by retired military leaders must have been influential because it was discussed during oral argument and then cited by Justice O’Connor in both her opinion for the Court and her verbal summary thereof.34 Some of the 33 amicus briefs in the Texas sodomy case

27 See Supreme Court of the United States, Frequently Asked Questions, http://www. supremecourt.gov/faq.aspx#faqgi9. 28 Timothy Bishop, ‘Opposing Certiorari in the US Supreme Court’, http://www.appellate. net/articles/oppcertinsc.asp; Ryan Black and Ryan Owens, ‘Agenda Setting in the Supreme Court: The Collision of Policy and Jurisprudence’, 71 Journal of Politics 1062 (2009); Gregory Caldeira and John Wright, ‘Organized Interests and Agenda Setting in the US Supreme Court’, 82 American Political Science Review 1109 (1988); Stephen Shapiro, ‘Certiorari Practice: The Supreme Court’s Shrinking Docket’ (‘Multiple amicus curiae briefs can reinforce the impression that the case is exceptionally important, especially when they come from different sectors of the public’), http://www.appellate.net/articles/ certpractice.asp. 29 Gregory Caldeira, John Wright and Christopher Zorn, ‘Organized Interests and Agenda Setting in the US Supreme Court Revisited’, version 1.1 (15 July 2012) (cited with permission). 30 Adam Chandler, ‘Cert-Stage Amicus “All Stars”: Where Are They Now?’, http://www. scotusblog.com/2013/04/cert-stage-amicus-all-stars-where-are-they-now/. 31 Caldeira et al, above n 29. 32 Kearney and Merrill, above n 16, 749 (finding ‘that small disparities of one or two briefs for one side with no briefs on the other side may translate into higher success rates but larger disparities do not’); Linda Sandstrom Simard, ‘An Empirical Study of Amici Curiae in Federal Court: A Fine Balance of Access, Efficiency, and Adversarialism’, 27 Review of Litigation 669, 689 (2008) (‘The Supreme Court respondents unanimously indicated that the number of amicus briefs filed tends to have zero influence on their considerations of the case’). 33 Laroche, above n 26, 704; Simard, ibid, 681. 34 Kelly Lynch, ‘Best Friends?: Supreme Court Law Clerks on Effective Amicus Curiae Briefs’, 20 Journal of Law and Politics 33, 34 (2004). See also Simard, above n 32, 696–697 (Justice Ginsburg said that this brief, and that of business professionals who valued diversity, were particularly helpful).

236 Stephen Calkins were widely credited with teaching the Court a different version of history than it had understood only 17 years earlier.35 There is other evidence as well. A survey of former Supreme Court law clerks shows that almost every amicus brief is at least skimmed.36 Amicus participation was seen as most helpful with respect to technical and specialized areas and in complex regulatory cases.37 A survey of judges similarly finds that amicus participation can be helpful, whether to make legal arguments otherwise missing, to focus attention beyond the immediate dispute, to help a party not adequately represented, to emphasize points needing greater attention or (although this is more controversial) to offer relevant factually information.38 Particularly revealing examples can result from the making public of a Justice’s private papers (the seminal antitrust decision Continental TV, Inc v GTE Sylvania Inc39 is noteworthy in this respect).40 One metric is whether the Court cites amicus briefs. Citation does not prove influence, but it would be surprising if amicus briefs had influence yet were never cited. The record shows that amicus briefs are cited, and at increasing rates. In 1946–55, 18 per cent of cases with one or more amicus briefs resulted in an opinion (of any kind) citing one or more briefs; by 1986–95 that number had increased to just under 37 per cent.41 The Solicitor General has been increasingly likely to be cited, with the rate rising from 26.47 per cent of cases in which the SG filed a brief in 1946–55 to 46.97 per cent in 1986–95.42 And during the 1994–2003 terms, the Court’s majority opinions cited an amicus in 38 per cent of the cases in which an amicus was filed.43

35 Rick Perlstein, ‘What Gay Studies Taught the Court’, Washington Post, 13 July 2003, http://www.glapn.org/sodomylaws/usa/usnews089.htm. 36 Lynch, above n 34. 37 Ibid. 38 Simard, above n 32. 39 433 US 36 (1977). 40 A review of the opinion and the record suggests that the amicus brief filed by the Motor Vehicle Manufacturers Association (MVMA) was particularly influential. See Calkins, ‘Antitrust Conversation’, above n 2, 640–641. We now know that Justice Powell, who authored the opinion, wrote a laudatory comment about the brief to his law clerk: ‘My recollection is that the [MVMA] brief … is the single most helpful brief in this case. No doubt you have drawn on it heavily. If not, I commend it to you’). Andrew Gavil, ‘A First Look at the Powell Papers: Sylvania and the Process of Change in the Supreme Court’, Antitrust, Fall 2002, 9, text at note 12 (Gavil observed that ‘[t]he amicus brief thus proved to be influential, perhaps owing to the fact that it was so well done, but perhaps also because, as was true of the relevant Chicago School commentary, it advocated an analytical framework that was consistent with Powell’s views and facilitated their expression’). 41 Kearney and Merrill, above n 16, 757. 42 Ibid, note 49. 43 Owens and Epstein, above n 14, 130.

The Antitrust Conversation (Continued) 237

2.2 For good or for ill? If amicus activity has increased, and if it has an influence, is that a good or a bad thing? Although there are concerns and views are not unanimous (see below), the consensus view supports amicus participation in adjudication.44 In part, this stems from the special role of the Solicitor General, whose participation is seen as singularly helpful and influential.45 The US legal system does not have an advocate general, so the SG is the most neutral and authoritative voice available to the Court. Filings by states are also given careful attention in the Supreme Court, here because of states’ role in the US system of government.46 Even apart from government participation, moreover, there is general support for amicus activity. Linda Simard’s survey of judges found widespread judicial agreement ‘that there is no need to clamp down on amicus activity’.47 Presumably the judges who participated in leading cases that relied on amicus participation believe that the participation was beneficial. The same reasons that amicus activity can be influential—by making legal arguments that are otherwise missing, by focusing attention beyond the immediate dispute, by helping a party that is not adequately represented, and by emphasizing points needing greater attention—are reasons why amicus activity can be beneficial. There is also a sense that amicus participation can help confer legitimacy.48 There are noteworthy examples of courts affirmatively encouraging amicus activity, either as a general matter or in a particular case.49 Judge Judith Kaye of the New York Court of Appeals once lamented that the court was relatively ‘friendless’, and elaborated on why this was a concern: Amici—because they have a wider perspective, or simply a different perspective— can be of inestimable value to courts in discharging their responsibility that extends beyond the litigants. An amicus can alert the court, as the parties perhaps cannot or would not wish to, that a large issue lurks in an appeal; that a case of seeming insignificance has potentially wide ramifications and will likely have major impact when a ruling is applied in other factual settings; or that a case of obvious major significance could conceivably have wholly unanticipated effects. Amici can sensitize

Kearney and Merrill, above n 16, 745. Lynch, above n 34, 46 (Supreme Court clerks ‘reported that amicus briefs from the Office of the Solicitor General were given a higher level of consideration than those of any other advocate’). 46 Ibid, 48–49. 47 Simard, above n 32, 700. 48 Omari Scott Simmons, ‘Picking Friends from the Crowd: Amicus Participation as Political Symbolism’, 42 Connecticut Law Review 185, 209 (2009). 49 Laroche, above n 26, 754 (‘the Court of Appeals has encouraged amicus filings in court opinions, official documents distributed by the court, and publications by the judges themselves’). 44 45

238 Stephen Calkins the court—when it may be irrelevant to the litigants, whose objective is to win—to the appropriateness of narrowing or limiting a holding, and with other factual situations in mind they can suggest alternate rationales for achieving results urged by the parties.50

Only this June, the New York Court of Appeals (New York’s highest court) publicly invited ‘amicus participation from those qualified and interested’ in helping the Court decide a commercial law question (whether a bank and its customer could agree to shorten the period for reporting a forged signature from the statutory one year to 14 days).51 Commentators also express the view that amicus participation may be of value not just to courts and the adjudicatory process, but also to the participating amici and democratic values of participation.52 Are there concerns about amicus activity? Of course. Common concerns include burden, abuse, and issues of reliability and tensions with the adjudicatory process.

2.3 Burden As is common, Judge Posner is the author of the most striking language complaining about burden: ‘The vast majority of amicus curiae briefs are filed by allies of litigants and duplicate the arguments made in the litigants’ briefs, in effect merely extending the length of the litigant’s brief. Such amicus briefs should not be allowed’.53 Judge Posner explained his concerns as follows: [J]udges have heavy caseloads and therefore need to minimize extraneous reading; amicus briefs, often solicited by parties, may be used to make an end run around court-imposed limitations on the length of parties’ briefs; the time and other resources required for the preparation and study of, and response to, amicus briefs drive up the cost of litigation; and the filing of an amicus brief is often an attempt to inject interest group politics into the federal appeals process.54

50 Judith Kaye, ‘One Judge’s View of “Friends of the Court”’, New York State Bar Journal, April 1989, 8, 13, quoted in Laroche, above n 26, 701. 51 New York Court of Appeals, Notice to the Bar: Amicus Curiae Participation (5 June 2013) (concerning Clemente Bros Contracting Corp v Hafner-Milazzo), http://www.courts. state.ny.us/CTAPPS/news/nottobar/Nottobar060513.pdf. 52 Simmons, above n 48, 205–207; Reuben Garcia, ‘A Democratic Theory of Amicus Advocacy’, 35 Florida State University Law Review 315 (2008). 53 Ryan v Commodity Futures Trading Commission, 125 F3d 1062, 1063 (7th Cir 1997) (Posner, J) (denying motion for leave to file brief) (‘After 16 years of reading amicus curiae briefs the vast majority of which have not assisted the judges, I have decided that it would be good to scrutinize these motions in a more careful, indeed a fish-eyed, fashion’). 54 Voices for Choices v Illinois Bell Telephone Co, 339 F3d 542, 544 (7th Cir 2003) (motions for leave to file amicus briefs denied).

The Antitrust Conversation (Continued) 239 Posner explained that briefs are most likely to be appropriate: in a case in which a party is inadequately represented; or in which the would-be amicus has a direct interest in another case that may be materially affected by a decision in this case; or in which the amicus has a unique perspective or specific information that can assist the court beyond what the parties can provide.55

Judge Posner’s view was rejected by Judge Alito, who adhered to an open door approach,56 and it is the Alito view that has generally prevailed. That said, concerns about burden are not limited to Judge Posner. Supreme Court law clerks, in a survey, ‘repeatedly emphasized that most amicus briefs filed with the Court are not helpful and tend to be duplicative, poorly written, or merely lobbying documents not grounded in sound argument’.57 More over, amicus participation continues to increase, apparently in all US courts. This last term alone saw 16 Supreme Court cases with a large number (19 or more) of amicus briefs, and six with more than 45. There is a sense in which amicus briefs tend to encourage more amicus briefs, with the burden steadily increasing.

2.4 Abuse? In the view quoted above, Judge Posner noted that amicus briefs are often solicited by parties, and of course this is true. Soliciting amici support is part of what Supreme Court practitioners do. Concern about abuse resulted in the Supreme Court revising the applicable rule in 1997, and Rule 37.6 now provides as follows for briefs from private amici: a brief filed under this Rule shall indicate whether counsel for a party authored the brief in whole or in part and whether such counsel or a party made a monetary contribution intended to fund the preparation or submission of the brief, and shall identify every person other than the amicus curiae, its members, or its counsel, who made such a monetary contribution.

But Supreme Court practitioners do not read this rule to bar ‘coordination or discussion between petitioner’s counsel and amicus counsel’, the circulation of draft submissions or a party’s counsel’s reviewing a draft amicus brief ‘in order to identify inaccuracies or avoid repetition’ and then offering advice ‘to correct, delete, or add limited explanation or clarification’.58 The Court is

Ibid, 545. Neonatology Associates, A v Commissioner of Internal Revenue, 293 F3d 128 (3d Cir 2002) (Alito, J). 57 Lynch, above n 34, 45. 58 Shapiro, above n 28 (‘The common sense of the new rule is that while general comment and correction are permitted, ghostwriting in any form is forbidden, without disclosure’). 55 56

240 Stephen Calkins forced to rely on the integrity of practitioners to draw lines fairly and then to adhere to them. One trend has been an increase in briefs by academics. It is now commonplace that antitrust cases stimulate briefs by academics, sometimes law professors, sometimes economists, sometimes both, and often conflicting. Richard Fallon has observed that the academic amicus brief is now commonplace— and he is uneasy about it. His experience is that such briefs are usually written not by an academic, but by a practicing lawyer, who circulates them close to the filing deadline and drums up as much support as possible on short notice.59 He worries that professors may affix their names too casually, exhibiting much less rigor than they would in academic work.60 Both liberals and conservatives have grumbled about the make-up of the community of amici. Justice Scalia has noted that briefs tend to come only from well-organized interest groups.61 One leading Supreme Court practitioner based in a multinational law firm wrote in 1999 that only certain types of associations appear very often as friends of the court. Business groups, for example, file fewer amicus briefs than the issues warrant … For decades, public interest groups, usually of a liberal political outlook, have made their views known to the Court through amicus briefs.62

On the other hand, data analysis shows that conservative, anti-regulatory interests now dominate the lists of amici supporting petitions for certiorari, with the Chamber of Commerce leading the way with 54 briefs in a little over three years and a remarkable 32 per cent success rate (second to the conservative Mountain States Legal Foundation’s astonishing 50 per cent success rate in 10 cases).63 Conservative groups have dramatically increased their amicus filing in support of certiorari. As Adam Chandler observes, ‘It is difficult to explain the dramatic boosts in several organizations’ tallies (e.g., Cato, the Center for Constitutional Jurisprudence, DRI) without imagining that, at some point in the last five years, they fundamentally reassessed how they could most effectively influence the Court’.64 Of course, participation—even carefully coordinated participation—is not abuse, but it is important to remember that repeat participants can be expected to behave strategically.

59 Richard Fallon Jr, ‘Scholars’ Briefs and the Vocation of a Law Professor’, 4 Journal of Legal Analysis 223, 230 (2012). 60 At the conference, Daniel Crane observed that a larger question is posed by the way that parties to long-term or ongoing litigation can see it as in their self-interest to sponsor research and writing that they hope will benefit their cause(s). 61 Jaffee v Redmond, 518 US 1, 35–36 (1996) (Scalia, J, dissenting). 62 Shapiro, above n 28. 63 Chandler, above n 30. Of course, amicus activity on the prevailing side may say more about correlation between the general views of the current Court majority and the amicus institution than about causation. 64 Ibid.

The Antitrust Conversation (Continued) 241

2.5 Reliability One somewhat controversial way that amicus participation can influence outcomes is by introducing factual information not otherwise before the Court—meaning not in the record.65 By definition, all of this information escapes the reliability tests that are built into a judicial fact-finding process. Experts testifying at trial need to prove their expertise and subject themselves to the discovery process; experts communicating through an amicus brief need only submit the brief. Especially, but not exclusively, when it is social science that is being submitted, there are concerns about lax quality controls.66 We will see that many of the same issues have played out in US antitrust cases.

3. American antitrust amici: an update A little over a decade ago, I wrote a review of and commentary on antitrust and trade regulation amicus efforts—an essay that ‘unabashedly celebrate[d]’ them.67 In particular, the article commented on ways that briefs had stood out and possibly made a difference: (i) because a repeat player such as the government took a position against interest or that was otherwise surprising; (ii) through providing legal research; (iii) by highlighting egregious errors; (iv) by supplying special emphasis; or (v) through setting out important analytical constructs. Some of those patterns have continued, but there have been some noteworthy changes. At a basic level, amicus briefs continue to be commonly filed in US competition cases. The FTC has an amicus brief web page with links to the 91 briefs files from 1998 to 2013.68 In the same period, the Justice Department’s Antitrust Division filed 70 amicus briefs.69 (Many of the briefs are jointly filed.) This is now a key part of the agencies’ missions.

Simard, above n 32, 695, 704–706. Andrew Morriss, ‘Private Amici Curiae and the Supreme Court’s 1997–1998 Term Employment Law Jurisprudence’, 7 William & Mary Bill of Rights Journal 823 (1999); Michael Rustad and Thomas Koenig, ‘The Supreme Court and Junk Social Science: Selective Distortion in Amicus Briefs’, 72 North Carolina Law Review 91 (1993). See also William Katt, ‘Roper and the Scientific Amicus’, 49 Jurimetrics Journal 253 (2009) (‘science brief’ stretched the studies on which it relied and miscolored the truth). 67 Calkins, ‘Antitrust Conversation’, above n 2. 68 See http://www.ftc.gov/ogc/briefs.shtm (the first four briefs bear my name as General Counsel). 69 See http://www.justice.gov/atr/public/appellate/index.html#page=page-1. 65 66

242 Stephen Calkins When that 2001 article noted some antitrust cases with a substantial number of briefs, it mentioned cases from the 1980s and 1990s with numbers such as 17, 15, 14, 12, 11 and 10.70 The past decade has continued to see competition cases with a substantial number of merits amicus briefs, including: FTC v Actavis, Inc (2013):7127 Verizon Communications Inc v Law Offices of Curtis V Trinko, LLP (2004):7219 Illinois Tool Works Inc v Independent Ink, Inc (2006):7317 F Hoffman La Roche Ltd v Empagran SA (2004):7416 American Needle, Inc v National Football League (2010):7514

3.1 Amici for certiorari What is more striking, however, are the number of amicus briefs—and important amicus briefs—filed in favor of certiorari in cases lost below by the defendants. The Roberts Court is a conservative Court,76 and the most important question for antitrust defendants has been whether they could persuade the Court to hear cases they had lost in the courts of appeal. Perhaps inspired by the scholarship suggesting that amici can make a difference at the certiorari stage, a string of cases over the past decade have featured important amicus participation on behalf of defendants at the certiorari stage: —— Pacific Bell Telephone Co v linkLine Communications, Inc, No 07-512 (2009): four amicus briefs supporting certiorari, including briefs by (i) the Solicitor General for the United States; (ii) Virginia and nine other states; (iii) Verizon Communications Inc and the National Association

70 The number of amicus briefs in Eastman Kodak Co v Image Technical Servs., Inc, 504 US 451 (1992), Texas Indus Inc v Radcliff Materials, Inc, 451 US 176 (1980), Monsanto Co v Spray-Rite Serv Corp, 465 US 752 (1984), Brown v Pro-Football, Inc, 518 US 231 (1996), Hartford Fire Ins Co v California, 509 US 764 (1993) and Texaco Inc v Hasbrouck, 496 US 543 (1990), respectively. 71 570 US 756 (2013). 72 540 US 398 (2004). 73 547 US 28 (2006). 74 542 US 155 (2004). 75 560 US 183 (2010). 76 Lee Epstein, William Landes and Richard Posner, ‘How Business Fares in the Supreme Court’, 97 Minnesota Law Review 1431, 1471 (2013) (‘five of the ten Justices who, over the span of our study (the 1946 through 2011 Terms), have been the most favorable to business are currently serving’).

The Antitrust Conversation (Continued) 243

——

——

——

——

——

of Manufacturers; and (iv) Robert H Bork and J Gregory Sidak for 18 ‘professors and scholars in law and economics’.77 Leegin Creative Leather Products, Inc v PSKS, Inc, No 06-480 (2007): three amicus briefs supporting certiorari, including briefs by (i) the National Association of Manufacturers and (ii) Amici Curiae E conomists—a particularly diverse group of 25 economists, including nine who had served as chief economist at the Antitrust Division or the FTC.78 Credit Suisse Securities (USA) LLC v Billing, No 05-1157 (2007): four amicus briefs supporting certiorari or vacatur, including briefs by (i) the Solicitor General for the US, the SEC and the FTC (supporting vacatur) and (ii) the Chamber of Commerce, the Securities Industry Association and the Bond Market Association. Bell Atlantic Corp v Twombly, No 05-1126 (2007): eight amicus briefs supporting certiorari, including briefs by (i) the Chamber of Commerce, CTIA-the Wireless Association, the Alliance of Automobile Manufacturers, Northwest Airlines, Inc and United Airlines; (ii) EI du Pont de Nemours and Co and the National Association of M anufacturers; (iii) the Washington Legal Foundation; and (iv) a diverse group of 25 economists.79 Weyerhaeuser Co v Ross-Simmons Hardwood Lumber Co, No 05-381 (2007): seven amicus briefs in supporting certiorari, including briefs by (i) the Solicitor General for the US and the FTC; (ii) the Chamber of Commerce and the American Forest and Paper Association; (iii) the Business Roundtable and the National Association of Manufacturers; and (iv) the Washington Legal Foundation. Texaco, Inc v Dagher, No 04-805 (2006): five amicus briefs supporting certiorari, including briefs by (i) the Solicitor General for the US and the FTC; (ii) the Chamber of Commerce and the National Association of Manufacturers; (iii) the Washington Legal Foundation; and (iv) ‘antitrust scholars’—a brief filed by Professor Robert Pitofsky in his capacity

77 William Baumol, Robert Bork, Robert Crandall, George Daly, Harold Demsetz, Jeffrey Eisenach, Kenneth Elzinga, Gerald Faulhaber, Franklin Fisher, Charles Goetz, Robert Hahn, Jerry Hausman, Thomas Jorde, Robert Litan, Paul MacAvoy, J Gregory Sidak, Pablo Spiller and Daniel Spulber. 78 William Baumol, Tomothy Bresnahan, Robert Crandall, David Evans, Franklin Fisher, Luke Froeb, Richard Gilbert, Jerry Hausman, Andrew Joskow, Benjamin Klein, G Franklin Mathewson, Paul Milgrom, Kevin Murphy, Philip Nelson, Janusz Ordover, Thomas Overstreet, Robert Pindyck, David Scheffman, Frederic Scherer, Richard Schmalensee, Carl Shapiro, Bruce Snapp, Pablo Spiller, David Teece and Richard Warren-Boulton. 79 William Baumol, Michael Boskin, Robert Crandall, Kenneth Elzinga, David Evans, Gerald Faulhaber, Franklin Fisher, Luke Froeb, Richard Gilbert, Paul Joskow, Michael Katz, Paul Milgrom, Thomas Moore, Robert Pindyck, Robert Porter, Frederic Scherer, Richard Schmalensee, Marius Schwartz, Vernon Smith, Edward Snyder, Michael Spence, Pablo Spiller, Alan Sykes, David Teece and Michael Whinston.

244 Stephen Calkins as of counsel to Arnold & Porter (the other signatories were University of Virginia economist Kenneth Elzinga and George Mason law professor Ernest Gellhorn). —— Verizon Communications Inc v Law Offices of Curtis V Trinko, LLP, No 02-682 (2004): three briefs supporting certiorari, including briefs by the Solicitor General for the US and the FTC. —— Nynex Corp v Discon, Inc, No 96-1570 (1998): three briefs supporting certiorari, including briefs by (i) the Solicitor General for the US and the FTC and (ii) the Association of the Bar of the City of New York. All of those grants of certiorari resulted in reversal and the issuance of an opinion in favor of the defendant(s). In contrast, the SG declined to support the losing defendant in the controversial bundled discount case, 3M v LePage’s Inc, and the Court let the lower court decision stand.80 (There are numerous examples of the SG, as amicus, successfully opposing petitions for certiorari filed by antitrust plaintiffs.81 Most notably, the SG successfully opposed several plaintiff requests for certiorari in ‘reverse payments’ cases82—once when the plaintiff was the FTC—before (also s uccessfully) supporting certiorari in FTC v Actavis, Inc.)

3.1.1 Making a difference on the merits Although it seems likely that several of those amicus-supported petitions for certiorari would not have been granted had it not been for amicus participation, it is harder to point with confidence to merits decisions where the other side likely would have prevailed but for the work of amici. That said, there were several cases where amicus participation could have made a difference because (i) a repeat player took a position against its narrow interest, (ii) an amicus advanced an analytical construct the Court found appealing, (iii) an amicus supplied special emphasis or (iv) an amicus provided research assistance.83

542 US 943 (2004). See McFarling v Monsanto Co, 552 US 1096 (2008) (with Patent & Trademark Office) (patent tying); Dee-K Enterprises, Inc, v Heveafil SDN BHD, 539 US 969 (2003) (with FTC) (foreign commerce); Statoil ASA v Heeremac VOF, 534 US 1127 (2002) (with FTC) (foreign commerce); CSU, LLC v Xerox Corp, 531 US 1143 (2001) (aftermarket refusal to deal); Armstrong Surgical Center, Inc v Armstrong County Memorial Hosp, 530 US 1261 (2000) (with FTC) (boycott). 82 Joblove v Barr Laboratories, Inc, 551 US 1144 (2007); FTC v Schering-Plough, Inc, 548 US 919 (2006); Andrx Pharmaceuticals, Inc v Kroger Co, 543 US 939 (2004) (with FTC). 83 The most important amicus continues to be the Solicitor General. Leah Brannan’s and Douglas Ginsburg’s count shows that the SG’s amicus filing was on the winning side of 84%, 79% and 91% of antitrust cases during the decades starting 1977, 1987 and 1997, respectively (33% during 1967–1976). Leah Brannon and Douglas Ginsburg, ‘Antitrust Decisions of the US Supreme Court, 1967 to 2007’, Competition Policy International (Autumn 2007). 80 81

The Antitrust Conversation (Continued) 245 Positions against interest Until this past term, almost all of the Supreme Court antitrust cases since 2001 fit the same pattern: a defendant lost a case in a court of appeals, persuaded the Court to grant certiorari and then prevailed.84 And in each such case, the Solicitor General filed a brief asking the Court to reverse or vacate the lower court’s decision. It simply has to be helpful to a defendant to have the nation’s chief law enforcer in its corner. The point is perhaps best illustrated by Leegin Creative Leather Products, Inc v PSKS, Inc.85 The majority and the dissent were in agreement that resale price maintenance (RPM) could have either pro-competitive or anticompetitive effects; they differed on whether existing law (which made it per se unlawful) should remain unchanged under the rule of stare decisis. The (5–4) majority chose to reverse, pointing to economics literature and the government amicus: It is also significant that both the Department of Justice and the Federal Trade Commission—the enforcement agencies with the ability to assess the long-term impacts of resale price maintenance—have recommended that this Court replace the per se rule with the traditional rule of reason. In the antitrust context the fact that a decision has been ‘called into serious question’ justifies our reevaluation of it.86

The other case in which a government amicus—state governments—played a particularly noteworthy role was the past term’s FTC v Phoebe Putney Health System, Inc.87 The lower court had held that a hospital authority’s acquisition was immune from an FTC challenge because the acquisition was pursuant to a state policy to displace competition, a conclusion that the authority said could be drawn from the broad powers the state had given it. The crucial amicus brief was filed by 20 states, asking the Court not to draw any such inference.88 Justice Kagan specifically challenged the authority’s lawyer on that very point during oral argument.89 When someone is purporting to 84 American Needle, Inc v National Football League, 560 US 183 (2010), is an oddity: the defendant NFL prevailed below but then supported certiorari, only to lose on the merits. The SG (with the FTC) first opposed certiorari and then supported the plaintiff–petitioner American Needle, although with an analysis different from the one adopted by the Court. 85 551 US 877 (2007). 86 Leegin, 900 (citations omitted). 87 568 US ___ (2013). 88 Brief of Amici Curiae States of Illinois, Arizona, California, Colorado, Connecticut, Delaware, Hawaii, Idaho, Maryland, Michigan, Minnesota, Nevada, New Hampshire, New Mexico, New York, North Carolina, Oregon, Pennsylvania, Tennessee and West Virginia in Support of Petitioner. The hospital authority’s state (Georgia) participated in the case with the FTC at trial, but then withdrew. 89 Transcript, 26–27: ‘JUSTICE KAGAN: Mr Waxman, we do have a brief from quite a number of states, and the brief basically says: We do this all the time; we set up these local authorities, and then we give them powers because they have to act in the world. We give them normal powers, like the ability to make contracts and the ability to buy property.’

246 Stephen Calkins defend state sovereignty, it is awkward to be opposed by the only states to be heard. Supplying an analytical construct The single example of an amicus supplying an analytical construct is Pacific Bell Telephone Co v linkLine Communications, Inc.90 In this price squeeze case, the key to the Court’s analysis was a two-step approach: first, ask whether the wholesale price is too high (and answer that there is no cause of action unless there is an underlying duty to deal); second, ask whether ‘the defendant’s retail prices are “too low”’91 (and answer using predatory pricing law). This approach is uncannily similar to that recommended by the Solicitor General.92 Emphasis In American Needle, Inc v National Football League,93 the Court addressed whether a joint venture that marketed NFL logos for sportswear was a single entity, and held that it was not. Here, the Court declined to follow the SG’s analytical construct, but its opinion also failed to track closely petitioner’s reasoning. Instead, the Court reverted to a test enunciated in its Copperweld opinion:94 ‘The question is whether the agreement joins together “independent centers of decisionmaking”’.95 Ironically, this distinction was most emphasized by an amicus supporting the (losing) defendant NFL.96

‘And when we do that, we don’t mean that they can do anything they want notwithstanding the antitrust laws. And to construe these very normal powers that we would give to a state entity in order to allow it to operate as a permission to violate the antitrust laws is not at all consistent with our own intentions.’ 555 US 438 (2009). Ibid, 451 (emphasis in original). 92 Compare SG Brief with Petitioners Brief. Rebecca Haw makes the same point. Rebecca Haw, ‘Amicus Briefs and the Sherman Act: Why Antitrust Needs a New Deal’, 89 Texas Law Review 1247, 1276 (2011) (criticizing the Court’s decision). 93 560 US 183 (2010). The NFL was represented by Covington & Burling, for whom I served of counsel at the time (but for whom I am not speaking). 94 Copperweld Corp v Independence Tube Corp, 467 US 752, 769 (1984). 95 130 S Ct. at 2212 (quoting Copperweld Corp v Independence Tube Corp, 467 US 752, 769 (1984)). 96 Brief for ATP Tour, Inc, WTA Tour, Inc, Major League Soccer, LLC and National Association for Stock Car Auto Pacing, Inc as Amici Curiae in Support of Respondents, 13 (‘they are single entities because they are single, economic producers in the marketplace by virtue of the nature of a league/circuit product. There is no need to test for § 1 concerted action using state law based structures (such as the requirements of a ‘joint venture’ or ‘effective merger’) when the market-based analysis taught by Copperweld and Dagher most directly—and accurately—distinguishes ‘concerted’ and ‘unilateral’ activity by asking whether the collective decision of ‘separate entities’ actually raises relevant ‘ antitrust dangers’ by depriving the market of otherwise competitive decisionmakers’). 90 91

The Antitrust Conversation (Continued) 247 Research assistance One of the more pronounced trends in recent cases has been the use of amici to provide research assistance. The most striking example was provided by the Ninth Circuit Court of appeals, which publicly invited amicus briefs addressing the proper treatment of bundled discounts.97 At the Supreme Court level, three cases stand out: F Hoffman-La Roche Ltd v Empagran SA,98 FTC v Actavis, Inc99 and Leegin. The cases serve both to illustrate the importance of amici and to raise questions about whether there can be excessive reliance on them. Hoffman-La Roche addressed the extent to which the Sherman Act should be applied to acts outside the United States. When Stephen Shapiro, counsel for the defendants-petitioners, was arguing that the Court could choose a statutory interpretation that avoids ‘antagonizing our allies’, a Justice asked, ‘How do we know what’s consistent with not antagonizing our allies?’100 Mr Shapiro’s answer: ‘Well, on the latter, we have amicus briefs from seven of our … most significant trading partners’.101 And, indeed, he was supported by amicus briefs from (i) Canada, (ii) Germany and Belgium, (iii) Japan and (iv) the United Kingdom, Ireland and the Netherlands. Illustrating the challenge of fact-finding by reliance on amici, a Justice rejoined, ‘But surely … there are other partners who have not been heard from’.102 Mr Shapiro responded by observing that no foreign government amicus brief took the other position.103 The Court must have been persuaded, because in its Opinion it noted that ‘several foreign nations have filed briefs here arguing that to apply our remedies would unjustifiably permit their citizens to bypass their own less generous

97 Cascade Health Solutions v Peacehealth, 479 F3d 726, 726 (2007) (‘The court invites supplemental briefs by any amicus curiae addressing the following issue raised in this appeal: whether a plaintiff who seeks to establish the predatory or anticompetitive conduct element … by showing that the defendant offered bundled discounts … must prove that the defendant’s prices were below an appropriate measure of the defendant’s costs. If so, what is the appropriate measure of costs and how should the trial court instruct the jury on the matter of costs? If not, what standard should the trial court instruct the jury to use …?’). The court received nine amicus briefs. 502 F3d 895 (2007). 98 542 US 155 (2004). 99 570 US 756 (2013). 100 Transcript, Hoffman-La Roche, 13. 101 Ibid, 13–14. 102 Ibid, 14. 103 Ibid (‘That’s true, but all of the foreign nations that have spoken up here agree with the United States that this is contrary to their ability to regulate commerce in their own nations’).

248 Stephen Calkins remedial schemes’ and that this ‘would undermine foreign nations’ own antitrust enforcement policies’.104 FTC v Actavis is an example of amicus briefs providing information that the Justices find useful. This was the pay-for-delay case, in which the majority ruled that agreements in which incumbents allegedly paid to keep out generic pharmaceuticals were subject to antitrust review even if the arrangement stayed within the ‘scope of the patent’. Justice Breyer for the Court wrote that these payments may yield the generic manufacturer even more profit than had it entered—citing a law review article and an amicus brief filed by 118 professors, with the Court adding that the professors’ brief estimated that this was true of the payment at issue.105 Chief Justice Roberts, writing for the three-Justice dissent (Justice Alito was out), relied on amici for information to a greater extent than the majority. He explained that there are multiple ‘first applicants’, as they are called, which was an important factual point for which Chief Justice Roberts cited the Generic Pharmaceutical Association amicus brief.106 He wrote that it is very risky for generics to take on branded firms, and they do so only after carefully weighing the risks and benefits—points for which he cited the same brief. The Court’s decision will lessen the chances of settlement and thus decrease the likelihood of entry by generic firms, he said—citing another amicus brief as authority.107 Leegin is the best example of the Court relying on amici for information. Each of the majority and the dissent cited amicus briefs 10 times. At issue was whether to maintain the per se prohibition of minimum resale price maintenance. The 5–4 majority opinion led with the declaration that ‘economics literature is replete with procompetitive justifications for a manufacturer’s use of resale price maintenance’—with the first citations for the proposition

104 542 US at 168. Further using amici for research, the Court reported that the Solicitor General and Petitioners had not found a case ‘in which any court applied the Sherman Act to redress foreign injury in such circumstances’, 542 US at 169. It also pointed to dueling amicus briefs on the consequences of a decision for deterrence of global cartels to show that it would be hard for lower courts to resolve such questions on a case-by-case basis. Ibid, 169 (‘Even in this relatively simple price-fixing case, for example, competing briefs tell us (1) that potential treble-damages liability would help enforce widespread anti-pricefixing norms (through added deterrence) and (2) the opposite, namely, that such liability would hinder antitrust enforcement (by reducing incentives to enter amnesty programs). Compare, eg Brief for Certain Professors of Economics as Amici Curiae 2–4 with Brief for United States as Amicus Curiae 19–21’). 105 133 S Ct at 2235 (citing Brief Amici Curiae of 118 Law, Economics, and Business Professors and the American Antitrust Institute in Support of Petitioners, 25). 106 Brief for the Generic Pharmaceutical Association as Amicus Curiae Supporting Respondents, 23–24 (citing FTC data). 107 Brief of Medication and Negotiation Professionals as Amici Curiae in Support of Respondents.

The Antitrust Conversation (Continued) 249 being amicus briefs.108 Even ‘those more skeptical of resale price maintenance acknowledge it can have procompetitive effects’—again, with the first citation to an amicus brief.109 And, as previously noted, on the question of stare decisis, the Court pointed to the views of the SG; it also wrote that ‘there is now widespread agreement that resale price maintenance can have procompetitive effects’—citing the economists’ amicus brief in support of petitioner.110 In response to the suggestion that the strictures of the RPM per se rule can be avoided by what are known as Colgate policies (taking advantage of broad rights to choose customers and refuse to deal with discounters), the Court wrote that the threat of antitrust liability ‘can lead, and has led, rational manufacturers to take wasteful measures’.111 Its only authority for this factual assertion was a claim in the amicus brief of a maker of high-end golf products.

3.1.2 Additional observations Three other developments since 2001 stand out: (i) scholars’ amicus briefs have proliferated, with the scholars sometimes becoming a focal point; (ii) amici have played notably active roles; and (iii) there is a new source of amicus offerings: the American Antitrust Institute. Scholars’ amicus briefs As is true of amicus practice more generally, and as suggested in the discussion above about amici as sources of information, scholars’ amicus briefs have become a regular part of the process. Examples include the following: —— American Express Co v Italian Colors Restaurant (2013):112 (i) Professors of Civil Procedure in Support of Respondents; (ii) Antitrust Scholars in Support of Respondents; and (iii) Distinguished Law Professors in Support of Petitioners. —— FTC v Actavis, Inc (2013):113 (i) Antitrust Economists in Support of Respondents; (ii) Professors David W Opderbeck and Erik Lillquist in Support of Respondents; (iii) Professors Gregory Dolin, Kent Bernard, et al. in Support of Respondents; (iv) Health Economics and Law Professors in Support of Respondents; and (v) 118 Law, Economics, and

108 551 US at 889 (citing Brief of Amici Curiae Economists in Support of Petitioner) (23 economists) and Brief for the United States as Amicus Curiae Supporting Petitioner). 109 Ibid (citing Brief for William Comanor and Frederic Scherer as Amici Curiae Supporting Neither Party, 3). 110 Ibid, 900 (citing Brief of Amici Curiae Economists in Support of Petitioner, 16). 111 Ibid, 903 (citing Brief of PING, Inc as Amicus Curiae in Support of Petitioner). 112 570 US __ (2013). 113 570 US 756 (2013).

250 Stephen Calkins

—— —— —— —— —— —— —— —— ——

Business Professors and the American Antitrust Institute in Support of Petitioners. FTC v Phoebe Putney Health System, Inc (2013):114 Economics Professors in Support of Petitioner. American Needle, Inc v NFL (2010):115 (i) Economists in Support of Respondents and (ii) Economists in Support of Petitioner. Pacific Bell Telephone Co v linkLine Communications, Inc (2009):116 Professors and Scholars in Law and Economics in Support of Petitioners. Leegin Creative Leather Products, Inc v PSKS, Inc (2007):117 (i) Professors William Comanor and Frederic Scherer Supporting Neither Party and (ii) Economists in Support of Petitioner. Bell Atlantic Corp v Twombly (2007):118 (i) Legal Scholars in Support of Respondent; (ii) Legal Scholars in Support of Petitioners; and (iii) Economists in Support of Petitioners. Weyerhaeuser Co v Ross-Simmons Hardwood Lumber Co (2007):119 (i) Law Professors in Support of Petitioner and (ii) Economists in Support of Petitioner. Illinois Tool Works, Inc v Independent Ink, Inc (2006):120 (i) Professor FM Scherer in Support of Respondent and (ii) Professors Barry Nalebuff, Ian Ayres, and Lawrence Sullivan in Support of Respondent. Verizon Communications Inc v Law Offices of Curtis V Trinko, LLP (2004):121 (i) Economics Professors in Support of Respondent and (ii) Law Professors Supporting Respondent. Nynex Corp v Discon, Inc (1998):122 Law Professors in Support of Respondent.

There is every reason to believe that the Court considers this information. Examples include those discussed above, and more. During oral argument in Illinois Tool Works, for instance, Justice O’Connor raised the following point with the representative of the SG: Mr Hungar, one of the amicus briefs for the respondent was submitted by a professor … named Barry Nalebuff … which took the view that the Court should, in any

568 US ___ (2013). 560 US 183 (2010). 555 US 438 (2009). 117 551 US 877 (2007). 118 550 US 544 (2007). 119 549 US 312 (2007). 120 547 US 28 (2006). 121 540 US 398 (2004). 122 525 US 128 (1998). 114 115 116

The Antitrust Conversation (Continued) 251 event, retain the presumption [of market power] where a patent is being used to impose a variable or a requirements tie. Do you have any comment on that view?123

In the end, the Court discussed the Nalebuff brief’s point only to disagree,124 but it certainly took note of it. In American Express Co v Italian Colors Restaurant,125 a case about arbitration of antitrust claims, Respondents’ lawyer Paul D Clement was arguing that the inevitably high litigation costs required class treatment: ‘if you look at the Hovenkamp amicus brief, it makes clear that you can’t bring this type of claim’, with Justice Breyer interjecting, ‘Well, that sounds expert to me. Now, Hovenkamp would be the person I would hire as the arbitrator.’126 When Justice Breyer referred again to the late Philip Areeda, Clement responded: ‘Justice Breyer, none of us can know for sure what Professor [Areeda] would say. But we know what [his co-author] Professor Hovenkamp says, and he says to bring these claims you need an expert.’127 The most striking example, however, is Leegin, where, as noted above, both sides relied on amici to argue economics. The unusually diverse group of economists listed on an amicus brief supporting certiorari included FM Scherer, a distinguished economist known to support energetic competition enforcement. Petitioners highlighted this: ‘It is also significant to note that Professor Scherer is one of the economists who submitted a brief as amici curiae urging the court to grant Leegin’s petition and overturn the per se rule against resale prices maintenance’.128 This prompted Professor Scherer to file a separate merits brief ‘in part to set the record straight on Scherer’s views’.129 Then, during oral argument, when Theodore Olson, for the petitioner, argued that economists supported his client’s position, Justice Breyer challenged him with respect to Scherer. Olson responded by referring to ‘the vast majority of the economist[s]’, prompting Justice Breyer to ask, ‘We’re supposed to count economists? … Is that how we decide it?’130

123 Transcript, 24 (referring to Brief of Professors Barry Nalebuff, Ian Ayres, and awrence Sullivan as Amici Curiae in Support of Respondent). During argument, L Respondents’ lawyer also invoked amici professors: ‘Can metering be procompetitive? … The briefs of Professors Nalebuff and Professor Scherer, the only economist briefs submitted in the case, show how metering is not necessarily efficient’. Transcript, 39. 124 547 US at 44. 125 570 US __ (2013). 126 Transcript, 27–28 (referring to Brief Amici Curiae of Antitrust Scholars in Support of Respondents). 127 Transcript, 29. 128 Reply Brief, 6. 129 Brief for William Comanor and Frederic Scherer as Amici Curiae Supporting Neither Party, 1. 130 Transcript, 7–8 (‘Which economists? … Professor Scherer is an economist, isn’t he? Worked at the FTC for a long time. A good expert in the field … What that sounds like is that if at least he, who is an economist, thinks if you get rid of Dr Miles, every American will pay far more for the goods that they buy at retail. Now that’s one economist, of course. There are others who think differently. So how should we decide this?’).

252 Stephen Calkins Amici playing active roles Although it is routine that the SG is allowed to participate in Supreme Court oral arguments, dividing time with counsel for the supported side, other amici are almost never heard. It is quite remarkable, therefore, that three times in the past 16 years the Court has allowed other amici to argue antitrust cases. Two of the cases involved resale price maintenance.131 It was almost as though the Court reasoned that, with the SG taking the side of the defendants, it was appropriate to allow another amicus—each time the State of New York—to speak up for the plaintiff. The third case was even more unusual. In Pacific Bell Telephone Co v linkLine, the plaintiffs (who had won below) unsuccessfully resisted certiorari, abandoned their position and decided not to defend the decision below (preferring that its complaint be dismissed with leave to amend).132 Justice Breyer, with Justices Stevens, Souter and Ginsburg, thought that that should end the matter: ‘I would accept respondents’ concession that the Ninth C ircuit majority’s “price squeeze” holding is wrong, I would vacate the Circuit’s decision, and I would remand the case’.133 The majority disagreed. It justified continuing because ‘[t]wo amici have submitted briefs defending the Court of Appeals’ decision on the merits, and we granted the motion of one of those amici to participate in oral argument’.134 (In truth, issues relating to the case’s merits were poorly joined at argument, since far too much time was spent on procedural confusion.) The American Antitrust Institute The amicus that quite remarkably argued linkLine was the American Antitrust Institute (AAI). The AAI is ‘an independent non-profit organization based in Washington D.C. with a mission to increase the role of competition, assure that competition works in the interests of consumers, and challenge abuses of concentrated economic power in the American and world economy’.135 The AAI was founded in 1998, and claims that until then ‘there was no public interest organization principally dedicated to supporting a more aggressive antitrust agenda’.136 The AAI has filed more than 60 amicus briefs since 2001 and can point to a series of cases, especially in the courts of appeal, where

131 Leegin (26 March 2007) (Barbara Underwood for New York); State Oil Co v Khan, No 96-871 (7 October 1997) (Pamela Harbour for state amici curiae). 132 Brief for Respondents, linkLine. 133 555 US at 457 (Breyer, J, concurring in the judgment). 134 555 US at 447. 135 See http://www.antitrustinstitute.org/. I am a member of the AAI Advisory Board and a former Senior Fellow, but I am not speaking on its behalf. 136 See http://www.antitrustinstitute.org/content/about-us.

The Antitrust Conversation (Continued) 253 its participation has made a difference.137 No other country has this kind of organization with its consistent voice for energetic competition enforcement.

4. Outside the US Nowhere in the world do amici play the same prominent role that they play in the United States. There are amici in countless jurisdictions, of course, but not on a scale comparable to the US either in general or in competition cases. There are many reasons for this: history and tradition explain much; the association with common law jurisdictions explains part; and public attitudes toward courts explains part. America lacks an Advocate General, so the Supreme Court can hear from the closest substitute (the solicitor general) only through an amicus appearance. With respect to competition law, much is explained by the fact that in the US competition law overwhelmingly consists of court decisions resolving private disputes, which means that government enforcers can shape the law only through amicus activity and other guidance. One important point is the distinction between an amicus and an intervener. In the US, a party with an interest in litigation is likely to participate as an amicus, which means that its role is limited to filing a brief and, very rarely, being granted leave to participate in argument. It has no other rights— in particular, no special rights to be copied on pleadings and no rights of appeal. In contrast, when non-party Member States wish to be heard by the European Court of Justice, they will typically intervene, meaning that they join the litigation, receiving documents and participating in the hearing. (Intervening Member States and institutions of the European Union before the General Court may also bring appeals.)138 On both sides of the Atlantic the objective is the same—to be heard by a court deciding a case of importance to a government—but the title and associated rights are different. In the UK, too, much of what would involve amicus participation in the US involves attempts to intervene.139 A consequence of addressing third party concerns by allowing intervention is that the judicial system simply must limit them, since it is not practical to allow scores of parties to participate in hearings. Another consequence is that

See http://www.antitrustinstitute.org/content/activities-amicus-program. Protocol (No 3) on the Statute of the Court of Justice of the European Union Article 56. 139 ‘To Assist the Court: Third Party Interventions in the UK, a Justice Report’ (26 October 2009), http://www.justice.org.uk/resources.php/32/to-assist-the-court. 137 138

254 Stephen Calkins third parties may seek to intervene as much to gain access to documents as to be heard, which is further reason to limit the numbers that can participate.140 But intervention, including as an amicus, is on the rise, even if not close to the US level.

4.1 International bodies Amicus participation in international bodies has become commonplace.141 NGOs actively participate in international litigation.142 This can also be seen in international arbitration,143 and at the WTO, NAFTA and ICSID.144 Indeed, NGOs have been instrumental in encouraging international bodies to accept amicus briefs.145

4.2 European Court of Justice The European Court of Justice’s 2012 annual report lists seven competition cases, excluding state aid cases.146 Two have participation by Member States, but only one by a non-party private entity—a trade association that had been allowed to intervene at the Court of First Instance. This lack of participation by private third parties is not unusual. The exception, rather than the norm, is provided by the 2010 judgment in Akzo Nobel Chemicals Ltd v Commission.147

140 Gianni De Stefano, ‘Access of Damage Claimants to Evidence Arising out of EU Cartel Investigations: A Fast-evolving Scenario’, 3 Global Competition Litigation Review 95 (2012) (discussing Case C-596/11 P(I) Schenker v KLM and Commission, EU:C:2012:334). 141 Anna Dolidze, ‘Making International Property Law: The Role of Amici Curiae in International Judicial Decision-Making’, 40 Syracuse Journal of International Law and Commerce 119 (2012) (especially discussing the European Court of Human Rights). 142 Dinah Shelton, ‘The Participation of Nongovernmental Organizations in International Judicial Proceedings’, 88 American Journal of International Law 611 (1994) (the International Court of Justice, the European Court of Justice, the European Court of Human Rights and the Inter-American Court of Human Rights). 143 Eugenia Levine, ‘Amicus Curiae in International Investment Arbitration: The Implications of an Increase in Third-Party Participation’, 29 Berkeley Journal of International Law 200 (2011). 144 Francesco Francioni, ‘Access to Justice, Denial of Justice and International Investment Law’, 20 European Journal of International Law 729 (2009). 145 Steven Kochevar, ‘Amici Curiae in Civil Law Jurisdictions’, 122 Yale Law Journal 1653, 1657, 1664–1665 (2013). 146 See Court of Justice of the European Union, Annual Report 2012, 29–32, http://curia. europa.eu/jcms/upload/docs/application/pdf/2013-04/192685_2012_6020_cdj_ra_2012_ en_proof_01.pdf. 147 Opinion of Advocate General Kokott in Case C-550/07 P Akzo Nobel Chemicals Ltd v Commission [2010] ECR I-8301, paras 26–29 (General Court had granted leave to intervene to the Council of the Bars and Law Societies of the European Union, the Algemene Raad

The Antitrust Conversation (Continued) 255 Leave to intervene was addressed in Schenker v KLM and Commission.148 Schenker AG sought review of the General Court’s dismissal of its application for leave to intervene. The Court of Justice reviewed Article 40 of the Statute of the Court of Justice, which provides that ‘Member States and institutions of the Union may intervene in cases before the Court of Justice’, as can (with certain exceptions) ‘any other person which can establish an interest in the result of a case’. The Court explained that an ‘interest in the result of a case’ means ‘a direct, existing interest in the ruling on the forms of order sought’.149 The Court of Justice noted that the General Court recalled instances in which intervention had been allowed—such as by competitors alleging abuse of a dominant position, or certain associations—but wrote that to allow intervention by every customer purchasing from members of a cartel could undermine the effectiveness of the courts.

4.3 The Commission as amicus When Regulation 1/2003 gave the national courts a wider role in enforcing Articles 101 and 102, it sensibly set up a structure to facilitate coordination with the Commission.150 Article 15 authorizes (i) national courts to request the Commission’s ‘opinion on questions concerning the application of the Community competition rules’ and (ii) the Commission to submit ‘written observations’ to national courts (and to make oral observations with a court’s permission) ‘[w]here the coherent application of Article 81 or Article 82 of the Treaty so requires’.151 (Article 15 also authorizes national competition authorities to submit written observations—oral with permission—relating to the application of those Articles.)

van de Nederlandse Orde van Advocaten (General Council of the Netherlands Bar), the European Company Lawyers Association, the American Corporate Counsel Association— European Chapter and the International Bar Association, and they took part in the appeal. However, the General Court had denied leave to the European Council on Legal Affairs and the Section on Business Law of the International Bar Association, and the President of the Court of Justice refused to grant leave to intervene to the following associations: the Association of General Counsel and Company Secretaries of the FTSE 100, the Chamber of Commerce of the United States of America, the International Chamber of Commerce, the American Bar Association, the Law Society of England and Wales, and the United States Council for International Business). Above n 140. Ibid, para 10. 150 See Emil Paulis, ‘Coherent Application of EC Competition Law in a System of Parallel Competences’, in Claus-Dieter Ehlermann and Isabela Atanasiu, eds, European Competition Law Annual 2000: The Modernization of EC Antitrust Policy (2001); Kathryn Wright, ‘European Commission Interventions as Amicus Curiae in National Competition Cases: The Preliminary Reference in X BV’, 30 European Competition Law Review 309 (2009). 151 See Case C-429/07 Inspecteur van de Belastingdienst v X BV [2009] ECR I-4833. 148 149

256 Stephen Calkins The Commission has made cooperation with national courts a priority.152 Its website reports numerous opinions responding to court requests for information or an opinion, where the national court gave permission for publication.153 The website also reports ‘written observations’ made to national courts (again limited to where the court approved of publication).154 The Commission’s ‘observations’ have met with considerable success.155

4.4 National courts Putting aside participation by governmental bodies, the safe generalization is that participation in national courts by amici (or interveners playing a similar role) is often allowed in appropriate cases but is dramatically less common than in the US.156 Originally associated with common law jurisdictions, it is becoming more recognized in civil law systems.157 Frequently it is a question of the court appointing an amicus to assist it (not a party).158 The UK emphasized the point by renaming the position ‘advocate to the court’.159 Ireland provides an example of judicial reluctant acceptance. In 2003 the Supreme Court declared that ‘it does have an inherent jurisdiction to appoint an amicus curiae where it appears that this might be of assistance in determining an issue before the court’.160 The same Opinion also declared, however,

152 See Commission Notice on the co-operation between the commission and the courts of the EU Member States in the application of Articles 81 and 82 EC, [2004] OJ C101/54. 153 See http://ec.europa.eu/competition/court/antitrust_requests.html. 154 See http://ec.europa.eu/competition/court/antitrust_amicus_curiae.html. 155 For discussion of the Commission’s role in the Irish BIDS case, see Victoria Balaguer, ‘Beef Industry Development Society: A Warning on Industry Restructuring’, 33 European Competition Law Review 296 (2012). 156 For instance, in an OECD filing, the Netherlands explained that ‘so far’ its competition authority ‘has hardly made use of its amicus-powers’. OECD Directorate for Financial and Enterprise Affairs, Competition Committee, Institutional and Procedural Aspects of the Relationship between Competition Authorities and Courts, and Update on Developments in Procedural Fairness and Transparency (DAF/COMP(2011)122) (OECD Update), Contribution of the Netherlands (footnote omitted) (amicus brief filed in only one case, where the civil judge requested it). For a review of common law jurisdictions, see ‘To Assist the Court’, above n 139; Lise Johnson and Niranjali Amerasinghe, ‘Protecting the Public Interest in International Dispute Settlement: The Amicus Curiae Phenonenon’, Center for International Environmental Law, 2009, http://www.ciel.org/Publications/Protecting_ ACP_Dec09.pdf. 157 Kochevar, above n 145. 158 For instance, see the practice in New Zealand, described in Law Commission, ‘Review of the Judicature Act 1908: Towards a Consolidated Courts Act’, chapter 15 (‘Representation and other participants’) (2012), http://ip29.publications.lawcom.govt.nz/ chapter+15+-+representation+and+other+participants/other+participants. 159 See Notice (2002), http://www.lawgazette.Couk/news/notice-4. 160 HI v Minister for Justice Equality and Law Reform [2003] IR 197, 4. See also Doherty v South Dublin County Council [2007] 1 IR 246 (Equality Authority joined as amicus);

The Antitrust Conversation (Continued) 257 that this is ‘a jurisdiction which should be sparingly exercised’161—and that is what has happened. Mr Justice Kelly of the Irish High Court recently refused to allow ‘Digital Rights Ireland Limited’ to be appointed as amicus curiae in a copyright dispute. Quoting previous decisions, Justice Kelly explained that the power should be exercised where a case has ‘a public law dimension’ in favor of an applicant that ‘has a bona fide interest and is not just acting as a meddlesome busy body’, and that is ‘largely neutral and in a position to bring to bear expertise in respect of an area which might not otherwise be available to the court’.162 Digital Rights Ireland Limited was excluded because the Court did not regard it ‘as a neutral party’ and did not believe it would provide ‘a perspective on matters of principle or public importance which would not otherwise be available to it’.163 Probably the leading European exception to the pattern is Germany.164 This was set out in a submission to the OECD, where Germany explained that The Bundeskartellamt participates in every proceeding before the Federal Court of Justice by way of oral statement and before the courts of lower instance by way of written statements in leading cases, in cases linked to on-going cases of the Bundeskartellamt and upon the request of the courts.165

Germany explained that ‘[t]his allows the Bundeskartellamt to help to safeguard a coherent development in the public and private enforcement of competition law’.166 The agency’s interventions are highly valued by the courts.167

O’Brien v Personal Injuries Assessment Board (No 1) [2005] 3 IR 328 (Law Society of Ireland joined as amicus). HI v Minister of Justice, ibid, 4. [2013] IEHC 204, paras 47–48 and 50 (quoting O’Brien v Personal Injuries, above n 160 (Finnegan P) and Fitzpatrick v FK [2007] 2 IR 406 (Clarke J)). Justice Kelley also considered that amicus participation was more appropriate at the appellate level. Ibid, para 54. 163 Ibid, paras 65 and 67. To add insult to injury, Digital Rights Ireland later faced a demand for approximately €26,000 in costs. E-mail from TJ McIntyre to Stephen Calkins (7 October 2013). 164 The author is grateful to Konrad Ost for emphasizing this point in conversations at the EUI Workshop and for translating part of a speech on the issue. 165 OECD Update, above n 156, Contribution of Germany, 48 (emphasis in original). 166 Ibid, 47–48 (‘Finally, the Bundeskartellamt also has the opportunity to be involved in private enforcement proceedings. It is informed of such private antitrust proceedings by the respective courts. According to Section 90 ARC and Article 15 Council Regulation (EC) No 1/2003, the courts are required to give this information. The Bundeskartellamt can participate as amicus curiae in the civil proceedings resulting from private enforcement actions … The Bundeskartellamt is to be informed of all private enforcement actions arising before courts and upon request can be sent all briefs, records, orders and decisions. Members of staff have a right to take an active part in the court proceedings by way of written or oral statements. In private antitrust proceedings the parties have the possibility to appeal the case in points of fact and law and finally to the Federal Court of Justice on points of law’). 167 Speech by Prof Dr Günter Hirsch, President of the Federal Court of Justice (15 January 2008), http://www.bundeskartellamt.de/wDeutsch/download/pdf/Publikationen/ Hirsch.pdf. 161 162

258 Stephen Calkins

5. A look to the future The number of amicus efforts and interventions outside the US seems almost sure to increase.168 Success breeds success. Once an entity has filed an amicus brief (or intervened), the next one is easier. Once one trade association or other NGO participates (and communicates about it169), others are likely to follow the lead. And once a tactic proves useful for one group of litigants, it is likely to be followed by another. (Recall, in this respect, the proliferation of ‘scholars’ briefs’ in the US.) Indeed, with the rise of global law firms, it may be the same law firm active in one jurisdiction that imports the practice to another. One change, in particular, is sure to increase the amount of amicus/ intervention activity: increased private enforcement. Private enforcement has been steadily increasing and the Commission seems committed to continue this trend.170 Follow-on actions now occur routinely in Europe—although not everywhere—and more seem sure to occur, even without a boost from the Commission. With the new initiative, the trend will accelerate. Private litigation is the great instigator of amicus activity. When government enforcers are the only show in town, government agencies obviously do not need to file amicus briefs or intervene. As the Commission so wisely recognized in 2003, if competition law is being developed in cases in which it is not a party, it needs a way to participate. As private litigation spreads, national enforcement authorities, too, will increasingly seek to participate. And as courts become more accepting of participation from government amici, they may well become more accepting of private ones as well. One final wrinkle deserves mention: private litigation can also serve to make diverse private amicus participation possible. The more litigation there is, the more there are lawyers with an interest in shaping the law through litigation—including through amicus activity. Moreover, private litigation means there are law firms working for both plaintiffs and defendants. Both groups of law firms can encourage amicus work (for instance, by working pro bono on scholars’ briefs). It is also noteworthy that the American Antitrust Institute, which has filed more than 60 amicus briefs, has an advisory board with a score of private practitioners, almost none of whom are from the world’s largest law firms but many of whom are from law firms that have

168 For instance, during 2010–11, Romania modified its competition law and gave its competition authority the power to ‘issue observations’, on its own initiative or at the request of courts, when competition rules are being applied. OECD Update, above n 156, Contribution of Romania, 229. 169 See, eg http://www.inta.org/Advocacy/Pages/amicus.aspx. 170 See http://ec.europa.eu/competition/antitrust/actionsdamages/index.html.

The Antitrust Conversation (Continued) 259 represented a significant number of plaintiffs.171 Although the AAI is careful not to be any kind of voice of the plaintiffs’ bar—the large majority of advisory board members are not part of that bar—it seems doubtful that the AAI could have thrived, and filed those scores of briefs, had the US not had a robust system of private antitrust enforcement. Just as one hopes that Europe can learn from the US’s experience and achieve the best of private enforcement, one hopes it can learn from the US’s experience with amici. Amici have brought great benefits to the US legal system. (As only one example of the kind of special contributions that a micus activity makes possible, note that the 49 amicus briefs filed in the recent case in which the Supreme Court decided that isolated DNA is not patentable included one by Nobel laureate James D Watson, co-discoverer of the double helix structure of DNA.)172 But the above review suggests some cautions: —— Burden is an issue, even if participation is currently at a modest level. Amicus/intervention participation breeds more participation, especially but not only if it is successful. —— Parties with the greatest resources inevitably have greater opportunity to employ resources in amicus/intervention. —— Once amicus participation is the norm, courts may find themselves relying on it, whether to play a role traditionally reserved for a party or a government official (as in linkLine), or by drawing conclusions from the absence of amici. (In the RPM case Leegin, Justice Scalia pointedly asked why the big discounters such as WalMart had not filed briefs.)173 —— Courts may also find themselves relying on amici for research and a kind of extra-record fact-finding in some uneasy tension with an adversarial system of adjudication, such as in determining the views of foreign nations (Hoffman-La Roche), the profitability of a business strategy (Actavis), the burdensomeness of Colgate policies (Leegin) or the consensus view of economists (Leegin). None of this is reason why amicus/intervention activity should not increase. It should and it will—increasing private litigation makes this inevitable. But this friend of amicus activity advises care.

See http://www.antitrustinstitute.org/~antitrust/content/advisory-board. Brief of James Watson, PhD as amicus curiae in Support of Neither Party. 173 Transcript, 31–32 (‘I mean, we talk about the Wal-marts and the Targets. They’re not here on amicus briefs because they’re—what they’re selling is cheap … So they’re not going to be harmed by the fact that some manufacturers want to provide not just the low price … but service’). 171 172

260

4 INFRINGEMENT PROCEDURES AND PUBLIC MEASURES DISTORTING COMPETITION Chair: Giorgio Monti Speakers: José Luis Buendía Sierra Daniel Crane Damien Gerard Ian Forrester Alexander Italianer

Ioannis Lianos Philip Lowe Konrad Ost Mario Siragusa Joshua Wright

Italianer: I’ll speak about Article 106. I am not a specialist on the subject because I haven’t had to deal with Article 106 cases and the question is, of course, why is that the case? First, let’s go back a bit to the origins of Article 106. Already in 1948 there was the problem in the context of trade liberalisation that state-controlled companies or companies that enjoy special or exclusive rights can distort trade because of that particular position. Therefore, in Article XVII of the GATT you have a non-discriminationprinciple that applies in relation to such companies. And, of course, when the European Economic Community was set up in 1958 this had to be respected, and the problem at the time was that there were very different ‘economic constitutions’ in the Member States as regards state ownership of companies— and, by the way, there still are. So, the question was how to deal with that in the Treaty, and actually if you look at the way the Treaty was constituted it’s something that one might call a creative ambiguity because, on the one hand, you have Article 37—it was Article 37 in the original Treaty and miraculously it still is Article 37—which instructs the Member States to adjust commercial state monopolies to avoid discrimination, so it’s a kind of response to the requirement set in Article XVII of the GATT. Then you have Article 106 in the context of public undertakings, or undertakings enjoying special or exclusive rights, and Article 106 prohibits such undertakings from infringing the competition rules. Then there is Article 345, which states that the Treaty is neutral as concerns the national systems of property ownership. So, on the one hand, state monopolies have to be adjusted; on the other hand, the DD

262 Part I: Effective Enforcement of Competition Law Treaty is neutral with regard to public ownership; and in between there is this fringe provision, Article 106. Now, Article 106 is in itself a compromise provision or a bizarre provision because it addresses itself on the one hand to the Member States and on the other hand to undertakings. And what does it do? In its first paragraph there is a prohibition, but it is always in combination with the other rules contained in the Treaty. It makes particular reference to the competition provisions, but it also refers to Article 18, which is the general non-discriminationrule, and again this echoes the GATT, but in practice Article 106 has also been used in combination with the articles on free movement. Secondly, Article 106 can be applied directly by national courts and actually quite a lot of the case law has derived from preliminary reference cases. And finally, and I think José Luis Buendía Sierra will come back to this, the Commission has full discretion as to whether to pursue an Article 106 case. There have not been that many decisions, especially in relation to paragraph 1. Going back to the jurisprudence of the Court of Justice in the 1970s and 1980s there was also the ‘effet utile’ case law, which looked at the general obligation of the Member States to respect the provisions of the Treaty, or in any case not to go against them, and one of the provisions relevant to that case law was the old Article 3 of the EC Treaty. The words that appeared in that provision—the reference to a system of undistorted competition—has meanwhile been relegated, thanks to former French President Sarkozy, to Protocol 27 of the Treaties. But protocols of the Treaties are just as important as the Treaty Articles, so I think the provision is still fully applicable. That seems to be the position of the Court of Justice in its post-Lisbon jurisprudence.1 I mention this because it reminds me a bit of the powers of the Office of Fair Trading (OFT). The OFT has to refer a market investigation to the Competition Commission because it allows the competition authorities to interfere in the market structure more generally if there is a competition policy problem. It has not been used, but I’ll come back to that at the end of my short presentation. Now I will not say much about the second paragraph of Article 106 because that has mainly to do with services of general economic interest, and there the Commission has been more active. Not long ago we revised what was originally called the Monti Package, and so there are some new elements in the various rules that govern services of general economic interest in the package that was adopted in December 2011. In particular, we have included the net avoided cost methodology. We have introduced a greater alignment with the procurement rules because Article 106 refers to all Treaty provisions, and therefore also to the procurement rules. We are also asking authorities to include and to foresee incentives for providers of general

See Case C-52/09 Konkurrensverket v Teliasonera Sverige AB [2011] ECR I-527, para 20.

1

Part I: Effective Enforcement of Competition Law 263 economic interest services to become more efficient. So we have been evolving in applying that paragraph as well. Coming back to the first paragraph, how has that provision been used in combination with Article 102? Which concerns have been addressed in the past? I will not make an extensive list, but one of the concerns is the inefficient monopolist. In other words, you assign a monopoly to a particular state enterprise—or private enterprise, for that matter—but this enterprise is not able to satisfy all the demand and then there is a risk that it goes beyond what is acceptable for Article 106. Or there can be a conflict of interest, for instance if you have an entity that regulates entry into a particular activity, for example motorcycle competitions, and at the same time is also active in the downstream market through the organisation of motorcycle competitions.2 Then there is a risk of extension of dominance to a secondary market, or if the company is forced because of its dominance to impose excessive prices. Now, in practice, almost all these cases face very difficult judgement calls, and this is exemplified in the ruling of the General Court in Greek Lignite, which has meanwhile been appealed and therefore is under legal scrutiny, so I will not comment on that case in this session.3 Finally, what can one say about the challenges for the future of Article 102? First of all, in paragraph 3, the Article provides that the Commission can address decisions to individual companies or Member States, but it can also issue directives, and actually the Article has been used to liberalise markets in an earlier stage of market liberalisation, especially in the telecoms sector, and since then this general liberalisation trend has been done through internal market instruments, whether it’s in telecoms, postal services or energy. So, this begs the question whether the use of Article 106 is actually still needed in the sense that market liberalisation has been accomplished through internal market rules. Well, that is indeed the case to a large extent, but there still can be a role for Article 106 today. First of all, we are still facing strong incumbents that remain in place. For example, Greek Lignite is a case concerning a monopoly on lignite extraction. Secondly, exclusive rights are still being granted in new areas, for example in waste management or waste collection. Thirdly, the definition of special rights is actually quite wide, and therefore even in a relatively small geographical area this may find its place. Finally, we still have public undertakings active in a large number of sectors, and therefore that may create potential problems. More generally, there is increased attention at the EU level on enhanced economic governance provisions that go far beyond the

2 See Case C-49/07 Motosykletistiki Omospondia Ellados NPID (MOTOE) v Elliniko Dimosio [2008] ECR I-4863. 3 Case T-169/08 Dimosia Epicheirisi Ilektrismou AE (DEI) v Commission, EU:T:2012:448, set aside and remanded: C-553/12 P Commission v Dimosia Epicheirisi Ilektrismou AE (DEI), EU:C:2014:2083.

264 Part I: Effective Enforcement of Competition Law enforcement of the competition rules by looking at structural reforms in the Member States with a view to unlocking the potential of the internal market. So, there are other instruments coming on stream which are being used in the context of what is called the ‘European semester’, where specific recommendations are addressed to the Member States.4 That reminds me a little bit of these market investigation references in the UK which exhort the Member States to engage in structural reforms, and these are becoming more and more precise. They may thus address some competition issues more generally even if policies or instruments are not specifically caught by Article 106 or specifically related to public undertakings or undertakings with special or exclusive rights. Therefore, I think the ‘effet utile’ doctrine, even though it was largely abandoned by the Court of Justice,5 may come back in another way. Finally, there is, of course, the state aid nexus, where José Luis will perhaps say that there is a close relationship between special or exclusive rights and state aid. Of course, in legal terms, special or exclusive rights may not always entail state aid, such as, for instance, if there are no state resources involved directly. We have cases like the PreussenElektra, where there are no state resources involved but which nevertheless give rise to competition concerns.6 Therefore, state aid provisions and enforcement may be another angle from which to look at Article 106 in the future as well. Thank you. Jose Luis Buendia Sierra: The explanation by Alexander creates a very good introduction to the topic. As he has explained, Article 106 is a complex provision; it’s part of competition law, or at least we think it is, but it’s not exactly a standard provision.7 As I say in my paper, it’s in the family, but it’s not often invited to the table. My objective in this short intervention is to try to explore why, and to try to see whether this is a good situation in general, or if we can do something in order to improve it. We are not speaking about state aid, we are speaking about other kinds of state-created distortions, and, of course, the way to deal with it is not the same as with antitrust. I am going to speak above all about direct enforcement by the European Commission. DD

4 http://ec.europa.eu/economy_finance/economic_governance/the_european_semester/ index_en.htm. 5 See Case C-2/91 Wolf W Meng [1993] ECR I-5751. 6 Case C-379/98 PreussenElektra AG v Schhleswag AG [2001] ECR I-2099. 7 For a more comprehensive study of Article 106, see José Luis Buendía Sierra, Exclusive Rights and State Monopolies under the EC Law: Article 86 (Former Article 90) of the Treaty, Oxford University Press, 1999. A new edition of the book is in preparation. For a more limited presentation of the topic, see José Luis Buendía Sierra, ‘Article 106— Exclusive Rights and Other Anti-competitive State Measures’, in Jonathan Faull and Ali Nikpay, eds, The EU Law of Competition, 3rd edn, OUP, 2014, chapter 6. For the application of Article 106 in the context of state aid cases, see José Luis Buendía Sierra, ‘Finding the Right Balance: State Aid and Services of General Economic Interest’, in EC State Aid Law: Liber Amicorum Francisco Santaolalla, Kluwer, 2008, chapter 10.

Part I: Effective Enforcement of Competition Law 265 Also, I have to mention there is also the possibility of going to the national courts and to ask them to ask the ECJ for preliminary rulings, and this has led to a number of interesting cases. In the past, the most interesting cases arose from the preliminary reference procedure, and we are still seeing new cases that emerge in that procedural context. Now, in this talk I will try to answer the question: is the enforcement of Article 106 effective? Also, is this enforcement legitimate? When we look at Article 106, we first see that there is a special channel and a special procedure under the third paragraph. You say: what is special about it? Well, not very much if you compare it with antitrust or state aid; the ‘specialty’ of this procedure is the fact that it is the Commission who decides if there is an infringement, and it’s up to the Member State to go to the Court and try to annul the decision, so it’s exactly the same situation that we normally have in antitrust or state aid but not at all the normal situation compared to any other infringement of European rules by the Member States. So, is it similar to antitrust and state aid? Well, not really. There is a very important difference, which is the fact that a ‘complainant’ cannot really challenge the position taken by the Commission, so the Commission might decide to act or might decide not to act.8 The Commission has full discretion in this respect. We have a couple of cases which say this very clearly, in particular max.mobil,9 and so if the Commission says ‘I don’t think this is an important case or we don’t have time’, there is very little we can do; we can’t effectively contest it before the EU Courts. What I can say is that very few decisions have been adopted—four or five, depending on how you want to count the decisions, which is an average of one case every two years. That’s not very impressive, but this is already being nice because if I had taken the last four years, the result would be zero decisions, and so the average would be even worse. The other side of the image is that the vast majority of complaints invoking this provision have been rejected for an alleged lack of European interest. So, it’s very difficult to monitor what is not in the picture because the rejections do not leave a lot of traces. But we can at least look at the four or five decisions, and there we have a couple of cases about mail in Germany and in Slovakia, and also the Greek Lignite case, where the Commission is trying to police the borders of liberalisation. So in sectors where there has been directive opening up the markets and the situation is not working well because there is an incumbent extending its dominant position to neighbour markets, the Commission is applying this provision as an effective way to answer. I will come back to the Greek Lignite case in a moment, but I will also refer to the other big Article 106 case, called

8 Case C-141/02 P Commission v T-Mobile Austria GmbH, formerly max.mobil Telekommunication Service GmbH [2005] ECR I-1283. 9 Ibid.

266 Part I: Effective Enforcement of Competition Law Livret A/Livret Bleu, which was a French case.10 This is a very interesting case because it shows how difficult it can be to distinguish between state aid and an Article 106 case. This was a state aid case when I was a young official at the Commission, and it was difficult to manage and difficult to quantify, and indeed it went to the Court of Justice, which annulled it for problems with quantification. So, like the main character in Kafka’s Metamorphosis, this case thought it was a state aid case but one day it woke up and thought: ‘Wait, I am not a state aid case, I am something different—I am a different animal’. And this different animal was an Article 106 case, which was similar but different: not only the lack of ‘standing’, so to speak, for complainants, but also the lack of recovery, which is a very important difference. So, we have a quasi-state aid decision, but without recovery. So, if I can qualify it in another way, it is like this series of small language books, le français sans peine or l’anglais sans peine—but this time it’s state aid without pain. This is one decision, and the other most well-known decision on Article 106 is the Greek Lignite case, a very important and very interesting case.11 The situation there was that the Greek government granted exclusive rights for the exploitation of lignite to the incumbent in the electricity market. Electricity has been liberalised in theory, but in practice if you want to enter that market you need a cheap source of energy and the only cheap source of energy was lignite, so the Greek government, by granting a monopoly on lignite, was de facto killing the electricity liberalisation process. So I think the Commission rightly reacted and applied a long-standing line of cases based on the illegal extension of a dominant position, and it said ‘okay, this has to be stopped’. It is also interesting in another respect because the Commission imported some techniques, like negotiating with a Member State to find remedies, as if it were a market test. All this makes it a very interesting case, but then we have judicial review and here I don’t think anybody would say that the General Court was particularly nice or deferential to the Commission; it was, rather, to the contrary. The General Court tried to apply the previous case law, but in my view the previous case law was more ambitious or more far-reaching. So they minimised it in a way, saying there was something missing in the decision, namely a precise description of the alleged abuse that could potentially be carried out. Well, if that is the problem, then the good news is that it is not a very difficult problem to solve. I think it a little bit strange that this led to an annulment of such an important decision, but that’s the way it works. But

10 Commission Decision of 10 May 2007 pursuant to Article 86(3) of the EC Treaty, on the special rights granted to La Banque Postale, Caisses d’Epargne and Crédit Mutuel for the distribution of the livret A and livret bleu, C(2007) 2110 final, http://ec.europa.eu/ competition/liberalisation/livret_a_en.pdf. 11 Greek Lignite, above n 69. As indicated there, the Court of Justice effectively reversed the judgment of the General Court.

Part I: Effective Enforcement of Competition Law 267 if the General Court’s judgment is upheld, the Commission can simply make some adjustments to its decision, and it is not so difficult. So, just to say a few words about the follow-up, the question when looking at enforcement is: are the above cases the only infringements of Article 106? In my opinion, it is quite obvious that they are not the only ones, and that they are not the main ones. I have some examples in mind—some of them you can find in my paper. Here we have a problem not of over-enforcement but perhaps of under-enforcement, and probably asymmetrical enforcement, because you see some cases treated and others not treated, and you don’t understand why. In my view, the conclusion is that it is very difficult to say that direct enforcement is effective at this moment. Perhaps the reason for that is that the Commission has some reservations about the legitimacy of this instrument. But if that is the case, then I don’t understand why—considering that the Commission does not seem to have any problem with pursuing state aid matters, which are just as difficult and delicate as Article 106 cases. The Commission itself is saying that it is more or less the same in the French case I just cited, and furthermore we have had quite a few reviews of the Treaty by all the Member States, and there was a discussion of whether Article 106 should stay or not, and the conclusion was clear: it has to stay. So if you have it, please use it, because if you don’t use it who do you think will? It is not by nature a provision that you can really imagine will be enforced by Member States or by national competition authorities. So, in my view, the problem is that the Commission needs some encouragement to do this powerful job, and the problem is that, because of the case law, the Commission has full discretion, and when the Commission has full discretion in an area like this we know what happens—very little if anything. Dan Crane: Thanks very much to Giorgio, Mel and Philip for the invite; it’s a pleasure to be here. The perspective that I would like to provide in this paper is less of a comparative perspective than a theoretical perspective. This is an extension of work I’ve been doing for some time now, thinking about judicial review, especially of anticompetitive state action and problems of judicial review, sort of thinking of different models of how courts can address state provisions that distort competition without becoming overly politically involved in those decisions.12 So, the overall framework would be some umbrella provision—it could be a constitution, a treaty, a federal statute that has pre-emptive effect that favours competition in some way, and then courts are called upon to decide whether governmental actions of some kind or another conform to the umbrella DD

12 Daniel Crane, ‘Judicial Review of Anticompetitive State Action: Two Models in Comparative Perspective’, 1 Journal of Antitrust Enforcement 418 (2013).

268 Part I: Effective Enforcement of Competition Law rovision in favour of competition. So, think of the Parker doctrine in the p US,13 and some of the Article 106 discussions that we’ve already heard this morning or the freedom of goods and services line of cases—so you can sort of think about a number of different kinds of provisions in this context. In prior work I’ve looked at what I consider the two leading models, which are roughly associated with the US and the EU, of how judicial review can function in this context, and I have tried to highlight some difficulties with those models.14 Here I want to recap those two models very briefly, and to articulate some of the difficulties they raise, and then I’ll propose a third model that I call ‘hard look review’. This is associated with a particular form of judicial review of administrative agency action in the United States, and I think that this model of ‘hard look review’ lies somewhere between the substantive review model and the representation reinforcement model that I have written about.15 The backdrop assumption I have is that public choice theory teaches us that capture and corruption of the political regulatory process is frequent. Particularly in the competition context, firms will engage in rent-seeking behaviour to try to be protected from competition by regulation or law, and therefore there is a need for some kind of scrutiny of regulation; we should not assume that it is all in the public interest. But also, many legitimate and public-minded regulations will have anticompetitive side-effects, so the question is: what kind of analytical framework can courts use to try to determine when a particular regulation of some kind favours the public interest or when it’s really a product of rent-seeking behaviour and therefore should be struck down? Now, the two models that I’ve discussed before and which I’ll recap now are, first of all, what I call a representation reinforcement model. I associate this with the Midcal test in the United States for state action.16 Essentially, the Midcal test requires that the action must be affirmatively expressed in state policy, and here we have the Phoebe Putney case.17 Second, it has to be actively supervised by state actors. So behind the representation reinforcement model is the idea that the government can anticompetitively regulate so long as the government itself conspicuously takes responsibility for the outcomes of the

See Parker v Brown 317 US 341 (1943). See ‘Judicial Review’, above n 78. 15 The term ‘hard look’ review was originally coined by US Circuit Court Judge Judge Harold Leventhal of the DC Circuit. See Pikes Peak Broad. Co v FCC, 422 F2d 671, 682 (DC Cir 1969); Greater Boston Television Corp v FCC, 444 F2d 841, 851 (DC Cir 1970). See further Matthew Stephenson, ‘A Costly Signaling Theory of “Hard Look” Judicial Review’, 58 Administrative Law Journal 753, 754 n 1 (2009); Cass Sunstein, ‘Constitutionalism after the New Deal’, 101 Harvard Law Review 421, 463–474 (1987). 16 California Retail Liquor Dealers Ass’n v Midcal Aluminum, Inc, 445 US 97 (1980). 17 Cited above n 56. 13 14

Part I: Effective Enforcement of Competition Law 269 regulatory process. The government can’t simply hide behind a fiction that this is all market outcome-driven; the government has to conspicuously take responsibility for all these regulatory decisions, and when it does so it will pay a political price if the results are unpopular. So this is again a view of how courts can reinforce political and democratic processes by making the government take responsibility for anticompetitive state action. Now, I think that this has a lot of problems in theory and in practice, and I’ll just mention two of them.18 One is the problem of cost externalisation—the export cartel problem. If you go back to Parker v Brown itself in the 1940s, this was the California agricultural cartel statute where 95 per cent of the raisins at issue that were subject to the California cartel were sold outside the state of California. So the costs were externalised on nonvoters, people who don’t actually have a say in the political process. So even if the state action is made conspicuous, and even if the state of California takes ownership, the chance that there will be a political repercussion is actually relatively small. A second major problem with representation reinforcement model is the asymmetry between the concentrated benefits to producers from anticompetitive regulations and the diffuse harms to consumers. So, often what happens—and this is sort of classic public choice theory—is that a very small group of producers stand to gain a lot from anticompetitive regulation, and so they will exercise the political capital to ensure those regulations are adopted, whereas the harm is spread across perhaps many, many millions of consumers who don’t have the same incentive to politically mobilise to fight back. There are other problems, but those are the two ‘push-backs’ against representation reinforcement as a model of what courts should look for. The second model, which I associate more with some aspects of EU jurisprudence, is what I call the substantive review model. The idea here is that the government can act to regulate in an anticompetitive way so long as it is justified by some important or compelling public interest and the regulation in some sense is sufficiently narrowly tailored to advance an interest without overstepping and harming competition too much relative to what is necessary. So this model actually does provide a check on some strong anticompetitive problems of regulation. The problems here are issues of political legitimacy because the model is highly anti-majoritarian and it potentially calls on judges to make core political decisions. So, just look at the elements: what interests are important? Well, it’s difficult in a vacuum for courts to contradict what a regulator or a national legislature has decided and to decide for itself that one

18 There is plenty of related literature. See notably Merrick Garland, ‘Antitrust and State Action: Economic Efficiency and the Political Process’, 96 Yale Law Journal 486 (1987); John Shepherd Wiley, Jr, ‘A Capture Theory of Antitrust Federalism’, 99 Harvard Law Review 713 (1986).

270 Part I: Effective Enforcement of Competition Law particular interest is less or more important than another. As to the narrow tailoring prong, this is particularly difficult, as its often rather easy to point out an alternative way in which something could have been accomplished when in fact that would not have been an alternative solution that would have been politically feasible given the different interests at stake. Political outcomes are inherently features of political compromise, and therefore what is most expedient from a regulatory perspective may simply not have been an alternative that was politically feasible at all. So, courts using a narrow tailoring approach may simply unwind political deals that were the only feasible deal on the table to get a problem addressed. Turning now to the third model that I proposed as residing somewhere between full substantive review, which raises these counter-majoritarian difficulties and representation reinforcement, which is perhaps too light a touch in terms of judicial review, is a doctrine called the ‘hard look doctrine’, and again I am borrowing here from US administrative law, where courts sometimes review administrative decisions or regulations for the purpose of ensuring that full consideration has been made on a public record but without necessarily substantively second-guessing the decisions of regulators.19 So, what are some of the features of hard look review that could be employed in a competition context? First of all, reasoned decision-making should be on the public record. There can’t be concealed reasons supporting anticompetitive regulation; it has to be articulated in a reasoned and open fashion as to what the justifications were in suppressing competition. Second, there should be an explicit consideration of alternatives: did the regulator or the legislator explicitly think about other alternatives? In the competition context, this would mean market-based alternatives. If there are market failures, did the regulator consider market-based approaches to solving those failures as opposed to adopting a more anticompetitive regulatory scheme? A third element of hard look review is the explicit consideration of public comment. Were stakeholders allowed to participate in the process? Were stakeholders allowed to make comments that were reflected in the record of decision-making by the agency, and were they taken seriously into account? And a final element of hard look review is the contemporaneousness of the assertion of a justification for a

19 See also Kathryn Watts, ‘Proposing a Place for Politics in Arbitrary and Capricious Review’, 119 Yale Law Journal 2, 34–35 (2009) (hard look review was ‘one of the main tools that the courts developed to ensure that agencies were looking at the statute and the evidence and were choosing answers that served the public good rather than special interests’); Richard Stewart, ‘The Reformation of American Administrative Law’, 88 Harvard Law Review 1667, 1758 (1975) (‘the requirement that agencies give adequate consideration to all affected interests, and in particular, the interests of the intended beneficiaries of an administrative scheme, has been utilized by the courts with increasing frequency to redress perceived agency favoritism to organized interests’).

Part I: Effective Enforcement of Competition Law 271 regulation and its promulgation. In other words, the justifications cannot simply be made up in litigation by lawyer teams who come in after the fact. If the state is going to regulate in a way that dramatically suppresses competition, it’s going to need to think about justifications for doing that at the time it regulates, which I think in the competition context means involving teams of economists and industry specialists who are best positioned to provide input. So I think this third model is promising in some ways. It provides more of a check than the representation reinforcement model and it presents perhaps less of a counter-majoritarian difficulty than some of the substantive review models because it leaves less room for courts to substitute their own judgement as to what is important or sufficiently tailored. So, ultimately, although the hard look review gives agencies and legislatures permission to regulate in anticompetitive ways so long as they are explicit in doing it and so long as they provide evidence of a reasoned and considered decision, that alone can have a deterrent effect on most anticompetitive regulatory measures. To conclude, and to be clear, I am not proposing the hard look model as a one-size-fits-all solution for all jurisdictions. In part, the motivation for this project is that jurisdictions all across the world are looking at the problem of state measures distorting competition, and it’s helpful to articulate in a sort of abstract way what kind of models are available. I wouldn’t suggest that the hard look approach is appropriate for all jurisdictions; it may be appropriate for some contexts and not for others. But I do think it’s a promising alternative to some of the problems with the representation reinforcement and substantive review models. Monti: Many thanks, and again I am just going to collect points as you raise your flags. As you do that, I’ll abuse my privilege and make a few observations of my own. If you think about Dan Crane’s presentation, one of the implications of this would be to constitutionalise competition law. That is to say that if you pass any legislation that has anticompetitive effect, just as you check it for human rights compliance, you now check it for competition compliance, and I think we can agree about the extent to which one might want to go down that road. Another point, from Alexander’s presentation, related to the ‘European semester’. In reference to that point, another important element to look at in terms of economic governance is the set of Troika measures that are imposed when a Member State wants financial assistance. So if you look at the Greek Lignite case, some of the issues may now be moot because liberalisation of the energy markets in Greece is a condition for receiving funds. Another interesting thing with the Troika requests is that they must reform their competition authorities and strengthen competition enforcement, so there is another dimension to strengthening competition enforcement which comes from economic governance—a perhaps unexpected result. Mario, you were first. DD

272 Part I: Effective Enforcement of Competition Law Siragusa: This is simply a piece of information. One of the reforms adopted by Mario Monti in his last act of government in Italy was to grant the Italian Antitrust Authority the power to take action against administrative acts that have a restrictive effect on competition. It is a very powerful decentralisation of some of those principles which you have discussed in relation to Article 106. So it is a very interesting example, and there have been cases where the authority has already begun to take action before the administrative courts. The Authority can first issue an opinion stressing why they consider that an administrative act has restrictive effects, and then, if no corrective action is taken, they can seek to the annulment of that act by the court. DD

Wright: One thought on the paper presented by Stephen Calkins. There is a way to know very concretely how amicus briefs affected the Supreme Court, and that is because former members of the Court have released collections of their papers into the public domain.20 On the basis of these papers, you can tell very specifically in some of the files how particular amicus briefs moved them. What is most interesting from their papers is that it is the brief that focuses on a pivotal assumption in the proceedings below, a pivotal assumption that is demonstrably wrong or shown to be doubtful. In the Kodak case,21 an enormously impressive and effective brief was one that said: here are 1,001 ways that after-markets can be subject to manipulation, and it was not from the states but from academics who said: ‘In case you’re interested, here is a list’. Maybe the nicest example is in Sylvania,22 and Justice Lewis Powell’s papers point this out. Powell used to make notes with a red pen and you see the original on the copy of the brief that Donald Turner filed for the motor vehicle manufacturers that said: ‘This one’s worth reading’—and Powell underlined that with his pen. You can see directly from the papers how it made a difference. Related to that, there is a lesser form of amicus brief, where academics who anticipate the emergence of matters write articles and if they come from fancy law schools they send a draft to the clerks of the Court. This has happened on a number of occasions, and in one case it provided the intellectual foundation for the famous Brunswick decision that deals with injury to competition. On a different note, let me pose a question for Dan Crane: my blog colleagues23 would say that the ‘hard look’ doctrine is one of the most manipulable concepts in administrative law. They’d sat it can become a cloak for fundamentally political decision making. How do I define hard look? How do I apply it? If you consider Mario’s DD

20 See http://www.scotusblog.com/2013/08/accessing-the-papers-of-supreme-court-justicesonline-other-resources/. 21 Cited above n 58. 22 Continental TV, Inc v GTE Sylvania Inc, 433 US 36 (1977). 23 See www.truthonthemarket.com.

Part I: Effective Enforcement of Competition Law 273 comment about how the antitrust authority can challenge other authorities and Giorgio’s observation about the Troika, is there some point at which you provoke a real political crisis? If you give competition agencies power to strike down any anticompetitive measures, do they ultimately start walking through labour law? Or through environmental law? How broad a mandate do they assume? At what point do you get a severe political backlash that puts the competition system in a much smaller box? Forrester: Two thoughts on the subject of the amicus and the role of the advocate, the public advocate for competition. It’s very interesting to hear what you say about the impact of amicus briefs in the Supreme Court. The European Courts have a different test and a slightly different role for the intervening party. So if you say: ‘I am worried about this case, this would be a very bad decision, or this would be a very good decision and I want you to hear why by reference to my situation’, the Court of Justice says: ‘Mmm, okay, you can come in’. But then this person becomes a party. It is heard, it makes observations, and that means that the other parties have to be given the chance to answer those observations. So the threshold for being admitted is different from the US threshold, and perhaps this complicates the case a little bit. It certainly slows down the handling of the case in Luxembourg. But it sometimes happens that in very large cases, where you have several intervening parties on one side and several intervening parties on the other side, the interventions can help focus the issues. The other thought I had concerns the role of the Commission itself as a competition advocate. I would have thought that, at a time of institutional crisis and economic crisis, the role of the Commission as a voice in support of competition is of enormous importance. Even if you did not have Articles 101, 102 and 106, the Commission can engage on issues and keep the Member States honest, so to speak. DD

Lowe: I’ll underscore that a bit, Ian. On amicus interventions, the Commission’s role is slightly different from any normal amicus brief, in the sense that it has a regulatory obligation to comment in certain instances. In other circumstances, like preliminary hearings and cases in national courts—and in contrast to the situation in Germany, where I think the BKA would systematically be asked by the court to appear—there is discretion for the Commission to judge whether to devote resources to that case given the enforcement priorities of the European Union. We’ll come back to agency priorities later, but frankly the burden of automatically filing amicus briefs at the Commission would be extremely high. On the other hand, as you say, there are important actions by Member States which need to be commented on and, as Giorgio and Alexander remarked, the present dialogue with programme countries and DD

274 Part I: Effective Enforcement of Competition Law countries with sovereign debt issues is one which is being levered effectively by the Commission to secure reforms which wouldn’t otherwise occur. But apart from that leverage possibility, why don’t we use Article 106 sufficiently? I think it was used at a time when liberalisation was at its inception in many sectors but, in the present state of most post-liberalisation markets, the issues involve several strands of public policy and not just competition versus lack of competition. I think it’s a judgement call for the Commission: is this something where the Commission should take action? Or is it better to let the democratic and political process play itself out? The choice over the last 10 years has generally been to defer to the democratic process involving both the Member States themselves and the Parliament. That may be wise in some cases because moving into this territory without looking very, very closely at the intended and the unintended consequences in a number of sectors may be dangerous. Furthermore, just as we may look at policies and regulation which restrict competition, policies and regulations which actually promote competition deserve to be looked at as well. In this respect, it is best to avoid a silo mentality. A more holistic view is needed in order to arrive at a superior solution to the problem. Is it regulation or a lack of it? Is it antitrust? Or is it direct application of state aid rules or is it Article 106? Frankly, both within the Commission and outside the Commission, we don’t necessarily take that holistic view, and you can see it in cases where the proposal is to regulate in order to promote competition. I didn’t mean to imply that we would go beyond our vires if we accepted voluntary commitments; that’s up to the Member States and up to companies. At the same time, we have a situation now where, for example, there are proposals to regulate inter-bank fees, and yet Alexander and his staff have been taking cases which solve the particular problem in the particular instance concerned. Another situation on the state aid side is the well-known role of DG competition in restructuring cases for banks and yet last week the Commission adopted a proposal to regulate that through a banking resolution mechanism,24 thus involving the Member States and the Parliament in things that could, in principle, be decided exclusively by the Commission. So yes, there are a good deal of judgement calls. Lianos: I have a question for Dan about his very interesting presentation. In Europe, obviously because of the federal, or quasi-federal type of model, state action is actually national, and I was wondering how you could apply the hard look doctrine in this context. Most of the Member States now have developed regulatory impact assessment processes for any type of regulation, so basically these are public documents most of the time, and you can see there the consideration of different options. But the problem I have is that that does not really solve the problem of externalities because, in regard to impact DD

See http://ec.europa.eu/internal_market/finances/banking-union/index_en.htm.

24

Part I: Effective Enforcement of Competition Law 275 assessments, the focus is on domestic effects. They don’t really n ecessarily take into account the possible effects on competition in other jurisdictions and, unless you have a requirement to that effect, I can’t see how you could apply the hard look review in that particular context. A proposal in that direction was made here at the EUI by former Advocate General Maduro in his PhD thesis, ‘We the Court’.25 That was not in the context of competition law but in context of the free movement rules. But in practical terms it’s very difficult to apply. And beyond that difficulty, even if I agree that in principle it is an attractive idea, judges will not necessarily have the legitimacy to get into this type of analysis. But here we also need to take into account comparative institutional perspectives in the sense that judges might have imperfections but maybe other institutions are burdened with even more issues. Gerard: I have a quick comment on Article 106. José Luis, you wondered if that provision belonged to the competition family; you said it was rarely invited to the table. For my part, I doubt that it really belongs to the family. Alexander Italianer just reminded us of the affiliation between Article 106 and the GATT, in the sense that Article 106 sort of embodies a discipline that is more related to issues of trade law than to competition. My concern is that when the Commission applies Article 106 together with Article 102, it gives an interpretation of Article 102 that is not consistent at all with the interpretation that this provision receives in ‘stand-alone’ cases. It is a much looser concept of Article 102, and of the concept of abuse in particular. So my question is whether it would not make more sense, in those cases, to apply the internal market provisions of the EU Treaties as opposed to the competition rules. In most of those cases you see that in fact what is at stake is very often an impediment to cross-border trade, or at least you can construe it that way. You also have public policy considerations that come into play, which are also very hard to factor into a competition law analysis. So I would like to hear your views about whether that provision should be removed from the competition law family. DD

Ost: I’d like to come back to amicus briefs. The Bundeskartellamt rarely writes briefs, but it does have the right to stand up and be heard before the German Federal Court of Justice. This may imply a trade-off between scarce resources on the one hand and the interest of the authority to help shape the law on the other. Roughly speaking, I think we have 5–10 interventions a year before the Federal Court of Justice. We also have a member of the Court here, who might confirm this. DD

25 See Miguel Poiares Maduro, We the Court: The European Court of Justice and the European Economic Constitution, Hart Publications, 1998.

276 Part I: Effective Enforcement of Competition Law Monti: Thanks very much. There were quite a number of questions and I would like to give the speakers a chance to respond. Let’s go in the order of the speakers. So Alexander, would you like to begin? DD

Italianer: On amicus interventions, I would just say that of course we have to prioritise these, together with our friends from the legal service. The two main criteria on which we intervene are: first, it must be at an advanced stage in the legal review process. It has to be close to the highest level of the court and not at the lower courts. So when we come in it is almost at the end of the process. Secondly, there must be a general issue at stake. We do not want to make an appearance in a case to ask the court to rule in favour of one party or another. As an example, the Irish beef case raised general issues regarding crisis cartels, so in the national court we expressed our views on how crisis cartels can be analysed from a competition point of view.26 On the ‘hard look review’ model, many of the elements that Dan proposed in that model are indeed present in the regulatory impact assessment process that we follow at the Commission. And yes, that process does include consideration of any potential effects on international trade. Even in the review of the rules of services of general economic interest we included elements of that model, although without seeking to apply any particular model. If Member States under the new SGEI rules want to use or indicate a service of general economic interest, they have to consult on that and they have to publish it, so some elements of the hard look model are actually already being introduced. Then there was Damien’s question: should Article 106 be taken out of the family? Well, you know the child has been born. [Laughter] I wouldn’t want to resort to euthanasia. Finally, on market studies, and this is one of our advocacy activities—we have set up a food task force because we were particularly concerned about the review of the common agricultural policy, and so, together with our colleagues from DG Agriculture, we have been extremely active in the review of the Common Agricultural Policy. That is the only sector policy in the EU where competition rules apply on an opt-in basis. The Treaty rules do not apply automatically, so the EU legislator has to opt in, and there was a big risk that part of these rules would be excluded. In particular, there was a danger that producer organisations would be allowed to collectively negotiate prices. So, through this advocacy action, what we actually obtained was that there has to be some kind of efficiency. So producers can sit together, and in limited circumstances they can act collectively, including on price negotiation; but there has to be an efficiency rationale, as in the case of a cooperative. So, from an advocacy perspective, that is a big achievement, and it is also in line with what Philip was saying: you sometimes have to take DD

See http://ec.europa.eu/competition/court/antitrust_amicus_curiae.html.

26

Part I: Effective Enforcement of Competition Law 277 a more holistic view of markets and corresponding legislation because if you want to act using the competition rules exclusively you may miss part of the picture. Cyprus is the most blatant example: the rescue operation in the case of the Cypriot banks did not involve state aid. There was a bail-out of the creditors but there was no state aid, so we could not impose any of the state aid rules on the restructuring of the Cypriot banks, and that is why the proposal on the single resolution mechanism will come in handy. I think it is a very clear demonstration of why you need complementarity of competition enforcement rules and of relevant framework rules. This may explain a little bit more why Article 106 is becoming, as far as the Commission’s activity is concerned, a bit more obsolete than it was in the past. Buendia Sierra: I have read the paper by Dan Crane and I listened to him, and I think it’s very interesting to see these three models. He recognises that it is not at all obvious that you can translate that to a different context like the EU. In the EU we have a very different issue that complicates things. We are not just weighing public interest; we are weighing different public interests, then a national one and then a European one. I think it is very hard from a theoretical point of view to depart from the substantive review approach, but I think as you recognise there are mixed elements, and we are seeing more and more mixed elements, with more deference by the Court and the Commission to the Member States. But if we go too far in that direction we might have a problem because, in an internal market, you cannot simply let Member States decide what the common interest is. Every Member State has its own common interest at the national level. Let me also react to Philip Lowe’s comment. What I am saying here is that I don’t understand why individual cases under Article 106 are treated so different compared to state aid cases, which are at least equally as important or sensitive, if not more so. In the state aid arena the Commission does not have any problem, at least not in the country where I come from, and I am sure in many others. I therefore suggest that the Article 106 system should be implemented in one way or another, and my point is that the only sensitive and effective way to implement it is through a more active role of the Commission. That allows me to answer the comment by Damien Gerard. Damien says that the application of Article 106 with Article 102 leads to a very primitive application of Article 102. If the Commission goes forward with a modern economic reading of Article 102, what would prevent the Commission from applying that Article combined with Article 106 in an economically intelligent way? That would just be a matter of ensuring a consistent policy. But as for the question of whether Article 106 is a member of the competition law family or not, well frankly I don’t care. What is important is that it is implemented. You say why apply Article 102 and not internal market rules? Well, both apply—and both can and must be applied, according to the Treaty. I don’t see a strict borderline. By the way, I would remind all of you that, since the last amendment of the DD

278 Part I: Effective Enforcement of Competition Law Treaty in 2009, the system of rules that ensures undistorted competition— which includes all the rules we’ve been talking about—is explicitly considered an essential element of the internal market. The competition rules of the Treaty are internal market rules by definition. Crane: Two quick points, one on cost externalisation. There is a paper by Rubinfeld that actually proposes heightened scrutiny for state action that suppresses competition when there is disproportionate cost externalisation to non-voters.27 I think that would be a sensible element of any kind of a regulatory system. Then quickly on Stephen’s paper and the question of amicus briefs. If we think about the role and influence of amicus briefs in the last 10 or 15 years, they seem to be one element of a wider, sort of ‘PR’ strategy, so to speak, that today includes party-sponsored academic papers and partysponsored op-eds and blog posts. So there is a wide range of ways for people interested in litigation outcomes to try to influence judges and advocates involved in the case. Amicus briefs are in this sense one very important part of a broader set of outlets for argumentation that are available and being increasingly deployed, at least in the US. DD

Monti: OK, thanks to everybody. You’ll be glad to know its lunchtime. We’ll reconvene in an hour. DD

27 Robert Inman and Daniel Rubinfeld, ‘Making Sense of the State Action Doctrine: Balancing Political Participation and Economic Efficiency in Regulatory Federalism’, 75 Texas Law Review 1203, 1207 (1997).

José Luis Buendía Sierra* Enforcement of Article 106(1) TFEU by the European Commission and the EU Courts

1. Introduction EU competition law not only consists in rules addressed to undertakings, such as Articles 101 and 102 TFEU, but also includes certain rules addressed to Member States in order to control and limit (at least to some extent) their interventions in the marketplace. The best known of these rules are Articles 107 and 108, which concern state aid. However, even if the concept of state aid has become increasingly elastic, it does not embrace all possible state measures which distort competition. For instance, it does not cover those state measures which distort the market but do not provide a financial advantage to certain undertakings. Nevertheless, the Treaty contains another provision relating to anticompetitive measures adopted by Member States: Article 106 TFEU.1 This chapter will review the application of Article 106 in the last few years by the European Commission and the European Courts. The common denominator in all cases is the application of competition rules in a context where state measures (other than state aid) distort competition. This review does not include the many recent state aid decisions in which Article 106(2) is used to justify the compatibility of the financing of services of general economic interest (SGEI).2

* Partner at Garrigues, Brussels; Visiting Professor at King’s College London. Some sections of this chapter are based on a previous article of the author for e-Competitions. See Buendía Sierra, ‘Exclusive or Special Rights under Article 106 TFEU: An Overview of EU and National Case Law’, e-Competitions No 44436 (20 March 2012). The author wishes to thank Maria Muñoz de Juan and Jose Manuel Panero Rivas for their assistance in the drafting of certain sections below. 1 The present text refers exclusively to the Treaty provisions by the new numbers applicable following the Treaty of Lisbon. Article 106 TFEU therefore corresponds to former Article 86 EC (which corresponded to Article 90 of the EEC Treaty). 2 For the application of Article 106 in the context of state aid cases, see José Luis Buendia Sierra, ‘Finding the Right Balance: State Aid and Services of General Economic Interest’, in EC State Aid Law: Liber Amicorum Francisco Santaolalla, Kluwer Law International, 2008, chapter 10.

280 José Luis Buendía Sierra Article 106 is a complex provision.3 On the one hand, Article 106(1) prohibits Member States from adopting, with regard to public undertakings or undertakings which have received exclusive or special rights, any measure that would lead to an infringement of another TFEU rule, including Articles 101 and 102 and the free movement provisions. On the other hand, Article 106(2) contains, in favour of SGEI, an exception from the different Treaty rules, including the competition and free movement provisions, to the extent that this is necessary for the achievement of the missions entrusted to them. Finally, Article 106(3) allows the Commission to adopt directives or decisions in order to implement the rules contained in the two previous paragraphs. Paradoxically, while Article 106 is recognised as a member of the EU competition law family, in practice it is not often invited to the family table. As we will see, at present its application by the European Commission is very limited. Moreover, the Commission’s policy as regards communicating the few decisions adopted in this area can hardly be hailed as a triumph of transparency, something which obviously makes it more difficult to understand this provision.

2. European Commission cases The Commission can apply Article 106 TFEU either by deciding on individual cases or by adopting general legislative, or quasi-legislative, measures addressed to all Member States. Article 106(3) expressly provides for both possibilities. We will now look at the way the Commission has acted in both areas.

2.1 Individual Article 106 TFEU cases The Commission seems to consider Article 106(1) infringements as an area in which it has full discretion to decide whether or not to act. This means that complaints against state measures of the type coming within Article 106(1) would only be pursued if the Commission considers them to be not only legally founded but also (politically) opportune. If not, the c omplainant

3 For a more comprehensive study of this provision, see Buendía Sierra, Exclusive Rights and State Monopolies under the EC Law: Article 86 (Former Article 90) of the Treaty, Oxford University Press, 1999. A new edition of the book is under preparation. For a more limited presentation of the topic, see Buendía Sierra, ‘Article 106—Exclusive Rights and Other Anti-competitive State Measures’, in Jonathan Faull and Ali Nikpay, eds, The EU Law of Competition, 3rd edn, Oxford University Press, 2014, chapter 6.

Enforcement of Article 106(1) TFEU 281 would face an implied or express refusal by the Commission to handle the complaint. Based on the current case law, it is almost impossible for an individual to have locus standi to challenge such a refusal before the General Court.4 The only theoretical option would then be for the complainant to have recourse to a national court and invoke the direct effect of Article 106(1) to set aside the state measure in question, often requesting the judge to refer a preliminary issue to the Court of Justice. Suffice to say that this is not an easy task in the real world. Clearly, then, the situation of potential complainants in the field of Article 106 is worse than in all other areas of EU competition law. It is true that the Commission often refuses to deal with antitrust cases on the basis of a lack of EU interest, but a comprehensive network of reasonably competent and independent authorities and courts exists at the national level to decide these matters. Moreover, in the more closely related area of state aid, the Commission cannot simply refuse to rule on a complaint; it has to adopt a decision which is normally subject to judicial review based on the legal merits. On the contrary, with respect to Article 106, the Commission considers itself to have total freedom to act or (more often) not to act, and this may be easily influenced by reasons of (political) opportunity. This is exactly what has always happened with all infringements of the non-competition-based Treaty rules, where the Commission has a rather pragmatic and policy-oriented approach. This failure to act automatically in all cases has obvious implications for the legal community’s perception of the predictability of the rules in question. In fact, one may even wonder whether the Commission really considers Article 106 as fully being part of the competition rules. In any event, given the discretion which the Commission currently enjoys in this area, it is particularly interesting to examine the small number of cases where the Commission has indeed decided to act. At the very least, this is easier than attempting to detect the numerous cases where the Commission has not acted and understanding why it chose to proceed in this way. During the last 10 years, the Commission has only adopted five individual decisions based on Article 106(3) concerning four cases of exclusive and special rights (one of the cases, on Greek lignite, was dealt with twice), a distinctly unimpressive average of approximately one decision every two years. However, despite the Commission’s discretion to act in this area, neither the number nor the identity of the cases shows any clear establishment of priorities. Two of the cases relate to the liberalisation of the postal sector. First, in 2004, the Commission acted against Germany for discriminating against mail

4 Case C-141/02 P Commission v T-Mobile Austria GmbH, formerly max.mobil Telekommunication Service GmbH [2005] ECR I-1283.

282 José Luis Buendía Sierra preparation providers.5 Although the preparation of mail was a value added activity that in theory had been liberalised by the Postal Directive, the German regulatory framework established a tariff system in which the incumbent, Deutsche Post, could not give private postal intermediates a discount because they were preparing the mail themselves and therefore reducing the cost of the service. As a result, private operators were at a commercial disadvantage compared to the incumbent, which was able to extend its legal monopoly from the reserved area to the mail preparation area. This extension of a dominant position was the result of the German law approving the tariffs and not the autonomous behaviour of Deutsche Post. The Commission therefore concluded that the state measure in question implied the extension of the dominant position of the incumbent in violation of Article 106(1) in conjunction with Article 102 and requested its removal. The alleged justifications under Article 106(2) were examined and rejected. Four years later, on 7 October 2008, the Commission adopted a very similar decision as regards a new Member State, Slovakia, regarding the extension of the previously existing postal monopoly to hybrid mail, a category of services not initially included in the exclusive rights of the incumbent. The extension was the result of a new law, which was considered by the Commission to imply the extension of the dominant position of the incumbent in violation of Article 106(1) in conjunction with Article 102. As with the German case, the alleged objective justifications presented by the beneficiary as regards the cost of the universal service were not accepted. In particular, the Commission rejected the methodology of the studies submitted by Slovakia and asked it to abolish the new law and let firms operate again in the hybrid mail market.6 The main lessons from the above two decisions relate to the interaction between Article 106 and harmonisation through EU directives.7 In these

5 Commission Decision of 20 October 2004 on the German postal legislation relating to mail preparation services, in particular to the access of self-provision intermediaries and consolidators to the public postal network and related special tariffs (BdKEP—Restrictions on mail preparation), http://ec.europa.eu/competition/elojade/isef/case_details.cfm?proc_ code=1_38745. For a summary, see Manuel Martinez Lopez and Silke Obst, e-Competitions No 36869, October 2004, www.concurrences.com. 6 Commission Decision of 7 October 2008 in Case COMP/39.562—Slovakian postal legislation relating to hybrid mail services, C(2008) 5912, C(2008) 5912, http://ec.europa. eu/competition/antitrust/cases/dec_docs/39562/39562_689_1.pdf. See Thomas Brunhes, e-Competitions No 35135, October 2008, www.concurrences.com. 7 By contrast, in an investigation conducted in France by the Conseil de la Concurrence, the competition authority considered that La Poste, the French postal incumbent, had taken autonomous decisions to leverage its market power into the mail preparation and hybrid mail markets. For this reason, it applied Article 102 instead of Article 106. (See La Poste, Decision 05-D-63 of 17 November 2005, summarised in Juliette Goyer and Laurianne Lépine-Sarandi, eCompetitions No 495, November 2005, www.concurrences.com.) The French case is curious because in 2001, in very similar circumstances, the Commission had already issued an Article 106(3) decision against France, finding a breach of A rticles 102

Enforcement of Article 106(1) TFEU 283 cases, the Commission has clearly used Article 106 to ‘police the borders’ of the reserved area under the EU directive, probably because this instrument is legally more robust and/or more effective than the traditional infringement actions for incorrect transposition. Moreover, it is clear from both decisions that the Commission considers that there is a presumption that services included in the ‘reserved area’ by the Postal Directive are necessary for the equilibrium of the SGEI and acceptable under Article 106(2). Those outside the reserved area, however, do not benefit from such a presumption and, theoretically, the Member State must prove the necessity of extending protection to them. However, these two cases show that, in the Commission’s view, it is extremely difficult to prove the actual necessity of any exclusive rights outside the area defined in the directives. Indeed, since mail preparation services fell outside the reserved area by the German decision and the tariffs of the latter area were supposed to be calculated so as to cover the SGEI’s costs, said necessity was de facto automatically excluded. In the Slovak case, the Commission also questioned the reliability of the method used by the Member State with a surprising degree of rigour compared to its traditional flexibility in this area (for example, the French postal case). Moreover, the decision also found that some intangible benefits to the undertaking had not been taken into account, such as the benefit to its image as a result of having been entrusted with managing the SGEI. This latter approach, if taken consistently, may have wider implications in many other cases related to SGEI sectors. In any event, the result of the above two postal cases is not surprising in view of the implicit deal at the heart of the liberalisation of the postal and other sectors. Since both the scope of the reserved area and that of the universal service have been negotiated, it would be extremely problematic from a political point of view to allow a Member State to unilaterally extend the exclusive rights. The same would probably apply if the Commission ever wanted to use the Treaty to reduce the scope of the exclusivity agreed in the directive. While this may be legally possible, politically it would undoubtedly be dangerous. On 10 May 2008,8 the Commission adopted an Article 106(3) decision against the exclusive rights granted by France to certain financial institutions (Banque Postale, Caisses d’Epargne and Crédit Mutuel) for the distribution of the Livret A/Livret Bleu, two types of account with fiscal advantages for account holders. For this reason, the latter were inclined to use these public

and 106. See Commission Decision of 23 October 2001 in Case COMP/37133—La Poste [2001] OJ L120/19. 8 Commission Decision of 10 May 2007 pursuant to Article 86(3) of the EC Treaty, on the special rights granted to La Banque Postale, Caisses d’Epargne and Crédit Mutuel for the distribution of the livret A and livret bleu, C(2007) 2110 final, http://ec.europa.eu/ competition/liberalisation/livret_a_en.pdf.

284 José Luis Buendía Sierra or semi-public banks not only with regard to Livret A/Livret Bleu, but for all their banking activities. Moreover, the state paid fees to the banks for them to manage certain public service obligations (using the funds collected from Livret A/Livret Bleu for financing social housing). For these reasons, in 1991 private banks complained about this exclusivity, mainly invoking state aid rules. The history of this case is particularly long and convoluted. For a number of years, the Commission treated it as exclusively concerning state aid and in 2002 it adopted a negative decision stating that Crédit Mutuel had been overcompensated and sought the recovery of the unlawful aid.9 The decision was ambiguous as regards the scope of the aid and did not specify whether the exclusive rights were considered as such or not. In 2005 the EU Court of First Instance set aside the decision for failure to give sufficient reasoning.10 The Commission then apparently abandoned any attempt to decide whether the exclusive rights amounted to state aid and to quantify their economic value. Instead, on 10 May 2007, it adopted an Article 106 decision claiming that the exclusive rights enjoyed by Livret A/Livret Bleu were in breach of Article 106(1) in conjunction with the Treaty rules on the free provision of services and freedom of establishment without being justified by any legitimate objective. In particular, the Commission examined the facts under Article 106(2) and rejected the arguments that the missions entrusted to the entities had a public service nature. Thus, not only was the SGEI nature of these services open to doubt, but also, the exclusive rights were not required to guarantee the objective at stake. As a result, the Commission asked the French government to abolish the exclusivity in the future in order to bring the infringement to an end. Of course, in line with its Article 106(3) practice, in its 2007 decision the Commission did not try to quantify or recover the economic advantage enjoyed by the entities through the use of the exclusive rights. This case raises interesting questions about the blurred dividing line between exclusive and special rights on the one hand and state aid on the other. It is clear that exclusive and special rights normally have an economic value, and it follows that where they are granted this may involve the granting of an advantage within the meaning of Article 107(1). This is indeed the line taken (in theory, at least) by the Commission in its latest review of the Altmark package (now called Almunia’s package).11 The logical consequence

See Press Release IP/02/67 of 15 January 2002. Case T-93/02 Confédération nationale du Crédit mutuel v Commission [2005] ECR II-143. 11 See http://ec.europa.eu/competition/state_aid/legislation/sgei.html. In particular, see the Communication from the Commission on the application of the European Union State aid rules to compensation granted for the provision of services of general economic, [2012] OJ C8/4. 9

10

Enforcement of Article 106(1) TFEU 285 of this would be that the advantage enjoyed should be quantified and the beneficiary served with a recovery order. However, such a difficult course of action was avoided in Livret A/Livret Bleu.12 This ambiguity makes the current procedural situation of Article 106 cases versus Article 107 cases even more absurd. Not only do exclusive/special rights have similarities with state aid; in many cases exclusive/special rights are state aid. It is submitted that, for this reason, the Commission should no longer treat the two categories as mutually exclusive. On 5 March 2008, the Commission adopted its Greek lignite decision.13 This decision found that the grant by Greece of exclusive rights for the extraction of lignite to PPC, the state-owned incumbent electricity producer, was a breach of Article 106(1) in conjunction with Article 102. Once again, the reasoning followed the theory of the extension of a dominant position: since lignite is the cheapest source for producing electricity in Greece and since its transport over long distances is not viable, access to local lignite is essential for firms wishing to enter the electricity market. As the government had granted all exploitation rights to PPC, this public undertaking had clearly received an essential advantage. This allowed PCC to maintain its dominant position in the electricity market despite the fact that this market had, in theory, been liberalised. Greece apparently did not even try to justify the measure as necessary to maintaining the equilibrium of the SGEI entrusted to PPC and the Commission therefore did not examine the possible application of the exemption contained in Article 106(2). The Commission finally found that the measure was contrary to Article 106(1) in conjunction with Article 102, but, surprisingly, did not specify the remedy which Greece had to adopt. Instead, it issued a number of indicative targets and suggestions to Greece, such as reassigning lignite extraction rights to newcomers and/or price caps. The national government was therefore given the opportunity to ‘negotiate’ with the Commission the remedies that would put an end to the infringement within a given deadline. The result of the negotiation was then formalised in a new Commission decision adopted on 4 August 2009, which accepted the ‘commitments’ made by Greece to ensure

12 Apart from the exclusive rights examined in the 2007 decision (see above n 8), the Commission later adopted a state aid decision in which it concluded that another measure, namely the fees paid by the state for the distribution of the Livrets, did not overcompensate the entity and therefore did not constitute state aid. See Commission Decision of 24 May 2011 on state aid C 88/97 implemented by France in favour of Crédit Mutuel (notified under document C(2011) 3436), [2011] OJ L309/23. 13 Commission Decision of 5 March 2008 on the granting or maintaining in force by the Hellenic Republic of rights in favour of Public Power Corporation SA for extraction of lignite, http://ec.europa.eu/competition/antitrust/cases/dec_docs/38700/38700_517_3.pdf. For a summary, see Philippe Chauve and Polyvios Panayides, e-Competitions No 35237, March 2008, www.concurrences.com.

286 José Luis Buendía Sierra fair access to lignite deposits.14 Greece had actually undertaken to tender the exploitation of four lignite deposits to newcomers so that they could obtain about 40 per cent of the market. However, in 2011, Greece proposed replacing the measures for opening up of lignite exploitation with an alternative approach: access to 40 per cent of drawing rights concerning electricity produced in existing lignite-fired PPC plants. Before deciding on this proposal, on 14 January 2011 the Commission published a notice in order to market test whether these new solutions were workable and resolved the infringements detected in 2008.15 The Commission has never published or reported the result of this market test. The two Greek lignite decisions led to interesting litigation before both the General Court and the Court of Justice. At first, the General Court annulled both decisions in 2012. The first judgment (Case T-169/08) reopened the debate on Article 106(1) TFEU in a very relevant way, as will be described below. As a consequence, the judgment in Case T-421/09 also annulled the other Commission decision adopted on 4 August 2009, stating that, since the legal basis of the second decision is the first decision, the annulment of the latter necessary implies the annulment of the former. As anticipated, the judgment in Case T-169/08 opened a very interesting debate on the application of Article 106(1) TFEU. According to the previous case law, at least according to my reading of it, while applying Article 106(1) in combination with Article 102, the Commission was only requested to demonstrate that the measure placed the concerned company in the dominant position that had led it to abuse such position. The Commission obligation was limited to demonstrating that the state measure created an extension of a dominant position from one market to another, the effects of which were equivalent to those of abuse.16 As a consequence, the Commission did not have to prove the existence of an abuse (whether real or potential). However, the General Court interpreted the EU case law in a very different way and concluded in its judgment that it was not sufficient to state that the state measure had led to the creation of unequal opportunities (ie effects) for the Commission to declare there was an infringement, but it also had to identify the actual or potential abuse (behaviour) to which the measure led

14 See Press Release IP/09/1226 of 6 August 2009; Commission Decision of 4 August 2009 establishing the specific measures to correct the anti-competitive effects of the infringement identified in the Commission Decision of 5 March 2008 on the granting or maintaining in force by the Hellenic Republic of rights in favour of Public Power Corporation SA for extraction of lignite, http://ec.europa.eu/competition/antitrust/cases/dec_ docs/38700/38700_521_3.pdf. 15 Press Release IP/11/34 of 14 January 2011. For the commitments proposed by Greece, see http://ec.europa.eu/competition/antitrust/cases/dec_docs/38700/38700_716_14.pdf. 16 This was the Commission’s position. See Case T-169/08 Dimosia Epicheirisi Ilektrismou AE (DEI) v Commission, EU:T:2012:448, para 61.

Enforcement of Article 106(1) TFEU 287 or could lead the dominant company.17 Since the contested decision did not demonstrate the existence of a real or potential abuse, the General Court annulled the Commission’s decision. In my opinion, before this case, the General Court had never been asked to prove the existence of a real or potential abuse. On the contrary, previous judgments, such as RTT, accepted that the mere granting of an exclusive right to a public undertaking previously enjoying a dominant position was, in itself, contrary to Article 106(1) combined with Article 102 TFEU. This conclusion was based on the effects of the measures. Similarly, AG Wathelet underlined in his Opinion that a breach of Articles 106 and 102 may occur even in the absence of any abusive behaviour; it is enough to prove that the effects of the state measure may distort competition.18 According to him, the creation of a situation in which certain competitors receive a competitive advantage by the mere fact of using their special rights would fall under Articles 106 and 102, this without any need to refer to any specific abusive behaviour by the undertaking.19 AG Wathelet rightly pointed out that this interpretation confirms that the combination of Articles 106 and 102 has its own area of application different from that of Article 102 alone.20 The judgment on said appeal, adopted by the Court of Justice on 17 July 2014, has put an end to the above discussion.21 In this important and interesting judgment, the Court of Justice set aside the judgment of the General Court and referred the case back to it. Indeed, the Court of Justice has made clear that the combined application of Articles 106 and 102 does not require scrutiny of the behaviour of the undertaking, and that it may be enough to look at the anticompetitive effects of the state measures. The Court of Justice has thus followed the suggestions of AG Wathelet and rejected the restrictive interpretation of the General Court. The Court thus upholds the traditional case law on Article 106, confirming that the extension of a monopoly from one market to another by a state measure is, due to its anticompetitive effects, automatically prohibited.

17 As the General Court states in para 118: ‘It does not therefore appear that the caselaw relied on by the Commission permits it to ignore the case-law cited in paragraph 94 above and to base its argument solely on the question whether the inequality of opportunities between economic operators, thereby distorting competition, is the result of a State measure. Thus, the Commission cannot maintain that it was not required to identify and establish the abuse of a dominant position to which the state measure in question lead, or could lead, the applicant.’ 18 See Opinion of Advocate General Wathelet in Case C-553/12 P Commission v Dimosia Epicheirisi Ilektrismou AE (DEI), EU:C:2013:807, paras 55–65. 19 Ibid, paras 61 and 64. 20 Ibid, para 59. 21 Case C-553/12 P Commission v Dimosia Epicheirisi Ilektrismou AE (DEI), EU:C:2014:2083.

288 José Luis Buendía Sierra I submit, however, that the Court of Justice has gone one important step further: it has not merely confirmed the traditional case law; it has also complemented it by leading the ‘effects doctrine’ to its logical consequence. Indeed, the judgment states that any state measure producing anticompetitive consequences would infringe Articles 106 and 102 TFEU. Contrary to previous judgments, this statement is not confined to the ‘extension of a dominant position’ from one market to another caused by an exclusive or special right. It also applies to the original grant of a stand-alone exclusive or special right to a company. The point is clearly summarised in paragraph 46 of the judgment: 46. It follows from the matters addressed in paragraphs 41 to 45 above that, as the Advocate General states in point 55 of his Opinion, infringement of Article 86(1) EC in conjunction with Article 82 EC may be established irrespective of whether any abuse actually exists. All that is necessary is for the Commission to identify a potential or actual anti-competitive consequence liable to result from the State measure at issue. Such an infringement may thus be established where the State measures at issue affect the structure of the market by creating unequal conditions of competition between companies, by allowing the public undertaking or the undertaking which was granted special or exclusive rights to maintain (for example by hindering new entrants to the market), strengthen or extend its dominant position over another market, thereby restricting competition, without it being necessary to prove the existence of actual abuse. (emphasis added)

In doing so, the Court of Justice seems to go back to the logic of the Corbeau judgment:22 exclusive and special rights are presumed to be illegal under Articles 106 and 102 unless they can be objectively justified (eg because they refer to an activity which is a natural monopoly) or unless they are necessary to guarantee the effective carrying out of a service of general economic interest. This has already been the case for many years as regards the application of the free movement rules to exclusive or special rights, but the Court had been hesitant, until now, to apply the same logic as regards Articles 106 and 102. The DEI judgment finally seems to confirm this logical convergence.23 This confirmation may have arrived late, but it is nevertheless welcome. It has arrived late because, by now, the Commission and even the Member States consider it uncontroversial that exclusive rights must be justified in order to be legal. The judgment merely says clearly what most people considered to be the logical consequence of the case law. It is nevertheless a useful clarification and an even more useful reminder of the enormous possibilities of Article 106. Now that the Court of Justice has

Case C-320/91 Paul Corbeau [1993] ECR I-2533. I had previously advocated such a solution in Buendia Sierra, ‘Article 106’ in Faull and Nikpay, eds, above n 3, specifically points 6.89 and 6.90. 22 23

Enforcement of Article 106(1) TFEU 289 finally confirmed the far-reaching content of this provision, it would be desirable for the Commission to start applying it in a more coherent and determined way. The Greek lignite case is also interesting because it has shown the increased influence of antitrust law on procedural issues. For instance, the use of socalled commitments which are accepted by the Member State allows for greater flexibility. The reliance on market tests for checking the suggested remedies is another interesting technique imported from the antitrust field. Procedurally speaking, then, such an approach is far removed from the rigidity of traditional infringement cases against the Member States. The main point in common with infringement cases is the Commission’s absolute discretion to decide whether or not to pursue the case. In this respect, Article 106 cases are unlike antitrust or state aid cases, in which the Commission has limited discretion. Currently, the Commission only pursues Article 106(3) decisions when it wishes to do so.24 The active involvement of the Commission in the Greek lignite case is very much in contrast to the total passivity in other seemingly more important cases which never led to a Commission decision.25 As noted above, it is submitted that the Commission should be legally obliged to act rather than doing so at its own complete discretion, since this would create much more consistent decision-making practice. The substantive approach taken by the Commission (until it was called into question by the 2012 judgment of the General Court in Case T-169/08) shows a marked preference for finding the extension of a dominant position from one market to another. This approach is taken in the vast majority of cases and is consistent with the role of these decisions as a supplementary instrument for policing the frontiers of the liberalised sectors.26 The only exception is the reliance by the Commission on the free movement rules in the Livret A/ Livret Bleu case. The Commission has barely relied on any other possible legal

max.mobil, above n 4. In 2008, the Spanish company Iberdrola filed a complaint against the exclusivity granted by France to EDF for the production of electricity from nuclear energy. The case seems to be very similar to the Greek lignite case except that it does not concern an almost unconnected network such as that existing in Greece, but rather a country at the heart of the EU grid. It also seems at first glance that nuclear energy in France is no less important to competition than lignite is in Greece. Despite all this, the Commission apparently decided that this case was a lower priority than the Greek one and, to my knowledge, has apparently taken no action at all. With respect to the filing of the complaint, see http:// www.expansion.com/2008/05/01/empresas/energia/1118426.html. 26 The ECJ also found that exclusive rights beyond the area reserved by the Postal Directive are incompatible with the Treaty. See Case C-220/06 Asociación Profesional de Empresas de Reparto y Manipulado de Correspondencia v Administración General del Estado [2007] ECR I-12175, paras 80–88. 24 25

290 José Luis Buendía Sierra arguments, such as the conflict of interest created by the exclusive right,27 unlike the Court of Justice, which has continued to apply this routinely.28 Naturally, given the Commission’s low profile in this area, it did not take the more controversial approach of considering that the exclusive right in itself automatically implies abuse of a dominant position and is therefore illegal under Articles 106 and 102.29 However, in the Livret A/Livret Bleu case, its approach under Article 106(1) in conjunction with the free movement rules leads to exactly the same result: exclusive rights are considered to infringe the Treaty unless they can be justified under Article 106(2) or the harmonisation directives. It follows from the above considerations that the legal community cannot but regret the reluctance of the Commission to apply Article 106(1). More active behaviour could therefore be encouraged. The situation could also be improved as regards the transparency of Article 106(3) decisions. Indeed, it is unfortunate that such decisions are no longer published in the Official Journal. Although never a legal requirement, traditionally they used to be published in Series C of the Official Journal. As a result, the text of decisions is now only informally available via a link on the Commission webpage and usually only in English. This limited enthusiasm for transparency as regards Article 106 has recently spread to the Annual Report on Competition Policy, since, after many years of continuous presence, the section on ‘State measures (public undertakings/undertakings with exclusive and special rights—Article 106 TFEU)’ has simply disappeared from the latest reports. It follows from the above that the Commission considers that it has almost absolute discretion with respect to the Article 106 procedure. Moreover, its practice suggests that it considers this instrument as largely interchangeable with other legal bases. Thus, quite often it relies on traditional infringement procedures to challenge certain exclusive rights that could also be the subject matter of Article 106 decisions. This is the case, for instance, with certain actions against exclusive rights in the broadcasting sector or the action related to the Maltese petroleum monopoly.30 We have already seen in Livret A/Livret

27 The ENI case seems to be based on the logic of a conflict of interest leading to an abuse. The Italian gas company ENI allegedly blocked its competitors’ access to its transport infrastructure. In 2010, the Commission obtained a commitment from ENI to divest all of its shares in all transport pipelines. See Press Release IP/10/1197 of 29 September 2010. 28 Case C-49/07 Motosykletistiki Omospondia Ellados NPID (MOTOE) v Elliniko Dimosio [2008] ECR I-4863, paras 51–53. 29 Corbeau, above n 22. 30 On the latter case, see Phyllis Aquilina, e-Competitions No 14006. See also Case C-419/07 Commission v Sweden, EU:C:2008:387 (order of 8 July 2008, in French and Swedish) and the Greek broadcasting monopoly case, Press Release IP/06/1401 of 16 October 2006.

Enforcement of Article 106(1) TFEU 291 Bleu that there seems to be a substantial overlap between Article 106 and state aid procedures. Even some recent antitrust or merger decisions include remedies in the form of structural measures by the Member States that should normally fall within the scope of Article 106. One can, of course, understand why the Commission would be interested in such flexibility. However, this should not be at the expense of the rights of Member States and third parties.

2.2 Quasi-legislative acts adopted under Article 106(3) TFEU The Commission can use Article 106(3) as a legal basis not just for individual decisions but also for general instruments of a legislative or quasi-legislative nature. Indeed, in the past, the Commission used Article 106(3) directives to introduce certain transparency obligations as regards public undertakings and to liberalise the telecommunications sector. This use was somewhat controversial at the time but its legality was clearly confirmed by the Court of Justice. Subsequently, the Commission has regularly threatened to rely on Article 106 in order to liberalise certain markets—for example, whenever the proposals of liberalisation directives, like those relating to electricity or postal services, were blocked by the political opposition of certain Member States at the Council.31 In 2005, the Commission also invoked the possibility of using such powers to liberalise certain professional services.32 However, none of the above threats have materialised in the shape of new liberalisation directives under Article 106(3); its use in the recent liberalisation agenda has been a pure political one. The Commission has, however, recently used Article 106(3) as a quasilegislative instrument, though in a different context: that of state aid in the so-called ‘Almunia package’. It has therefore progressively changed from using this provision as a liberalisation instrument to its current main use as a clarification instrument in the area of state aid to SGEIs.33 In particular, the Commission adopted Decision 2012/21/EU under Article 106(3) in order to clarify the conditions under which state aid in the form of public service compensation could be considered compatible with Article 106(2).34 This decision

Commission, Report on Competition Policy 2006, paras 163 and 381. Commission, Report on Competition Policy 2005, para 140. A similar situation happened as regards individual cases. There are now very few Article 106(3) decisions relying on the prohibition of Article 106(1), but there are many state aid decisions examining the compatibility of the measures under Article 106(2). For the application of Article 106 in the context of state aid cases, see Buendía Sierra, ‘Finding the Right Balance’, above n 2. 34 Commission Decision of 20 December on the application of Article 106(2) of the Treaty on the Functioning of the European Union to state aid in the form of public service 31 32 33

292 José Luis Buendía Sierra is remarkable in two respects. First, despite being a decision, it is addressed to all Member States and therefore has a quasi-legislative nature. Secondly, this is the first time that the Commission has used this legal basis not to clarify the scope of the prohibitions under Article 106(1) but to clarify the scope of the exemptions allowed under Article 106(2). It is worth noting that Commission Directive 2006/111/EC, regarding the transparency obligation between Member States and public companies, which is also based on Article 106(3), continues to apply.35 One can conclude from the above that the Commission has not forgotten at all about the existence of Article 106(3) directives. Instead, the use (or nonuse) of this instrument seems to be dictated by political opportunity. There is nothing to object to in this, given that, unlike individual decisions, we are dealing with general legislative actions, not individual decisions based on specific violations of the Treaty rules.

3. ‘Indirect enforcement’ of Article 106(1) by the Court of Justice: case law on Article 106(1) TFEU arising from preliminary rulings The definition of the boundaries of Article 106(1) TFEU by the Court of Justice is not only driven by its enforcement by the Commission. A significant number of cases in which the substantive limits of the provision are tested stem from preliminary rulings resolved by the Court of Justice on questions submitted by national courts under the mechanism provided by Article 267 TFEU. These rulings allow domestic courts to enforce Article 106 TFEU at the national level, providing them with guidance on how to interpret the provision, leading to a sort of indirect enforcement of the provision by the Court of Justice. Remarkably, the use made by the Commission of its discretion on whether or not to pursue a case under Article 106, combined with the restrictive criteria on locus standi of interested parties for challenging the Commission’s decisions (or non-decisions), have resulted in these preliminary rulings being the bulk of the work of the Court of Justice when defining Article 106(1) TFEU.

compensation granted to certain undertakings entrusted with the operation of services of general economic interest, [2012] OJ L7/3. 35 Commission Directive 2006/111/EC of 16 November 2006 on the transparency of financial relations between Member States and public undertakings as well as on financial transparency within certain undertakings, [2006] OJ L318/17.

Enforcement of Article 106(1) TFEU 293 As Article 106 TFEU itself is placed on the same footing as other provisions, not all preliminary rulings in which a potential ‘Article 106(1) case’ is brought before the Court expressly refer to the provision. For instance, directives issued under Article 106(3) TFEU can be seen as specifications of Article 106(1) TFEU. However, when a preliminary ruling is brought before the Court in these cases, it is relatively uncommon to see a reference to Article 106(1) TFEU in the judgment, as national courts tend to specifically refer to the particular provision of the directive at hand. A similar phenomenon occurs with the combined application of Article 106(1) with Articles 49 and 56 TFEU, cases in which national courts tend to obviate the reference to Article 106(1) TFEU. Therefore, in my view, the cases I am going to refer to in this section should not be considered as the only cases in which Article 106 could have been applied. I will now review the last few years of preliminary rulings on Article 106(1) TFEU. My suggested approach is as follows: (i) first, to focus on a basic element of Article 106(1) TFEU, which is the recent case law of the ECJ on the concept of undertaking; (ii) secondly, to review the preliminary rulings concerning the combined application of Articles 106(1) and Article 102; and (iii) thirdly, to review the combined application of Article 106(1) with Articles 49 and 56 TFEU. I will end this section with some remarks on preliminary rulings concerning the interpretation of Article 106(2) TFEU when invoked as an exception to Article 106(1) or to the Treaty provisions on state monopolies of a commercial character.

3.1 Basic concepts (the concept of an undertaking) The concept of ‘undertaking’ is an essential element for the application of Article 106(1) TFEU in combination with other provisions of the Treaty. Three interesting recent cases should be recalled here. Two of them refer to the assessment of whether or not an entity managing a social security scheme should be considered an undertaking—and they provide contradictory responses—and the third refers to a very different area of activity— authorisation and organisation of motorsports. The cases referring to social security schemes are Kattner36 and AG2R.37 In both cases the ECJ invoked its previous case law by considering the fact that the entities were non-profit-making entities and pursuing a social objective

36 Case C-350/07 Kattner Stahlbau GmbH v Maschinenbau- und Metall- Berufsgenossenschaft [2009] ECR I-1513. 37 Case C-437/09 AG2R Prévoyance v Beadout Père et Fils SARL [2011] ECR I-973.

294 José Luis Buendía Sierra to be insufficient in itself to exclude the possible consideration of the entities managing the scheme (MMB and AG2R) as undertakings for the purposes of application of Articles 106 and 102 TFEU.38 Following its previous jurisprudence, the Court also analysed the extent to which the schemes at hand could be regarded as applying the principle of solidarity and to what extent the entities were supervised (or controlled) by the state.39 While in both cases the Court was satisfied with the degree of solidarity permeating the schemes,40 it arrived at different conclusions as regards the level of control of the entities by the state. In Kattner, the Court assessed that the degree of control by the state was sufficient to avoid MMB being considered an undertaking. The reasons for this were that the amount of contributions and the amount of benefits were subject to state control,41 and the fact that the state acted as the supervisory authority of the insurance association.42 In AG2R, the ECJ reached the opposite conclusion. The Court considered that, while some elements of the scheme restricted the margin of manoeuvre for AG2R,43 other characteristics pointed to the conclusion that the entity—and the social partners who agreed the given scheme, in turn validated by the government—enjoyed a certain degree of autonomy. Therefore, it concluded, AG2R could still be considered an undertaking for the purpose of Articles 106(1) and 102.44 Interestingly, the elements assessed concerning this last condition at AG2R do not exactly match those identified by the ECJ in Kattner. In the German case, the Court was looking at whether contributions and benefits were subject to the control of the state; in the French case, the debate was on the degree of autonomy of the social partners—or, conversely, the lack of involvement of the state—in the design of the system and whether operators of a different type to the one selected could provide similar services. The elements considered by the Court in AG2R allowed it to avoid the assessment of whether the action of the French state validating the agreement

See paras 35 and 38–41 of Kattner, and paras 43 and 44 of AG2R. See Case C-218/00 Cisal di Battistello Venanzio & C. Sas v Istituto nazionale per l’assicurazione contro gli infortuni sul lavoro (INAIL) [2002] ECR I-691, paras 38–44. 40 See Kattner Stahlbau, above n 36, paras 51–59; AG2R Prévoyance, above n 37, paras 48 et seq. 41 The minimum and maximum amounts of earnings to be taken into account in calculating contributions and benefits were established by law. However, the maximum amount could be established, where appropriate, by statutes of employers’ liability insurance associations. See Kattner Stahlbau, above n 36, para 63. 42 Ibid, para 64. 43 See AG2R Prévoyance, above n 37, para 58. 44 These elements were that: (i) supplementary collective guarantees can be established in different ways, being the use of collective agreement a choice of the social partners; and (ii) providential operations may be entrusted not only to provident societies and mutual insurance associations, but also to insurance companies. 38 39

Enforcement of Article 106(1) TFEU 295 between social agents would be equivalent to direct action by the state in establishing it at its discretion. It did not assess the actual degree of involvement of the state in the design of the system. It seems that the Court relied on form rather than on substance—whether the German system is not the result of the will of the sector and whether the French system does not respond to the will of the state—in order to draw the dividing line between the two situations. In the facts of these cases, it seems that the distinction was clear. However, other, less clear-cut, cases might require the Court to go into greater depth on this issue. In a different sector of activity, MOTOE also provided interesting remarks on the concept of undertaking for the purposes of application of EU competition rules.45 The question posed to the ECJ by the national court consisted of whether ELPA (the Automobile and Touring Club of Greece), a non-profit-making legal entity representing the International Motorcycle Federation (FIM) in Greece, which was legally entrusted with the powers to authorise motorcycle races in Greece, should be considered an undertaking for these purposes. Importantly, besides its powers of authorisation, ELPA also organises motorcycling events in Greece and in that connection enters into sponsorship, advertising and insurance contracts for the commercial exploitation of these events. The background of the case refers to an implicit refusal by ELPA to grant authorisation to MOTOE (itself a non-profit-making association organising motorcycle races in Greece) for the purposes of organising a series of races. MOTOE complained to a national court about the lack of conformity of the Greek legal provisions defining the powers of ELPA, as a potential violation of Articles 106 and 102 TFEU. As a first step in its assessment, the Court evaluated whether ELPA should be considered as an undertaking. Here, the Court recalled the basic principles that any activity consisting in offering goods or services on a given market is an economic activity,46 and the fact that an activity has a connection with sport does not hinder the application of the rules of the Treaty.47 As regards the possible effects of the exercise of public powers by ELPA on its classification as an undertaking, the Court noted that the fact that an entity is vested with public powers for the exercise of part of its activities does not, in itself, prevent it from being classified as an undertaking for the remainder of its activities,48 and that the classification of each activity should be carried out separately.49

MOTOE, above n 28. Ibid, para 22. Ibid. 48 Ibid, para 25. 49 Ibid. 45 46 47

296 José Luis Buendía Sierra The fact that ELPA (and also MOTOE) are non-profit-making associations does not alter this assessment since, as the Court points out, their offers coexist (or potentially coexist) with other associations that do seek profits.50 On these bases, the Court concluded that ELPA should be considered an undertaking for the purposes of EU competition law.51

3.2 Combined application of Article 106(1) with Article 102 TFEU One of the provisions most frequently applied in combination with Article 106(1) is Article 102 TFEU. While we are no longer in 1991 (when Höfner,52 ERT,53 Merci Convenzionali54 and RTT55 were ruled on by the Court), there have been some remarkable cases in this area. In addition to DEI (already mentioned),56 preliminary rulings have also provided interesting elements. The above-mentioned AG2R case raised two interesting issues: first, that in this case the Court assimilated the exclusive right held by AG2R to receive and manage the contributions paid by employers and employees in a given sector (traditional bakery sector in France) with the existence of a dominant position within the meaning of Article 102 TFEU. Specifically, the Court stated: Since, because of those exclusive rights, undertakings in the French traditional bakery sector cannot pay contributions to a scheme of supplementary reimbursement of healthcare costs managed by a different body, AG2R has a statutory monopoly in a substantial part of the common market and may be regarded as occupying a dominant position within the meaning of Article 102 TFEU.57

However, the creation of a dominant position in the sense of Article 102 TFEU by the granting of a special right in the sense of Article 106 TFEU does not automatically imply a breach of these Articles. The Member State would only contravene them if, by merely exercising the exclusive rights it holds, the undertaking is led to abuse its dominant position or when such

Ibid, para 28. Ibid, para 29. 52 Case C-41/90 Klaus Höfner and Fritz Elser v Macrotron GmbH [1991] ECR I-1979. 53 Case C-260/89 Elliniki Radiophonia Tiléorassi AE and Panellinia Omospondia Syllogon Prossopikou v Dimotiki Etairia Pliroforissis and Sotirios Kouvelas and Nicolaos Avdellas and others [1991] ECR I-2925. 54 Case C-179/90 Merci convenzionali porto di Genova SpA v Siderurgica Gabrielli SpA [1991] ECR I-5889. 55 Case C-18/88 Régie des télégraphes et des téléphones v GB-Inno-BM SA [1991] ECR I-5941. 56 See Section 2.1 above. 57 AG2R Prévoyance, above n 37, para 67. 50 51

Enforcement of Article 106(1) TFEU 297 rights are liable to create a situation in which the undertaking is led to commit such abuse. In this case, and far from the semi-automatic notion of abuse the Court supported in other cases, the ECJ considered that nothing in the documents supplied by the national court showed that the services supplied by AG2R did not meet the requirements of the undertakings concerned.58 This reluctance of the Court to apply the theory of the semi-automatic abuse it had previously supported can also be seen in the subsequent DEI case.59 The MOTOE case mentioned above also provides interesting insights into the reasoning of the ECJ on these points. First, regarding the relationship between exclusive rights or special rights and the creation of a dominant position, the Court held that an undertaking can be put in [a dominant position] when it is granted special or exclusive rights enabling it to determine whether and, as the case may be, in what conditions, other undertakings may have access to the relevant market and engage in their activities in that market.60

But in this case, once the dominant position of ELPA was established, the Court did not hesitate much in identifying the potential abuse it could commit. It considered that, to entrust a legal person such as ELPA, which itself organizes and commercially exploits motorcycling events, the task of giving the competent administration its consent to applications for authorisation to organize such events, is tantamount de facto to conferring upon it the power to designate the person authorised to organize those events and to set the conditions in which those events are organised, thereby placing that entity in an obvious advantage over its competitors. Such a right may therefore lead the undertaking which possesses it to deny other operators access to the relevant market.61

Therefore, this case should be directly included among the cases in which the infringement of Articles 106(1) and 102 TFEU derives from the existence of a conflict of interest at the undertaking to which the special rights are granted (RTT,62 ERT63 and Silvano Raso64). In the facts of this case, since ELPA was granted the right to authorise motorcycling events in Greece while at the same time being an organiser of events itself, the Court held that Articles 106(1) and 102 TFEU precluded a national rule which confers on a legal entity that organises motorcycle competitions (and enters, in that connection, into

Ibid, para 72. For further discussion of this issue, see Section 2.1 above. 60 MOTOE, above n 28, para 38. 61 Ibid, para 51. 62 RTT, above n 55. 63 ERT v Dimotiki, above n 53. 64 Case C-163/96 Silvano Raso [1998] ECR I-533. 58 59

298 José Luis Buendía Sierra s ponsorship, advertising and insurance contracts) the power to give consent to applications for authorisation to organise such competitions, without that power being subject to restrictions, obligations or review. Interestingly, the Court emphasised the lack of restrictions or controls on the actions of the undertaking—an element that was also mentioned in Albany.65 Whether, and to what extent, this element is a necessary condition for a finding of a breach of Article 106 TFEU in connection with Article 102 is not entirely clear. Neither RTT nor ERT required it, while Albany considered it to be a distinctive factor as regards RTT. Should this element be necessary, it is also unclear whether the necessary degree of ‘strictness’ or powers of review would be required. Another interesting case in which the Court was invited to rule on the combined application of Articles 106 and 102 TFEU was UPC.66 The case concerned a challenge by several cable TV operators operating in the region of Brussels-Capital relating to their legal obligation to transmit simultaneously and in their entirety television programmes broadcast by the public service broadcasters falling under the powers of the French community and the Flemish community. The claim on the basis of Articles 106(1) and 102 TFEU consisted of the challenge of the grant of must-carry status to certain private operators (arguably a special right), some of them having a dominant position in certain markets or sub-markets. After a brief recap of the basis of its case law on Articles 106 and 102 TFEU, the Court considered that, with the limited elements it held, it did not have all the necessary information it needed to determine whether the conditions relating to the existence of a dominant position or abusive conduct for the purposes of Article 102 were satisfied. This was particularly the case with regard to the determination of what the relevant markets were and in what way private broadcasters held an individual or collective dominant position within them.67 In this judgment, the Court is far from its practice of equating the existence of an exclusive right with the existence of a dominant position. Clearly, the Court could not (or did not want to) determine by itself the relevant markets in this case and the extent to which the undertakings holding special rights could individually or collectively exercise market power. In UPC, the Court was also very far from its theory of semi-automatic abuse. Of course, in this case the rights at stake were special rights and not exclusive rights, and the market power of the operators was not entirely clear.

65 Case C-67/96 Albany International BV v Stichting Bedrijfspensioenfonds Textielindustrie [1991] ECR I-5751, paras 116–121. 66 Case C-250/06 United Pan-Europe Communications Belgium SA and Others v Belgium [2007] ECR I-11135. 67 Ibid, para 21.

Enforcement of Article 106(1) TFEU 299 Neither was it entirely clear from the arguments of the parties what form the alleged abuse of the must-carry operators would take. Under these circumstances, it is possible to understand the Court’s position in refusing to make a ruling in the case.68

3.3 Combined application of Article 106(1) with Articles 49 and 56 TFEU As mentioned in the introduction to this section, Article 106(1) can also act as a point of entry for provisions of the Treaty other than those mentioned in Chapter I of Title VII of the TFEU (‘Rules on Competition’). This is particularly the case concerning the fundamental freedoms contained in Articles 49 and 56 TFEU. However, in many cases, national courts and/or the ECJ decide to take the ‘direct route’ and assess the provision directly under Articles 49 or 56 TFEU. A good example of this is the UPC case discussed above. In that case, the Court performed a separate assessment of whether the—arguably—special rights granted to the must-carry operators respected the fundamental freedoms, but included no mention whatsoever of Article 106(1) TFEU.69 In order to avoid the impression that I am giving Article 106(1) a greater role than the one assigned to it by the Court, I will limit myself to those few cases in which the combined application of Articles 106(1) and 49 or 56 TFEU is expressly mentioned by the ECJ.

68 For other cases where the Court refused to make the assessment, see Case C-451/03 Servizi Ausiliari Dottori Commercialisti Srl v Giuseppe Calafiori [2006] ECR I-2941, paras 21–26; Case C-295/05 Asociación Nacional de Empresas Forestales (Asemfo) v Transformación Agraria SA (Tragsa) and Administración del Estado [2007] ECR I-299, paras 41–45. 69 The Court considered that the restrictions established by the special rights could be justified, if proportionate, under the general interest aim of maintaining the pluralism of television programmes available in this country. Going back to other cases already mentioned, the Kattner case provides a good example in which an exclusive right—the one granted to MMB—was also considered by the Court under the rules on freedom to provide services as established in Article 49 TFEU. The Court assessed that the system may constitute a restriction on the fundamental freedom to provide services. However, it might well be justified for the overriding reason of ensuring the financial equilibrium of a branch of the social security. Concerning the proportionality of the measure, the Court did not give a final ruling—that was left to the national court—but it did indicate that two elements that definitely appeared to support the lawfulness of the measure were to be considered, namely: (i) that the statutory system offers minimal coverage, thus leaving room for supplementary insurance; and (ii) (also concerning the coverage—and given that the system is operated under the principle of solidarity) the need to include some activities in the coverage that would be more profitable than others.

300 José Luis Buendía Sierra The ANAV70 and Acoset71 cases are examples of this. In both cases, the question posed to the Court referred to national legislation that permitted the award of a public service concession in circumstances in which the existence of competition was disputed. The ANAV case concerned Italian legislation which set no limits on the freedom of the public authorities to decide to award the contract as a result of a public tender procedure or to directly award the contract to a company wholly owned by the authorities. In giving its response, the Court first pointed out that the principles of equal treatment and non-discrimination imply a duty of transparency on the authorities. They must ensure, for the benefit of any potential tenderer, a degree of advertising sufficient to enable the services concession to be opened up to competition and the impartiality of procurement procedures to be reviewed.72 Then the Court considered that theoretically a complete lack of any call for competition in the case of the award of a public service concession does not comply with the requirements of Articles [49 and 56 TFEU] any more than with the principles of equal treatment nondiscriminationand transparency. Furthermore, it follows from Article [106(1)] that the Member State must not maintain in force legislation which permits the award of public service concessions without their being put out to competition, since such award infringes Article [49 and 56 TFEU] or the principles of equal treatment, nondiscrimination and transparency.73

However, the Court applied the ‘in-house doctrine’ and stated that, nevertheless, in the field of public service concessions, the application of these principles is precluded if two cumulative conditions are met: (i) if the control exercised over the concessionaire by the concession granting authority is similar to that which the authority exercises over its own departments; and (ii) if that entity carries out the essential part of its activities with the controlling authority.74 On that basis, the Court assessed that the grant of the concession without a public tender to a company such as the one in question did not infringe Article 106(1), 49 or 56 TFEU.75

70 Case C-410/04 Associazione Nazionale Autotrasporto Viaggiatori (ANAV) v Comune di Bari and AMTAB Servizio SpA [2006] ECR I-3303. 71 Case C-196/08 Acoset SpA v Conferenza Sindaci e Presidenza Prov Reg ATO Idrico Ragusa and Others [2009] ECR I-9913. 72 See also Case C-458/03 Parking Brixen GmbH v Gemeinde Brixen and Stadtwerke Brixen AG [2005] ECR I-8585, para 49. 73 ANAV, above n 70, paras 22–23. See also Parking Brixen, ibid, paras 50 and 52. 74 ANAV, ibid, para 24; Parking Brixen, ibid, para 62. 75 ANAV, ibid, para 33.

Enforcement of Article 106(1) TFEU 301 An interesting remark should be made in this regard: the Court also considered that the initial intention (then disregarded) of the Italian authority was to open the capital of the wholly owned subsidiary to a private party. Since the initial intention did not materialise, the Court did not rule on its conformity with EU law. However, it pointed out that if, for the duration of the contract at issue … the capital of [the subsidiary] is open to private shareholders, the effect of that situation would be the award of a public services concession to semi-public company without any call for competition, which would interfere with the objectives pursued by [EU] law.76

In that scenario, the subsidiary would no longer be able to benefit from the in-house doctrine.77 The latter scenario (the opening of the capital of a wholly owned company that could initially benefit from the ‘in-house doctrine’) is closely related to what was analysed by the Court in the Acoset case. The question posed to the Court in this second case was: Is the model of a semi-public company formed specifically to provide a particular public service … to which that service is awarded directly, the private ‘industrial’ and ‘operational participant’ in the company being selected by means of a public and open procedure, after verification of the financial and technical requirements and of the operational and managerial requirements specific to the service to be performed and the specific services to be provided, consistent with [EU] law and in particular with the obligations of transparency and free competition referred to in Articles [49, 56 and 106 TFEU]?

Somehow, in contrast to the indications that the Court seemed to have given in ANAV, in the Acoset case it considered that Articles 106(1), 49 and 56 would not be infringed if the lack of a tender for the award of services in an institutionalised public–private partnership (IPPP) were ‘rectified’ by the selection of the private participant by means of a public tender.78 The reason that seems to justify this for the Court is that the use of a double tender in this case (first to select the partner in the IPPP, then to select the IPPP as the grantee of the concession) would deter private parties from entering into IPPPs because of the lengthy periods necessary for forming IPPPs and the uncertainty attached to the grant of the concession.79 At least, the Court established that the IPPP must retain the same corporate purpose throughout the duration of the concession and that a new tender process would be required if there were any material amendment to the contract.80

Ibid, para 30. Ibid, paras 31–32. Acoset, above n 71, para 63. 79 Ibid, para 61. 80 Ibid, para 62. 76 77 78

302 José Luis Buendía Sierra Other cases in which the issue was similar (if not identical) to the one mentioned above (and with an outcome that qualifies the case law discussed) were resolved by the Court without mentioning Article 106(1) TFEU (and without offering any sound rationale for ignoring it).81 In a partially similar context, one of the questions posed to the ECJ in the Wall case82 refers to whether an undertaking awarded a concession is bound, when subcontracting, by Article 106(1) and should respect the principle of transparency flowing from Articles 49 and 56 TFEU. The Court imposed two conditions for such undertaking to be considered a public authority for these purposes: (i) the undertaking in question must be controlled by the state; and (ii) it must not compete in the market. In the facts of the case, the ECJ disregarded that possibility.83 Another important case concerning the relevance of the existence of a tender in connection with Article 106 TFEU is ASM Brescia.84 The question raised in the case concerned the conformity of the extension of concessions for the distribution of natural gas, which were originally granted without any previous tender with Articles 49, 56 and 106(1) TFEU and the principles of equal treatment, non-discrimination and transparency. The Court noted that even if the grant of concessions falls outside the scope of the directives on public contracts, Member States are nevertheless bound by the fundamental rules of the Treaty and the general principles.85 In this case, the Court ruled that the deferment of the award of the new concession constitutes a difference in treatment to the detriment of the undertakings which might be interested in the concession and which are located in a Member State other than that to which the contracting authority belongs.86 Nevertheless, this difference in treatment could be justified by objective circumstances, such as the need to comply with the principle of legal certainty.87 This element was important in the case, as the contested concession was granted in 1984, that is, before the Court established its interpretation that the failure to tender for contracts with a cross-border interest could infringe the Treaty.88 Although the Court left the question of assessing the importance of the element up to the national court,89 it is clear that it would not object to any

81 See Parking Brixen, above n 72; Asemfo-Tragsa above n 68; Case C-573/07 Sea Srl v Comune di Ponte Nossa [2009] ECR I-8127. 82 Case C-91/08 Wall AG v La ville de Francfort-sur-le-Main and Frankfurter Entsorgungsund Service (FES) GmbH [2010] ECR I-2815. 83 Because the public stake in the undertaking was 51% (and a majority of three quarters of the votes was required for a resolution by the general shareholders’ meeting) and more than half of its turnover came from other contracts (see paras 49–57). 84 Case C-347/06 ASM Brescia SpA v Comune di Rodengo Saiano [2008] ECR I-5641. 85 Ibid, para 58. 86 Ibid, para 63. 87 Ibid, para 64. 88 Ibid, para 70. 89 Ibid, para 72.

Enforcement of Article 106(1) TFEU 303 of the possible solutions. This may contrast with the ECJ case law concerning retroactivity of new interpretations of EU primary law by the ECJ.90 The APERMC case is another interesting case concerning the need to tender for public contracts in which the Court applied Article 106(1) in combination with Articles 49 and 56 TFEU.91 Here the ECJ had to rule on whether (i) it is necessary to launch a tender for the selection of the operator entrusted with the universal postal service (and its attached reserved areas in order to ensure the equilibrium of the system) when the entity entrusted with the operation of the system is a wholly state-owned entity and whether (ii) in the non-reserved areas of the system, the public authorities could also use the same entity without the need for any tender process. Concerning the first question, the response provided by the Court was that EU law does not preclude legislation of a Member State allowing public authorities to entrust, without regard to the rules governing the award of public service contracts, the provision of postal services reserved, in a manner consistent with Directive 97/67, to a public limited company whose capital is wholly state-owned and which, in that state, is the provider of the universal postal service.92 The response given by the Court to the second issue was that, with regard to the non-reserved postal services, the authorities were obliged to apply the rules governing the award of public service contracts, as the two simultaneous conditions set out for the application of the ‘in-house doctrine’ were not fulfilled.93

3.4 Preliminary rulings on Article 106(2) TFEU justifying exclusive or special rights falling within the scope of Article 106(1) As is well known, Article 106(2) TFEU may be used to justify exclusive or special rights falling within the scope of Article 106(1) as long as those rights are needed for the exercise of an activity considered as a service of general economic interest (SGEI).94 Some recent preliminary rulings have provided further light in this respect. First, in the APERMC case mentioned above, the Court very clearly established, with respect to the postal sector, that services that do not fall within

90 See, eg Case T-289/03 British United Provident Association Ltd (BUPA), BUPA Insurance Ltd and BUPA Ireland Ltd v Commission of the European Communities [2008] ECR II-81, para 158. 91 APERMC, above n 26. 92 Ibid, para 41. 93 Ibid, para 83. 94 See, inter alia, Case C-340/99 TNT Traco SpA v Poste Italiane SpA and Others [2001] ECR I-4109.

304 José Luis Buendía Sierra the concept of reserved services in the sense of the Postal Directive95 cannot benefit from the protection of Article 106(2).96 Another case in which the relationship between Articles 106(1) and 106(2)— and their relationship with other provisions of the Treaty—was clarified was Servatius.97 The case derived from a national proceeding in which Servatius, a Dutch applicant, incorporated two subsidiaries in Belgium with the aim of developing a public housing project in the country. When it applied for authorisation from the Dutch authorities (as required by Dutch legislation), the minister refused it on the basis that the project was located in Belgium. Therefore, one of the points on which the referring national court asked for guidance was whether a prior authorisation scheme by the Dutch government for any cross-frontier activity of an undertaking in the public housing sector would be considered a special right falling within the meaning of Article 106(1) and whether this right could, nonetheless, be covered by the exception contained in Article 106(2) on the basis of the general economic interest the measure (the need for authorisation) aims to protect. The Court refused to rule on this issue, since it considered that the Dutch system did not serve the purpose of providing an SGEI.98 Moreover, the Court considered that the national proceeding did not concern the grant of a special or exclusive right to the foreign company which applied for (and was refused) the authorisation. In the Court’s view, the national procedure solely concerned the lawfulness of the restriction to which Servatius was subject, which took the form of the requirement to follow a prior administrative authorisation procedure.99 Finally, the Hanner case also contains interesting remarks on Article 106(2) TFEU.100 Strictly speaking, this is a case in which the provision is used to justify a possible infringement not of Article 106(1) but of Article 37 TFEU, which refers to state monopolies of a commercial character. The background of the case concerned a criminal proceeding in which the Swedish state monopoly for the retail sale of medicines was called into question. Here the Court followed the previous case law established in Franzén.101 It recalled that, although Article 31 does not require a total abolition of state monopolies of a commercial character, it requires them to be adjusted in such

95 Directive 97/67/EC of the European Parliament and of the Council of 15 December 1997 on common rules for the development of the internal market of Community postal services and the improvement of quality of service, [1998] OJ L15/14. 96 APERMC, above n 26, para 83. 97 Case C-567/07 Minister voor Wonen, Wijken en Integratie v Woningstichting Sint Servatius [2009] ECR I-9021. 98 Ibid, paras 45–47. 99 Ibid, para 46. 100 Case C-438/02 Krister Hanner [2005] ECR I-4551. 101 Case C-189/95 Harry Franzén [1997] ECR I-5909.

Enforcement of Article 106(1) TFEU 305 a way as to ensure that no discrimination regarding the conditions in which goods are procured and marketed exists—in law or in fact—between Member States. The Court then examined whether the specific conditions set out by it in Franzén were met in this case.102 As long as the Swedish monopoly did not fulfil one of the conditions (the existence of a purchasing plan or a system of call for tenders in which producers whose products were not selected would be allowed to be apprised of the reasons or contest the decision), the Court assessed that the way in which the Swedish monopoly was operated amounted to a violation of Article 37 TFEU. Several issues are interesting here: first, the Court made it very clear that, unless (actual or potentially) discriminatory elements were removed from the operation of the system, it could not benefit from the ‘protection’ of Article 106(2) TFEU. Secondly, there was an almost automatic application by the Court of the Franzén case law; the Court did not minimally assess whether in fact the monopoly had disadvantaged foreign products. The Opinion of Advocate General Léger in this case is remarkable, since it invited the Court to set aside the Franzén case law.103 The Advocate General pointed out that the previous judgment was based on a misinterpretation of Article 37 TFEU. He considered that the court should not take a ‘piecemeal approach’ when assessing state monopolies, referring its assessment to the rules for operation of the monopoly and whether those rules were discriminatory, and not assessing the monopoly as a whole. Such overall examination should take account of the restrictions on the free movement of goods created by the state measures. The Court, considered the Advocate General, should depart from a restrictive interpretation of the concept of discrimination that considers that Article 37 prohibits only discrimination against products from other Member States, since it should also consider discrimination between nationals of Member States regarding the conditions under which goods are procured and marketed. In other words, Article 37 is not only aimed at protecting the free movement of goods as such, it is primarily aimed at protecting traders who participate in that movement.104 I have always found this approach to be a sensible one. However, clearly the Court considered that both Mr Léger and I—and, incidentally, a number of cases contradicting Franzén105—were wrong: Franzén is—unfortunately—still good law.

Hanner, above n 100, paras 37–41. See Opinion of Advocate General Léger in Franzén, above n 101, paras 45 et seq. 104 Ibid, paras 56 et seq. 105 See cases cited in ibid, para 68. 102 103

306 José Luis Buendía Sierra

4. Conclusions The above review of recent Article 106 cases shows that the European Commission has only a limited interest in its direct enforcement and often avoids acting on complaints. Clearly, this has negative implications for a wider understanding of the provision. Given the nature of Article 106 as a provision setting the limits for Member States, it is clearly up to the European institutions to take the lead in their application. The Court of Justice more or less plays its part through the preliminary ruling mechanism. However, the European Commission, apart from a few salient decisions, is clearly avoiding its full responsibility as regards Article 106. Indeed, in this area of competition law, the ‘watchdog’ seems to be doing a lot more watching than barking, and very little biting. The DEI litigation may explain, at least in part, the Commission’s reluctance in recent years to rely on Article 106. Indeed, the General Court’s 2012 annulment of one of the very few Article 106 decisions that have been adopted—the Greek lignite decision—may have raised some doubts in that respect. However, the judgment of the Court of Justice in the DEI appeal,106 which quashed the General Court’s decision, should completely remove any such doubts. We now know that this provision offers tremendous possibilities for the Commission in the area of anticompetitive state measures. The Commission’s passive approach is made possible by the ECJ judgment on admissibility in max.mobil. Indeed, under the current case law, it is almost impossible for an individual to have locus standi to challenge the Commission’s refusal to act before the General Court.107 The Commission seems to view Article 106(1) as an area in which it has complete discretion to decide whether to act. This means that complaints are only pursued where the Commission considers not only that they are based on solid legal grounds, but also that it is opportune to act. This obviously means that potential complainants under Article 106 are in a far worse situation than in all other areas of EU competition law. The Commission’s passivity in Article 106 cases increasingly contrasts with its activism with respect to Article 107. This is striking since, from a functional point of view, special rights and financial aid are often interchangeable instruments, leading to similar distortions. This suggest that the Commission’s ‘light touch’ policy as regards Article 106 may be less acceptable in the future, particularly if protectionism continues to increase, money continues to be scarce and special rights becomes a popular surrogate for state aid.

See Commission v DEI, above n 21. max.mobil, above n 4.

106 107

Daniel A Crane* Hard Look Review of Anticompetitive State Action

In previous work,1 I examined two models of judicial review of anticompetitive state action that are broadly representative of the existing forms of judicial review. The first is a representation reinforcement model under which courts do not consider the substantive justifications for the anticompetitive state action but only insist that the decision to take the action was made in a way that ensures political accountability. The second is a substantive review model under which the state must justify its anticompetitive act as reasonably related to some important state interest. Aspects of these two models are presented in different standards used by legal systems to address state action that impairs free competition. The representation reinforcement model largely corresponds to the system currently utilized in the United States to determine whether state regulations are pre-empted by the federal antitrust laws. Aspects of a substantive review approach can be seen in European decisions concerning restrictions on the free movement of goods and services and the provisions of Article 106(2) of the TFEU concerning compliance with competition principles by statesponsored undertakings. Both of the primary models have significant shortcomings. The representation reinforcement model suffers from an important empirical deficit insofar as it relies on the assumption that consumers affected by anticompetitive state regulations will respond politically. Often, anticompetitive costs are externalized outside the jurisdiction where officials face political accountability. Even if they are internalized within the accountable jurisdiction, the harms to consumers are usually diffuse compared to the rents that concentrate on a few motivated producers, which creates asymmetries of political motivation between the consumer and producer classes. The substantive review model faces challenges of its own, mostly in the category of judicial legitimacy. Courts face a countermajoritarian difficulty when they undertake to decide which state interests should trump the interest in competition and which should not—a quintessential political question. American courts abandoned

* Associate Dean for Faculty and Research and Frederick Paul Furth, Sr, Professor of Law, University of Michigan School of Law. 1 Daniel A Crane, ‘Judicial Review of Anticompetitive State Action: Two Models in Comparative Perspective’, 1 Journal of Antitrust Enforcement 418 (2013).

308 Daniel A Crane this mode of analysis after a widely discredited experiment with economic ‘substantive due process’ during the early decades of the twentieth century. This chapter extends my earlier work by raising the possibility of a third model—hard look judicial review—positioned somewhere between the representation reinforcement and the substantive review models. The components of this hard look model are borrowed from a mode of judicial review utilized by American courts to scrutinize regulatory decisions by federal administrative agencies.2 Its features include a requirement of reasoned justification based on facts in a public record, consideration of alternatives to the chosen regulation, consideration of public comments, and contemporaneousness between an assertion of purpose and the promulgation of the challenged regulation.3 While not all of these features are easily transferable to every context in which courts scrutinize anticompetitive regulations for consistency with some meta-principle of free competition, the core insights of the hard look approach offer a promising middle ground between the relatively lax representation reinforcement and rather strict substantive review approaches. The remainder of this chapter proceeds as follows. Section 1 describes the fundamental problem of judicial review of anticompetitive state regulations. It then introduces the two leading models—representation reinforcement and substantive review—and briefly analyses their shortcomings. Section 2 sketches the key components of a hard look approach and contemplates its extension to judicial review of anticompetitive state regulations. Section 3 compares the three models. Section 4 briefly concludes.

1. The state action problem and leading responses 1.1 Conceptualising the problem State action problems involving judicial review of anticompetitive regulations arise in many different contexts. The set of problems considered in this chapter concern a paradigm with a few consistent features, including an 2 The term ‘hard look’ review was originally coined by US Circuit Court Judge Judge Harold Leventhal of the DC Circuit. See Pikes Peak Broad Co v FCC, 422 F2d 671, 682 (DC Cir 1969); Greater Boston Television Corp v FCC, 444 F2d 841, 851 (DC Cir 1970). See generally Matthew Stephenson, ‘A Costly Signaling Theory of “Hard Look” Judicial Review’, 58 Administrative Law Journal 753, 754, footnote 1 (2009); Cass Sunstein,’Constitutionalism after the New Deal’, 101 Harvard Law Review 421, 463–474 (1987). 3 See Kevin Stack, ‘Interpreting Regulations’, 111 Michigan Law Review 355, 378–379 (2012); Thomas Miles and Cass Sunstein, ‘The Real World of Arbitrariness Review’, 75 University of Chicago Law Review 761, 761–762 (2008); Kevin Stack, ‘The Constitutional Foundations of Chenery’, 116 Yale Law Journal 952, 972 (2007); Merrick Garland, ’Deregulation and Judicial Review’, 98 Harvard Law Review 505, 509 (1985).

Hard Look Review of Anticompetitive State Action 309 enforceable meta-principle supporting competition, state action that substantially reduces the competitiveness of one or more markets, and jurisdiction for courts to review the conformity of the state action with the pro-competition meta-principle. Let us briefly consider each of these elements in turn. The first element is some binding legal meta-principle favoring competition that is juridically capable of pre-empting or overriding the challenged state regulation. This element implies a hierarchical relationship between two legal competences, as for example the relationship between a constitution and an ordinary statute or between a treaty and domestic regulation. Practically speaking, the two most salient examples are: (i) the relationship between federal statutory law and state statutes of regulations in the United States where duly enacted federal law pre-empts inconsistent state laws pursuant to the Constitution’s Supremacy Clause; and (ii) the relationship between Member State domestic laws or regulations and the binding provisions of the TFEU. In both cases, subordinate regulatory decisions that conflict with a pro-competitive meta-principle (such as federal antitrust law in the US and Articles 101, 102, and 106 in the TFEU) may be pre-empted. The second element involves some legislative or regulatory intervention by a subordinate legal body (subordinate in the sense that it cannot lawfully override the meta-principle) that substantially impairs competition. ‘Substantially’ is an important qualification because many statutes and regulations have the effect of disadvantaging or discouraging some potential competitors without impairing the overall competitiveness of the market. For example, a regulation requiring an advanced education degree for primary school teachers may reduce the number of applicants for teaching jobs but, so long as the advanced degrees are widely available and the pool of applicants large compared to the number of available spots, it would seem unnecessary to subject the degree requirement to any sort of competition law scrutiny. The kinds of restraints on competition addressed here are those that have serious anticompetitive consequences. Examples include the establishment of monopoly franchises, regulatory schemes that enable joint pricing or market allocation, and restrictions that limit the entry of new firms or suppliers. The final element is jurisdiction for a court to consider whether the subordinate regulation or legislation is consistent with the pro-competitive metaprinciple, as is true of federal courts and state regulations in the US and European courts and Member State regulations in the EU. The paradigm under consideration excludes evaluation of an inferior body by a superior body within an administrative hierarchy, as might happen, for example, under the Chinese Anti-Monopoly Law.4 Although such circumstances may often 4 For discussion, see Mel Marquis, ‘Abuse of Administrative Power to Restrict Competition in China: Four Reflections, Two Ideas and a Thought’, in Michael Faure and Xinzhu Zhang, eds, Regulation and Competition Policy in China: New Developments and Empirical Evidence, Edward Elgar, 2013, 73 et seq.

310 Daniel A Crane involve similar questions, the paradigm in question assumes the interaction of courts and legislative or administrative bodies within a generally democratic framework.5

1.2 Representation reinforcement 1.2.1 The model The representation reinforcement model finds expression in the American state action doctrine or Parker v Brown6 doctrine. In Parker, the Supreme Court rejected both negative commerce clause and Sherman Act challenges to a California agricultural proration scheme that sharply limited the volume of raisins produced and marketed in the state. The Court found ‘no suggestion of a purpose to restrain state action in the [Sherman] Act’s legislative history’.7 Later cases fleshed out the scope of Parker immunity for state action. In Midcal,8 the Court announced a two-part test to determine whether Parker immunity applies. Under the first prong, the ‘challenged restraint must be one clearly articulated and affirmatively expressed as state policy’.9 Under the second prong, the state itself must ‘actively supervise’ the implementation of the competition-displacing policy.10 Parker immunity is conceptually predicated on a political accountability model of regulatory design.11 The basic idea is that, if a government wants to displace the usual rules of competition with an alternative scheme, it must do so obviously and conspicuously (clear articulation and affirmative expression) and it must take full responsibility for the way in which the alternative policy is implemented (active supervision). In this way, the government itself will become politically responsible for the displacement of competition. If things go poorly, the electorate can trace the decision back to the responsible government and punish it appropriately. This political accountability approach to anticompetitive state action reflects one of the most popular constitutional theories of the last half century: John Hart Ely’s theory of representation reinforcement. In Democracy

5 See Crane, ‘Judicial Review of Anticompetitive State Action’, above n 1, 3–7 (discussing application of judicial review models to administrative hierarchies). 6 317 US 341 (1943). 7 Ibid, 351. 8 California Retail Liquor Dealers Ass’n v Midcal Aluminum, Inc, 445 US 97 (1980). 9 Ibid, 105 (citations omitted). 10 Ibid. 11 See, eg William Page and John Lopatka, ‘State Regulation in the Shadow of Antitrust: FTC v Ticor Title Insurance Co’, 3 Supreme Court Economic Review 189, 229 (1993).

Hard Look Review of Anticompetitive State Action 311 and Distrust,12 Ely advanced the argument that the courts should not override the will of the majority in service of values higher than majority rule. Rather, the courts should strive to make democracy work better—to break down impediments to meaningful majority and minority participation in the political process and hence promote the legitimacy of the political system. The representation reinforcement model provides insights into the implicit theory of Parker immunity. Parker was decided in an era in which the courts were retreating from their project in the early twentieth century of invalidating socio-economic legislation (both state and federal) under substantive due process principles, which required courts to decide substantively which economic policies were legitimate and which were not.13 When the Supreme Court began to retreat from economic substantive due process in the late 1930s, it asserted that courts could not legitimately decide questions of socioeconomic policy and that those questions should be left to legislators.14 Having made this turn away from judicial creation of economic policy under the guise of the due process clause, the court was hardly prepared to take on a similar role under the antitrust laws. But this did not mean that the courts would abdicate on state economic policies altogether. Rather than substantively second-guessing state economic policies, the courts would inquire into the political legitimacy of the processes that produced the anticompetitive scheme. Hence, the Court has not hesitated to strike down anticompetitive state policies that fail the Midcal test— ones where the state does not take clear political responsibility for the scheme.15 In this way, the Parker doctrine ostensibly strengthens the political processes through which economic regulation flows.

1.2.2 The model’s shortcomings The representation reinforcement model is premised on the assumption that when governments are forced to take visible responsibility for their r egulatory

12 John Hart Ely, Democracy and Distrust: A Theory of Judicial Review, Harvard University Press, 1980. 13 Paul Verkuil, ‘State Action, Due Process, and Antitrust: Reflections on Parker v Brown’, 75 Columbia Law Review 328, 330–341 (1975). 14 See, eg United States v Carolene Prods Co, 304 US 144, 152 (1938) (‘[T]he existence of facts supporting the legislative judgment is to be presumed, for regulatory legislation affecting ordinary commercial transactions is not to be pronounced unconstitutional unless in the light of the facts made known or generally assumed it is of such a character as to preclude the assumption that it rests upon some rational basis within the knowledge and experience of the legislators’); Ferguson v Skrupa, 372 US 726, 730 (1963) (examining historical rejection of economic substantive due process and explaining that ‘courts do not substitute their social and economic beliefs for the judgment of legislative bodies, who are elected to pass laws’). 15 Daniel A Crane, The Institutional Structure of Antitrust Enforcement, Oxford University Press, 2011.

312 Daniel A Crane decisions, the democratic process will hold them accountable when the costs of displacing competition with regulation exceed the benefits.16 But this political accountability story does not work as to anticompetitive regulations whose benefits are captured mostly by local producers and costs externalized to consumers who cannot vote in the jurisdiction that imposed the anticompetitive regulation. A striking example of externalization of monopoly costs can be seen in Parker, the case that created the modern doctrine of antitrust immunity for state regulations.17 At the time of Parker, half of the world’s raisins and 95 percent of raisins sold in the United States came from California.18 Further, more than 90 percent of the raisins grown in California were shipped outside of the state.19 Hence, California raisin producers had significant market power and were able to export most of the costs of their cartel to consumers who had no say in California politics. Nonetheless, the Supreme Court held that the California scheme was categorically immunized from antitrust challenge. Perversely, Parker, Midcal and their progeny create an even worse problem than the ordinary export cartel.20 For a state regulatory scheme to qualify for Parker immunity, it must be affirmatively expressed as state policy and actively supervised by agents of the state. This means that it is usually not enough for a state to passively allow anticompetitive conduct to take place. It must require it—or at least come very close to requiring it21—and then actively monitor it. The upshot is that states often play the role of cartel ringmaster, thus solving the cartel’s perennial problem of preventing defection. Thus, unlike mere private market cartels, which may eventually collapse under their own weight, state-run cartels are backed by the coercive power of the state and hence are stable and perniciously effective.

16 Robert Inman and Daniel Rubinfeld, ‘Making Sense of the State Action Doctrine: Balancing Political Participation and Economic Efficiency in Regulatory Federalism’, 75 Texas Law Review 1203, 1207 (1997). Inman and Rubinfeld would ask: ‘1) Does a state regulation generate significant monopoly spillovers onto nonresidents?; and 2) Was the state regulation decided without political participation of the affected nonresidents as evidenced by the lack of interstate regulatory agreement? If the answer to both questions is yes, then the state regulation fails the spillover test for economic efficiency, and a Sherman Act review of the regulation is appropriate.’ Ibid. 17 Parker, above n 6. 18 Ibid, 344. 19 Ibid. 20 See Frank Easterbrook, ‘Antitrust and the Economics of Federalism’, in Richard Epstein and Michael Greve, Competition Laws in Conflict: Antitrust Jurisdiction in the Global Economy, AEI Press, 2004, 189 et seq. 21 A state regulatory policy can qualify for Parker immunity even if the state does not compel market actors to adhere to the regulatory scheme. Southern Motor Carriers Rate Conf, Inc v US, 471 US 48, 61 (1985). Nevertheless, state compulsion is often considered the ‘best evidence’ that the policy of displacing competition is clearly articulated and affirmatively expressed as state policy. Ibid.

Hard Look Review of Anticompetitive State Action 313 Even when the effects of an anticompetitive regulatory scheme are internalized within the regulating jurisdiction, the assumption that adversely affected consumers will pressure the responsible official to rethink the anticompetitive policy often does not hold. A body of public-choice theory suggests that anticompetitive schemes may be the product of regulatory capture by powerful industries and the diffusion of costs over a wide body of consumers who do not have a sufficient individual interest to mobilize politically and resist.22 For example, imagine a regulatory scheme that severely restricts competition in the funeral casket business.23 The costs of the regulatory scheme—in the form of increased casket prices—are spread over the entire population. The benefits are concentrated on a handful of funeral homes that enjoy a regulatory oligopoly position. No individual consumer has much of an interest in mobilizing to defeat the regulatory scheme (indeed, most consumers will be unaware of its existence). By contrast, the funeral homes have a very strong interest in mobilizing to support the policy’s continuation. Hence, many policies will remain in force despite imposing large costs on consumers as a class.

1.3 Substantive review 1.3.1 The model In contrast to the representation reinforcement model, the substantive review model requires courts to balance the public interest in competition against the regulating entity’s interest in pursuing a conflicting regulatory objective. State regulations that restrict competition sometimes undergo substantive review by courts in the EU, although not necessarily under competition law principles. Under current EU law, the courts and European Commission are at least sometimes directed to answer the questions that Parker and its progeny have directed courts to eschew: whether the state action is justified by substantive policy considerations. Some of these cases arise under Article 56 TFEU (ex-Article 49 EC), which prohibits unjustified restrictions on the provision of cross-border services.24 Others, like the Slovakian hybrid mail case discussed below, arise under Articles 106(2) and 102, which directly cover competition law.

22 See, eg Merrick Garland, ‘Antitrust and State Action: Economic Efficiency and the Political Process’, 96 Yale Law Journal 486 (1987); John Shepherd Wiley Jr, ‘A Capture Theory of Antitrust Federalism’, 99 Harvard Law Review 713 (1986). 23 A real-world example is discussed in Crane, ‘Judicial Review of Anticompetitive State Action’, above n 1, 426–428. 24 See, eg Joined Cases C-94 and 202/04 Cipolla and Macrino [2006] ECR I-11421 (rejecting a challenge under Article 10 EC and Article 81 to the Italian statutory fee scale for

314 Daniel A Crane An example of a case decided under Article 56 TFEU is the Italian legal fee regulation cases,25 where the ECJ explained that restrictions on competition that produce anticompetitive effects may sometimes be justified given a sufficiently important state interest and a reasonable relationship between the restriction and the state’s important goal: In that respect, it must be pointed out that, first, the protection of consumers, in particular recipients of the legal services provided by persons concerned in the administration of justice and, secondly, the safeguarding of the proper administration of justice, are objectives to be included among those which may be regarded as overriding requirements relating to the public interest capable of justifying a restriction on freedom to provide services … on condition, first, that the national measure at issue in the main proceedings is suitable for securing the attainment of the objective pursued and, secondly, it does not go beyond what is necessary in order to attain that objective.26

The Slovakian hybrid mail case27 involved competition law principles and hence a different set of doctrinal and analytical questions but, at its core, a comparable inquiry into the substantive justifications for the state’s restrictions of competition. The European Commission held that the Slovak Republic’s extension of its postal monopoly to hybrid mail (mail electronically transmitted to a service provider and printed, put into an envelope and delivered by the provider) conflicted with Article 106(2)’s requirement that Member States comply with EU competition rules—in this case, Article 102’s prohibition on abuse of a dominant position. The Commission first found that Slovenská Pošta held a dominant position within the meaning of Article 102.28 It then observed that the Commission’s 1998 Postal Notice stressed that ‘the use, without objective justification, of a dominant position in one market to obtain market power on related or neighboring markets which are distinct from the former or at the risk of eliminating competition on those markets’ would be contrary to Article 106 in conjunction with Article 102.29 The Commission further noted that this would only be the case ‘in the absence of specific justification, if the functioning of services in the general economic

lawyers but upholding the same challenge under free movement principles); Case C‑212/08 Zeturf Ltd v Premier Ministre [2011] ECR I-5633 (holding under Article 49 EC that ‘[n]ational legislation [creating gambling restrictions] is appropriate for ensuring attainment of the objective pursued—combating criminal and fraudulent activities and protecting society—only if it genuinely reflects a concern to attain it in a consistent and systematic manner’). Cipolla and Macrino, ibid. Ibid, para 64. 27 Commission Decision of 7 October 2008 in Case COMP/39.562—Slovakian postal legislation relating to hybrid mail services, C(2008) 5912. 28 Ibid, para 113. 29 Ibid, para 114. 25 26

Hard Look Review of Anticompetitive State Action 315 interest was not previously endangered’.30 Having found that Slovenská Pošta had used its dominant position in one market to obtain a dominant position in another, it then placed the burden of justifying the state-sponsored restraint on competition on Slovenská Pošta. Slovakia argued, among other things, that [Slovenská Pošta’s] profits would be reduced … to such an extent that they would not cover the costs of providing the postal universal service and that competition would impact negatively its revenues due to the diversion of large volumes of bulk mail items to hybrid mail.31

The Commission considered this assertion at length, and found that Slovakia had failed to substantiate its claim that allowing private hybrid mail delivery would undermine the state’s ability to provide universal postal service. In other words, the Commission engaged in a substantive review of the anticompetitive regulatory scheme, ascertaining whether it was necessary to advance the state’s legitimate interests. This sort of substantive review of anticompetitive state action resonates with US constitutional doctrines that require state action that burdens some freedom or interest (such as free speech or gender equality) to meet a two-part test: (i) that the state action serves some compelling or at least important state interest; and (ii) that the state action is narrowly tailored to meet the state’s interest.32 Such tests can be phrased in varying degrees of strictness (ie ‘compelling’ or merely ‘important’) reflecting the degree of prima facie hostility the judiciary will manifest to the state action. In the United States today, such tests are generally reserved for non-economic questions involving personal freedom, equality or social standing.

1.3.2 The model’s shortcomings The substantive review model has its own serious challenges, which generally fall under the heading of what Yale law professor Alexander Bickel called ‘countermajoritarian difficulties’.33 When a court invalidates subordinate legislation on the theory that it violates a pro-competition meta-principle, it is trumping the democratically determined public will based on the opinions of unelected judges. The challenges of political legitimacy facing judges

Ibid. Ibid, para 161. 32 The EU’s ends–means test is called a ‘proportionality’ test, consistent with the general principle of proportionality in EU jurisprudence. See, eg Damien Chalmers, Gareth Davies and Giorgio Monti, European Union Law: Cases and Materials, 2nd edn, Cambridge University Press, 2010, 368–369. 33 Alexander Bickel, The Least Dangerous Branch, 2nd edn, Yale University Press, 1986, 16–23. 30 31

316 Daniel A Crane e mploying a substantive review approach fall analytically into two categories, consistent with the two-part test generally applicable to state actions burdening protected individual rights. The first is to decide what state interests are sufficiently important to justify displacement of baseline competition principles. The second is to decide when the state regulatory scheme is sufficiently tailored to meet the state’s important objectives. The first prong in a substantive review test usually requires an inquiry into the legitimacy and magnitude of the state interest at issue. In the Italian legal fee regulation case, the ECJ acknowledged that safeguarding consumers against sharp practices by lawyers and preserving the integrity of the justice system are ‘overriding requirements relating to the public interest’.34 This formulation of the public interest component of substantive review implicitly contains two subcomponents—legitimacy and magnitude. First, the state interest must promote public, as opposed to private, welfare. Second, the state interest must be an important one in relation to society’s interest in free competition. Both of these subcomponents have their difficulties. First, the distinction between the public interest and the private interest is slippery at best.35 Read narrowly, the public interest could be understood as just the interest of the state in the administration of governmental programs (for example, integrity in public works programs). But that understanding is certainly too narrow to comprehend the usage of the public interest requirement in EU law. Recall that, in Italian legal free regulation, the relevant public interests included the administration of the justice system—which would meet the ‘public’ requirement in the narrow sense just articulated—but also the interests of clients in not being fleeced by lawyers, which could happen inside or outside of the public adjudication system. States frequently displace competition with regulation in order to protect the interests of ‘private’ constituencies such as consumers or producers. If the public interest is not simply the interests of the government in administering its own programs, then what is it? Given the usual assumption that competition laws are primarily oriented toward the promotion of consumer welfare, one might posit that the ‘public interest’ in the state action context simply means the interests of consumers as opposed to the interests of producers. But that approach may not be terribly appealing normatively because it would require disregarding the interests of producers, no matter how large in comparison to the interests of consumers. For example, it would prevent a state from deciding that a particular group of producers would benefit considerably from a regulatory scheme that made competition somewhat less intense and only caused a slight loss of consumer welfare. Equating the

Cipolla and Macrino, above n 25. See, eg Alexander Meiklejohn, Political Freedom: The Constitutional Powers of the People, Harper & Brothers, 1960, 80–81. 34 35

Hard Look Review of Anticompetitive State Action 317 ‘public interest’ with consumer welfare would prohibit states from following a total social welfare maximization approach to regulatory decision-making and require them to focus narrowly on consumer welfare in cases where the regulation might impair the competitiveness of the market. Alternatively, one could read the ‘public interest’ requirement as a rule that state actors are not allowed to bestow privileges on narrow special interest groups at the expense of the general population. But such a rule would be difficult to implement because there is no clear dividing line between ‘special interest groups’ and other constituencies deserving of state favor. Special interest group is little more than an epithet that one applies to political adversaries. Consider, for example, the Italian matches case,36 in which Italian law required Italian match producers to join a consortium and to set sellers’ quotas—a case closely reminiscent of Parker v Brown.37 Are match makers a special interest group? If so, then isn’t every neighborhood association that lobbies against a polluting factory, every union that seeks a more favorable labor law and every group of inventors that seeks to strengthen the patent laws a special interest group as well?38 As James Madison famously explained in ‘The Federalist No. 10’,39 self-interested factions are an inherent feature of democracy and governmental structures should seek to prevent any one faction from growing too powerful rather than seeking to eliminate factional intrigue. Despite these difficulties, the public interest or legitimacy criterion has at least the potential to be intelligible. By contrast, the magnitude or ‘importance’ criterion does not—at least, not as a judicial criterion. The problem is that judges have no comparative advantage over legislators or regulators to decide what interests are important. Further, making the lawfulness of a regulation turn on the judges’ views on the comparative importance of different state interests seriously threatens the legitimacy of the judicial process. Of course, judges make value judgments all of the time, but such judgments become highly sensitive when the judge is substituting her valuations for those of state actors in the legislative or executive—and hence more democratically accountable—branches. And although modern constitutional d octrine

Case C-198/01 Consorzio Industrie Fiammiferi [2003] ECR I-8055. See Eleanor Fox, ‘State Action in Comparative Context: What if Parker v Brown Were Italian?’, in Barry Hawk, ed, International Antitrust Law & Policy: Fordham Corporate Law Institute 2003, Juris Publishing, 2004, chapter 19, 473. 38 The historical evidence suggests that the bakeshop act invalidated by the US Supreme Court in Lochner v New York, 198 US 45 (1905), which is considered the peak of the Court’s now-rejected substantive due process jurisprudence, was the product of special interest agitation. See David Bernstein, Rehabilitating Lochner: Defending Individual Rights Against Progressive Reform, University of Chicago Press, 2011, 3. 39 Publius [James Madison], ‘The Federalist No. 10: The Utility of the Union as a Safeguard Against Domestic Faction and Insurrection (Continued)’, Daily Advertiser, 22 November 1787. 36 37

318 Daniel A Crane requires judges to determine the importance of asserted state interests in a variety of contexts, such as racial or gender classifications and abridgements of free speech, determining the importance of the state’s interests in economic regulation is a very different matter. In the personal rights context, courts are not so much deciding what interests are important in the abstract as they are deciding whether the state’s asserted interests are inconsistent with a moral principle embodied in a constitutional limitation. For example, when gender classifications are at issue, the courts examine the state’s proffered justifications to determine whether they are inconsistent with the constitutional principle that state classifications should not be predicated on gender stereotypes or the subordination of women.40 Proffered state interests are usually not rejected because they are insufficiently weighty but because they are pretextual or categorically illegitimate. There is no comparable categorical illegitimacy move available in competition law. Competition law is not premised on moral considerations about the dignity or worth of individuals and usually does not contain categorical prohibitions on the kinds of assumptions states may make or the values they may seek to promote. To the extent that competition law does contain categorical prohibitions, they are decisional rules for streamlining litigation, not constraints on the range of options that states can consider. For example, the rule of per se illegality for price fixing and other naked horizontal restraints prohibits parties from arguing in defense of a restraint that competition was ‘ruinous’ and therefore needed to be curtailed. But this rule does not rest on a legislative or judicial judgment that competition could never be ruinous and in need of curtailment. Rather, it is a probabilistic rule designed to minimize litigation costs and errors. Given that competition is usually beneficial to society and that parties who want to suppress competition will invariably argue that competition was deleterious, it is better to make a categorical judgment for purposes of antitrust litigation that competition is virtuous. Nothing in this reasoning, however, would prohibit a state from limiting competition in order to promote other social or economic values. Asking courts to determine the strength of the state’s interest in curtailing competition would be a judicially unmanageable task, requiring courts to engage in a sort of open-ended cost–benefit analysis that would sometimes result in the judges substituting their view of what interests are important for those in the executive or legislative branches. The second step in a substantive review model is to determine whether the restriction on competition is a reasonable means to pursue the state’s

40 See, eg United States v Virginia, 518 US 515, 533 (1996) (explaining that state interests in gender classification context ‘must be genuine, not hypothesized or invented post hoc in response to litigation. And it must not rely on overbroad generalizations about the different talents, capacities, or preferences of males and females’).

Hard Look Review of Anticompetitive State Action 319 regulatory goals—the fit between means and ends. The central challenge with any means/ends test is identifying the required tightness of the fit. In the state action context, the ECJ has expressed the means/ends criterion as requiring that the restraint on competition ‘is suitable for securing the attainment of the objective pursued and … does not go beyond what is necessary in order to attain that objective’.41 This formulation suggests two components to the means/ends evaluation stage. First, there is a question of alternatives— was there a different means of achieving the state’s regulatory goals without harming cross-border competition? Second, there is a question of breadth— even if the means chosen was reasonable, did it contain competition-harming features beyond the necessary scope of the chosen means? Both of these components could prove tricky for judges to address. On the alternatives question, it is almost always the case that the state has alternative means of achieving its ends that do not involve an equally obvious impairment of competition. In the Slovak hybrid mail case, for example, S lovakia could have allowed hybrid mail delivery and made up for the lost revenue streams through other devices, such as raising postage rates or general taxation. But although those other devices might not impair competition to the same degree, they might not be politically feasible or might impair other socio-economic objectives of the mail system, such as progressive wealth transfers or subsidization of rural areas. Judges are in a very poor position to make decisions about these kinds of trade-offs, given their systemic effects. Similarly, the scope question will often be difficult to answer in the state action context. Many statutory or regulatory schemes are explicable only as compromises between different interest groups. Hence, statutes and regulations are often populated with quirks—carve-outs, exemptions, special grants—that seem extraneous to the central workings of the statutory or regulatory scheme. But even if they look like chaff during post hoc judicial inspection, they are often essential points of compromise that secure the political capital necessary for the passage of the bill or regulation. Judicial invalidation of restrictions on competition as being overly broad for the function of a statutory or regulatory scheme may excise politically indispensable features.

2. The hard look alternative 2.1 Background and context Although often labeled a ‘doctrine’, hard look review is less a principle of US administrative law than a description of the scrutiny that courts sometimes Cipolla and Macrino, above n 25.

41

320 Daniel A Crane impose on administrative rule-making and other agency action. Certain features or sources of authority for hard look review can be justified as statutory, deriving either from an agency’s organic statute or from the general requirements of the Administrative Procedure Act (APA). Others derive from a common law process of judicial engagement with regulatory decision-making. The precise boundaries and components of hard look review are contestable, although there is a wide consensus as to the core elements of the hard look approach. The spirit of hard look review is to insist that agencies engage in thorough, reasoned decision-making without involving the courts in replacing the agencies’ technical and political judgments with their own. Judge Leventhal, one of the hard look doctrine’s most distinguished expositors, described the doctrine as combining ‘judicial supervision with a salutary principle of judicial restraint’.42 As discussed in greater detail in the following section, hard look review involves a high degree of judicial engagement with regulatory decisions but also a more restrained role for judicial review than characterizes the sort of substantive review balancing described previously. The expression ‘hard look’ describes two different features of the interaction between courts and agencies. First, the doctrine requires agencies to take a ‘hard look’ at the regulatory question, considering all relevant data, alternatives and public comments. Second, courts take their own ‘hard look’ at the regulatory decision to ensure that the agency’s decision was sufficiently deliberate. The result is an iterative interaction between courts and regulators in which courts can halt the enforcement of an agency decision until the regulator has taken a harder look at the problem, or can ultimately vacate the regulatory decision altogether.

2.2 Primary features of hard look review 2.2.1 Reasoned decision-making on public record The most general and important feature of hard look review is judicial insistence that agencies explain their regulatory decisions in a thorough and reasoned fashion, with reference to factual predicates available in a public record. Courts invalidate agency action when the agency fails to explain itself through logically consistent reasoning and with reference to public facts can be verified for accuracy. Thus, for example, if an agency offers two reasons for a regulation—either one of which might be a valid basis for action

42 Harold Leventhal, ‘Environmental Decisionmaking and the Role of the Courts’, 122 University of Pennsylvania Law Review 509, 511 (1974).

Hard Look Review of Anticompetitive State Action 321 individually—and the two offered reasons contradict each other, a court may invalidate the agency action on the grounds of internal inconsistency in the explanation. Or, if agency action assumes undisclosed factual predicates, a court may require the agency to disclose the data or facts underlying its decision before the regulation can become operative. The requirement of reasoned decision-making on a public record can be justified as a mechanism for insuring that agencies act for public-interested reasons and not merely to patronize narrow special interests. When an agency’s true motivation is economic parochialism, the requirement that it justify its actions on public interest grounds may force the agency to come up with implausible, contradictory or empirically unsupported justifications, in which event a court may invalidate the agency action.43 Application of such a reasoned decision/public record requirement to instances of state action burdening competition could be valuable. At a high level of generality, the state action problem is about the difficulty in determining when a regulation that limits competition is necessary to achieve the interests of the general public and when it merely represents regulatory capture and rent-seeking behavior by the regulated entities. Hard look review serves as an information-forcing device, requiring the regulating entity to reveal, if not its actual cards, at least a hand capable of rationally supporting the regulatory decision.44 Since it would usually be politically unpalatable for agencies to justify anticompetitive regulations purely as bestowing economic rents on favored groups of producers, the reasoned decision/public record requirement could erect a barrier to anticompetitive state action that does not have some plausible public interest explanation. To be sure, the reasoned decision/public record requirement would not eliminate all instances of competitive suppression due to economic parochialism. An agency might justify its decision on plausible publicly minded grounds and survive hard look review even though the true purpose and effect of the regulation was parochial. Indeed, if the agency offered an explicitly parochial justification—ie this regulation is intended to benefit a particular class of particularly deserving producers—that would not fail hard look review unless the

43 Kathryn Watts, ‘Proposing a Place for Politics in Arbitrary and Capricious Review’, 119 Yale Law Journal 2, 34–35 (2009) (‘Hard look review was one of the main tools that the courts developed to ensure that agencies were looking at the statute and the evidence and were choosing answers that served the public good’ rather than special interests’); Richard Stewart, ‘The Reformation of American Administrative Law’, 88 Harvard Law Review 1667, 1758 (1975) (‘[T]he requirement that agencies give adequate consideration to all affected interests, and in particular, the interests of the intended beneficiaries of an administrative scheme, has been utilized by the courts with increasing frequency to redress perceived agency favoritism to organized interests’). 44 Stephenson, above n 2, 755 (discussing hard look review as an information-forcing device).

322 Daniel A Crane statute creating the regulator’s jurisdiction excluded the agency for acting for such reasons. Hard look review functions not by eliminating certain regulatory purposes as illegitimate or asking courts to balance competing interests, but by forcing the regulator to produce more information about its decision. The production of this information can have the salutary effect of deterring egregious special interest regulation in some cases.

2.2.2 Explicit consideration of alternatives A second facet of hard look review is a requirement that agencies give explicit consideration to alternatives to the regulatory decision taken.45 The consideration of alternatives requirement serves some of the same purposes as the narrow tailoring or ‘least restrictive alternative’ analysis employed in substantive review, but with the important difference that the reviewing court does not strike down a regulation just because it fails to achieve the regulatory objective in the most efficient or least restrictive way. Rather, the alternatives requirement merely forces the agency to explain why obvious alternatives (including those proposed through public comments, as discussed next) were inferior to the chosen solution. This ensures that the regulatory authority is seriously considering alternative solutions and that its reasoned justification for the chosen solution can be evaluated in comparison with other options. Analytical or factual flaws in the agency’s explanation for the rejection of alternatives can themselves lead to judicial invalidation of the agency’s action. In the competition context, application of the consideration of alternatives requirement could prompt regulatory authorities to consider regulatory approaches that created fewer barriers to competition. In particular, where a regulatory authority substitutes centralized planning for market-based determinations of production and distribution, courts could ensure that the regulator articulates reasons why market-based solutions were inadequate to meet the regulatory objective.46 This, in turn, would require the regulator to explain not merely the market failures that prompted the regulatory decision, but also why those failures could not be corrected through less intrusive regulatory actions.

45 The leading case on the consideration of alternatives requirement is Motor Vehicle Mfrs Ass’n v State Farm Mut Auto Ins Co, 463 US 29, 43, 50 (1983), in which the US Supreme Court required the Department of Transportation to consider technological alternatives to its proposed automobile safety regulations concerning seatbelts and airbags. 46 For example, US Presidential Executive Order 12866, issued in 1993 by President Clinton and continuing in force today, requires agencies to ‘identify and assess available alternatives to direct regulation, including providing economic incentives to encourage the desired behavior, such as user fees or marketable permits, or providing information upon which choices can be made by the public’, http://www.whitehouse.gov/sites/default/files/ omb/inforeg/eo12866/eo12866_10041993.pdf.

Hard Look Review of Anticompetitive State Action 323

2.2.3 Explicit considerations of public comments A third feature of hard look review—one closely tied to statutory features of American administrative law concerning citizen participation in the regulatory process—is the requirement that agencies explicitly consider public comments made during the regulatory process.47 This element is particularly important in the context of ‘notice and comment’ rule-making, where the agency is required to place its proposed rule in the public record with sufficient information to allow critical response by the public before the rule is promulgated. While the agency is not required to respond to every comment, it should respond to significant comments—as determined ex post by judicial review.48 Because it is sometimes hard to anticipate which comments a court will subsequently determine to be significant and how much of a response the courts will require for each significant comment, this requirement can incentivize regulatory authorities to respond broadly to public comments in the context of framing rules. As is true with respect to other features of hard look review, this requirement has been criticized for increasing the burdens and costs of regulating and hence ossifying existing regulation.49 It can also create opportunities for regulatory gamesmanship by incentivizing special interests to submit lengthy and frequent comments that impede regulatory progress. In the competition context, however, the consideration of public comments requirement could be beneficial insofar as it requires agencies to entertain comments from constituencies outside of the circle of favored producers—for example, potential new entrants, consumers or firms vertically related to the relevant market.

2.2.4 Contemporaneousness between assertion of justification and promulgation A final important feature of hard look review is the requirement that any justifications for the regulatory decision be presented at the time of the regulatory decision and not subsequently invented for litigation purposes.50

47 David Franklin, ‘Legislative Rules, Nonlegislative Rules, and the Perils of the Short Cut’, 120 Yale Law Journal 276, 283–284 (2010). 48 William Pedersen, ‘Formal Records and Information Rulemaking’, 85 Yale Law Journal 38, 75–76 (1975); ACLU v FCC, 823 F3d 1554, 1581 (DC Cir 1987). 49 See Jason Webb Yackee and Susan Webb Yackee, ‘Testing the Ossification Thesis: An Empirical Examination of Federal Regulatory Volume and Speed, 1950–1990’, 80 George Washington Law Review 1414 (2012) (exploring and critiquing the ossification thesis). 50 In the United States, this is known as the ‘Chenery doctrine’. See SEC v Chenery Corp, 318 US 80, 95 (1943) (holding that validity of agency discretionary action must rise or fall based on the validity of the agency’s contemporaneous explanation for its decision).

324 Daniel A Crane The contemporaneousness rule stands in contrast to ‘rational basis’ review, under which a regulatory action is upheld if it could be supported by any conceivable rational basis. Not only must the regulatory decision be empirically supported as opposed to merely rational, but the agency must think through the justifications upon which it will rely before promulgating the regulation. The basis for the regulation should be decided by agency personnel charged with the regulatory decision and not lawyers subsequently brought in to defend it. In the competition context, the contemporaneousness requirement could increase the likelihood that specialized agency staff—particularly economists and technological experts—will be involved in framing the objectives and bases of the regulatory decision. It would diminish the likelihood that regulators would act solely to insulate special interests from competition and then rely on legal arguments to defeat challenges to the anticompetitive regulatory decision. It would also diminish the likelihood that regulators would rely on theoretical or potential rather than documented market failures to justify measures that suppress competition. In short, the contemporaneousness requirement could prompt agencies to take a more careful look at the competitive effects of their decisions before taking actions that reduce market competitiveness.

3. Comparing the three models The hard look model lies somewhere between the representation reinforcement and substantive review models of judicial review. It involves a significantly greater degree of judicial scrutiny than the representation reinforcement model. Conspicuous state responsibility for the anticompetitive effect is insufficient—the state must affirmatively justify it to a level of factual and analytical rigor that satisfies the reviewing court. But the hard look model involves less intensive judicial involvement than full substantive review.51 The reviewing court does not balance competing interests, decide what interests are important or trivial, or insist that the agency pursue paths less restrictive of competition. In this sense, hard look review eschews some of the

51 On the other hand, some scholars have worried that a rigorous application of the hard look doctrine involves courts in a kind of substantive review akin to that employed during the earlier era of economic substantive due process. See, eg Martin Shapiro, ‘The Administrative Procedure Act: A Fortieth Anniversary Symposium: APA: Past, Present, Future’, 72 Virginia Law Review 447, 479 (1986) (‘Economic substantive due process as substantive review of the reasonableness (read correctness) of rules is alive and well in the DC Circuit’s clear error/hard look synopticism-demanding version of “arbitrary and capricious…”’).

Hard Look Review of Anticompetitive State Action 325 most poignant countermajoritarian difficulties and questions of d emocratic legitimacy that plague the substantive review approach. To the extent that one believes that the representation reinforcement and substantive review models are both problematic, a hard look type approach may offer a compelling middle ground. I do not argue that hard look review is unambiguously superior to alternative models; much depends on the legal, regulatory, political and institutional ecosystem in which courts and regulators operate. The hard look model is currently deployed in the administrative law context and involves a particular set of relationships—the triangular relationship between the legislature (Congress), federal administrative agencies (which may be part of the executive branch or independent) and generalist federal courts. It is unclear how easily or successfully aspects of hard look review could be transplanted to other contexts. For example, when the regulatory relationship at issue is between a reviewing court and a legislature (as opposed to an administrative agency), can a court demand the same sorts of procedures as it can in the administrative context? Can a court demand that a legislature create an evidentiary record for its decision, consider public comments and offer a unified rationale for a new statute?52 The transplantation of hard look review into other contexts would depend on the political feasibility and legal permissibility of this sort of ongoing engagement between courts and the regulatory body. There is also a question about the incentives and competence of courts to deploy any of the three models discussed in this text—or others. Courts may be relatively well equipped to determine when a regulator has sufficiently taken ownership of an anticompetitive outcome that it can be held democratically accountable. Locating the markers of such political ownership—clear statutory authority and routine interaction with the regulated parties—does not involve courts deeply in the technical details of the policy itself. Judges who feel insufficiently schooled in the regulated technologies and economic circumstances may thus be drawn to a highly proceduralist version of representation reinforcement review. On the other hand, judges may shy away from substantive review, not merely out of concerns over political legitimacy, but also because of a concern that they do not sufficiently understand the underlying technologies or market conditions to improve on the regulatory decision. Even if a court is concerned that a regulatory decision suppressing competition reflects special interest group capture of the regulator, it may be reluctant to invalidate the regulation

52 In some contexts, such as where the power of Congress to regulate is in doubt or the statute in question presents serious conflicts with constitutionally protected liberties, courts do sometimes rely on Congressional findings of fact and statements of purpose in deciding the constitutional question. See Daniel A Crane, ‘Enacted Legislative Findings and the Deference Problem’, 102 Georgia Law Journal 637 (2013).

326 Daniel A Crane for fear of making matters worse. Since regulatory d ecision-making often involves mediating the conflicting demands of warring special interests, invalidating a regulation because it conflicts with competition values may have the perverse effect of simply handing victory to a different special interest, with harmful implications for consumers. Judges may be reluctant to deploy competition principles to second-guess regulators out of a fear that alternative regulatory or deregulatory solutions will be even worse. How successful a hard look approach would be in responding to these concerns with the substantive review approach depends in large part on how ‘hard’ the look turns out to be. A successful hard look approach would merely involve courts in judging the internal consistency of the agency’s position— are the things the agency said internally consistent and supported by the record, did the agency thoroughly respond to what people said and the alternatives they proposed, has the agency maintained a consistent justification over time? By focusing on matters internal to a public record, the hard look approach reduces the likelihood that courts will exceed their competence or create unintended consequences by proposing ideas or solutions sua sponte. Again, this does not guarantee that state action harming competition will always be optimal, but has the potential in some circumstances to elevate the level of judicial engagement about the rudiments of representation reinforcement without entailing some of the larger risks of substantive review.

4. Conclusion This chapter has considered three models of judicial review of state action impairing competition. (There are surely others, including hybrids of the three discussed here.) All three models have advantages and vulnerabilities. The success of any of three depends crucially on the degree of the problem and the legal and political context in which both regulation and judicial review occur.

PART II Legitimate Enforcement of Competition Law Chair: Philip Lowe Speakers: Rafaer Allendesalazar Gerald Barling Stephen Calkins Bill Kovacic Wolfgang Kirchhoff Frédéric Louis Paul Lugard

Renato Nazzini Konrad Ost Christine Parker Jacqueline Riffault-Silk Mario Siragusa James Venit Joshua Wright

Philip Lowe: So, I think we’re all here, welcome back. I have to say that whether we realise this or not, this morning we were discussing effectiveness and this afternoon we’re discussing legitimacy, but I think that this morning’s discussion proves that the two are completely interlinked. My job first of all, before our coffee, is to introduce two subjects: one on parallel enforcement, fundamental rights and the rule of law; and the other on agency priorities and case allocation. Any time anybody mentions anything in Latin it becomes much more serious, so ne bis in idem is definitely going to be a source of great discussion between us. Rafael has the first word on this. DD

Rafael Allendesalazar: I will try to be brief to recover a bit of time. Parallel enforcement, and the principles of parallel enforcement, were established by the Court of Justice more than 20 years ago in Walt Wilhelm1 and other cases, and subsequently they were codified in Regulation 1/2003. I think these principles were established in a time in which parallel enforcement was absolutely exceptional, but after 10 years of Regulation 1/2003 it has become a very important topic because there are a lot of cases where we see parallel enforcement or cumulative enforcement in cartel cases. I would like to mention at least two cases: the Detergents case,2 which was both European and DD

Case 14/68 Walt Wilhelm and others v Bundeskartellamt [1969] ECR 58. Decision No 11-D-17 of 8 December 2011.

1 2

328 Part II: Legitimate Enforcement of Competition Law French;3 and there is also an Envelopes case,4 which in Spain has given rise to three decisions plus an investigation from the Commission, which I think raises a problem of cumulative application of EU and national law by the national competition authorities and the Commission. The first consequence is that the fines can exceed the famous 10 per cent limit, which is applicable both in Spanish competition law and in EU competition law. One of the ways to avoid problems would have been to apply in a flexible way the ne bis in idem principle. But even though that principle is enshrined in Article 50 of the Charter of Fundamental Rights, it has in practice been interpreted in a narrow way by requiring three ‘identities’ or ‘unities’: (i) identity of the facts; (ii) unity of offender; and (iii) unity of the legal interest protected.5 This third factor, in the view of Advocate General Ruiz-Jarabo, could be considered to apply when national and EU competition law are applied;6 but the Court does not seem to share that view, and the ne bis principle has never been an obstacle to the cumulative application of antitrust cases by national and competition authorities. We can go back to the Walt Wilhelm case in 1969, where the Court said that, as a general requirement of natural justice, any authority should take into consideration previous punitive decisions that have been adopted in similar cases.7 As I said, there is already a risk that companies can be fined above the 10 per cent ceiling with this splitting of cases. And ironically, while the Commission and the EU Courts and the national competition authorities have all used this idea of a single complex infringement to bundle several possible infringements into one, they are on the contrary now splitting infringements into several separate infringements that can therefore give rise to separate fines. There is a third problem, which is a problem of leniency applications. The system is almost unworkable: you would have to go to 29 different authorities to make a leniency application. I think it is impossible to understand that in merger control we have a one-stop shop, but in this area of law, which is penal law, we have not been capable of providing companies with clear rules of case allocation. This is quite striking. Let me conclude by saying that cumulative or parallel enforcement creates a number of specific challenges in order to maintain the effectiveness but also the fairness of the procedure. I think it is fundamental now to rethink these principles that were established 30 years ago at a moment in which parallel enforcement is no longer an exception but the day-to-day routine in many, if not most, cartel cases. Thank you. Commission Decision of 13 March 2011 in Case COMP/39579—Consumer Detergents. CNC Decision of 25 March 2013 (Expte S/0316/10, Sobres de papel). 5 Joined Cases 18 and 35/65 Gutmann v Commission [1966] ECR 103; Joined Cases C‑238/99 P, C‑244/99 P, C‑245/99 P, C‑247/99 P, C‑250/99 P to C‑252/99 P and C‑254/99 P Limburgse Vinyl Maatschappij and Others v Commission [2002] ECR I‑8375, para 59. 6 Opinion of Advocate General Ruiz-Jarabo Colomer in Case C-213/00 P Italcementi— Fabbriche Riunite Cemento v Commission, EU:C:2003:84, paras 91 and 94. 7 See above n 1. 3 4

Part II: Legitimate Enforcement of Competition Law 329 Renato Nazzini: I’ll be addressing a similar topic, and let me begin by thanking the organisers: I am delighted to be here. On ne bis in idem, I’ll perhaps take up what Philip said about things being more serious when they are expressed in Latin. The ne bis in idem right—because it is indeed a right—is serious. It is very serious, and the theme of my paper is that in the competition law field in particular, as a special field of EU law, it is nevertheless not taken seriously enough—by the legislature in the first place, and then by the EU Courts. Then I will try to suggest ways in which the EU case law should and must be developed so that this right can be taken seriously. As I said, competition law seems to be treated differently, because if you look at the EU jurisprudence, for instance under Article 54 of the Convention on the implementation of the Schengen Agreement, the ne bis in idem right is given very serious consideration. But for some reason this is not the case in our field. I’ll start by giving some examples of that, then I’ll move on to sketch out some elements of the way in which the jurisprudence should be clear and transparent about the undisclosed assumptions that are the foundation and the reason why certain arguments are accepted or rejected in this field. And then I will set out, at the end, my conclusions on six key issues in relation to the ne bis in idem right or the ne bis in idem principle: (i) what is the same offence; (ii) whether competition authorities are allowed to subdivide or even dissect the same offence and investigate and punish the offence in different sets of proceedings; (iii) what constitutes a decision which precludes a second prosecution and punishment; (v) the consequences of setting aside the first acquittal or conviction on procedural grounds; (v) the relationship between EU and Member State enforcement; and (vi) EU and non-Member State enforcement. First of all, what do you do when you design a system of parallel enforcement or potentially overlapping jurisdictions and powers? Well, you make sure that the risk of parallel proceedings is limited or eliminated as much as possible. However, when you look at Regulation 1/2003 and at the ECN Notice, exactly the opposite is enacted.8 There are almost no provisions in this regard, except perhaps in Article 16. The provisions are more concerned about preserving the primacy of EU law; there are no provisions really barring multiple proceedings, or regulating jurisdiction or powers. Of course, in the ECN Notice you have case allocation mechanisms that are by and large sensible, and their operation has avoided what might have been a very big problem, but these are completely left to the discretion of the competition authorities. Then you move to the problem of the fragmentation of the same offence, which again seems to fall within the discretion of the authorities: for DD

8 See Commission Notice on cooperation within the Network of Competition Authorities, [2004] OJ C101/43.

330 Part II: Legitimate Enforcement of Competition Law example, the Detergents case9 that Rafael mentioned, and to an extent the Mastercard cases,10 and possibly the LCD screen cases,11 where one case was about screens for personal computers and other cases were about screens for iPhones and smart phones and so on, but apparently most of the anticompetitive meetings were the same ones. What type of decision bars a second prosecution—not only a conviction, but also an acquittal? What system do we have under Regulation 1/2003? According the The Tele2 Polska judgment,12 national competition authorities are forbidden from making positive exemption decisions even if they are convinced, after a full investigation, that there was no infringement. If an NCA is convinced there was no infringement, it can only adopt a noninfringement decision under national law—so such decision can be adopted as far as the application of the EU competition rules is concerned. This is significant because, in a system of parallel enforcement, if NCAs are precluded from adopting exemptions the first limb of the ne bis in idem principle goes away: we can investigate you time and again, until we have enough evidence to convict you. Then there is the legal interest test in Toshiba.13 As Advocate General Kokott said in her Opinion in that case, NCAs are only allowed to apply national competition law in accordance with the delimitation of competences set out in Regulation 1/2003. Regulation 1/2003 itself must be interpreted in accordance with Article 50 EU Charter and the general principle of ne bis in idem.14 This apparently doesn’t apply to EU competition law, and so on and so forth, and you know the ne bis in idem principle does not apply in EU and non-EU enforcement and so on. The case law of the Court of Justice in Walt Wilhelm and Toshiba relies on the requirement that two offences are the same only if they protect the same legal interest. Because, as already argued, this requirement is in breach of Article 50 EU Charter. This case law lacks foundation and should be overruled at the earliest opportunity. The Court of Justice should reconsider the position and recognise that when Member States and the Commission apply national or EU competition law to infringements consisting of the same or inextricably linked facts and committed by the same undertakings, the ne bis in idem principle applies with full force and bars a second prosecution or punishment.

See above n 3. Case T-111/08 Master Card and Others v Commission, EU:T:2012:260; appeals dismissed: Case C-382/12 P, EU:C:2014:2201. 11 Case T-128/11 LG Display v Commission, EU:T:2014:88; appeal dismissed: Case C-227/14 P, EU:C:2015:258. 12 Case C-375/09 Prezes Urzędu Ochrony Konkurencji i Konsumentów v Tele2 Polska sp z oo, now Netia SA [2011] ECR I-3055. 13 Case C-17/10 Toshiba Corporation v Úřad pro ochranu hospodářské soutěže, EU:C:2012:72. 14 Opinion of Advocate General Kokott in Toshiba, EU:C:2011:552, paras 105–106. 9

10

Part II: Legitimate Enforcement of Competition Law 331 Now, here is what I think is a way forward. Clearly there are considerations and undisclosed assumptions that the EU Courts rely on, and these should be made explicit. They concern the theoretical foundations of the ne bis in idem principle in EU law, which is the protection of the right not to be tried or punished twice for the same offence. This is the ‘rights-based’ rationale. On a different level, there is the facilitation of free movement of persons. This is the ‘free movement’ rationale. Then there is the preservation of the legitimacy of the legal system as a whole and the competition enforcement regime in particular. This is the ‘legitimacy’ rationale. On substantive legitimacy grounds, the case law of the Court of Justice on the relationship between EU enforcement and non-EU enforcement (non-EU meaning also non-Member State), this is obviously quite right. The powers of the EU authorities to achieve effective enforcement of EU law cannot be limited by the exercise of powers by foreign authorities (South Korea or the United States, etc) that have different objectives. Finally, there is the desirability of achieving efficiency and respect for the rule of law in the investigation and prosecution of offences. This is the ‘disciplinary’ rationale. Now, for each of the issues that I set out in the beginning, the balancing of all these factors leads to a certain conclusion. Unfortunately, I do not have time to go through the whole reasoning, but I can set out briefly what my conclusions are. First of all, there is the matter of the same offence. The EU Courts apply a threefold test to the definition of the same offence, requiring the identity of the facts, the identity of the offender and the identity of the legal interest protected. This test has been completely out of line with the test of the European Court of Human Rights since Zolotukhin,15 and it must be overruled. What counts here is whether the facts are the same or inextricably linked, which is the test applied under the convention implementing the Schengen agreement. Otherwise, of course, the ne bis in idem principle would only ever apply when EU competition law is enforced, which is obviously not the purpose of the system. The reason for this is that both the free movement rationale and the rights-based rationale point very clearly in this direction. This is the same Court of Justice which says that, albeit under a different provision, the fact that these provisions are worded in a different way is obviously relevant but not conclusive. The European Court of Human Rights in Zolotukhin reached the conclusion that the legal classification of the offence is irrelevant.16 The fragmentation of the offence, such as what we have seen in the Detergents case, cannot be a means to enable multiple prosecutions to proceed. Once the first decision on the merits is reached, all other proceedings should be barred. It is very clear that the operation of the ne bis in idem principle cannot depend purely on prosecutorial discretion. Case No 14939/03, Zolotukhin v Russia, 10 February 2009, para 82. Ibid.

15 16

332 Part II: Legitimate Enforcement of Competition Law Third, what type of decision bars further proceedings? It is clear that if the legislature or a prosecuting authority has complete freedom in deciding what an ‘acquittal’ means, or when an acquittal decision can be made, the ne bis in idem rule protecting a person from further proceedings even following an acquittal would be completely obliterated. To achieve this, it would suffice for legislation to prohibit acquittal decisions and provide for mere case closures on non liquet grounds whenever the defendant’s guilt cannot be proved to the required standard, or to give the prosecution or court complete freedom to bring a case to an end without an acquittal whenever they think fit. Allowing this is basically to abolish one of the two elements of the ne bis in idem principle that bars further prosecution even if there has been an acquittal. It is not the prohibition of double punishment; it’s the prohibition of a second set of proceedings once there has been a decision on the merits. And does a finding of ‘no grounds for action’ under Article 5 of Regulation 1/203 bar further prosecution? Does any complaint or rejection of a complaint by the Commission bar any further prosecution? Of course not. In the paper, I suggest a test which is close to that proposed by Advocate General Sharpston in her Opinion in Gasparini, where she argued that a defendant should be entitled to be protected by the ne bis in idem principle when ‘he has de facto been placed in jeopardy’ and the ‘criminal proceedings have involved any significant consideration of the merits of the case’.17 If this has happened, the authority or any court or any decision-maker cannot avoid the application of the ne bis in idem principle by saying ‘Well, I am merely issuing a non liquet decision’. I already mentioned the relationship between EU and non-EU enforcement, and the case law has developed quite significantly from Boehringer Mannheim18 to the SGL Carbon cases.19 The early case law referred first to effects, then to the idea that the legal interest was not the same, and then, if we read the latest judgments carefully, the Court of Justice changes the language and refers to the different objectives pursued and the fact that the powers of the EU authorities cannot be curtailed in this way. This is a legitimacy rationale which is absolutely correct in this scenario. My last point concerns the setting aside of decisions on procedural grounds. Here, as we know, proceedings can be reopened—there is no acquittal or conviction, so there is no decision on the merits. This is true from a doctrinal viewpoint, and it’s also true that if the decision has been set aside for procedural reasons then again the effectiveness rationale, which I agree is part of

17 Opinion of Advocate General Sharpston in Case C-467/04 Giuseppe Francesco Gasparini [2006] ECR I-9199, para 96. 18 Case 7-72 Boehringer Mannheim GmbH v Commission [1972] ECR 1281. 19 Case C-328/05 P SGL Carbon AG v Commission [2007] ECR I-3921, para 28, confirming the approach in the earlier case law: Case C-308/04 P SGL Carbon AG v Commission [2006] ECR I-5977, paras 26–32.

Part II: Legitimate Enforcement of Competition Law 333 the legitimacy rationale, may well allow a reopening of the proceedings. In the paper, I set out a different possibility, a different test, which would allow for a better protection of the defendant’s rights because otherwise the defendant could potentially be exposed to endless incompetent prosecutions. Under this different test, proceedings could be reopened only if the effective enforcement of EU law so requires and only if the procedural error is not attributable to the prosecution at all. I will stop here and I hope I have not exceeded my time. Wolfgang Kirchhoff: I would like to focus on just particular aspects of my paper, and it was already mentioned by Konrad Ost this morning. I refer to the themes of effective parallel enforcement, fundamental rights and legal succession—in particular, from a German perspective. OK, we know certainly from all Member States we have the same substantive law as far as fundamental rights are concerned. However, effective enforcement also requires uniform powers of investigation and penalties. And from a German perspective a serious problem here is that, under German competition law, there is only a limited scope of culpability of undertakings for regulatory fines. A legal person may be fined only where its legal representative or person responsible for its operation on behalf of the management has committed the offence as a result of which duties of the legal person have been violated or where the legal person has been or was intended to be enriched. In contrast, Article 23(2) of Regulation 1/2003 stipulates fines for undertakings in the broad sense of economic entities, and these references result in difficulties with the efficient enforcement of EU competition law in Germany. In some cases, a group is restructured after an affiliate company has been fined for a competition law infringement. The fined company may have merged with another company, or its assets may have been sold. Efficient enforcement requires in such case that it should still be admissible to execute the fine. Under the concept of a single economic entity in EU law, this is possible. However, according to a recent decision of the German Federal Court of Justice,20 this concept does not apply in relation to sanctions by the Bundeskartellamt. Liability for regulatory fines imposed by the Bundeskartellamt by attribution of assets may only be based on an express statutory provision—which is missing in German law. Without a clear rule in statutory law, an extension of the regulatory fines liability to legal successors can only be accepted where there is an identity or near identity between the old and the new collection of assets. This only exists when the company continues business in the form of an unchanged or nearly unchanged new legal entity, and the other assets of the new entity are to be neglected. In contrast, there is a lack of economic DD

20 BGH, Decision of 10 August 2011—KRB 55/10, Wirtschaft und Wettbewerb 2012—Versicherungsfusion.

334 Part II: Legitimate Enforcement of Competition Law identity when companies of approximately the same size and nearly identical market shares merge and bring together their business activities. Any further extension of the liability of fines on legal successors would exceed the limits drawn to judicial construction of the law by the ‘no punishment without law’ principle found in Article 103, second paragraph of the German Basic Law but also in Article 7 ECHR and Article 49 of the Charter of Fundamental Rights.21 So it is the legislator’s task to introduce liability of legal successors. In fact, the tension between efficient enforcement and fundamental rights in Germany was mitigated, but not solved, by an amendment of German competition law at the end of June 2013. By virtue of that amendment, regulatory fines may now be imposed on a full or partial legal successor.22 In fine proceedings, the legal successor has to assume the legal status of the predecessor. Other forms of circumventing liability for fines without full or partial legal succession, eg by asset deals, will nevertheless remain. The German appellate court in Düsseldorf in December 2012 examined whether liability of the legal successor may follow from Article 5(2) of Regulation 1/2003.23 According to this provision, NCAs are authorised to enforce EU competition law by imposing fines or any other sanction provided for in their national law. In the understanding of the appellate court, the authorisation of NCAs by Article 5 is limited to punishing violations of EU competition law according to the procedural rules of their respective national laws. Article 5 does not transfer the same powers of investigation and sanctions to the NCAs that are granted to the Commission in the Regulation. The appellate court gave a number of reasons for this opinion, such as the wording, regulatory purpose and legislative history of Article 5 of the Regulation. Moreover, and in our context most interestingly, the appellate court argued that any other interpretation of Article 5(2) of Regulation 1/2003 would result in an infringement of the ‘no punishment without law’ principle—which, as I said before, is contained in Article 7 ECHR and Article 49 CFR. According to the case law of the European Court of Justice, the principle nulla poena sine lege implies that legislation must clearly define the nature of an offence and the corresponding penalties. That condition is met in the case where the individual concerned is in a position, on the basis of the wording of the relevant provision and with the help of the interpretative assistance given by the courts, to know which acts or omissions will make him criminally liable. The appellate court concluded from Germany’s constitutional tradition—which

21 According to this principle, ‘No one shall be held guilty of any criminal offence on account of any act or omission which did not constitute a criminal offence under national law or international law at the time when it was committed’. 22 According to § 30, para 2a of the Act, a regulatory fine may be imposed on a full or partial legal successor. 23 OLG Düsseldorf, judgment of 17 December 2012—V-1 Kart 7/12 (Owi), Wirtschaft und Wettbewerb 2013, 749—Silostellgebühren II.

Part II: Legitimate Enforcement of Competition Law 335 the ECJ takes into account—that the ‘principle of precision’ applies not only for the elements of fact which define an offence, but also for the threat of sanctions. Since Article 5(2) of Regulation 1/2003 neither provides for any specific sanctions nor establishes any framework for sanctions, it would be contrary to the principle of ‘no punishment without law’ to use that provision as a legal basis for the imposition of fines by the NCAs. In contrast, in Tele2 Polska, Advocate General Mazák took the view that Article 5 allows all NCAs to have consistent decision-making powers, without the need to wait for the implementation of Article 5 in national law.24 However, the ECJ in its judgment did not refer to this point in the Advocate General’s Opinion, at least not directly. In the opinion of the Düsseldorf appellate court, the interpretation of Article 5 is sufficiently obvious and has to be regarded as an acte claire. Consequently, no reference for a preliminary ruling according to Article 267 TFEU was made. The decision is presently under appeal on points of law to the Federal Court of Justice (my court), and so clearly I will not give any indications as to the merits of that appeal on points of law. However, the judgment of the appellate court is an instructive example of the frictions which may arise between efficient enforcement and fundamental rights in the system of parallel enforcement with the Commission and the NCAs, as established by Regulation 1/2003. Specifically in this case, the issue to be clarified is whether Article 5(1) of Regulation 1/2003 directly, without the need for an enabling provision in national law, authorises NCAs to impose fines on undertakings—in the broad sense of economic entities as defined by the case law of the ECJ, including legal successors in case of economic continuity. It is entirely open whether the Federal Court of Justice will seek advice in Luxembourg, and what the ECJ will decide should the Federal Court choose to do so. It would have to carry out a delicate balance between the principles of efficient enforcement and legal certainty as laid down in Article 7 ECHR and Article 49 CFR. The wording of Article 5(2) does not explicitly define the recipient of NCA decisions that impose fines by means of reference to the concepts of economic entity and economic continuity in EU competition law. On the other hand, one may argue that, by authorising the NCAs to apply Articles 101 and 102 TFEU, Article 5 refers indirectly to the ECJ’s interpretation of these Articles, including the concepts of economic entity and economic continuity. It may well be that the European legislator will have to intervene. Such action should not be limited to the problem of legal succession. In the internal market, which requires a level playing field for undertakings, serious differences in investigatory powers and sanctions between the Commission and

Opinion of Advocate General Mazák in Tele2 Polska, above n 12.

24

336 Part II: Legitimate Enforcement of Competition Law NCAs, as well as among NCAs, are unacceptable in the long run and contradictory to the principles of the system of effective parallel enforcement of EU competition law. In particular, one may doubt whether the NCA of a Member State that allows undertakings to escape penalties for competition infringement by simply changing their identities through restructurings, sales or other changes can still be considered well placed to deal with the case in the ECN. It is true that natural persons may be subject to substantially different types of sanctions in Member States. In contrast, sanctions imposed on undertakings are of the same type in all systems throughout the Union. Harmonisation might be easier here, and in particular harmonising instruments should define the addressees of NCA decisions that impose fines by reference to the concepts of economic entity and economic continuity as established in EU competition law. Thank you very much. Lowe: From a non-German perspective, this problem of succession raises a number of question marks. Certainly, if one company takes over or merges with another, there is a problem of due diligence. When the succession takes place, it must include, in principle, contingent liability and the allocation of that liability. Maybe some of the commentators around the room can say whether they believe this is due to the absence of harmonisation of procedural rules and enforcement in the application of competition rules, or whether there is a wider problem. Or course, if a company denies that it is liable by succession, it is a bit like the airline, when you get on an airliner and they apologise for it being two hours late and they say ‘Well, the reason for this is because of the late arrival of the incoming aircraft’. It leaves you without any grounds of liability whatsoever. Neither Alexander nor I would argue against a further approximation of procedural rules, but I’m going to invite comments at this point because when we move to Bill’s intervention it will be on a slightly different topic. DD

Konrad Ost: On the paper of Renato Nazzini and on the ne bis in idem issue, I think it might be very questionable whether the national authority really has the power to fine on the basis of effects on other jurisdictions. I doubt whether the Bundeskartellamt can do that. Therefore, I think the approach of Advocate General Kokott in Toshiba, which stresses need to consider effects as part of the assessment of ne bis in idem, is a correct one.25 There isn’t a major problem in the cases where you have parallel proceedings of several national authorities. A second point relates to leniency. You stress a little bit that parallel enforcement is very difficult and poses major problems, DD

Opinion of Advocate General Kokott in Toshiba, above n 14, paras 105–106.

25

Part II: Legitimate Enforcement of Competition Law 337 but I think the ECN—the Commission and the national authorities—saw the issue, and they made a huge effort and arrived at a good result with the Model Leniency Programme. Today, it is really not such a big task to file parallel leniency applications. So we shouldn’t overestimate that issue. Even if a case concerns a restriction of competition by object, you may assess the effects of it when you think of your jurisdiction. You assess the gravity of an infringement in your territory. For example, we only take into account the turnover affected within our territory, which is a clear indication that we try to have a proxy for the effects within our territory, and not in other territories. Stephen Calkins: In reading the papers, the example that came to mind, and I don’t know really what the solution is or should be, is that you can observe just from reading the newspapers that there will be multiple NCAs that are all investigating almost exactly the same conduct by almost exactly the same company at exactly the same time. Any one of those might lead to a conduct remedy, which could affect a firm’s operations across a national border. I haven’t been in Europe long enough to know how this is all neatly resolved so that we are neither wasting resources nor unduly burdening the company being investigated, or taking action that leads to inconsistent remedies or findings. But there are obviously serious issues of coordination here. DD

Lowe: Well, the time dimension and the jurisdictional dimension may be different, and I am sure that Fred has some comments on the implications regarding the leniency programme because it becomes very complex when you’re even dealing with the complexities of different jurisdictions and time frames. DD

Fred Louis: Yes indeed, and while some NCAs allow you to make the short-form application by phone and in English, others won’t. When you are trying to keep the circle of people in the know in the company as small as possible, and when you are trying to develop the evidence because the Commission and the US Department of Justice are on the starting blocks, you are forced in those critical hours to divert significant resources to a complete sideshow which is unnecessary. Why can’t we have a system where you make a central application to the Commission, and then the authorities discuss it in the ECN, then decide the best place and the case gets transferred? Why do we have to go to everyone? You know, it seems to be a completely unnecessary step. On ne bis in idem, you say ‘Well, we avoid it by only looking at turnover in Germany’. Well in the parallel investigations in the Mills case, the French authority based the fine imposed on the German companies on their sales in Germany, using for this the information given by the BKA, which it had obtained in the course of its own investigation of the conduct. And some of DD

338 Part II: Legitimate Enforcement of Competition Law the German companies that were involved were based 800 kilometres from the French border. They had never sold anything in France but, by using the type of theoretical approach that we’ve come to know at EU level, you say ‘Yes but you could have sold in France, therefore somehow your sales in Germany are a good proxy for what you could have achieved in France’. And where a product cannot be sold more than 200 kilometres from where it is produced, this is ludicrous. There are a lot of real problems with parallel application of the competition rules. Jim Venit: The leniency problem is such a simple thing to fix. Fred is right, it diverts attention at a serious point in time when it’s bad to have attention diverted. There is differentiation in the way it’s handled. For example, each authority tries to be user-friendly, but some succeed more than others. To my mind, it is completely unjustifiable to have that extra complication when there is such an easy solution to it. DD

Lowe: Well, to find a procedural solution for leniency solutions EU wide, we do have a notion of the authority best placed to deal with an infringement, so we should be able to get through that. Who else would like to intervene? DD

Renato Nazzini: I would like to go back on a couple of points, especially the point made by Konrad Ost on effects and multiple prosecutions in multiple Member States because effects are different. The first point is that I agree it’s not clear whether those effects can be fined, or whether fines take into account effects in different countries. This lack of clarity is part of the problem. I started my presentation and I started my paper by saying that one of the sources of the problem is that you have complete lack of clarity on this distribution of powers, and quite deliberately so. And there is a question mark, of course, as to whether this should be an issue for EU law or for national law. The second point is that there may well be infringements that have effects in different countries and therefore in theory can be prosecuted and fined in these different countries; but isn’t this precisely the reason why we have EU-wide ne bis in idem rights? Again, if you read the case law of the ECJ on the Convention of the Schengen Agreement, or the money laundering cases, you take money or cigarettes in one country, you commit a crime there, you cross the border and then you use the money or you sell the cigarettes, you do things, you go to the exchange, the bureau de change in the other country, well surely that’s a different effect? Indeed, different conduct, if you want, and also a different legal interest of another Member State because cigarettes are bought or imported from a non-EU country into Member State A and then they are taken to Member State B and other things happen. Well, the Court of Justice has said that this is the same offence. So ne bis in idem applies DD

Part II: Legitimate Enforcement of Competition Law 339 here, and this is precisely the type of EU-wide case ne bis in idem is supposed to deal with because if you exercise your freedom of movement, you cannot be punished or prosecuted again: you are exercising your freedom of movement. If someone can explain the difference clearly and convincingly I will be happy, because it means that the current law is fine. But I cannot see why this doesn’t apply to an undertaking which exercises its Treaty rights by selling in multiple jurisdictions. The ne bis in idem principle is there so that if I exercise my freedom of movement I am not penalised time and again. Finally, there is the efficiency point raised by Stephen Calkins. I didn’t have time to get into that, but that is very important as well. It’s probably not the core reason for the ne bis in idem principle, but it does impose discipline on national authorities. That is efficient, for example, because it avoids duplication of resources. Lowe: So we want the system to be effective in relation to leniency and sanctioning, but the ne bis in idem principle also has to be respected. The bottom line is that an authority, whether it is the Commission or one of the national authorities, should take responsibility for investigating and sanctioning a particular infringement that affects several Member States, which from the point of view of effects will lead us probably to the issue of the extent of the sanction. How far-reaching was the infringement, and how can one respect the ne bis principle while ensuring that the sanction is appropriate given the full range of the infringement? It seems to me that this reinforces the point Konrad and Judge Kirchhoff made: if we want the best-placed authority to take the decision, you’ve got to get the procedural rules sorted down to the national level because there will be all sorts of implications of one national authority taking a decision to sanction or remedy problems in another jurisdiction. We don’t want any situation of double punishment, but on the other hand we do want effective enforcement. Raphael? DD

Allendesalazar: On this idea of having the fine established based on where the effects materialise, as we all know, with mono-product firms we will reach 10 per cent, so if you have two cases you will be above that, and that has already been established except by the Spanish competition authority, which applied 10 per cent and said: ‘I know there is an investigation by the Commission but since I don’t know what’s going to happen, I will fine 10 per cent’. And probably, in the Envelopes case, where the Spanish authority has also hit 10 per cent, there is another procedure conducted by the Commission, and it won’t be surprising when the total exceeds 10 per cent. DD

Lowe: Which pleads again for a suspension of the decision until such time as all authorities concerned are on the same page and can ensure good coordination. Judge, you have the last word. DD

340 Part II: Legitimate Enforcement of Competition Law Kirchhoff: I would like to say at least a few words to your scepticism concerning our German issue here on legal succession. I can understand that that may be a bit difficult to understand from an English point of view, but our problem in Germany is that, according to our regulatory law, it is clear that only a legal person may be fined. The question is: what is the legal person, and how long is a legal person still alive? Certainly, we do have in Germany due diligence in acquisitions of companies, in particular for competition infringements and so on. This is an important aspect of due diligence, but that concerns cases where a company is sold—where the shares are bought and the legal person continues to exist. In that case, it is clear that the acquiring shareholder would assume the risk together with the company; that is clear. But if the legal person is dissolved because it is restructuring, and if this legal person is split up, for example, then that becomes an issue. DD

Lowe: Thanks very much, and I realise that issue of the definition of an undertaking is indeed a major challenge. We will now have to move on, and I am going to ask Bill to speak next. DD

Bill Kovacic: Thank you, Philip. How does the setting of priorities affect the legitimacy and the effectiveness of competition law? I am going to talk about three things: the benefits both to effectiveness and legitimacy; the obstacles that get in the way—why do competition agencies find this very difficult to do, even though the conceptual reasons to do so are compelling? And last, I will talk about the means with which to do this effectively, especially in light of the obstacles. Why does it help the quality of a competition agency’s programmes? First and foremost, a good programme that provides for the setting of priorities forces the agency to define what it is all about. This allows the agency to spell out its aims carefully, and to engage in what a commercial enterprise might call effective branding. If you can’t define your programme—if you can’t spell out your purposes clearly—you are probably not in a good position to establish a good reputation and design a programme for the agency. So a process that routinely compels you to set and defend priorities is a necessary first step for establishing what it is the agency ought to be doing. A second and closely related benefit is that, done well, this can focus attention on what’s more important. In effect, it puts the agency in the position of a manager of a securities portfolio where the critical question is: how can the portfolio be designed to generate the best rate of return in light of the expectations of the investors committing funds towards that endeavour? And third, having a clear idea of objectives and priorities is indispensable to picking good programmes; and if you have a clear definition of goals in mind, that helps you to do the following. DD

Part II: Legitimate Enforcement of Competition Law 341 First, very fundamentally, it helps you to tell your professional staff what you expect of them. This way, they will not just randomly hunt for projects but will focus their attention on the specific activities you think are most valuable to the agency and to its constituencies. Second, and quite critically, setting objectives and priorities enables you to match commitments to capabilities. One of the greatest weaknesses of an institution is to take on too much. There is no sin in trying to work at 110 per cent of capacity, that’s probably healthy. But trying to work at 200 per cent capacity is a formula for a meltdown, and a good priority-setting mechanism is an excellent way to check to make sure that as the agency is committing itself to the next project and the project after that. It has to ask basic questions about who is going to do it, when are they going to do it, and how will we know that it’s working. And the last ingredient, which is reflected in the title to our topic, concerns decisions about cases. First, are cases the best way for a particular agency to do its job? And when cases are appropriate, which cases should be pursued? This gives the agency a routine occasion to lay out the entire range of policy tools it has at its disposal, and to ask what strategy or collection of strategies will produce the best result. Let us now consider why all of this is good for legitimate competition enforcement. Meaningful disclosure, and by that I mean disclosure that candidly and completely describes what the agency is doing, is one of the most important ingredients of a legitimate competition system. That kind of disclosure—for example, the revelation of priorities and the justification for priorities—is crucial to facilitate intelligent monitoring by external bodies, including legislatures, academics and professional societies. If they are to do a good job in exercising their monitoring function, it is essential for them to have a meaningful idea of what it is the agency means to do and how it means to do it. In particular, where the agency is forced to make trade-offs, as all agencies are, disclosure makes it possible to understand what the trade-offs are. The typical agency has a multitude of responsibilities. And it tends not to have resources commensurate with those responsibilities—so it has to make choices. A clear identification of priorities is a good way to spell out how those trade-offs are made, and that in turn is the necessary foundation for having a useful public debate about whether those priorities are the right ones. Also, by spelling out in detail what the agency intends to do and the means to accomplish those ends, that sets the baseline by which you can conduct useful ex post evaluation of performance. Why don’t agencies do this? I would say that it is more often the case than not that agencies do not engage in the degree of priority setting, case selection and programme selection that one might desire. First, the typical competition law, especially as you look at newer systems, imposes on the agency a wide array of goals that are often internally inconsistent: promote employment, promote small and medium business enterprises, achieve high levels of efficiency, assure a fair distribution of wealth, and several other things. It is a list

342 Part II: Legitimate Enforcement of Competition Law which, if each were executed to the fullest, would produce a completely inconsistent and contradictory set of policies. So one reason agencies do not spell out their priorities is that they are expected to reconcile those and they try to do it quietly, or at least under the banner of broad slogans such as consumer welfare, which enable them to accomplish a variety of ends by using specific objectives that sound pleasing to the ear on the outside. So one reason that they do not spell out priorities is that to do so; for example, to say: ‘I don’t care about small enterprises, I don’t care about distribution, I care only about total welfare’—to say that with great clarity—would bring the wrath of God upon them. Second, the typical competition agency today is not just a competition agency. Of the 120 plus jurisdictions that have competition laws, well over half of those agencies are policy conglomerates. They do competition policy and procurement and consumer protection and intellectual property. They are not assigned only to do competition law, and the multiplicity of objectives greatly complicates the setting of priorities because, here again, legislators tend to add new commands without providing resources to carry them out. What they have done in a somewhat dishonest way is to put the agency in the position of the emergency room of the hospital that is performing triage. It has to choose which commands are going to die and which one has a chance of living, without making clear to the outside world how it is making those choices. For example, the FTC, my former agency, enforces over 50 statutes. These include quaint mechanisms, such as the Muhammad Ali Boxing Act, which gives the FTC a role in overseeing the way prize fights are conducted in state jurisdictions. You might wonder what the FTC’s expertise is in that area, and if you ask how many FTC staff members are assigned to that task, that would be exactly zero. The expectation is that, along with the Dolphin Protection Act and a host of other interesting grab-bag responsibilities, the Boxing Act does not constitute a serious demand—but the legislature has not told the FTC to forget about them. So one reason you see a fuzzy definition of priorities is that, to say we have 50 statutes and here are 10 to which we’re devoting no resources at all, that would provoke a very unfortunate legislative response. But there are also various features of agency design that make setting priorities difficult. Governance is a prime example. If you have a multi-member board, it can be very difficult to get a consensus on what those priorities ought to be, and to spell them out clearly. A single hierarchy has a somewhat better capacity to do that. A multi-member board is more or less like an automobile with five participants, only one of which is the driver. The other four may all think that they are better drivers than the person driving. Sometimes they just give advice, but now and then they try to grab the wheel away from the driver. That greatly complicates decisions on where the vehicle ought to go. And if you add multiple enforcement agents and other public agents, and you want the system as a whole to have some coherence, you are faced with the difficult decision of bargaining across agency boundaries as to what exactly it is that

Part II: Legitimate Enforcement of Competition Law 343 the system as a whole ought to accomplish. Any time you place two institutions in the same policy domain, you have an inherent source of tension. It can be resolved, but it definitely gets in the way of coherent results. Another feature is that, depending on the jurisdiction, some legislative commands tell the agency: you have no discretion over setting priorities; you must address and resolve every case that comes your way. You’re the public bus and you must take anyone who gets on board and drop them off at their preferred destination, and that makes it very difficult to say to a complainant: ‘I’m sorry, your complaint just isn’t worth our time, we’re not going to pay any attention to it’—much less ignore it altogether. In assessing agency leadership, which we touched upon a bit already, don’t invariably point in the direction of investing in this kind of agency prioritysetting process. What is the main command to individual leaders? How are they measured? In the university setting the student asks: what is the basis for the grade in this course? For competition law, the most heavily weighted variable is how much stuff the agency does. That means cases—with further extra credit for big cases, and even more extra credit for big, visible results. Bigger fines, higher standings in the GCR reviews. That is obviously not the only way to run an agency, but if you look at the impulses that press upon agency leadership, it is all about short term output. Relatively few of those impulses say: ‘Build a foundation for your successor to do a good job’. A last question: what means can be used to ensure that priorities and goals are sensibly identified and pursued? A useful starting place is to look at agencies that do this well, and then to benchmark them and ask them how they do it. There are two agencies I like a lot. First, the New Zealand Commerce Commission. If you read their annual reports and their statements of plans, you see a very clear statement of objectives and a clear statement of means for carrying them out. Second, the Fingleton–Collins OFT formula for setting priorities was also excellent. In that system, case handlers were pressed to answer the following questions in each instance. First, what are we going to gain from this? Second, what could we lose? Third, who is going to do it on the inside? Fourth, how long is it going to take? Fifth, how will we know it’s working? Any agency that presses its operating units to answer those five questions about every matter that comes before it is in a very good position to set priorities and select programmes effectively. The agency process in light of this can be refined to put in place a systematic routine process in which these questions are answered. Every year is part of a budgeting exercise. The agency asks: what’s the ideal portfolio for the overhead? How do we maximise the rate of return with respect to improvements in doctrine, net results for consumers, and experience that builds our team’s capability? Is it a portfolio that is properly balanced with respect to doctrinal risk, political risk and implementation? That is a formula that can push the agency in the direction of better programme design. And the last element is to engage the broader community in a process of norms d evelopment.

344 Part II: Legitimate Enforcement of Competition Law What is the norm that matters? Go back to the incentive I described in a somewhat pejorative way. We are all complicit in this, the private bar likes action most of all. A competition system that generates no cases also generates no revenues. So the bar and its consultancies want action. Academics want action because that’s what we write about—case books would be very thin if there were no cases. Lectureship opportunities would fall dramatically. The bar loves to complain about the case that is not brought properly. It says: ‘I want more cases—oh, but not that one’. But if you have a system that doesn’t have action, none of us can make a living. We need to move away from an incentive structure that encourages a lot of action for the sake of action, to move away from the habit of saying: ‘We’re going to measure the quality of the system by its net accomplishments’. That is wrong-headed thinking. My time is up, I’ll stop there. Lowe: Thank you very much, Bill. That was really stimulating. Now there are competition authorities represented around this table and there are also several ‘controllers’ of agencies, so to speak, including members of the courts. Let’s hear what they have to say. DD

Ost: Thank you. That is a very broad picture you are painting and it is quite demanding for an authority to fulfil all your ideas, but of course there is much truth in it. It’s important to set the priorities, and it’s very important to assess whether a case is worth doing. This is done internally within the Bundeskartellamt, of course, for those cases where we have discretion. In the merger business you can’t really choose whether you take up a case because it is just notified and you have to decide on it. I’m a little bit sceptical about setting detailed priority programmes, for example in cartel enforcement. For example, I’d hesitate to say: ‘This year we want to do food or energy in cartel enforcement’, because, among other things, you don’t know which leniency cases will pop up or which complaints come to you. So you can focus on cartel enforcement, but it may not always be possible to set targets in relation to particular sectors. DD

Gerald Barling: Bill is right. I am sure that you have to have cases, but you don’t hear cases, as he says, for the sake of academics or lawyers, or even for courts. The reason, surely, that you have to have a certain irreducible number of cases coming through the system is because without them you don’t have compliance. The people on the ground—companies, that is, ‘undertakings’—need to have the feeling that if you are going to have effective compliance, they could be in the frame. If you have too closed a definition of what your priorities are going to be, and if a fortiori your priority is that you only run very big cases, the danger is that the people below that level or outside that area of priority begin to feel a little too comfortable. They may feel DD

Part II: Legitimate Enforcement of Competition Law 345 that they are not going to be in the frame, and compliance may suffer. So you have to have a certain number of small cases, and you have to have perhaps a certain amount of uncertainty so that people do not feel too comfortable if you are really going to have effective enforcement and compliance. Jacqueline Riffault-Silk: I would like to add something about the making of the decision, and especially the choice of sanctions. Given the specialisation of the authority in economics and the fact that, at least in France, there are some economists but it’s not exactly a speciality of the judge, it’s very important for the decision to be very clear and nourished in economic elements to sustain the debate before the appellate courts, and to identify the particular issues from the economic side. From that point, it is much easier to get to the law. So arriving at an adequate understanding of the economics is a priority in a competition case so that the debate within the case can be well grounded and produce an appropriate outcome. DD

Josh Wright: This is going to come in the form of a question for Bill, rather than a comment. With respect to an agency’s branding efforts, one of the things we are buying is deference, and one of the things driving the rate of return is going to be deference. If I put on my economist’s hat, it strikes me, reading through the paper, that I am going to make some investments in reputational capital and I am going to get paid off with deference, but it seems like I’ve got two markets where I can get the deference. One is from courts— I can invest in high quality and I can get my rate of return in a court by having good decisions and we can talk about what it costs to do that and whether the rate of return justifies what I am getting back with the deference, and I’ll bet it does. The second market which you mentioned has to do with political oversight. One of the things you talked about briefly here but in more detail in the paper is the idea that it might well be that the things that buy me reputational capital from Congress either correlate positively but very roughly with doing good, or worse, they may be negatively correlated with doing good. It might just be that doing more, there is at least some hope in the answer that it is rough but positively correlated. Doing more isn’t always good but sometimes, and it depends on what you mean by more, but it could be that they are negatively correlated; I can certainly think of a few reasons why they would be. And if that’s the case, if I take the branding analogy to its logical end, then I’ve got a choice: I can successfully brand with one and not the other, but the one I can’t brand successfully will impact my budget. DD

Calkins: It is always great fun to hear from Bill. I have heard him use a marvellous analogy when talking about how important it is to keep score. Think of what the NBA, the National Basketball Association, would look like if they didn’t keep track of the number of assists that basketball p layers DD

346 Part II: Legitimate Enforcement of Competition Law had—that is a great comparison. That said, and without disagreeing with anything in particular, I just want to say a little word in favour of cases. I would point to the Staples/Office Depot merger case,26 which I worked on when I was General Counsel at the FTC. First, back when we were there, I remember lots of times when we were working on some consent order, or working on some study where we said this is what people will remember and it makes a difference, and all that sort of stuff. And you know, all that had a lot of truth to it, but if you did a little check as to what difference a consent order did in fact make, my guess is that bringing and winning the Staples/Office Depot merger case would have outranked all sorts of things that we thought were terribly important at the time just because cases become what US antitrust law is. That stimulates commentary and discussion, and follow-on cases and so on. So that’s point one. Point two is that, although Bill’s absolutely right that there are all sorts of incentives for action and such, I think it’s also fair to observe that there are also all sorts of incentives for not bringing particular cases which you are not willing to litigate. You know, the quiet life is certainly not helped by being told you need to go litigate 24/7 against some law firm that’s got all sorts of resources and fancy economists and things. And money is an issue: it’s going to cost a whole lot in terms of time and resources and outside experts. I must say that, in Ireland right now, one of the realities is that we have a merger case where we blocked the merger and it was taken to the High Court—this was before my time—and we lost. It’s now on appeal at the Supreme Court, and we are not only paying for all these barristers, we’re also on the hook for costs for the other side if we lose, and so you know that decision to risk litigation is really a massive financial risk for us, not to mention a big reputational risk if we lose. I remember that, with Staples/Office Depot, some were saying: ‘If we go into a courtroom we could have our butt handed to us, and this will really do some harm. It’s going to be costly, it’s going to be expensive and if we lose this will tarnish the agency’s reputation.’ So there were all sorts of incentives not to get out there and litigate. Whether that results in, if you will, cheap settlements, or whether it results in doing nothing at all, the point is that there are also some very powerful incentives against taking action, and in particular against assuming the risks of real litigation. Paul Lugard: It’s very hard to disagree with the position that agencies should focus on what they have accomplished in a past period because that will ultimately determine or co-determine the legitimacy of the agency’s enforcement actions. Now, if we agree that agencies should be under some type of obligation that they should explain the accomplishments to the general public and to stakeholders, the question is how we measure those accomplishments. DD

See http://www.cato.org/publications/commentary/antitrust-staplesoffice-depot-merger.

26

Part II: Legitimate Enforcement of Competition Law 347 Here there’s a difficulty because economists would in most cases disagree on the precise calculation of those benefits. Some would suggest that agencies tell the public: ‘We have generated 50 million euros’ worth of consumer surplus’. Other economists would say, ‘Well, your methodology is flawed or we simply don’t believe what you’re telling us’. But let’s assume that that problem can be solved. The other issue is how to know that the right choices have been made, because an agency may tell us: ‘We have generated that amount of surplus with the enforcement against a construction cartel’. But that same agency might have generated much more surplus if it had prosecuted a case against some dominant company for exclusionary conduct, or if it had intervened in the telecoms sector. So perhaps there should also be some mechanism that would explain to us what the agencies have not done, and why they have made that choice. That makes it extra complex, but maybe it could at the same time make the complexity more meaningful. I would add that perceived violations do not help the legitimacy of the agency because typically we wouldn’t know the harm to society of a hard-core cartel case. It may be a case that inflicts a lot of injury on society, but it may also be a case where the cartel had not operated efficiently, or maybe it has not even been implemented. A statement that the agency has imposed a fine of so many million simply does not help us to understand precisely how much welfare the agency has generated. So there are issues of measurement and choice, and the opportunities that the agency has foregone in taking enforcement actions. These issues have a direct relationship with legitimacy. Christine Parker: This discussion reminded me of the empirical work we’ve done trying to identify what makes an agency good in terms of responsive regulation theory. In the analysis of our surveys we have talked about strategic sophistication, and what we’ve found is that if I go and give a talk like this to the agencies everybody is really inspired and says ‘Yes, that’s good’, but when we go and actually ask the businesses we get a diversity of perceptions of whether it has actually been implemented or not. That doesn’t mean it’s not good to give the talk, but it still begs the question essentially because basically what you’re saying is that you need to be effective and legitimate, but the people hearing this are going to have different perceptions of what counts as effective or legitimate. In your slides you had a point about norm creation. I couldn’t find it in the paper just now, but obviously you understand what really has to happen here is that there has to be some sort of negotiation between all these different players about what legitimacy and effectiveness mean. So, it’s an interaction, and not just a game where an agency can unilaterally do the right thing and thereby create the right perceptions. Perceptions are created through a shared process. I disagree a little bit about why cases are important, for example. Cases are not just important as a deterrent signal, they are important as a legitimate, democratic means through which we create shared meaning about the law, potentially, or sometimes they may create DD

348 Part II: Legitimate Enforcement of Competition Law the opposite—that is, they may provoke opposing views about it. It would be great to talk about what other ways we have of creating shared meanings so that then these kinds of theories of what an agency could do would actually be perceived in a common way. Mario Siragusa: Very quickly, as we have heard, often agencies do not act on complaints, not because complains do not have merit but because of other priorities or because of a lack of resources. Now, to what extent should this choice be known and transparent? I say that because, very often, if you act as a plaintiff’s lawyer in an antitrust matter, generally you first try to involve an agency in your case, right? You file a complaint with an agency; if the agency does not react, then you go in front of a court, bringing that case to a judge. And very often what happens is the defendant will say: ‘Well, a complaint was filed, and the Commission was sleeping on this for a year’, or ‘The national authority has done nothing for months’. But then the judge is going to think, ‘Well if the administrative authority has not taken any action, why should I?’ So I have always wondered whether one could find a way where, if an agency decides not to act on a complaint because of a lack of resources or other priorities, this should somehow be communicated to the parties in order to avoid this very uncomfortable feeling. I have been in situations where contacting the authorities turned out to be a negative factor! I would have been better off going straight to the judge, rather than waiting six months first for the local authorities to do nothing. DD

Nazzini: I was at the OFT when John Fingleton implemented the system Bill referred to. And I remember Bill talking to us on more than one occasion. I agree with everything Bill said today, but I also see certain problems. First of all, in terms of prioritisation, we also have to distinguish among the different tasks of competition authorities. All that has been said about prioritisation works very well for policy work, for legislative changes, improving the framework, guidelines, and so on and so forth. When it comes to enforcement, first of all there is a preliminary issue I think has to be resolved: what is the role, and what is the function, of a competition authority? There can be many, but I see two. One is the role that I think Bill has in mind where you act in the general interest in order to prevent harm to total welfare. The other objective, which is in tension with the first one, is to resolve individual situations, and (oversimplifying) to act on complaints and to vindicate individual rights or to resolve issues as they arise. In a number of Member States the duty to act on complaints is precisely a reflection of that. Now, I am in favour of the first role of the competition authority, but then we have to have something to resolve anticompetitive issues that competition authorities cannot look at. This might be the case, for instance, with regard to some of the recent reforms in the UK that give more powers to the CAT through fast-track procedures DD

Part II: Legitimate Enforcement of Competition Law 349 open to small or medium-sized enterprises or consumers. And more broadly, private enforcement will make it easier to say competition authorities should have wide discretion to prioritise. The second point is just that there is also a need to prioritise the prioritisation exercise, because that exercise itself can be very costly and complex. First of all, you need information. You can’t just prioritise in the abstract and in a completely top-down manner; otherwise you are completely blind. So you need to gather market intelligence, information and so on. There may be consultation processes and evaluation processes, and this has to be done, I don’t know, every year? The OFT had an annual plan every year. Priorities did not change from one year to the next, but obviously this can be a very costly exercise. Lowe: Before Bill comes back, I’ll just make one or two remarks myself. Bill’s presentation is predicated on the agency having discretion, and every agency has a part of its work where it has no discretion. If there are legal or practical deadlines, whether it is merger control or leniency programmes, that’s going to take up a bit of your resources even before we start thinking about the overall objective. In areas not governed by deadlines, you have some discretion, but maybe you are also obliged to act at a certain point. And arguably we should have more legal deadlines, particularly on complaints— which would respond to Mario’s point. The second thing is that, when establishing objectives, if they’re too generic it’s going to be difficult to measure performance. They’ve got to be sufficiently related to benefits, presumably to consumers or to the business process; otherwise, it is quite difficult to monitor, and I am not really impressed by those colleagues who go to the finance committees of their parliaments and explain that this year they have generated six billion units of extra consumer welfare for their constituents because those figures can be contested. Nevertheless, there has to be some basic focus, as Bill says. And I tend to agree with Christine that it’s not entirely possible to define that focus in a strictly rational way without relating it to things that are going to capture the imagination of the public, which gives you legitimacy. Whether we like it or not, cases that tell stories confer legitimacy on public action. Whether it is construction cartels in the Netherlands or energy companies in Germany, you have to be prepared to explain your general objectives in relation to the specific portfolio of cases you’re dealing with. Otherwise an agency will not be able to explain to people how it has produced benefits for consumers. The next point is that, in the European Union, there is territory to be carved out; not in the imperialist sense, but in the sense of who is the bestplaced authority to do the job. I don’t mean just the Commission versus 28 national authorities, or national authorities between themselves. That is important too, of course. But the issue here is: what is the space between the competition authority and regulatory authorities? We now have in Spain a DD

350 Part II: Legitimate Enforcement of Competition Law unitary competition and regulatory authority, something many people may have some doubts about. In the Netherlands, we have an enlarged competition authority doing everything. One big advantage of those organisations, at least, is that they can say: ‘What is the best-placed authority to deal with this particular problem? Which one will deliver a satisfactory solution to this competition problem within a relevant time period?’ Antitrust takes a long time and legislation takes a long time, so sometimes the action of a regulator is the better solution. The capacity to actually work out who is doing what, and when, is something that has to be done nationally between the regulatory authorities and the competition authorities. And it has to be done between the European level and the national authorities. However, we could devote a whole day to this. So then, Bill? Kovacic: Thank you for giving me a whole day to reply to the excellent comments. [Laughter] Just a couple of replies, beginning with Konrad Ost’s comments: on cartels, yes, there are action-enforcing mechanisms in every system that reduce discretion. It was a conscious decision by the Department of Justice around 1993 to say: ‘What’s the best way to detect cartels? Let’s sweeten the prize for people to inform.’ That was a deliberate, conscious process of deciding how to go about improving a standard, mature product, namely, cartel enforcement. It was also a deliberate move to build a set of priorities around improvements in detection, and that’s how leniency has become effective. If they had not engaged in a process of looking at the programme and thinking of how it could be advanced, I doubt they would have gone beyond the 1970s version of leniency, which was very uncertain in its pay-off, and very incomplete. So even for a mature programme, I’d suggest that, even with action-enforcing deadlines, there is a place, as part of a regular setting of priorities, to ask: ‘How do we take this mature product and improve it dramatically?’ Gerald Barling points to an extremely important consideration. Isn’t there an irreducible minimum of cases that are crucial for legitimacy and effectiveness? I would say yes and yes. For effectiveness, if people don’t think you are going to be willing to enforce the law, the perimeter that you are defending starts to collapse around you, and the system loses its effectiveness. A colleague uses a navigation metaphor: if the ship isn’t going fast enough, turning the rudder makes no difference; you need a certain speed to guide the boat, you need a certain critical mass of activity. I don’t know what that is, but as an example, by having no abuse of dominance programme at all in the US, having a number of zero cases over a period of eight years, that was a setback. Zero is a number that is quite easy to understand. The system suffered, and there was a spillover that affected the FTC, when there was a possibility of doing things that were useful. So identifying the irreducible minimum is crucial, and that minimum is greater than zero—it has to be. J acqueline’s point relates to the habit of clearly articulating purposes, ends and rationales in the litigation of cases, as this facilitates effective judicial review. A clear statement DD

Part II: Legitimate Enforcement of Competition Law 351 of aims, a clear statement of rationales is a necessary element of transparency that makes judicial review effective. And that is entirely true. The discipline that comes from being able to go through that exercise is a useful way to make sure that the case is a good one. On Josh’s point, it helps to identify the cases that create political capital, that is valuable. But at the same time, one has to do it in a way that’s consistent with important economic performance goals. On the consumer protection side, when Tim Muris did the ‘Do not call’ programme, which limits telemarketing to the home, that put a huge amount of political capital in the bank that made it possible to fend off concerns that the FTC was not doing enough. The key is to find matters that lawmakerstakeholders will like. You point to those matters to show you are doing legitimate work, and exercising authority in a way that is consistent with consumer welfare. And by choosing matters wisely, an agency can acquire the political capital to say ‘No’ as well, when they press in the wrong direction. Coming to Stephen’s comment, you certainly have to have a portfolio of cases. What I call for is maybe a different approach in thinking about what the portfolio should be. How large should it be, what types of cases should be in it? The question is: are there incentives for the agency leadership to internalise the costs and benefits of bringing the cases? The internalisation pressures are greater for mergers because they are short term; they tend to fall within the tenure of the person who sets them in motion. If your merger case falls to pieces, you’re the chairman, or the bureau director, and that’s probably going to be on your score card. Abuse of dominance is a bit different, that is more like building—shall we say, aircraft carriers. There is more of a tendency to say: ‘I will get the credit for initiating the case, but my successors can clean up what happens later on’. So mergers and abuse of dominance are a little bit different. As Paul and others have mentioned, it’s hard to calculate the benefits, and that is undoubtedly true. But wouldn’t you be very nervous about any service or process where you ask an important public authority if they can you give you an idea of whether it works—and the answer is no, it’s too hard to figure it out? They say: ‘It’s too hard to calculate, but we assure you it’s very important, and we need more money to do it’. At some point, for purposes of legitimacy and accountability, you’d like to have a better approach. I think that greater transparency—and explanations about what the agency is leaving out, as Paul was suggesting—make lots of sense. On Christine’s comment, my approach in going to an agency is not saying ‘Here’s my menu’. My approach is to start by saying ‘Tell me what you do. Tell me how you decide what you do, tell me how you get these cases, tell me how you go about it, tell me the constraints that guide you in this direction.’ And then one may ask the agency: ‘You’ve been assembling things this way, but have you thought about this approach, or that approach?’ And in particular, since we’ve been talking about cases, among the things you can ask the agency is: ‘Did you ever think of putting those resources in a market study, instead of spending it on litigation?’ In many instances it may be a matter of adding other things to the case

352 Part II: Legitimate Enforcement of Competition Law agenda. I’d agree, if it is a shopping mall, cases are the anchor tenant—no anchor tenant, no mall. But then—no other shops and stalls, no mall either. So it’s a matter of diversification, and this is a matter of a shared discussion with the larger community. Fifteen years ago nobody was talking about evaluation. Evaluation now is part of John’s programme at the OECD. There are conferences now that look at it, and there is a big literature that focuses on it. That’s because a norm developed over time: a growing number of people began to say over and over that evaluation is a useful exercise. Two final thoughts. Renato, I agree completely with your concerns about the costs of doing prioritisation. But isn’t it also concerning if I said, ‘I’ve got so much to do. I’m going somewhere, I just can’t spend money to figure out where I am going. But I am going. I am going to go travelling but I just can’t spend the time to figure out where, or whether it’s the right place to go.’ I’d get very nervous if someone said, ‘I just can’t afford to figure out whether this direction is the right direction, and it has cost me too much, so I will just keep sailing in this direction’. Of course, you can’t spend all your time just thinking about what you’re going to do, but a degree of investment in that exercise has to be in the mix. And Philip, the questions you posed are precisely the questions I’d ask as part of the process of strategy-setting and prioritisation: who’s best suited to do this? Are they the right ones to make these types of calls? What is the right strategy? I agree again that, as a political scientist, which the good head of an agency has to be, you have to pick matters that have broad popular appeal, and if you’re not doing that as a conscious process, you’re falling behind. A good example is the Department of Justice after antitrust offences were reclassified from a misdemeanour to a felony in 1974. There was a deliberate decision to ‘popularise’ that process by bringing public procurement cases—where the defendant was someone, arguably, taking milk from the mouths of the little children in school, or taking uniforms off the backs of service personnel in farflung places, or taking roads from people who were in traffic jams. So the DOJ basically set out to make antitrust a popular policy. We’ll build this programme by picking initiatives that people understand. So that is a shining example of how priority-setting itself can be a highly effective tool. Lowe: Thank you, that covered a lot of ground. I don’t think the messages there are contradictory; you need clear focus, you need to use all the instruments available, you need to know whether you should pursue a case. And you need to consider different strategies: is a case appropriate or, as you say, should the problem be handled by way of a market investigation, or by a different public authority such as a regulator? You’ve got to measure what you do, but you also have to explain it in a way that makes sense and enhances your legitimacy. Now then, we will come back at 5 o’clock after a coffee. Thanks very much to Bill in particular for that stimulating talk. DD

Renato Nazzini* Parallel Proceedings in EU Competition Law: Rethinking Ne Bis In Idem as a Limiting Principle

1. The problem and the structure of the enquiry Articles 101 and 102 TFEU prohibit, under certain conditions, agreements, concerted practices and decisions of associations of undertakings and abuses of a dominant position, respectively. The effectiveness of these provisions rests on two pillars: public enforcement by the European Commission and the national competition authorities (NCAs); and private enforcement by claimants who consider themselves harmed by a breach of Article 101 or 102 TFEU. The focus of this chapter is on public enforcement, where the application of the ne bis in idem principle more frequently arises.1 Public enforcement of EU competition law is governed by Regulation 1/2003,2 which gives the Commission and NCAs the power to order that infringements of Articles 101 and 102 TFEU be brought to an end and to impose fines on undertakings that have committed an infringement. One of the objectives of the regime enacted by Regulation 1/2003 is to enhance the effectiveness of the enforcement of EU competition law by empowering the NCAs to apply Articles 101 and 102 alongside the Commission. Such a regime, however, gives rise to an acute risk of multiple proceedings against the same undertakings for the same infringement. This is for the following reasons: 1. Under Regulation 1/2003, the NCAs and the Commission all have concurrent jurisdiction to apply Articles 101 and 102 TFEU.3

* Professor of Law, King’s College London. I am grateful to Alice Galbusera, Silvia Massaro and Valerio Torti for their research assistance. 1 The application of the ne bis in idem principle in private enforcement is rare, but may be triggered, for example, by private law rules that allow the award of punitive damages when the defendant has already been punished for the same infringement by the Commission or an NCA. See Devenish Nutrition Ltd v Sanofi-Aventis SA (France) [2009] 3 All ER 27, [2008] EWCA Civ 1086, [2009] 3 WLR 198. 2 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. 3 See Article 5 of Regulation 1/2003.

354 Renato Nazzini 2. The powers of the NCAs and the Commission to impose fines are criminal in nature for the purpose of the ne bis in idem rules.4 3. There are no territorial limitations to the enforcement powers of the Commission or the NCAs in Regulation 1/2003. This means that, in principle, competition authorities in the EU have the power, under Regulation 1/2003, to investigate cases and impose fines for effects beyond their national markets.5 4. NCAs do not have the power to make a final non-infringement decision.6 5. NCAs retain the power to apply national law alongside EU law.7 6. In competition law, the EU Courts apply a threefold test to the definition of the same offence for the purposes of the ne bis in idem principle, requiring the identity of the facts, the identity of the offender and the identity of the legal interest protected.8 The requirement of the identity

4 The European Court of Human Rights adopted a substantive test for the definition of the autonomous meaning of ‘criminal charge’ under Article 6 of the European Human Rights Convention in Case No 5100/71, Engel v Netherlands, 8 June 1976. Administrative proceedings under Community law are considered criminal for the purposes of Article 6 of the European Human Rights Convention. See Case C-235/92 P Montecatini Spa v Commission [1999] ECR I-4539, paras 175–176. In principle, the same applies to proceedings in the Member States. In the UK, for example, see Napp Pharmaceutical Holdings Ltd v Director General of Fair Trading (Napp No 3) [2001] Comp AR 33, [2002] ECC 3, paras 68–76; Napp Pharmaceutical Holdings Ltd v Director General of Fair Trading (Napp No 4) [2002] ECC 13, paras 98–103. In Italy, see Case No 43509/08, Menarini Diagnostics srl v Italy, 27 September 2011. 5 This results from a number of considerations, including the following: (i) there is nothing in Regulation 1/2003 that limits in any way the territorial scope of the infringements that may be investigated and punished by national competition authorities; (ii) the Commission Notice on cooperation within the Network of Competition Authorities, [2004] OJ C101/43 envisages instances in which a single national competition authority is well placed to act in relation to infringements the effects of which are not limited to its own territory (see para 11 in particular); (iii) certain national competition authorities consider that they have the power to impose fines taking into account effects outside their own territory (see OFT Guidance as to the appropriate amount of a penalty, OFT423, September 2012, para 2.10, where the OFT states that it will take account of such effects through the its assessment of relevant turnover but when the relevant turnover includes foreign turnover and the market is wider than the United Kingdom the OFT will take such foreign turnover into account only with the express consent of the relevant Member State or national competition authority, as appropriate). There may, of course, be limitations on the jurisdiction of national competition authorities resulting from their own national law and, possibly, from general principles of EU law or public international law. A study of these limitations and their implications under EU law is outside the scope of this chapter. What matters for present purposes is that national competition authorities have the power, in principle, to investigate and punish infringements of EU competition law that have effects beyond their national territory and may take such effects into account when imposing fines on undertakings. 6 Case C-375/09 Prezes Urzędu Ochrony Konkurencji i Konsumentów v Tele2 Polska sp z oo, now Netia SA [2011] ECR I-3055. 7 Case 14/68 Walt Wilhelm v Bundeskartellamt [1969] ECR 1; Case C-17/10 Toshiba Corporation v Úřad pro ochranu hospodářské soutěže, EU:C:2012:72. 8 Joined Cases C-204/00 P, C-205/00 P, C-211/00 P, C-213/00 P, C-217/00 P and C-219/00 P Aalborg Portland A/S v Commission [2004] ECR I-123, para 338.

Parallel Proceedings in EU Competition Law 355 of the legal interest protected rules out the application of the ne bis in idem principle whenever the first and the second offence protect different legal interests. Since no other offences in EU or national law arguably protect the same interest as Articles 101 and 102 TFEU, the result is that the ne bis in idem principle only applies when the same undertaking is prosecuted or punished again for the same violation of EU competition law which is already the subject matter of a final decision on the merits by the Commission or an NCA concerning the same undertaking. 7. Regulation 1/2003 does not prohibit parallel proceedings by the Commission under EU law and the NCAs under national law. Article 3 of Regulation 1/2003 obliges NCAs to apply Articles 101 and 102 TFEU whenever they apply national competition law to conduct that has an effect on trade between Member States. Furthermore, save for the application of stricter unilateral conduct rules under national law, any conflict between EU and national law is expressly prohibited. This provision must be read against the background of the case law of the Court of Justice, which, since 1969, has allowed parallel proceedings by the Commission under EU law and by NCAs under national law provided that the primacy of EU law is respected and, if multiple sanctions are imposed on the same person under EU and national laws, the previous punishment is taken into account in determining the second sanction.9 8. When a competition authority has established an infringement and imposed fines on an undertaking, further proceedings are not barred if the previous decision is annulled by a court on procedural grounds. The case law concerning decisions of the Commission is clear: when a decision has been set aside by the EU Courts because of a procedural defect, there is no final conviction any more. No decision on the merits exists because the original decision has been annulled on procedural, rather than substantive, grounds and such an annulment is, therefore, not an acquittal.10 Thus, competition authorities can make serious procedural mistakes affecting the validity of an infringement decision without losing the power to resume the proceedings after the decision has been set aside by the courts.11 Furthermore, it may not always be clear when a decision has been set aside on procedural grounds and when there has been a decision on the merits instead. In a pending case, the applicant is challenging

Walt Wilhelm, above n 7, para 11. Joined Cases C-238/99 P, C-244/99 P, C-245/99 P, C-247/99 P, C-250/99 P to C-252/99 P and C-254/99 P Limburgse Vinyl Maatschappij and Others v Commission [2002] ECR I-8375, paras 59–62; Case T-24/07 ThyssenKrupp Stainless AG v Commission [2009] ECR II-2309, para 190. 11 The Commission normally readopts cartel decisions that have been annulled by the EU courts on procedural grounds: Gas insulated switchgear (summary decision) [2013] OJ C70/12; Alloy surcharge (summary decision) [2007] OJ L182/31; Steel beams (summary decision) [2008] OJ C235/4; PVC [1994] OJ L239/14. 9

10

356 Renato Nazzini the Commission’s decision to reopen cartel proceedings after its previous decision was partially set aside by the General Court. The applicant does so on the ground that the reopening of the proceedings infringes the ne bis in idem principle.12 9. As regards decisions in non-EU Member States, the ne bis in idem principle would appear not to apply at all, and the EU authorities are not required to take into account the penalties imposed in non-EU Member States.13 As a result, the protection of undertakings against multiple sanctions for the same conduct is left entirely to the exercise of self-restraint by competition authorities and courts around the world, ensuring that fines imposed in each given jurisdiction reflect only the harm caused by a global cartel in that jurisdiction. This system of EU and international concurrent jurisdictions finds a limit in the ne bis in idem principle, which, at the general level, may be described as prohibiting further prosecution or punishment of a person who has already been finally acquitted or convicted of the same offence. In EU competition law, the principle of ne bis in idem is a fundamental right recognised in Article 50 of the EU Charter of Fundamental Rights (EU Charter),14 which applies to EU institutions as well as to the Member States when they are implementing Union law.15 Even before the EU Charter was given the same legal value as the Treaties by the Treaty of Lisbon,16 the principle of ne bis in idem was a general principle of EU law17 and, as such, continues to apply alongside Article 50 EU Charter18 even if it has probably lost most of its autonomous 12 See Case T-240/12 Eni SpA v Commission [2012] OJ C217/27. Following the completion of this chapter, the General Court ruled by Order that the matter was devoid of object insofar as the Commission had in the end decided not to reopen proceedings. See Joined Cases C-240/12 and T-211/13 Eni SpA v Commission, EU:T:2014:132. 13 Case C-289/04 P Showa Denko KK v Commission [2006] ECR I-5859; Case C-308/04 P SGL Carbon AG v Commission [2006] ECR I-5977; Case C-328/05 P SGL Carbon AG v Commission [2007] ECR I-3921. 14 Article 50 EU Charter forbids a second criminal prosecution or punishment if a person has already been finally convicted or acquitted of the same offence within the Union in accordance with the law. 15 Article 51(1) EU Charter. The scope of application of the EU Charter is given a broad interpretation. See Case C-617/10 Åklagaren v Hans Åkerberg Fransson, EU:C:2013:105, paras 16–31. 16 Article 6(1) of the Treaty on the European Union, [2012] OJ C326/01. 17 The principle was expressed as precluding a person from being subject to proceedings or punished a second time for unlawful conduct in respect of which he has already been punished or exonerated by a previous decision that is no longer amenable to challenge. See Case T-224/00 Archer Daniels Midland v Commission [2003] ECR II-2597, paras 85–86; Joined Cases T-271/03 and T-245/03 Fédération nationale des syndicats d’exploitants agricoles (FNSEA) v Commission [2004] ECR II-271, para 340; Case T-39/06 Transcatab SpA v Commission [2011] ECR II-6831, para 254; Toshiba, above n 7, para 94. 18 Wouter Wils, ‘EU Antitrust Enforcement Powers and Procedural Rights and Guarantees: The Interplay between EU Law, National Law, the Charter of Fundamental Rights of the EU and the European Convention on Human Rights’, 34 World Competition 189, 207 (2011).

Parallel Proceedings in EU Competition Law 357 significance. The EU fundamental right not to be tried or punished twice must be interpreted in the light of the principle of ne bis in idem in Article 4, Protocol 7 to the European Convention on Human Rights and Fundamental Freedoms (ECHR).19 Well before the EU Charter came into force, the EU Courts have referred to the ECHR right not to be tried or prosecuted twice as reflecting a fundamental principle of EU law.20 Furthermore, there are other ne bis in idem rules in EU law, including, in particular, Chapter 3 of the Convention implementing the Schengen Agreement (CISA)21 and Article 3(2) of the Council Framework Decision of 13 June 2002 on the European arrest warrant and the surrender procedures between Member States, as amended (Framework Decision on the EAW).22 The case law of the Court of Justice on these provisions may be of persuasive authority in interpreting Article 50 EU

19 Convention for the Protection of Human Rights and Fundamental Freedoms (Rome, 4 November 1950), Protocol No 7, Article 4. This provision—Article 4, Protocol 7 to the ECHR—provides in the first paragraph for the right of a person not ‘to be tried or punished again in criminal proceedings under the jurisdiction of the same State for an offence for which he has already been finally acquitted or convicted in accordance with the law and penal procedure of that State’, thus limiting the geographical application of the principle to a single jurisdiction. Furthermore, paragraph 2 of Article 4 sets out the circumstances in which further proceedings and a second punishment are admissible, namely if, ‘in accordance with the law and penal procedure of the State concerned’, there is evidence of new or newly discovered facts or there has been a fundamental defect in the previous proceedings, which could affect the outcome of the case. Not all Member States have ratified Protocol 7, but this has not prevented the EU courts from referring to it as a ‘reference point’ or a source of inspiration of the general principle of EU law of ne bis in idem. See now, eg Article 52(2) EU Charter, which expressly provides that: ‘In so far as this Charter contains rights which correspond to rights guaranteed by the Convention for the Protection of Human Rights and Fundamental Freedoms, the meaning and scope of those rights shall be the same as those laid down by the said Convention. This provision shall not prevent Union law providing more extensive protection.’ 20 Limburgse Vinyl Maatschappij, above n 10, para 59; see also Wouter Wils, ‘The Principle of Ne Bis in Idemin EC Antitrust Enforcement: A Legal and Economic Analysis ’, 26 World Competition 131, 133 (2003). For a general overview of the problems arising from the application of the European Human Rights Convention to administrative proceedings by the Commission, see Alan Riley, ‘The ECHR Implications of the Investigation Provisions of the Draft Competition Regulation’, 51 International and Comparative Law Quarterly 55 (2002); Richard Wainwright, ‘Human Rights: What Have They to Do with Competition Law?’, in Mads Andenas, Michael Hutchings and Philip Marsden, eds, Current Competition Law Vol III, BIICL, 2005, 473 et seq. 21 Convention implementing the Schengen Agreement of 14 June 1985, [2000] OJ L239/19. 22 Council Framework Decision of 13 June 2002 on the European arrest warrant and the surrender procedures between Member States [2002] OJ L190/1, amended by Council Framework Decision 2009/299/JHA of 26 February 2009 amending Framework Decisions 2002/584/JHA, 2005/214/JHA, 2006/783/JHA, 2008/909/JHA and 2008/947/JHA, thereby enhancing the procedural rights of persons and fostering the application of the principle of mutual recognition to decisions rendered in the absence of the person concerned at

358 Renato Nazzini Charter and the general principle of ne bis in idem. Finally, the principle of ne bis in idem is also recognised in public international law but, unless an international convention provides otherwise, its application is limited to a second prosecution or punishment within the same state.23 Against the background of the risk of multiple proceedings and decisions against the same undertakings for the same competition law infringements, the thesis in this chapter is that discretionary prosecutorial restraint is not a sufficient safeguard against such a risk, whereas the principle of ne bis in idem can effectively perform the function of limiting vexatious and inefficient exercise of concurrent jurisdiction in competition matters within the Union. This limiting function of the ne bis in idem principle is analysed by focusing on the six elements by which the ne bis in idem principle is capable of regulating the exercise of concurrent jurisdiction within the EU: (i) what constitutes the same offence for the purpose of the principle of ne bis in idem in competition law; (ii) whether a prosecuting authority is allowed to bring separate proceedings concerning different elements of the same offence, for example by considering the same conduct as constituting separate offences, depending on the Member State where the anticompetitive effects occur; (ii) what types of decision bar a second prosecution or punishment; (iv) the consequences of the first acquittal or conviction being set aside on appeal or judicial review for the application of the principle of ne bis in idem; (v) the application of the EU principle of ne bis in idem in the relationship between EU law and national law proceedings within the Union; (vi) the application of the EU principle of ne bis in idem in an international context, namely when the first decision was made in a non-EU state.

2. Inadequacy of discretionary prosecutorial restraint as a limiting principle As explained in the previous section, the basic features of the EU competition enforcement system are, almost by design, conducive to multiple proceedings and multiple decisions. Any abuse may, of course, be avoided as a matter of prosecutorial restraint. Thus, within the Union, widespread abuse of

the trial, [2009] OJ L81/24, Article 3(2), which provides for a mandatory ground of nonexecution of a European arrest warrant ‘if the executing judicial authority is informed that the requested person has been finally judged by a Member State in respect of the same acts provided that, where there has been sentence, the sentence has been served or is currently being served or may no longer be executed under the law of the sentencing Member State’. 23 Opinion of Advocate General Tizzano in Case C-397/03 P Archer Daniels Midland Co v Commission [2006] ECR I-4429.

Parallel Proceedings in EU Competition Law 359 the system has been avoided by a sensible administrative allocation of cases within the European Competition Network (ECN),24 whereby the NCAs and the Commission aim to identify the authority best placed to act on each given case, thus avoiding multiple proceedings and, by implication, multiple fining decisions for the same conduct.25 However, prosecutorial restraint cannot be the answer to the problem of multiple prosecutions and decisions, particularly because the Commission and NCAs enjoy wide and largely uncontrollable discretion in allocating cases, defining the scope of the investigations and launching a second investigation, even if the same or another authority has already investigated the same conduct. First, instead of using the ECN case allocation mechanism to avoid duplication of proceedings, the Commission and NCAs may find it expedient, on efficiency grounds, to allocate different aspects of what may constitute a single infringing conduct to different authorities. One example of this practice may be the consumer textile detergents cartel, where it appears that the Commission and the French competition authority may have brought proceedings against the same parties concerning closely connected behaviours that could have amounted to a single overall infringement. The issue came to light because, in the French proceedings, Henkel applied to the Commission for permission to use certain documents obtained in the Commission investigation in order to prove that the two sets of proceedings were closely connected, but its application was rejected and the decision upheld by the General Court.26 It appears that a similar issue could arise in the Commission’s second probe into the LCD market. The second investigation would appear to relate to LCD screens for mobile devices, whereas the first decision concerns LCD screens for television sets and computer monitors. However, the meetings in which the alleged cartel conduct took place were possibly the same as regards both types of LCD screens.27 Finally, in July 2013, Orange made an application to the General Court challenging the validity of a decision by the Commission ordering Orange to submit to an inspection under Article 20(4) of Regulation 1/2003, claiming, inter alia, that the practices investigated by the Commission were very similar to conduct investigated by the French

24 The ECN is a network of public authorities formed by the Commission and the national competition authorities for the purposes, inter alia, of coordinating the enforcement of Articles 101 and 102 TFEU under Regulation 1/2003. The structure and functioning of the ECN is set out in the Commission Notice on cooperation within the Network of Competition Authorities, [2004] OJ C101/43 (Network Notice). 25 Network Notice, ibid, paras 5–42. 26 Case T-64/12 Henkel AG & Co KGaA and Henkel France v Commission, EU:T:2013:116. 27 Case T-128/11 LG Display v Commission, [2011] OJ C130/10. The issue of double jeopardy is raised under the fourth plea. Following the completion of this chapter, the General Court declined to accept the double jeopardy argument. See Case T-128/11 LG Display v Commission, EU:T:2014:88, para 242; appeal dismissed: Case C-227/14 P, EU:C:2015:258.

360 Renato Nazzini c ompetition authority only nine months earlier and in respect of which the latter authority found no evidence of infringement.28 While the Commission and the NCAs enjoy wide discretion at case allocation stage, with the power to consolidate or subdivide investigations as they find most expedient, parties have no or very limited scope to challenge case allocation decisions.29 As a result, a challenge can be made only ex post on ne bis in idem grounds. Indeed, the absence of the parties’ involvement in the case allocation procedure and of a right to challenge case allocation decisions calls for stronger ne bis in idem rules to regulate multiple proceedings and fines ex post. Furthermore, case allocation and decisions as to the scope of a case occur when an investigation commences and a case is notified to the ECN. There is no mechanism within the ECN to bar a second investigation once the first case has been closed. This problem has been exacerbated by the ruling of the Court of Justice in Tele2 Polska that NCAs do not have the power to ‘acquit’ an undertaking by way of a final non-infringement decision.30 Similarly, Regulation 1/2003 does not impose a duty on the Commission to make a non-infringement decision if an investigation has not established sufficient evidence to prove an infringement. Article 10 of Regulation 1/2003 provides for the discretionary power of the Commission to make a finding of inapplicability when the Union public interest so requires, but no such finding has been made so far and, in any event, the decision whether to do so is highly discretionary. On the other hand, Article 7 of Commission Regulation 773/2004 provides for the power of the Commission to reject a complaint, but such a decision is not a finding that there has been no infringement31 and, on the basis of a purely formalistic application of what constitutes a previous conviction, it would appear that it does not necessarily bar further proceedings.32 Therefore, the same undertaking can be investigated time and again in the Union by NCAs and the Commission and, as long as no sufficient evidence is

28 Case T-402/13 Orange v Commission [2013] OJ C313/28. The General Court subsequently dismissed the appeal and rejected Orange’s ne bis in idem argument on the basis of an NCA’s inability (given the ECJ’s judgment in Tele2 Polska, above n 6) to adopt a ‘constitutive’ decision finding that no infringement has occurred. See Case T-402/13 Orange v Commission, EU:T:2014:991, especially paras 30–31; see also the discussion below in the main text. 29 Nicholas Khan, Kerse & Khan on EU Antitrust Procedure, 6th edn, Sweet & Maxwell, 2012, 50–51. 30 See above n 28. 31 Commission Regulation 773/2004/EC of 7 April 2004 relating to the conduct of proceedings by the Commission pursuant to Article 81 and 82 of the EC Treaty, [2004] OJ L23/18. 32 On this general proposition, see Khan, EU Antitrust Procedure, above n 29, 86. The case law appears to be of more limited scope. See Case C-279/95 P Langnese-Iglo GmbH v Commission [1998] ECR I-5609, para 30 (in relation to a comfort letter); Case T-241/97 Stork Amsterdam v Commission [2000] ECR II-309, para 80 (in relation to a complaint rejected for lack of sufficient Union interest).

Parallel Proceedings in EU Competition Law 361 found or the authority decides not to make an infringement decision, further proceedings are not barred. It is clear, therefore, that even in an integrated system such as the ECN, prosecutorial restraint cannot be the answer to the problem of multiple investigations and convictions. Even less so in the international context, where there is no formal mechanism for allocating cases and coordinating investigations, and no rule of public international law that imposes on the Commission or an NCA a duty not to punish or investigate an undertaking that has already been finally acquitted or convicted for the same infringement in a non-EU country, or even to take into account previously imposed penalties.33 The only justiciable and, at least in principle, predictable way of limiting the proliferation of proceedings and decisions concerning the same competition infringements against the same undertakings is a robust and coherent application of the ne bis in idem principle. This function of the ne bis in idem principle is, however, currently impaired by its overly restrictive judicial interpretation in competition matters.

3. Ne bis in idem as a limit to parallel proceedings 3.1 Definition of the same offence The key governing concept for the application of the ne bis in idem rules is the definition of what constitutes the same offence. In competition law, the EU Courts apply a threefold test to the definition of the same offence, requiring the identity of the facts, the identity of the offender and the identity of the legal interest protected.34 The requirement of the identity of the legal interest protected comes into play to rule out the application of the ne bis in idem principle to offences committed by the same person and arising out of the same facts in different legal systems, as different legal systems are assumed to protect competition within their own national markets,35 or in violation of rules aimed at achieving different regulatory or policy objectives, such as, for example, competition rules, on the one hand, and rules in the areas of

33 Opinion of Advocate General Geelhoed in Case C-301/04 P Tokai Carbon v Commission [2006] ECR I-5915, paras 46–52; Opinion of Advocate General Mazák in SGL Carbon, above n 13, paras 39–41; Opinion of Advocate General Geelhoed in Showa Denko, above n 13, paras 74–75; Opinion of Advocate General Geelhoed in SGL Carbon, above n 13, paras 46–52; Opinion of Advocate General Tizzano in Archer Daniels Midland, above n 23, paras 86–115. 34 Aalborg Portland, above n 8, para 338. 35 Case C-150/05 Jean Leon Van Straaten v Staat der Nederlanden [2006] ECR I-9327.

362 Renato Nazzini telecommunications regulation36 or patent law,37 on the other. The Grand Chamber of the Court of Justice confirmed the requirement of the identity of the legal interest protected in Toshiba,38 even though Advocate General Kokott forcefully invited the Court to adopt a criterion based only on the identity of the facts, which the Court itself applies in other areas of EU law39 and which the European Court of Human Rights applies under Article 4, Protocol 7 to the ECHR.40 The Advocate General relied strongly, albeit implicitly, on the identity of the rationales for the ne bis in idem principle in different areas of EU law and in Article 4, Protocol 7 to the ECHR. First, the Advocate General emphasised the shared free movement rationale between the ne bis in idem principle in EU competition law and the same principle in Article 54 CISA and under the Framework Decision on the EAW.41 Then, she pointed to the need to ensure that the protection under EU law did not fall short of the minimum standard of protection under Article 4, Protocol 7 to the ECHR.42 Thus, it is the protection of the fundamental right of the undertakings concerned that requires homogeneity of results under EU law and Article 4, Protocol 7 to the ECHR. Such a protection cannot be lower under EU law than it is under the ECHR system. The Opinion of the Advocate General in Toshiba sets out the correct approach, and it is to be regretted that the Court of Justice refused to follow it and even confirmed the requirement of the identity of the legal interest, although it was not required to do so to decide the case. Indeed, the requirement of the identity of the legal interest protected reduces significantly the scope of protection enjoyed by the defendant and runs counter to the rightsbased foundation of the ne bis in idem principle. It is for this reason that the

36 Undertakings may be subject to telecommunications regulation as well as competition law. See, eg Case C-280/08 P Deutsche Telekom AG v Commission [2010] ECR I-9555, paras 56–110, where the applicant argued that the relevant telecommunications regulation did not leave it any scope to determine its competitive conduct autonomously and that its compliance with telecommunications regulation gave rise to a legitimate expectation that there was no breach of Article 102 TFEU in the circumstances. Conversely, if national telecommunications regulatory authorities had sanctioned the applicant for conduct that also amounted to margin squeeze, a ne bis in idem issue could have arisen. 37 See, eg Case C-457/10 AstraZeneca AB v Commission, EU:C:2012:770, paras 74–100, where the dominant undertaking could in theory have been exposed to fines under national law for providing false or misleading information to patent offices, which was also found to be an infringement of Article 102 TFEU. 38 Toshiba, above n 7, para 97, even if the Court then decided the case on the ground that the facts were not identical. See paras 98–103. 39 Van Straaten, above n 35, paras 48 and 53; Case C-436/04 Leopold Henri Van Esbroeck [2006] ECR I-2333; Case C-288/05 P Jürgen Kretzinger [2007] ECR I-6441, paras 29–31; Case C-367/05 Norma Kraaijenbrink [2007] ECR I-6619, paras 23–26. 40 Case No 14939/03, Zolotukhin v Russia, 10 February 2009. 41 Opinion of Advocate General Kokott in Toshiba, above n 7, EU:C:2011:552, paras 116–118. 42 Ibid, paras 119–123.

Parallel Proceedings in EU Competition Law 363 European Court of Human Rights rejected this requirement under Article 4, Protocol 7 to the ECHR.43 The European Court of Human Rights reached this conclusion notwithstanding the use of the word ‘offence’ in Article 4, Protocol 7,44 emphasising that the approach based on the legal characterisation of the offence ‘is too restrictive of the rights of the individual’ and ‘risks undermining’ the principle of ne bis in idem.45 These considerations apply with equal force to the principle of ne bis in idem under EU competition law. The EU Courts, of course, are not bound by the case law of the Strasbourg Court; they can depart from it if there is a good reason to do so. It is difficult to see, however, why the protection of the ne bis in idem principle should be lower under EU competition law than under the ECHR system. On the contrary, because of the internal market rationale for the EU principle of ne bis in idem, the need for a conduct-based interpretation of the concept of ‘offence’ is even stronger. The reason is clear: the characterisation of an offence is, by definition, different in different legal systems, as the legal interest protected changes when considered from the perspective of different sovereign states.46 This is why international definitions of the ne bis in idem principle, such as the one in Article 4, Protocol 7 to the ECHR or in Article 14(7) of the International Covenant on Civil and Political Rights, prohibit multiple prosecutions and convictions within one and the same state: it is inconceivable that one sovereign state could forgo its right to prosecute and punish offences under its own law because they are an infringement of its own legal order and, therefore, different in nature from offences under the law of other states that protect a different legal interest. But to extend this approach to the EU-wide principle of ne bis in idem would be to deprive it of most if its effectiveness in a cross-border context. The Court of Justice’s approach to competition cases thus results in a market fragmentation which, in other areas, the Court itself has consistently held to be incompatible with one single internal market. As Advocate General Ruiz-Jarabo said in his opinion in Van Esbroeck, ‘if, instead of the acts alone, account were taken of the offences or of the rights protected by the prohibition of the said acts, the ne bis in idem principle would never function at international level’.47 This is the reason why, in Van Esbroeck, the Court of Justice decided that Article 54 CISA applies regardless of the legal classification of the offence.48 The only requirement is the identity of material facts, understood as ‘a set of facts which are inextricably linked together in

Zolotukhin, above n 40, para 82. Ibid, para 80. 45 Ibid, para 81. 46 A point rightly made by the Court of Justice in Kretzinger, above n 39, para 31. 47 Opinion of Advocae General Ruiz-Jarabo Colomer in Van Esbroeck, above n 39, para 47. 48 Van Esbroeck, above n 39, para 38; Case C-467/04 Giuseppe Francesco Gasparini [2006] ECR I-9199, para 56; Van Straaten, above n 35, para 52. 43 44

364 Renato Nazzini time, in space and by their subject-matter’.49 As the Court rightly held, following the Advocate General, ‘a criterion based on the legal classification of the acts or on the protected legal interest might create as many barriers to freedom of movement within the Schengen territory as there are penal systems in the Contracting States’. This conduct-based definition is also consistent with the Framework Decision on the EAW,50 which provides that, for offences other than those listed in Article 2, paragraph 2 (essentially, a list of serious offences where dual criminality is presumed), ‘surrender may be subject to the condition that the acts for which the European arrest warrant has been issued constitute an offence under the law of the executing Member State, whatever the constituent elements or however it is described’. Clearly, the EAW adopts a conduct-based definition of same offence. The express exclusion from what constitutes an offence under the law of the executing Member States of the ‘constituent elements’ of the offence or its description demonstrate the precise, unequivocal intention to avoid an approach based on the legal characterisation of the offence and, even less, on the legal interest protected. Why this reasoning should not apply to ne bis in idem as a general principle of EU law or to Article 50 EU Charter is difficult to understand. The only principled argument that could be made in this respect is that Article 50 EU Charter refers to the ‘offence’ rather than to ‘same acts’ as Article 54 CISA. The Court of Justice in Van Esbroeck did rely on this literal argument.51 However, the arguments based on ensuring that the principle of ne bis in idem is given its full effect as a fundamental right and on the freedom of movement rationale are much more powerful and should be capable of overcoming the literal argument. This is even more so given that the use of the term ‘offence’ is in itself not conclusive: it is a neutral term that begs the question of its meaning in designing the boundaries of the ne bis in idem principle rather than answering it. The European Court of Human Rights in Zolotukhin had no difficulty in interpreting the term ‘offence’ under Article 4, Protocol 7 to the ECHR as a set of facts rather than a legal construct. In conclusion, the correct position under Article 50 EU Charter and the general EU law principle of ne bis in idem is that two infringements are the same offence when they are constituted by the same or inextricably linked facts. The case law of the EU Courts on the principle of ne bis in idem, which still requires the identity of the legal interest protected, should be reviewed and brought into line with the case law of the Court of Justice on Article 54

Kraaijenbrink, above n 39, paras 23–26; Kretzinger, above n 39, paras 29–31. Council Framework Decision 2002/584/JHA of 13 June 2002 on the European arrest warrant and the surrender procedures between Member States, [2002] OJ L190/1. 51 Van Esbroeck, above n 39, paras 27–28. 49 50

Parallel Proceedings in EU Competition Law 365 CISA and the jurisprudence of the European Court of Human Rights on Article 4, Protocol 7 to the ECHR.

3.2 Fragmentation of the same offence A question related to the definition of what constitutes the same offence is whether competition authorities can subdivide a set of facts which are inextricably linked together in time, in space and by their subject matter into several infringements of competition law which can be separately prosecuted and punished. If the separate violations arise from facts which are inextricably linked together in time, in space and by their subject matter, the ne bis in idem principle could bar a second prosecution or punishment even if the facts that were the subject matter of the previous decision clearly and unequivocally did not include the facts that are the subject matter of the new investigation. The classic example in which this could occur is an EU-wide cartel where the conduct, while consisting of individual agreements or concerted practices, has the same anticompetitive object and amounts, therefore, to a single overall infringement52 but which the Commission and the NCAs choose to prosecute as a set of separate infringements based on the countries where the effects of the conduct took place or the different products involved. Reasons of efficiency may favour this approach to prosecuting complex cartels but the question is whether an undertaking is entitled to the protection of the ne bis in idem principle whenever the offence is a single one, concerning inextricably linked facts. The right-based foundation of the ne bis in idem principle appears to require a strict application of this safeguard whenever a person is prosecuted again for the same offence. The same offence is defined as the same sets of facts or inextricably linked facts. This test is an objective one: the Court of Justice ruled that it does not depend on the subjective intention of the perpetrator.53 But equally, the test cannot depend on the discretion of the prosecuting authorities. If the prosecuting authorities were free to decide, in the definition of an offence, which facts are inextricably linked together and which are not, the guarantee of ne bis in idem would become totally ineffective. The argument follows, a fortiori, from the reasoning of the Court of

52 Case C-441/11 P Commission v Verhuizingen Coppens, EU:C:2012:778; Case T-15/89 Chemie Linz v Commission [1992] ECR II-1275, para 308; Case C-49/92 P Commission v Anic Partecipazioni SpA [1999] ECR I-4125, para 113; Joined Cases T-25 to 39, 42 to 46, 48, 50 to 65, 68 to 71, 87, 88, 103 and 104/95 Cimenteries CBR v Commission [2000] ECR II-491, paras 4025 et seq. 53 Kraaijenbrink, above n 39, para 29.

366 Renato Nazzini Justice in relation to Article 54 CISA54 and the European Court of Human Rights55 in relation to Article 4, Protocol 7 to the ECHR. The definition of what constitutes the same offence cannot depend on the legal classification of the offence, otherwise the national legislature would be free to limit the effectiveness of the principle as he sees fit. But if this applies to the legislature, the same applies with even more force to the prosecuting authorities, who have a strong incentive to prosecute cases and would not be bound, in the definition of the offence, by anything other than their own discretion. The freedom of movement foundation of the ne bis in idem principle in EU law also suggests that fragmentation of the same offence should not be allowed. If a person must be free to exercise his right of freedom of movement without the fear of being prosecuted or punished again for the same offence in different Member States, the principle should apply to all offences across all Member States that arise from the same or inextricably linked facts. Different legal characterisations or the possibility of dividing a single offence depending, for instance, on where the effects of the conduct are felt would make the ne bis in idem principle devoid of purpose and run directly counter to the freedom of movement rationale. Nor does it appear that the effective enforcement of EU competition law requires that competition authorities be given such a high degree of discretion to conduct multiple prosecutions and take multiple decisions concerning inextricably linked facts. The Commission and NCAs are well equipped to pursue single continuous infringements or infringements that have effects in several Member States. Indeed, Regulation 1/2003 and the Notice on cooperation within the network of EU competition authorities provide for specific mechanisms of case allocation, coordination and investigation designed to facilitate a single prosecution of multi-jurisdictional56 or complex infringements.57 Mere reasons of expediency and efficient resource allocation within the ECN could never result in the practical obliteration of an undertaking’s right not to be prosecuted or punished twice, which would be the inescapable consequence if competition authorities were permitted to define the factual scope of the offence at their discretion. The test for when the facts are the same or inextricably linked must be applied in the circumstances of each individual case. However, two areas are probably sufficiently typified in the case law. First, whenever a set of facts gives rise to a single overall infringement, there is a single offence under Article 101 TFEU. The very test for the establishment of a single continuous

54 Van Straaten, above n 35, paras 48 and 53; Van Esbroeck, above n 39; Kretzinger, above n 39, paras 29–31; Kraaijenbrink, above n 39, paras 23–26. 55 Zolotukhin, above n 40, para 16. 56 See Article 22(1) of Regulation 1/2003. 57 See Articles 17–21 and 22(2) of Regulation 1/2003.

Parallel Proceedings in EU Competition Law 367 infringement requires that the facts are closely connected and, therefore, inextricably linked for the purposes of the ne bis in idem principle.58 Second, whenever the effects of anticompetitive conduct are caused by the same behaviour or are dependent on each other, the facts are inextricably linked and there is, therefore, a single offence even if the effects in question are felt in different Member States. This is the case, for example, when the parties agree at the same meetings to raise the price of their products in several Member States or when the parties agree to raise the price of an input sold in one Member State, which has the further effect of significantly raising the price of an output which is sold in several Member States. Finally, the application of the test presupposes that the first prosecuting authority had jurisdiction in relation to all the elements of the conduct in question. No preclusion can arise if there was no jurisdiction to decide a matter and no issue of the application of the ne bis in idem principle can be even conceivable if a person has never been in jeopardy of being convicted. Therefore, single continuous infringements affecting new Member States can be prosecuted under national law for the period of time when Article 101 TFEU did not apply to them even if the Commission has adopted a decision on the same cartel.59

3.3 The nature of the decision giving rise to preclusion of further prosecution or punishment Only a final ‘acquittal’ or conviction within the Union gives rise to preclusion of further proceedings.60 The case law requires that, for the ne bis in idem principle to apply, there must have been a first decision that exonerated or penalised an undertaking and that is no longer amenable to challenge.61 Therefore, the principle of ne bis in idem applies not only when there has been a conviction or a penalty, but also when an undertaking has been acquitted or exonerated. But what constitutes an acquittal for the purposes of the ne bis in idem principle? It is clear that if the legislature or a prosecuting authority has complete freedom in deciding what an acquittal means or when an acquittal decision can be made, the ne bis in idem rule protecting a person from further proceedings after acquittal would be abolished and the right would be reduced to the right not to be prosecuted again or punished only after conviction but

58 Verhuizingen Coppens, above n 52; Chemie Linz, above n 52, para 308; Anic Partecipazioni, above n 52, para 113; Cimenteries, above n 52, paras 4025 et seq. 59 Toshiba, above n 7. 60 Article 50 EU Charter. 61 Transcatab, above n 17, para 254; Archer Daniels Midland, above n 23, paras 85–86; Toshiba, above n 7, para 94.

368 Renato Nazzini not after acquittal. To achieve this, it would suffice for legislation to prohibit acquittal decisions and provide for mere case closures on non liquet grounds whenever the defendant’s guilt cannot be proved to the required standard or to give the prosecution or court complete freedom to bring a case to an end without an acquittal whenever they think fit. It seems obvious that this runs counter to the right not to be prosecuted again or punished after acquittal— yet it appears that this is the current position in EU competition law. This results from three sets of rules: 1. Under Article 10 of Regulation 1/2003, the Commission has the power to make a finding of inapplicability of Article 101 or 102 TFEU in its discretion when the Union public interest so requires. It is never required to do so, and the wording of Article 10 suggests that the finding of inapplicability is an exception rather than the norm. 2. When the Commission considers that, on the basis of the information in its possession, there are insufficient grounds for acting on a complaint, it shall reject the complaint according to Article 7 of Commission Regulation 773/2004.62 The decision is not defined as a final acquittal, but simply as the rejection of a complaint, as it does not establish that no infringement has been committed. As a result, at least prima facie, further proceedings can be opened without necessarily engaging the ne bis in idem rules; 3. In Tele2 Polska, the Court of Justice ruled that, under Article 5 of Regulation 1/2003, NCAs do not have the power to make non-infringement decisions.63 Therefore, an undertaking is exposed to multiple investigations in the Member States as long as no NCA establishes sufficient evidence of an infringement. If no or insufficient evidence of infringement is found, an NCA is only allowed to decide that there are no grounds for action. Again, prima facie, it would appear that the ne bis in idem principle does not apply in these circumstances. The EU competition enforcement system appears to have been designed so that only the Commission, in exceptional circumstances, can make acquittal decisions precluding further proceedings. In all other cases, if there is no proof of an infringement, the proceedings will be closed by a non liquet decision, which allows for further proceedings against the same parties in relation to the same or inextricably linked facts. This is so even if the substance of the case has been fully explored in the first set of proceedings and the competition authority came to the conclusion that there is no infringement. The problem is well illustrated by Tele2 Polska itself. The Polish competition authority

See Article 7 of Commission Regulation 773/2004/EC. Tele2 Polska, above n 6.

62 63

Parallel Proceedings in EU Competition Law 369 in that case had come to the conclusion that there was no infringement of Article 102 TFEU or the equivalent national provision. It had also decided that there were ‘no grounds for action’ under EU law, and that there was no infringement under national law. However, it issued a non liquet decision instead of a decision on the merits.64 The question is whether a non liquet decision in such circumstances bars a second prosecution whatever its legal label under secondary EU legislation or, in a national context, under national law. The same question arises as regards decisions by the Commission rejecting a complaint under Article 7 of Commission Regulation 773/2004. Currently, both the Court of Justice and the European Court of Human Rights essentially defer to national law to determine whether a decision is final so as to bar a second prosecution. Under Article 54 CISA, the Court of Justice held in Vladimir Turanský that, in order to assess whether a decision is final for the purpose of the application of the ne bis in idem principle, it must be ascertained whether the decision is final and binding in the Member State concerned and whether it leads, in that Member State, to the protection granted by the ne bis in idem principle. A decision which does not definitively bar a second prosecution in the Member State in which it was adopted cannot prevent further proceedings or punishment in another Member State.65 This approach is consistent with the previous case law and, in particular, with Gözütok and Brügge, where the Court of Justice, while giving a wide interpretation of Article 54 CISA as applying also to decisions of the prosecutor to discontinue the proceedings based on the defendant performing certain obligations, clearly emphasised that it was material to this conclusion that the decision in question barred further prosecution under German law.66 This outcome, however, is justified in the light of the special features of the CISA, which recognises the preclusive effects of national criminal decisions and is in no way conditional on the harmonisation or approximation of the criminal laws of the Member States.67 Article 54 CISA is, therefore, based on the mutual trust of the Contracting States in their criminal justice systems.68 As a consequence, it is entirely logical, and indeed unavoidable, that the finality of a decision is assessed under the relevant national law. However, under Regulation 1/2003, the decisional powers of the Commission and the NCAs are directly governed by EU law. Secondary legislation must be interpreted in a way which is consistent with the ne bis in idem principle, which is a hierarchically superior norm. There is

Ibid. Case C-491/07 Vladimir Turanský [2008] ECR I-11039, paras 35–36. 66 Joined Cases C-187 and 385/01 Hüseyin Gözütok and Klaus Brügge [2003] ECR I-1345, paras 25, 27, 29–31 and 34. 67 Case C-297/07 Klaus Bourquain [2008] ECR I-9425, para 36; Gözütok and Brügge, ibid, para 32. 68 Bourquain, ibid, para 37; Gözütok and Brügge, ibid, para 33. 64 65

370 Renato Nazzini no issue of mutual trust in, or deference to, national legal systems. Quite the contrary: following the approach of the Court of Justice in Tele2 Polska, the area of the types of decision that NCAs must and are allowed to make has been fully harmonised under EU law. The European Court of Human Rights has not ruled on whether Article 4, Protocol 7 to the ECHR imposes a constraint on the Contracting States as to the circumstances in which criminal proceedings can be discontinued without giving rise to preclusion of further proceedings or punishment for the same offence. The jurisprudence of the Court generally holds that a decision is final ‘when no further ordinary remedies are available or when the parties have exhausted such remedies or have permitted the time-limit to expire without availing themselves of them’.69 Thus, the question as to whether a decision is final must be answered under national law.70 However, this concept of finality is procedural, as it merely denotes that the decision is no longer amenable to ordinary challenges. The test does not relate to the content of the decision, which is what matters in determining whether a decision constitutes an acquittal or is a mere case closure. Therefore, this case law is not relevant to the problem at hand. A further strand of the Court’s case law holds that there is no right to a formal acquittal under Article 6 ECHR. Proceedings can be discontinued without any decision on the merits.71 However, the question as to whether there is an infringement of the presumption of innocence if proceedings are discontinued without a decision on the merits is different from the question as to whether such a practice offends against the ne bis in idem principle. The rule that a prosecution may be discontinued without a formal acquittal would be fully compatible with the rule that, in certain circumstances, a prosecution discontinued without formal acquittal bars further proceedings. It appears, therefore, that the case law of the European Court of Human Rights does not provide any specific guidance on whether non liquet decisions may bar further proceedings under the ne bis in idem principle. Therefore, the question must be answered on first principles. And as the Court said in Zolotukhin, the Convention must be interpreted so as to give effect to the rights it aims to protect.72 To allow the Contracting States carte blanche to adopt non liquet decisions would be tantamount to cancelling the ne bis in idem right from Protocol 7. The dissenting opinion of Judge Zupančič in Sievert v

69 Zolotukhin, above n 40, para 107; Case No 50178/99, Nikitin v Russia, 20 July 2004, para 37; Case No 70982/01, Horciag v Romania, 15 March 2005. 70 Case No 75602/01, Sundqvist v Finland, 22 November 2005; Case No 28034/04 Müller v Austria (No 2), 18 September 2008. 71 Case No 70982/01, Niedermeier v Germany, 3 February 2009; Case No 59493/00, Withey v United Kingdom, 26 August 2003. 72 Zolotukhin, above n 40, para 16.

Parallel Proceedings in EU Competition Law 371 Germany73 appears to go in this direction. As the judge said, in deciding whether the protection against double jeopardy applies, the question is ‘what is the real nature of the decision taken by the prosecutor’74 and that ‘any decision based on any lack of evidence … is … of necessity a final decision’.75 It follows from the above analysis that, under the EU general principle of ne bis in idem and Article 50 EU Charter, it would be wrong to accept that ‘no grounds for action’ decisions by NCAs and complaint rejections by the Commission do not bar subsequent proceedings as, formally, they do not constitute acquittals on the merits. First, from a right-based perspective, there is no point in recognising a right not to be tried or punished twice while at the same time giving the legislature a free hand to permit proceedings to be discontinued by issuing non liquet decisions even if there has been a significant appraisal of the merits of the case. Second, from an internal market perspective, it seems that allowing EU NCAs and the Commission to conduct as many investigations against the same undertakings in relation to the same facts as they think fit in their discretion is precisely what the ne bis in idem principle should prohibit. Indeed, the problem that occurs because of the absolute freedom of the legislature to permit non liquet decisions in a national context becomes even more acute across the Union because the risk of multiple investigations increases considerably for undertakings engaging more actively in the provision of goods and services in different Member States. Third, to allow, without constraints, repeated fishing expeditions against the same undertaking in connection with the same facts is also inefficient as repeated investigations into the same alleged infringements result in a waste of public and private resources and lower the incentives of competition authorities to investigate and prosecute cases diligently and effectively. The only argument that can be made in favour of repeated investigations is based on the effective enforcement of EU competition law: the detection and punishment of competition infringements requires that competition authorities be allowed to test allegations and close an investigation without this precluding further proceedings. There may be good reasons to do so: new evidence could come to light or new resources could mean that a case which

73 Case No 29881/07, Sievert v Germany, 19 July 2012. The case concerned the principle of ne bis in idem only indirectly. The applicant complained that he had been unable to examine two witnesses because they availed themselves of the privilege against self-incrimination and refused to answer questions. There was an issue as to whether the decisions to discontinue the prosecutions of the two witnesses in question would bar a new prosecution. If it did, then the privilege against self-incrimination would not have been available. The majority did not address this point and decided the case without asking whether the privilege against self-incrimination had been rightly invoked. Ibid, paras 58–69. 74 Ibid, para 19 (emphasis in the original). 75 Ibid, para 29 (emphasis in the original).

372 Renato Nazzini previously could not be pursued as a priority can now be investigated. However, the correct approach would be to balance these considerations against the need to protect undertakings against multiple investigations. Thus, a distinction should be made between decisions to close a case that are based on a preliminary and summary investigation, which should not bar further proceedings, and decisions to close a case that follow a significant consideration of the merits of the case and are adopted in a procedure where the defendant has had a full opportunity to exercise his rights of defence, which should bar further proceedings whatever the formal classification of the final decision. This test is close to that proposed by Advocate General Sharpston in her Opinion in Gasparini, where she argued that a defendant should be entitled to the protection of the ne bis in idem principle when ‘he has de facto been placed in jeopardy’ and the ‘criminal proceedings have involved any significant consideration of the merits of the case’.76 The case law of the Court of Justice lends some support to this test. In Miraglia, the Court had to rule on whether a judicial decision to close the case following the public prosecutor’s determination to discontinue the prosecution on the sole ground that proceedings that had been commenced in another Member State gave rise to preclusion under Article 54 CISA.77 Not surprisingly, the answer was that further proceedings were not precluded because ‘no determination has been made as to the merits of the case’78 and ‘there had been no assessment whatsoever of the unlawful conduct with which the defendant was charged’.79 The Court then went on to consider the EU objective of ensuring free movement of persons in conjunction with appropriate measures to prevent and combat crime.80 Arguably, however, this objective would not have prevailed over the right of the individual if the case closure decision had involved a substantive assessment of the case. In Van Straaten, the Court had no hesitation in ruling that a final decision acquitting the defendant for lack of evidence precludes further proceedings because, unlike in Miraglia, such a decision is based on ‘a determination as to the merits of the case’.81 The objective of protecting the right to freedom of movement played a key role in the reasoning of the Court.82 The same strong freedom of movement rationale led the Court in Gasparini to hold that decisions acquitting the defendant because the prosecution is time barred give rise to preclusion of further proceedings.83 For the

76 Opinion of Advocate General Sharpston in Giuseppe Francesco Gasparini, above n 48, para 96. 77 Case C-469/03 Filomeno Mario Miraglia [2005] ECR I-2009, para 30. 78 Ibid. 79 Ibid, para 34. 80 Ibid. 81 Van Straaten, above n 35, paras 56–60. 82 Ibid, para 58. 83 Gasparini, above n 48, paras 22–33.

Parallel Proceedings in EU Competition Law 373 same reasons, in Gözütok and Brügge, the Court held that a decision of the prosecution giving effect to a plea agreement barred a second prosecution.84 This case law on Article 54 CISA is relevant under Article 50 of the EU Charter and for the development of ne bis in idem as a general principle of EU law because all these provisions and principles share the same foundations: safeguarding individual rights and ensuring free movement. Its logical development is that, at least where EU legislation is concerned, provisions allowing proceedings to be closed without a formal acquittal must be interpreted as precluding further proceedings if there has been a substantive appraisal of the case and the defendant has had the opportunity to exercise his rights of defence in full. The same test should apply to commitments decisions under Article 9 of Regulation 1/2003. Under Article 9(1), commitments decisions conclude that ‘there are no longer grounds for action by the Commission’. Recital 13 clarifies that commitments decisions do not determine ‘whether or not there has been or still is an infringement’ and are ‘without prejudice to the powers of competition authorities and courts of the Member States to make such a finding and decide upon the case’. However, if a commitments decision is adopted after a substantive appraisal of the case and after the undertakings concerned have had a full opportunity to exercise their rights of defence, the prohibition of further proceedings and convictions should apply. In Gözütok and Brügge, the Court of Justice clearly held that decisions discontinuing the prosecution because the defendants made certain commitments are capable of giving rise to preclusion.85 In that case, German law itself considered the decision as triggering the application of the ne bis in idem protection.86 As regards commitments decisions under Article 9 of Regulation 1/2003, while secondary legislation considers such decisions as not constituting a formal conviction, the application of the ne bis in idem principle derives directly from Article 50 EU Charter, which is a hierarchically superior norm.

3.4 Setting aside of an acquittal or conviction on procedural grounds In EU law, the setting aside by the EU Courts of an infringement decision by the Commission on procedural grounds does not preclude the Commission from resuming the proceedings. Since PVC (No 2), the case law justifies this conclusion on two grounds: (i) the annulment of a decision on procedural grounds does not constitute a decision on the merits and is not, therefore, an acquittal ‘within the meaning given to that expression in penal matters’; and

Gözütok and Brügge, above n 66, paras 36–38. Ibid, para 27. 86 Ibid, para 17. 84 85

374 Renato Nazzini (ii) the penalties imposed as a result of the second set of proceedings are not added to the previously imposed penalties but replace them.87 Both strands of reasoning are not fully convincing. As regards the effect of the judgment setting aside an infringement decision on procedural grounds, it is not obvious that this is the only relevant decision for the purposes of the ne bis in idem protection. The relevant decision may well be the annulled decision, which established the infringement and imposed a penalty and which, after the annulment, is no longer amenable to appeal or judicial review and, therefore, is final. The doctrinal objection to this argument is that, once a decision has been set aside, it no longer exists, and it is therefore impossible to say that there is a final acquittal or conviction within the Union according to Article 50 EU Charter. However, it would be equally plausible to argue that a conviction does not simply cease to exist because it has been annulled. While its legal effects are set aside retrospectively, its significance under the ne bis in idem principle cannot be cancelled. The reason is that the purpose of the ne bis in idem principle is to protect the right of the defendant not to be subject to a second prosecution or punishment once he has already been subjected to proceedings on the merits resulting in a final decision. In the case of the procedural annulment of a conviction, the defendant has actually been placed in jeopardy as there was a substantive assessment of the case in proceedings in which he had a full opportunity to exercise his rights of defence. Applying the test proposed in the previous section, it can be argued that such a decision should bar further proceedings or punishment. As regards the argument that the second penalty replaces the previously imposed one and is not added to it, it is important to underline that the ne bis in idem principle does not protect a right not to be punished twice because the preclusion also applies when the first decision is an acquittal. Therefore, the circumstance that, when the first decision is set aside, the second penalties are not added to the previous ones but replace them is irrelevant and based on an incorrect understanding of the principle. From the point of view of the defendant, what matters is that the jus puniendi of the state or the Union has already been exercised once in proceedings on the merits, resulting in a decision that is no longer amenable to appeal or review. The circumstance that the jus puniendi was invalidly exercised should not detract from the protection afforded to the defendant. If, from the perspective of the protection of the fundamental right of the undertakings concerned, there appears to be only a weak doctrinal justification for limiting the ne bis in idem principle when the first decision has been set aside, such a justification can be found in the need to ensure the effective enforcement of EU competition law: effective law enforcement requires that

87 Limburgse Vinyl Maatschappij, above n 10, para 62; Case T-24/07 ThyssenKrupp Stainless, above n 10, para 190.

Parallel Proceedings in EU Competition Law 375 procedural defects affecting convictions do not prevent reopening the case, provided that the defect can be cured. The defendant has a right to a procedure according to law and can enforce his right by mounting procedural challenges against a conviction. However, if a procedural challenge is successful, the ne bis in idem protection must give way to the public interest in effective law enforcement. Even on this reasoning, however, it could be argued that an unqualified, potentially open-ended possibility of resuming proceedings following the annulment of a conviction on procedural grounds is too draconian an inroad into the ne bis in idem right. Furthermore, disciplinary and efficiency considerations suggest that allowing the prosecution to reopen the case time and again notwithstanding serious procedural defects in the proceedings is not desirable from a public interest point of view. The above considerations suggest that a possible, and probably better, approach to the preclusive effect of a judicial annulment of a conviction on procedural grounds could be to frame a test that balances the protection of the right of the undertakings concerned, the need to safeguard the effective enforcement of EU competition law and the requirement to ensure that competition authorities act in accordance with the law and in an efficient way on a case-by-case basis. The analysis above suggests that a suitable test could be to allow proceedings to be resumed if: (i) the effective enforcement of EU competition law so requires; or (ii) the procedural defect of the decision was either not attributable to the prosecuting authority or, if it was so attributable, was neither in bad faith nor grossly negligent.

3.5 Relationship between EU and national law enforcement within the Union It is uncontroversial that infringement decisions by NCAs under EU competition law bar further proceedings under EU competition law by other NCAs or the Commission relating to the same facts and the same undertakings.88 This chapter has already demonstrated that, as regards the same infringement, the test should be whether the proceedings arise from the same or inextricably linked facts. The problem concerns, therefore, the possibility of parallel proceedings under EU and national competition law. In such a case, even if the NCA does not apply EU law, it is still bound by Article 50 EU Charter and by the general EU law principle of ne bis in idem. As Advocate General Kokott said in her Opinion in Toshiba, NCAs are only allowed to apply national competition law in accordance with the delimitation of competences set out in

88 Joined Cases T-14, 147 to 150 and 154/07 ThyssenKrupp Liften Ascenseurs NV v Commission [2011] ECR II-5129; ThyssenKrupp Stainless, above n 10, para 162.

376 Renato Nazzini Regulation 1/2003. Regulation 1/2003 itself must be interpreted in accordance with Article 50 EU Charter and the general principle of ne bis in idem.89 As a consequence, the application of national competition law must also comply with EU fundamental rights and general principles. In Toshiba, the Court did not address the issue specifically, and limited itself to the analysis of the temporal application of the ne bis in idem principle.90 However, implicitly, the Court accepted that the ne bis in idem principle applies when an NCA applies national competition law. This accords with the extensive interpretation of Article 51 EU Charter in Fransson91 and with the case law of the Court of Justice, which, since Walt Wilhelm,92 requires an NCA applying national competition law to take into account the fines imposed by the Commission on the same undertakings for the same facts. The question is then whether, at the current stage of development of EU law and of integration of the internal market, an EU competition authority should be prevented from applying EU or national competition law if the same infringement by the same parties is already the subject matter of a decision on the merits under national or EU competition law, respectively. In Walt Wilhelm, a case dating back to 1969, the Court of Justice ruled that parallel proceedings by the Commission under EU law and by NCAs under national law are allowed as long as the NCA respects the primacy of EU law93 and the previous punishment is taken into account in determining the second sanction.94 Walt Wilhelm was decided in a very different context, in which the Court of Justice noted that EU competition law and national competition law have different objectives, the former being concerned with the establishment of the single market and the latter with unspecified ‘considerations peculiar to it’.95 Furthermore, the Court considered it relevant that EU law had not (yet) regulated the relationship between EU and national competition law according to Article 103(2)(e). Both elements in Walt Wilhelm apply with much less force today. It is still true that purely domestic competition law may have objectives that are different from the objectives of EU competition law. However, the reality is that the national competition laws in the Member States are by and large modelled on EU law. Nor is it possible to say that EU competition law

Opinion of Advocate General Kokott in Toshiba, above n 41, paras 105–106. Toshiba, above n 7, para 95. Åkerberg, above n 15, paras 16–31. 92 Walt Wilhelm, above n 7; Toshiba, above n 7, para 81; Case C-550/07 P Akzo Nobel Chemicals Ltd v Commission [2010] ECR I-8301, para 103; Joined Cases C-295 to 298/04 Vincenzo Manfredi v Lloyd Adriatico Assicurazioni SpA [2006] ECR I-6619, para 38. 93 Walt Wilhelm, above n 7, paras 3–9. 94 Ibid, para 11. 95 Ibid, para 3. 89 90 91

Parallel Proceedings in EU Competition Law 377 is exclusively concerned with the establishment of the internal market while national competition laws pursue wholly different objectives. While the wellfunctioning of the internal market is the primary objective of EU competition law, the internal market itself has further objectives, including the economic well-being of the peoples of Europe, economic growth and productivity, which are therefore, albeit indirectly, also objectives of EU competition law.96 Broadly, it is difficult to argue that facilitating trade, promoting economic growth and productivity, and enhancing the economic welfare of the society are not objectives of the national competition laws of the Member States. Furthermore, Article 3 of Regulation 1/2003 now regulates the relationship between EU competition law and national competition law. In particular, the possibility of parallel proceedings under EU and national law is significantly limited by the duty of NCAs to apply Articles 101 and 102 TFEU whenever they apply national competition law to conduct that has an effect on trade between Member States. Save for the application of stricter unilateral conduct rules under national law, any conflict between EU and national law is expressly prohibited, thus achieving a significant degree of convergence.97 Notwithstanding the considerations above, in Toshiba, the Grand Chamber of the Court of Justice confirmed the Walt Wilhelm approach to parallel EU and national proceedings. In that case, the Czech competition authority had fined the parties to a cartel for the period prior to the accession of the Czech Republic to the EU. The Commission had imposed fines on the same parties for the same cartel. Following the Advocate General, the Court confirmed that EU and national competition law may apply in parallel as they consider restrictions of competition from different angles and have different scopes of application.98 This, combined with the requirement that the ‘legal interest’ protected must be the same for two offences to be the same for the purposes of the ne bis in idem principle, means that an undertaking can be subject to two fining decisions, one under EU law and one under national law, with the only requirement being that the second decision must take into account the previously imposed sanction. The Court in Toshiba even went as far as saying that Article 16(2) of Regulation 1/2003 allows NCAs to apply EU and national law after the Commission—and, it would appear to follow a fortiori another NCA—has already made a decision, with the only limit that a subsequent national d ecision

96 Renato Nazzini, The Foundations of European Union Competition Law. The Objective and Principle of Article 102, Oxford University Press, 2011, 113 and 152. 97 Frédéric Louis and Gabriele Accardo, ‘Ne Bis in Idem, Part “Bis”’, 34 World Competition 97, 104 (2011). 98 Toshiba, above n 7, paras 81–83.

378 Renato Nazzini c annot run counter to a previous decision by the Commission.99 On this reasoning, an undertaking could be fined by the Commission for a cartel and then fined again by an NCA for the same cartel under national law as long as the national decision takes into account the fine imposed by the Commission. The same would apply in the case of abuse of dominance, with the additional consideration that, because, under Article 3 of Regulation 1/2003, national law may apply stricter rules than Article 102 to unilateral conduct, it could be argued that a national law decision may even follow a Commission’s decision of inapplicability. The case law of the Court of Justice in Walt Wilhelm and Toshiba relies on the requirement that two offences are the same only if they protect the same legal interest. Because, as already argued, this requirement is in breach of Article 50 EU Charter, this case law lacks foundation and should be overruled at the earliest opportunity. The Court of Justice should reconsider the position and recognise that when Member States and the Commission apply national or EU competition law to infringements consisting of the same or inextricably linked facts and committed by the same undertakings, the ne bis in idem principle applies with full force, thus barring a second prosecution or punishment. This conclusion is subject to the sole proviso that preclusion only arises if the authority which made the first decision had jurisdiction to make a finding of infringement in relation to the facts that constitute the offence prosecuted by the second authority. Therefore, in the case of accession states, a decision under national law is not precluded by a decision of the Commission concerning the same parties and inextricably linked facts for the period prior to accession. It was on this simple basis that the Court of Justice should have decided the reference in Toshiba.

3.6 Relationship between proceedings within and outside the EU As regards the relationship between proceedings in the Union and previous acquittals or convictions outside the Union, it is settled case law that the principle of ne bis in idem does not apply. The rationale for the jurisprudence of the EU Courts has, however, changed significantly over time. Initially, a first strand of case law represented by Boehringer Mannheim v Commission appeared to rely on the different factual matrix between cartels prosecuted by foreign authorities and cartels prosecuted within the Union. In particular, the case law seemed to establish a presumption that the geographical scope of the cases was different and any punishment imposed in foreign

Ibid, paras 84–87.

99

Parallel Proceedings in EU Competition Law 379 proceedings would not take into account the anticompetitive effects of the conduct within the Union.100 The subsequent case law moved away from this approach. In Kyowa Hakko Kogyo v Commission, the Court of First Instance construed Boehringer Mannheim as ruling only that a hypothetical duty of the Commission to take into account a foreign penalty when setting the level of a fine under EU law could only arise if the facts were identical in the two sets of proceedings. Since the applicant had not established that it was so in that case, the argument failed. Boehringer Mannheim had not decided that the Commission had a duty to take into account fines imposed by foreign authorities.101 On the issue of the application of the principle of ne bis in idem, the Court relied on an argument a fortiori drawn from the Walt Wilhelm case law: that if the ne bis in idem principle does not apply in the relationship between EU proceedings and proceedings under the national laws of the Member States, then it cannot apply with respect to foreign proceedings. If EU and national proceedings within the Union pursue different ends, so must EU and foreign proceedings, which aim at protecting different markets.102 However, the rule that an EU competition authority must take into account the fines already imposed by another EU competition authority on the same undertaking for the same conduct cannot be extended to the relationship between the Commission and foreign authorities. The distinguishing factors are, first, that the relationship between proceedings by the Commission and proceedings of the NCAs is characterised by a special system of concurrent jurisdiction in relation to conduct on the same territory; and, second, that there is a close interdependence between the national markets of the EU Member States. Both factors are lacking when it comes to the relationship between proceedings in the Union and foreign proceedings.103 Then, when the case law developed a ne bis in idem test requiring not only the identity of the facts and the identity of the defendant but also the identity of the legal interest protected by the two offences,104 a third strand of case law emerged holding that, as regards infringements of foreign competition law and infringements of EU competition law, the requirement of the unity of the legal interest is, by definition, not fulfilled and, therefore, the ne bis in idem principle does not apply.105 Advocate General Tizzano in Archer Daniels Midlands v Commission appears to have followed this approach: after explaining that

Case 7-72 Boehringer Mannheim GmbH v Commission [1972] ECR 1281, paras 4–8. Case T-223/00 Kyowa Hakko Kogyo Co Ltd v Commission [2003] ECR II-2553, paras 106–114. 102 Ibid, paras 100–101. The argument is now undermined a fortiori by the implicit acceptance in Toshiba that the ne bis in idem principle applies in national competition law proceedings. See Toshiba, above n 7, para 95. 103 Kyowa Hakko Kogyo, above n 101, paras 108–111. 104 Aalborg Portland, above n 8, para 338. 105 Archer Daniels Midland, above n 17, para 63. 100 101

380 Renato Nazzini there is no public international law rule that imposes on the Union institutions a duty to take into account penalties inflicted by foreign authorities, the Advocate General went on to say that, in any event, EU and foreign proceedings aim at protecting different legal interests.106 The Court of Justice, however, decided the case on a narrow ground, reminiscent of Boehringer Mannheim: the applicants had not shown that the facts in the two investigations were identical. Therefore, their appeal failed on this simple ground.107 The Court of Justice changed its jurisprudence significantly in Showa Denko v Commission and in SGL Carbon v Commission (No 1). In his Opinions in both cases, Advocate General Geelhoed had adopted the same approach as Advocate General Tizzano in Archer Daniels Midlands, relying, first, on the absence of a duty under public international law to take foreign penalties into account and, in any event, on the lack of identity of the legal interest protected.108 The Court did not follow this reasoning, but relied on what appears to be a more fundamental argument: the ne bis in idem principle does not apply at all when the previous decision is made by a non-EU authority because foreign and EU competition laws pursue different ends within different geographical scopes of application. In protecting free competition in the internal market, the Union cannot be constrained by decisions of foreign authorities pursuing different objectives. The inapplicability of the principle of ne bis in idem in an international context was reaffirmed by the Court of Justice in SGL Carbon v Commission (No 2):109 proceedings aimed at safeguarding a fundamental objective of the Union cannot be barred by proceedings in non-EU Member States that pursue different interests relevant under foreign law.110 The case law of the Court of Justice is correct. In the absence of a public international law norm binding the Union to take into account penalties imposed for competition law infringements outside the Union, the matter is one for Union law. Article 50 EU Charter is limited to proceedings within the Union and so is, and should be, the general principle of ne bis in idem. The effective enforcement of EU competition law plays a fundamental role in this respect: the Union cannot be prevented from pursuing its objectives by the exercise of jurisdiction by foreign authorities. There is a strong presumption—based on the doctrine of international comity—that foreign authorities do not take over the role of punishing and deterring infringements occurring outside their respective jurisdictions.111 But even if they do, 106 Opinion of Advocate General Tizzano in Archer Daniels Midland, above n 23, paras 92–104. 107 Archer Daniels Midland, ibid, paras 46–76. 108 Showa Denko KK, above n 13; Opinion of Advocate General Geelhoed in SGL Carbon, above n 13, paras 46–62. 109 SGL Carbon, above n 13, paras 26–32, confirming the approach in the earlier case law. 110 Ibid, paras 26–27. 111 This appears to be implied in Boehringer Mannheim, above n 100, paras 4–6.

Parallel Proceedings in EU Competition Law 381 the remedy must be found within the legal systems of those states exercising extraterritorial jurisdiction or in international law norms forbidding the exercise of exorbitant jurisdiction. The exercise of jurisdiction by a non-EU competition authority cannot deprive the Commission or the NCAs of their power and duty to punish and deter competition infringements within the Union.

4. The function of the ne bis in idem principle in limiting multiple proceedings and decisions in EU competition law This chapter argues that the risk of multiple proceedings and convictions in EU competition enforcement is inadequately addressed by the current interpretation of the ne bis in idem principle in competition matters. The review of the case law of the EU Courts in the light of this model leads to the following conclusions: 1. The correct position under Article 50 EU Charter and the general EU law principle of ne bis in idem is that two infringements are the same offence when they are constituted by the same or inextricably linked facts. The case law of the EU Courts on the principle of ne bis in idem, which still requires the identity of the legal interest protected, should be reviewed and brought into line with the case law of the Court of Justice on Article 54 CISA and the jurisprudence of the European Court of Human Rights on Article 4, Protocol 7 to the ECHR. 2. A single infringement consisting in the same or inextricably linked facts cannot be subdivided and investigated separately by the Commission and NCAs in the Union because the first acquittal or conviction will bar further proceedings. This applies in particular: (i) whenever a set of facts gives rise to a single overall infringement; (ii) whenever the effects of anticompetitive conduct are caused by the same behaviour or are dependent on each other; and (iii) provided always that the first authority had jurisdiction to investigate the facts that are the subject matter of the second set of proceedings. 3. At least when the procedure is governed by EU law, provisions allowing proceedings to be closed without a formal acquittal, such as those under Article 5 of Regulation 1/2003 (no grounds for action decisions by NCAs) or Article 7 of Commission Regulation 773/2004 (complaint rejections by the Commission), must be interpreted as precluding further proceedings if there has been a substantive appraisal of the merits of the case and the defendant has had the opportunity to exercise his rights of defence in full.

382 Renato Nazzini 4. The EU case law on the reopening of the proceedings following the judicial annulment of a conviction on procedural grounds may be justified by the need to ensure the effective enforcement of EU competition law prevailing over the defendant’s right and efficiency considerations. An alternative approach could be to frame a test that balances the protection of the right of the undertakings concerned, the need to safeguard the effective enforcement of EU competition law and the requirement to ensure that competition authorities act in accordance with the law and in an efficient way on a case-by-case basis. A suitable test could be to allow proceedings to be resumed if: (i) the effective enforcement of EU competition law so requires; or (ii) the procedural defect of the decision was not attributable to the prosecuting authority or, if it was so attributable, was neither in bad faith nor grossly negligent. 5. The settled case law holding that NCAs may apply national law even if the Commission has already imposed a fine on the same undertakings for the same facts is untenable. Given the substantial degree of convergence between the national competition laws of the Member States and EU competition law achieved in particular after the entry into force of Regulation 1/2003, the correct position under Article 50 EU Charter and the general principle of ne bis in idem appears to be that, when Member States and the Commission apply national or EU competition law to infringements consisting in the same or inextricably linked facts and committed by the same undertakings, the ne bis in idem principle applies with full force, thus barring a second prosecution or punishment. This conclusion is subject to the sole proviso that preclusion only arises if the authority which made the first decision had jurisdiction to make a finding of infringement in relation to the facts that constitute the offence prosecuted by the second authority. 6. The case law of the EU Courts is correct in holding that the ne bis in idem principle does not apply when the first acquittal or conviction occurred outside the Union. In conclusion, the ne bis in idem case law of the EU Courts in competition law is in need of urgent review. This chapter suggests possible solutions that could allow the principle of ne bis in idem to perform more effectively its function of limiting vexatious and inefficient multiple proceedings against the same undertakings for the same alleged infringements of competition law within the Union.

Wolfgang Kirchhoff * Reflections on Parallel Enforcement, Fundamental Rights and the Rule of Law in the Competition Law Context

In this chapter I will report on two German cases which highlight the delicate relationship between efficient enforcement of competition law on the one hand and fundamental rights under the European Convention of Human Rights (ECHR) and the EU Charter of Fundamental Rights (CFR) on the other. Moreover, a recent Swiss judgment shows what may be a surprising impact of the possible accession of the EU to the ECHR—although ECJ case law now appears to make accession unlikely—for the system of direct applicability of Article 101(3) TFEU established by Regulation 1/2003.

1. Legal succession in fines and fundamental rights Regulation 1/2003 establishes a system of parallel enforcement of EU competition law by the Commission and the NCAs of the EU Member States.1 The substantive law applied against infringements of EU competition law is the same. However, effective enforcement goes beyond the harmonisation of substantive law. It also requires uniform powers of investigation and penalties. In this context, two recent judgments in Germany show the limits of effective enforcement created by fundamental rights and the rule of law. Under German competition law there is only a limited scope of culpability of undertakings for regulatory fines according to Section 30 of the Act on Regulatory Offences. A regulatory fine may be imposed on a legal person or an association of persons only where a legal representative of such entity or a person responsible of its operation on behalf of the management has committed a criminal or regulatory offence as a result of which duties of the legal person or association of persons have been violated, or where the legal person or association of persons has been enriched or was intended to be enriched. This is different from Article 23(2) of Regulation 1/2003, which stipulates fines for undertakings—in the broad sense of economic entities—where they

Judge at the Federal Court of Justice, Germany. See Articles 4 and 5 of Regulation 1/2003.

* 1

384 Wolfgang Kirchhoff infringe EU competition law. The differences between EU and German competition law create difficulties for the efficient enforcement of EU competition law in Germany. In some cases, a group is restructured after an affiliate company has been fined for a competition law infringement. The fined company may merge with another company, or its assets may be sold. Efficient enforcement requires, in such a case, that it should still be possible to execute the fine. Under the concept of an economic entity in EU law, imposing a fine in those circumstances is allowed. However, according to the German Federal Court of Justice,2 this concept does not apply in relation to sanctions by the Federal Cartel Office (FCO). An extension of the regulatory fines liability to legal successors of the entity for whom the offender has acted can only be accepted in the exceptional case where there is an identity or near-identity between the earlier and the new connection of assets from a commercial perspective. Any further extension of the liability of fines on legal successors would not be covered by the wording of Section 30 of the Act on Regulatory Offences. Consequently, it would exceed—according to the Federal Court of Justice—the limits drawn to judicial construction of the law by Article 103, paragraph 2 of the G erman Basic Law, the Grundgesetz. The latter provision is equivalent to the ‘no punishment without law’ principle (nulla poena sine lege) found in Article 7 ECHR and Article 49 CFR. According to that principle, ‘[n]o one shall be held guilty of any criminal offence on account of any act or omission which did not constitute a criminal offence under national law or international law at the time when it was committed’. According to German law, an identity or near-identity of the legal successor only exists where the company continues business in the form of an unchanged or nearly unchanged new legal entity and the other assets of the new entity are to be neglected. For the purpose of regulatory fines, this is the only case in which the same legal entity results from the restructuring under corporate law. In contrast, there is a lack of economic identity when companies of approximately the same size and nearly identical market shares merge and bring together their business activities. Thus, German law does not stipulate a ‘group liability’ for regulatory fines. The individual subsidiaries are independent legal entities in relation to each other and in relation to their parent company. Liability for regulatory fines imposed by the FCO by attribution of assets may only be based on an express statutory provision, which is missing in German law. The Federal Court of Justice was aware that respecting these limits derived from the fundamental right of ‘no punishment without law’ can lead to

2 BGH, decision of 10 August 2011—KRB 55/10, Wirtschaft und Wettbewerb 2012, 81—Versicherungsfusion.

Reflections on Parallel Enforcement 385 nfortunate consequences. It enables the circumvention of sanctions by u intentionally employing certain business structures under corporate law. However, under the division of powers, it is the task of the legislator to introduce further liability of legal successors for regulatory fines imposed by the FCO. The tension between efficient enforcement and fundamental rights in Germany is mitigated, but not solved, by an amendment to the Act on Regulatory Offences in July 2013. According to Section 30, paragraph 2a of the Act, a regulatory fine may be imposed on a full or partial legal successor. In fines proceedings, the legal successor enters into the legal status of the predecessor. Other forms of circumventing liability for fines without full or partial legal succession, eg by asset deals, will nevertheless remain. On the basis of present German law, the Appellate Court of Düsseldorf delivered a judgment in December 2012 in a cartel case concerning producers of dry cement (Trockenmörtelhersteller).3 In this case, the conditions for exceptional liability of the legal successor according to the legal principles established for German law by the Federal Court of Justice were not met. It was found that there was no identity or near-identity between the succeeding company and the entity for which the offender acted. The Appellate Court found an infringement not of EU, but only of national, competition law. It held that the behaviour in question had no effect on the trade between Member States. However, the Appellate Court did not terminate its analysis with this finding. Instead, it proceeded to examine whether liability of the legal successor follows from the second sentence of Article 5 of Regulation 1/2003. According to this provision, NCAs are authorised to enforce EU competition law by imposing fines or any other sanction provided for in their national law. In the understanding of the Appellate Court, the authorisation of NCAs by Article 5 is limited to punishing violations of EU competition law according to the procedural rules of their respective national law. Article 5 does not transfer the same powers of investigation and sanction to the NCAs that are granted to the Commission in the Regulation. The Appellate Court gave a number of reasons for this conclusion: 1. Article 4 of Regulation 1/2003 provides that the Commission ‘shall have the powers provided for by this Regulation’. In contrast, Article 5 limits the competence of NCAs to apply Articles 101 and 102 TFEU in individual cases and limits their powers to certain measures, like imposing fines. In this context, the powers of the Commission under the Regulation are not mentioned. Moreover, the wording ‘or any other penalty provided for in their national law’ indicates that ‘fines’ within the meaning

3 OLG Düsseldorf, judgment of 17 December 2012—V-1 Kart 7/12 (Owi), Wirtschaft und Wettbewerb 2013, 749—Silostellgebühren II.

386 Wolfgang Kirchhoff

2.

3.

4.

5.

of Article 4 of Regulation 1/2003 are only those fines which are provided for by national law. The Purpose of the Regulation is to grant the Commission and NCAs congruent powers to apply EU competition law. However, the European legislator did not stipulate congruent powers to prosecute and sanction infringements. The legislative history of Article 5 of Regulation 1/2003 also shows that the authorisation of NCAs is limited to punishing infringements of EU competition law by imposing fines according to the provisions of their respective national law. Finally—and in our context most interestingly—the Appellate Court argued that any other interpretation of the second sentence of A rticle 5 of Regulation 1/2003 would result in an infringement of the ‘no punishment without law’ principle laid down in Article 7 ECHR and in Article 49 CFR. According to the case law of the European Court of J ustice, the nulla poena sine lege principle implies that legislation must clearly define offences and the penalties they attract. That condition is met in the case where the individual concerned is in a position, on the basis of the wording of the relevant provision and with the help of the interpretative assistance given by the courts, to know which acts or omissions will make him criminally liable.4 The Appellate Court concluded from the constitutional tradition of Germany, which the ECJ takes into account in the context of its fundamental rights jurisprudence, that the ‘principle of precision’ applies not only for the elements of fact which define an offence, but also for the threat of sanctions. Since the second sentence of Article 5 Regulation 1/2003 neither states specific sanctions nor provides a framework for sanctions, it would infringe the principle of ‘no punishment without law’ if it were used as a legal basis for the imposition of fines by the NCAs. The Appellate Court was well aware of the ECJ’s judgment in Tele2 Polska. In that case, the ECJ clarified that the second paragraph of Article 5 of Regulation 1/2003 is directly applicable. It is clear from the Court’s judgment that Article 5 precludes the application of a national law which would require a procedure relating to the application of Article 102 TFEU to be brought to an end by a decision stating that there has been no breach of the latter provision.5 By contrast, NCAs may decide that there are no grounds for action on their part where the c onditions on

4 See Case C-303/05 Advocaten voor de Wereld VZW v Leden van de Ministerraad [2007] ECR I-3633, paras 49 et seq; Case C-352/09 P ThyssenKrupp Nirosta GmbH v Commission [2011] ECR I-2359, para 80. 5 Case C-375/09 Prezes Urzędu Ochrony Konkurencji i Konsumentów v Tele2 Polska sp z oo, now Netia SA [2011] ECR I-3055, para 35.

Reflections on Parallel Enforcement 387 rohibition are not met on the basis of the information in their posp session. The third sentence of Article 5 of Regulation 1/2003 directly authorises the NCAs to adopt such a ‘no grounds for action’ decision.6 Advocate General Mazák, in his Opinion in Tele2 Polska, took the view that Article 5 of Regulation 1/2003 allows all NCAs to have consistent decision-making powers, without the need to wait for the implementation of Article 5 in national law.7 However, the ECJ did not—at least, not directly—refer to that Opinion in its judgment. According to the Appellate Court of Düsseldorf, the Tele2 Polska case was irrelevant for the question of legal succession in cartel fines. In the view of the Appellate Court, it was not the direct applicability of Article 5 that was crucial in the Düsseldorf case but its regulatory content. Based on these considerations, the Appellate Court concluded that the authorisation of NCAs by Article 5 is limited to sanctioning violations of EU competition law according to the procedural rules of their respective national laws. Furthermore, for the Appellate Court, the meaning of Article 5 was sufficiently obvious and had to be regarded as an acte claire. Consequently, it was unnecessary to request a preliminary ruling from the ECJ according to Article 267 TFEU. One may doubt whether the application of the acte claire doctrine is convincing in the context of this case. Moreover, the decision is presently under appeal on points of law to the Federal Court of Justice—so, clearly, I will not give any indications as to the merits of that appeal on points of law. However, the judgment of the Appellate Court of Düsseldorf in any event gives an instructive example of the frictions that may arise between efficient enforcement and fundamental rights in the system of parallel enforcement by the Commission and by the NCAs as established by Regulation 1/2003. The Federal Cartel Office will suggest that the Federal Court of Justice should ask for a preliminary ruling from the ECJ on the question of whether the first sentence of Article 5 of Regulation 1/2003 directly, without need for an enabling provision in national law, authorises NCAs to impose fines on undertakings—in the broad sense of economic entities as defined by the case law of the ECJ, including legal successors in case of economic continuity.8 It is entirely open whether the Federal Court of Justice will seek advice in Luxemburg, and what the ECJ would decide if the Federal Court chose to do so. It would have to carry out a delicate balance between the principles of efficient enforcement and legal certainty as laid down in Article 7 ECHR and A rticle 49 CFR. The wording of Article 5 of Regulation 1/2003, s econd

See ibid, para 32. Opinion of Advocate General Mazák in Tele2 Polska, above n 5, para 57. 8 See C-280/06 Autorità Garante della Concorrenza e del Mercato v Ente tabacchi italiani—ETI SpA and Others [2007] ECR I-10893, paras 41–43. 6 7

388 Wolfgang Kirchhoff s entence, does not explicitly define the recipient of NCA decisions that impose fines by means of reference to the concepts of economic entity and economic continuity in EU competition law. On the other hand, one may argue that, by authorising the NCAs to apply Articles 101 and 102 TFEU, Article 5 of Regulation 1/2003 refers indirectly to the ECJ’s interpretation of those Articles, including the concepts of economic entity and economic continuity. Against this background, the possibility that the problem of legal succession in cartel fines will at some point be addressed by the ECJ cannot be excluded. It may also transpire that the European legislator will intervene to establish further principles in relation to the enforcement and sanctioning powers of NCAs. Any such intervention should not be limited to the problem of legal succession. In the past, compelling political reasons may well have prevented the European legislator from stipulating uniform powers of investigation and penalties for NCAs equal to those of the Commission. But in an internal market which requires a level playing field for undertakings,9 serious differences in investigative powers and sanctions between the Commission and NCAs as well as among NCAs are unacceptable in the long run and contradictory to the principles of a system of parallel enforcement of EU competition law. If such differences continue to exist, the system of efficient case allocation within the European Competition Network10 and the network of NCAs may be harmed considerably. In particular, it is questionable whether an NCA that allows undertakings to escape penalties for competition infringements by simply changing their corporate identity through restructurings, sales or other legal or organisational changes can really be considered ‘well placed’ to deal with a case in the network.11 It is true that natural persons may be subject to substantially different types of sanctions across the various jurisdictions in Member States. In contrast, sanctions imposed on undertakings are of the same type in all systems throughout the Union.12 Harmonisation might be easier here. Since it seems unlikely that the Member States will harmonise the powers of investigation and penalties for NCAs on their own, a Commission initiative might be appropriate. However, it should consider limiting such an initiative to powers of investigation and penalties concerning undertakings, and in particular to defining the addressees of NCA decisions imposing fines for infringements of Articles 101 and 102 TFEU by reference to the concepts of an economic entity and of economic continuity as established in EU competition law.

See Recital 8, second sentence, of Regulation 1/2003. See Recital 15 Regulation 1/2003. See Commission Notice on cooperation within the Network of Competition Authorities, [2004] OJ C101/43, paras 5–15. 12 See Recital 16 of Regulation 1/2003. 9

10 11

Reflections on Parallel Enforcement 389

2. ThyssenKrupp Nirosta Another instructive case for a possible clash between effective enforcement and the principle of ‘no punishment without law’ is the ThyssenKrupp Nirosta case. In 2006, the Commission took a decision in which it fined that company €3.2 million for fixing steel prices. That decision was based on the ECSC Treaty, which expired in 2002. The company argued that this was a violation of the principle of ‘no punishment without law’. However, the ECJ confirmed the Commission’s decision. The Court stated it would be contrary to the objectives and the coherence of the Treaties … if the Commission did not have jurisdiction to ensure the uniform application of the rules deriving from the ECSC Treaty which continue to produce effects even after the expiry of that treaty.13

In support of this understanding of the effective enforcement principle, the Court referred to its Lucchini judgment of 2007.14 Following the 2011 judgment of the ECJ in ThyssenKrupp Nirosta, the Finnish company Outokumpu filed a complaint with the German Federal Constitutional Court against the ECJ’s judgment. Outokumpu claimed that the judgment had infringed its fundamental rights.15 However, according to the case law of the Federal Court, it does not review decisions by the EU institutions as long as they ensure a protection of fundamental rights similar to those guaranteed by the German Basic Law.16 If the Federal Constitutional Court dismisses Outokumpu’s complaint, the company may file an application to the European Court of Human Rights. Unlike the constitutional courts of the Member States, the Strasbourg Court has to exercise its judicial review of EU decisions under the ECHR unconstrained by the principle of primacy of EU law. It is not sufficient for the Commission to ensure fundamental rights only similar to those protected by the ECHR before taking a decision, which would in particular fulfil the requirements of German constitutional law; the Commission must comply with the ECHR standard in full. Although it clearly should be avoided, it

ThyssenKrupp Nirosta, above n 4, para 78. Case C-119/05 Ministero dell’Industria, del Commercio e dell’Artigianato v Lucchini SpA [2007] ECR I-6199, para 41. 15 BVerfG 2 BvR 2752/11. See Karin Matussek, bloomberg.com, 21 February 2013; Stefano Berra, Global Competition Review, 25 February 2013. Following the completion of this chapter, the Federal Constitutional Court rejected the complainant’s request for interim measures. See BVerfG, decision of 30 August 2013—2 BvR 2752/11, WuW/DE-R 4081—ThyssenKrupp Nirosta. 16 See BVerfG, decision of 4 April 2001—2 BvR 2368/99, Neue Juristische Wochenschrift 2001, 2705, 2706; BverfG, decision of 27 April 2010—2 BvR 1848/07, para 13, Gewerblicher Rechtsschutz und Urheberrecht Internationaler Teil 2011, 69. 13 14

390 Wolfgang Kirchhoff cannot be excluded that there might be slight but decisive differences in the understanding of certain fundamental rights between the Strasbourg Court and the Union Courts in Luxembourg.

3. The Swiss Publigroupe judgment and Regulation 1/2003 In June 2012, the Federal Court of Switzerland decided the Publigroupe case.17 This judgment contains interesting considerations on the impact of the ECHR on Swiss competition law, which may be relevant for EU competition law as well. The prohibition of abuse of a market dominating position under Article 7 of the Swiss Cartel Law (Kartellgesetz—KG) is quite similar to Article 102 TFEU. Article 7 KG also contains the notions ‘market dominating position’ and ‘abuse of a market dominating position’, which are indefinite legal concepts to be interpreted by courts. Fines for abuse are stipulated in Article 49a KG. According to the third paragraph of this provision, no fine may be imposed if the undertaking makes an announcement of the intended practice to the Competition Commission five months before it takes effect and the Commission does not open an investigation. In view of the Swiss Federal Court, this possibility of ‘announcement in advance’ is a necessary compensation, from the perspective of Article 7 ECHR, for the use of indefinite legal concepts in the Swiss provisions on sanctions.18 On the basis of this reasoning, questions concerning the compatibility of the system of decentralised application of Article 101(3) under Regulation 1/2003 with the ECHR may arise. Article 10 of Regulation 1/2003 provides for a finding of inapplicability by the Commission acting on its own initiative if so required by the public interest of the Union. And according to the ECJ in Tele2 Polska, national law may not provide for a proceeding of an NCA to be brought to an end by a decision stating that there has been no infringement of EU competition law.19 Thus, unlike in a system of prior notification and exemption as established by Regulation 17/62, undertakings are no longer able to achieve legal certainty as to the lawfulness of an agreement and to obtain immunity from fines. Instead, they must rely on their own assessment of their practices regarding their compliance with competition law.20

17 BGH, judgment of 29.6.2012—2 C_484/2010—Publigroupe, Bundesgericht Entscheidungen (BGE) 139 I 72. 18 Ibid, 90. 19 See Tele2 Polska, above n 5, para 35. 20 See, eg U Böge and A Bardong in Günther Hirsch, Frank Montag and Franz Jürgen Säcker, eds, Competition Law: European Community Practice and Procedure, Sweet & Maxwell, 2008, 4-1-009, 4-1-022.

Reflections on Parallel Enforcement 391 In order to make such an assessment, the undertakings need to apply indefinite legal concepts, eg ‘market dominating position’ or ‘abuse’. The crucial question is whether it is sufficiently foreseeable for an undertaking that a certain type of conduct amounts to a violation of competition law that justifies the imposition of a fine. Following the judgment of the Swiss Federal Court in Publigroupe, this question—already discussed before adoption of Regulation 1/200321—needs to be considered and answered again. According to the case law of the European Court of Human Rights, the notion of law in Article 7 ECHR implies qualitative requirements, notably those of accessibility and foreseeability. The scope of the notion of foreseeability depends to a considerable degree on the content of the text at issue, the field it is designed to cover, and the number and status of those to whom it is addressed.22 A law may still satisfy the requirement of foreseeability even if the person concerned has to take appropriate legal advice in order to assess, to a degree that is reasonable in the circumstances, the consequences which a given action may entail.23 This is particularly true in relation to persons carrying out professional activities, who are used to having to proceed with a high degree of caution when pursuing their occupation. On this account, they can be expected to take special care in assessing the risks that such activity entails.24 In their self-assessment under Article 101 TFEU, undertakings and their advisers can refer to the Commission’s Block Exemption Regulations and guidelines, and to the case law of the European Courts.25 Moreover, they may seek clarification from relevant NCAs by means of a decision stating that there are no grounds for action, and they may refer to the case law of national courts. Nevertheless, some doubts still remain as to whether these information sources will always give enough certainty to know where a ‘grey area practice’ crosses the borderline and constitutes an infringement. It can be expected that the Commission and the Union Courts will utilise the notion of negligence, and the Commission will use its power of discretion to impose fines in these grey areas cautiously, so that conflicts with Article 7 ECHR are avoided. The same is true for established NCAs and for national courts. However, caution is advised when applying EU law in new Member States without a long-standing tradition in competition law.

See ibid, 4-1-009. See Case 10890/84, Groppera Radio AG and Others v Switzerland, 28 March 1990, para 68. 23 See Case 18139/91, Tolstoy Miloslavsky v the United Kingdom, 13 July 1995, para 37. 24 See Case 17862/91, Cantoni v France, 11 November 1996, paras 29 ff. 25 The conditions for a guidance letter by the Commission are so demanding that they can only be satisfied in a rather small number of cases. See Böge and Bardong, above n 20, 4-1-024; Commission Notice on Informal Guidance, [2004] OJ C101/78. 21 22

392 Wolfgang Kirchhoff Against this background, new concerns about the legitimacy of the selfassessment system established by Regulation 1/2003 in view of the Publigroupe judgment of the Swiss Federal Court do not seem to be well founded. An inter-state claim against Regulation 1/2003 under Article 33 ECHR is very unlikely, if not excluded. An individual application (Article 34 ECHR) requires a natural or legal person claiming to be the victim of a violation of the rights set forth in the Convention by the EU or a Member State. In the given context, such violation must be related to the application of EU competition law. The Strasbourg Court may only deal with the matter once all domestic remedies have been exhausted (Article 35 ECHR). Under these circumstances, an EU competition case decided by the Commission or an NCA can only violate the principle of legal certainty found in Article 7 ECHR under very exceptional circumstances. As a result, the application of Regulation 1/2003 in an individual case may be illegitimate. However, this does not concern the system of self-assessment as such.

William E Kovacic* Creating a Respected Brand: How Competition Agencies Signal Quality

1. Introduction The modern expansion in the number of jurisdictions with competition laws— from roughly 20 in 1990 to over 120 today—has focused increased attention on how the design and management of competition agencies can improve their effectiveness.1 The establishment of new systems has led agency officials, practitioners and researchers to examine more closely the links between institutional characteristics and substantive policy results. Various international organisations have placed greater emphasis on policy implementation and performance measurement.2 One determinant of an agency’s effectiveness is its reputation, or ‘brand’. Much like a commercial enterprise, a competition agency develops a brand that signals quality to various observers. A good reputation can help the competition agency recruit skilled personnel, gain deference from courts, build credibility with business managers and build popular support that can yield larger budgets and enhancements to its powers. An agency with a strong brand stands a greater chance of being effective than one with a weak brand. This chapter considers how branding can affect agency performance. It analyses how investments in building a good brand enable the competition agency to signal quality to various observers—insiders such as agency staff and outsiders such as businesses, consumer groups, courts and legislators. The chapter emphasizes that the establishment of a good brand occurs mainly by

* Global Competition Professor of Law and Policy, George Washington School of Law and Non-executive Director, UK Competition and Markets Authority. 1 William E Kovacic and David A Hyman, ‘Competition Agency Design: What’s on the Menu?’, 8 European Competition Journal 527 (2012). 2 This is evident in the work plans of organizations such as the International Competition Network (ICN), the Organization for Economic Cooperation and Development (OECD) and the United Nations Commission on Trade and Development (UNCTAD). The ICN has formed a working group on agency effectiveness, which addresses a wide range of implementation issues. The OECD performs peer reviews of its members and observer jurisdictions, and its Competition Committee has undertaken a long-range project concerning performance measurement. UNCTAD has its own program of peer reviews, which devote considerable attention to institutional arrangements.

394 William E Kovacic design, not by accident. It underscores how strong mechanisms to set strategy and priorities provide indispensable foundations for successful branding. The examination of agency branding has several purposes. One aim is to improve our understanding of how public agencies build a reputation, and to study the role of reputation in determining effectiveness. A closely related goal is to give public officials a better understanding of how they should approach the task of deciding what their agencies must do to prosper. A further aim is to underscore the impact of institutional design and managerial incentives on agency performance and to illuminate how design choices and incentive schemes influence the development of a well-respected, coherent agency brand. Various design choices—for example, to give the competition agency a single function or a multi-purpose substantive mandate, to govern the agency by a single executive or by a board, to integrate the tasks of prosecution and adjudication in a single body or to unbundle them among distinct entities—affect the capacity of the agency to enhance the quality of its brand. Incentives that give incumbent leaders reason to make investments in long-term agency capacity and quality have the same effects. The chapter first defines the concept of a brand for public agencies. It then discusses why an agency’s brand can be important to its effectiveness, and identifies what types of agency activities either enhance or degrade an agency’s brand.

2. Brands and public institutions Public institutions, such as competition agencies, build brands that the agency’s own employees and external observers associate with the agency. Brands perform two functions for the public agency. The first is informational: a good brand conveys a clear sense of what an agency does. It communicates, at least in a general way, the scope of the agency’s responsibilities and the aims that motivate the agency in the exercise of its powers. For example, for many years, the United Kingdom’s Office of Fair Trading—succeeded now by the Competition and Markets Authority—sought to achieve these ends with the phrase ‘Making Markets Work for Consumers’. A brand also serves as a signal for institutional quality. For a competition agency, the foundations for a good brand are sound substantive programs (eg cases, regulations, reports), sound procedures (eg meaningful disclosure of information, rigorous testing of evidence, regular assessment of outcomes), strong capabilities (eg deep expertise in economics and law) and a healthy culture (eg thoughtfulness, integrity, courage, a commitment to continuous improvement). As described more fully below, a good brand advances the agency’s interests in a number of ways.

Creating a Respected Brand 395 Deference from judges. Courts form views about the quality of public agencies—especially if the same court reviews a number of agency decisions over time and develops a general sense of how well the agency functions. An agency with a good reputation is more likely to gain the benefit of the doubt (the ‘margin of discretion’) from a reviewing tribunal than an agency with a weak reputation. Where judges exercise the discretion inherent in the interpretation of statutes, they take account of the agency’s reputation for doing strong substantive work and using sound methods to make policy. Deference from legislators. A good brand can help the agency persuade legislators to approve generous budgets and provide desired expansions of the agency’s authority. Respect from legislators (and other budgetary gatekeepers) can yield needed increases in resources. Deference from other public regulators. A good brand can enhance the power of advocacy before other government bodies—national ministries, local officials—which make regulatory decisions that affect the quality of competition. An agency with a reputation for extensive expertise is more likely to be seen as a valuable source of guidance than an agency believed to have little idea of what it is doing. Respect from and credibility with non-government groups. A competition agency’s effectiveness depends substantially on the reputation it develops among various groups beyond the courts, legislatures, ministries and other government policy-making institutions. A good brand builds credibility with the general public, advocacy groups, universities, the media, professional societies, trade associations and individual businesses. For example, the agency’s reputation among these groups can influence the views of legislators about whether the agency deserves greater resources and more authority. The experience of the US Federal Trade Commission (FTC) from the late 1970s through the early 1980s illustrates the importance of branding to an agency’s effectiveness. In the late 1960s, two major studies of the FTC concluded that the agency had performed poorly for most its history. One blue-ribbon committee recommended that, if the agency did not undertake immediate, drastic improvements, Congress should dismantle it and allocate its functions to other government bodies. The US Congress endorsed this view, and the FTC responded with an ambitious program of competition and consumer protection initiatives. Two problems accompanied the FTC’s new agenda. First, the agency did not account for the political consequences of taking on an especially broad and powerful collection of commercial interests simultaneously. The Commission’s programs galvanized strong, bipartisan political opposition that threatened its very existence by the end of the 1970s and into the early 1980s. Second, the agency experienced a severe mismatch between its program commitments and its capacity to deliver. Many high-profile initiatives

396 William E Kovacic s uffered because the agency could not mobilize the top-quality talent to carry them off successfully. Well before Ronald Reagan came to Washington, DC in January 1981, the FTC’s brand had disintegrated in the eyes of various significant external observers. In March 1978, the Washington Post—a key barometer of liberal political sensibilities—scorned the FTC for proposing to restrict television advertising directed towards children. An editorial in the newspaper called the proposal ‘a preposterous intervention that would turn the agency into a great national nanny’. The ridicule from a left-of-center publication ordinarily disposed to favor rather than oppose regulatory intervention meant that it was open season on the FTC. In the spring of 1980, during legislative debates about measures to curb the FTC’s powers, one prominent member of the House of Representatives called the agency ‘a king-sized cancer on our economy … a rogue agency gone insane’. Another said ‘[t]he way to kill a rattlesnake is to cut its head off. That is what we ought to do today.’ On the day before the 1980 presidential election, Vice-President Walter Mondale attacked the FTC for pursuing a case that sought to break up the four leading US cereal producers into smaller firms. Mondale informed a political rally in Battle Creek, Michigan (Kellogg’s headquarters) that ‘it is inconceivable to me and to many independent experts that divestiture would be pursued. Neither President Carter nor I would support such an action.’ Mondale added that, if re-elected, he and Carter ‘certainly would’ support legislation to bar the FTC from obtaining a divestiture remedy. In a number of instances, the courts also made statements that suggested a low regard for the agency’s substantive analysis or its way of doing business.3

3. Forces that determine the quality of a competition agency’s brand A variety of entities inside and outside a competition agency play roles in defining the agency’s brand. The discussion below identifies the internal and external forces that shape the brands of public agencies.

3 See American Medical Association v FTC, 638 F2d 443, 454 (2d Cir 1980) (Mansfield, J, dissenting) (‘[T]he Commission insisted on pressing for its pound of flesh. In my view, the FTC is engaged in the futile business of beating a dead horse’).

Creating a Respected Brand 397

3.1 Effective orientation: strategy, priorities, program selection The quality of an agency’s brand depends on the strength of its processes for establishing a strategy, setting priorities, and selecting programs to achieve its ends. An effective process does not consist of mechanically doing what the agency did the year before. Without a clear statement of ends and means, it is impossible for the agency to provide useful guidance to its own employees or to explain its programs coherently to external constituencies. A number of obstacles impede the formulation and articulation of strategy and priorities. One of the most basic hurdles is the tendency of legislatures to command the agency to achieve a multiplicity of goals. These can be, and often are, internally inconsistent (eg they improve efficiency but preserve the position of small and medium-sized enterprises). A faithful effort to recite these goals would reveal the essential contradictions of the legislative scheme. A fully transparent attempt to rationalize the goals structure by elevating some goals and subordinating others would attract criticism from elected officials who favor the subordinated goals. These conditions tend to press agencies towards the adoption of goals statements that obscure, rather than clarify, their aims and methods.

3.2 Coherent assignment of functions The clarity of an agency’s brand depends on the nature of the agency’s responsibilities. Over half of the jurisdictions with competition laws assign the competition agency functions beyond competition agencies. In addition to competition law, these multi-purpose agencies enforce consumer protection laws, oversee public procurement and set access prices for regulated s ectors, such as energy and telecommunications. As an agency’s policy portfolio expands and its responsibilities become more diverse, it is likely to encounter a greater level of difficulty in defining its purpose and building its brand. This difficulty is diminished if, at least conceptually, the multiple policy duties are strong complements. A mix of competition and consumer protection could be said, in general terms, to satisfy this condition by bringing supply-side policy goals (increase the number and quality of choices available to consumers) and demand-side policy goals (ensure that consumers can make well-informed choices among alternatives without distortions created by fraud or misrepresentation) under the same regulatory house. Branding is more difficult if the policy duties are substitutes or entirely unrelated. An agency assigned to enforce a competition law and to impose anti-dumping penalties is likely to experience recurring episodes of

398 William E Kovacic schizophrenia.4 Other agencies can become a legislature’s default destination for regulatory duties not easily located elsewhere. The Federal Trade Commission, for example, enforces over 50 statutes that encompass a broad range of competition, consumer protection and data protection mandates. Some of these measures run far afield from what might be regarded as the core of the FTC’s responsibilities.5

3.3 Meaningful disclosure Each of an agency’s public statements—by means of a decision in a case, a report, a guideline, a speech or testimony before a government body—is an occasion to shape its brand. The clarity and quality of the brand depend on the agency’s ability to spell out its priorities and link individual initiatives to the attainment of its stated aims. Each public statement, from high officials to the most humble case handlers, should reinforce the agency’s major themes. In part, this simply involves the solution of familiar problems of coordination, communication and monitoring that arise in many interactions between principals and their agents. The principal must select a suitable agent, spell out her preferences clearly and give informative instructions to the agent, then monitor fulfillment of the assignment. These tasks grow in difficulty as the size of the institution under the principal’s supervision grows in size and complexity.

3.4 Agency structure and policy coherence Agency design can determine the difficulty of the basic task of setting and articulating the agency’s priorities. Suppose that, instead of a single executive hierarchy, the agency’s governance mechanism consists of a college of decision makers. Well over half of the world’s competition systems are governed by multi-member boards. Achieving coherence in external messaging is likely to be more difficult with a board than with a single-person executive. Branding in the multi-member board setting is a matter of team production, as the overall impression conveyed to outsiders by the board depends on the commitment of individual members to build and reinforce a coherent brand. In some cases, the

4 In its first years, Panama’s competition system united competition, consumer protection and anti-dumping safeguards under the same roof. Panama later spun off the antidumping mandate to a separate agency. 5 One notable outlier is the Mohammed Ali Boxing Act, which gives the FTC a role in regulating professional boxing matches.

Creating a Respected Brand 399 board’s members may share a strong sense of common cause and align their messages to state priorities clearly and coherently and to emphasize agreedupon themes in all public statements. For various reasons, this result is not inevitable. Divergent preferences among board members may be so accentuated as to frustrate the attainment of consensus.6 Non-chair board members may chafe at their subordinate status relative to the chair, and may use speeches or other statements to stake out positions that will attract attention at the cost of agency coherence.7 The interaction of board members is probably best understood as a repeated game in which possibilities for retribution and tit-for-tat retaliation discourage extreme deviations. Yet these sanctioning mechanisms may be only partially effective in discouraging sharp variations, especially when an individual member derives substantial utility from playing a visible role as a maverick. This has important implications for the selection of individuals who will serve on a board. Ideally, in addition to superior technical skills, board nominees should each possess personal sensibilities that facilitate effective collegial interaction and decision making. Moreover, the board’s chair should have a demonstrated ability to form consensus and to dispense credit freely to others. These attributes are difficult to screen for in the appointments process, where it is often difficult enough to obtain appointees with consistently high technical skills. It may be that a system should trade off some level of technical skill in exchange for a person’s evident capacity to work agreeably in a college. An individual with merely adequate (not superior) accomplishments in competition law who also has good personal sensibilities might be a better board member than a person who has exceptional experience and knowledge in competition policy but whose preferred style in dealing with colleagues is hand-to-hand combat.

3.5 Location of the competition authority Other institutional design characteristics can impede the kind of coherence in expression that improves the brand. Most competition agencies are standalone entities, but others (such as the Directorate General for Competition of 6 This is particularly true where the statutory criteria for appointment include a political diversification requirement. In the United States, for example, no more than three of the FTC’s five board members can belong to the same political party. 7 In my experience as a member of the board of the FTC and from discussions with officials in dozens of competition agencies over the past 20 years, I have noticed how frequently internal strife stemming from acute policy disagreements, personality clashes, jealousy and spite have undermined the effectiveness of multi-member boards. I suppose this human condition should surprise nobody above a certain age, yet it is rarely taken into account in considering how collegial governance mechanisms function and why they sometimes perform poorly.

400 William E Kovacic the European Commission and the Antitrust Division of the US Department of Justice, DOJ) are subdivisions of larger complex institutions. A competition agency’s branding efforts can suffer when it resides within a larger institution whose stature lags behind the competition agency’s reputation— a circumstance similar to that of a first-rate academic department located within a second-rate university. The weaker reputation of the larger enterprise can spill over onto the brand of a superior subdivision, such as a competition agency.

3.6 Importance of internal agency structure The internal structure of the competition agency can also complicate coherent priority setting and branding. Some competition systems establish high levels of separation between case-handling units and the agency’s political leadership. For example, in Germany’s Federal Cartel Office (Bundeskartellamt, or BKA), the case-handling chambers enjoy great autonomy vis-à-vis the President of the agency. The President has no voice in the initiation or conclusion of the agency’s cases. Thus, in a fundamental sense, the President does not set priorities or define the agency’s objectives. He can offer suggestions or make recommendations to the chambers, but he cannot dictate what they will do. When the President speaks on behalf of the agency, he can provide a synthesis of past and current activities and interpret what they mean for future policy making. The President can offer predictions about future work (based on past decisions by the BKA’s case-handling chambers), but he cannot directly dictate the course of policy for the institution.8

3.7 Work product volume and quality A competition agency determines its brand mainly through the volume and quality of its work product. Presented below are some positive observations about how an agency’s work product influences its brand.

8 The President of the BKA has some influence over the chambers, but it is indirect. The President selects the heads of the chambers and can use this appointment authority to choose individuals who seem to share the President’s views about what the agency should do. These appointments, however, come with the equivalent of tenure for the selected officials. Once the appointments are made, the President loses influence over the chamber decision makers.

Creating a Respected Brand 401

3.7.1 Activity levels are proxies for quality For the most part, groups external to the agency tend to equate higher levels of activity with higher quality. An agency that does more things typically is seen to be a better agency than an agency that does fewer things. This principle can be derived from reading the proceedings of international organizations such as the OECD and UNCTAD,9 the ratings of publications such as the Global Competition Review, and academic and popular commentary on what competition agencies do. The most highly valued form of activity consists of cases, and the initiation of law enforcement proceedings involving high profile enterprises tends to receive a higher weighting. By contrast, the filing of reports and the performance of various functions grouped under the category of ‘advocacy’ receive less attention and count for less. It is important not only to bring a certain level of cases, but to win a sufficient number of them (not all) to indicate that the agency can succeed in court. A small number of highly visible litigation victories can accomplish this result. A corollary to this principle is that higher levels of fines or sanctions are often taken as signs of higher quality. An agency with law enforcement activities that yield a larger amount of monetary recoveries is deemed to be more effective than one with smaller recoveries.

3.7.2 A minimum level of activity is necessary to create an image of effectiveness As a practical matter, an agency with an activity rate falling below a s pecific level runs a risk of being deemed ineffective, and its program will be depicted by external observers and in media accounts as being deficient. This is true regardless of whether the socially optimal decision in a specific case would be to stand aside or whether a program of little or no enforcement for certain types of offenses might improve economic performance. In this framework, decisions not to prosecute tend to receive little or no credit. To withstand criticism of individual decisions not to prosecute, it is important for the agency to accumulate some reputational capital in the form of previous decisions to intervene. Past prosecutions give the agency capital it can spend in future matters when agency leadership concludes that a case is not warranted. For new leadership, it is ideal to have some early matters in

9 The OECD and UNCTAD both perform peer reviews of competition agencies. A common metric used in these exercises to determine the effectiveness of a competition agency is whether it has been ‘active’—ie is the agency bringing a substantial number of cases, and is the number increasing over time?

402 William E Kovacic which the agency can take action and thus establish its willingness to act and be tough when necessary. DOJ enforcement of the US prohibitions on dominant firm misconduct provides a possible illustration. DOJ Antitrust Division leaders appointed under George W Bush initiated no cases involving monopolization or attempted monopolization. This figured prominently in attacks made by Barack Obama against the Bush Administration’s antitrust program d uring the 2008 presidential campaign.10 This number contrasts with seven such cases begun under leadership appointed by President Bush to head the FTC. As a matter of policy and strategy, the decision to bring no dominant firm misconduct cases arguably robbed the Bush Antitrust Division of an important measure of political and reputational capital. Zero is an easy number to understand. One would not want to encourage a law enforcement body to bring cases simply for the sake of bringing cases. Instead, the Bush Administration Antitrust Division leadership might have been well advised to develop at least a small number of cases, which could have been done consistently with the incumbent administration’s philosophy (eg dominant firm exclusion accomplished through misuse of the regulatory process).

3.7.3 Queuing effects: the order of potential enforcement events can influence the decision to prosecute The queue of matters that come before an agency is partly determined exogenously and partly determined endogenously. Mergers provide an illustration. From the perspective of the competition agency, a major cause of merger rates is the state of the economy. In periods of growth and ascending stock values, firms are more likely to undertake mergers than when economic conditions are bleak. The decision to merge, however, also depends on an endogenous factor. The agency’s enforcement record and its statements of enforcement intentions shape perceptions of potential merging parties and their advisors about whether to proceed. The order in which matters come before the agency may affect what the agency decides to do. In the mid-1990s, under the leadership of Robert Pitofsky, the FTC achieved important litigation merger victories in transactions involving office supplies (Staples/Office Depot) and pharmaceutical distribution systems (Cardinal/Brunswick). Later in the decade, the FTC allowed Boeing to purchase McDonnell Douglas without restrictions and permitted several large mergers of petroleum companies (most notably,

10 President Obama’s Antitrust Division leadership repudiated a DOJ study that set out a narrow view of improper dominant firm conduct and announced plans to increase enforcement in this area. The yield to date has been a single settlement—involving exclusive dealing by a hospital system in Wichita Falls, Texas.

Creating a Respected Brand 403 Exxon’s purchase of Mobile) with some divestitures. Consider how the FTC might have evaluated Boeing/McDonnell Douglas or Exxon/Mobil if one of those transactions had occurred earlier in Pitofsky’s tenure. Would the FTC Chairman, who had criticized enforcement policy under the Reagan administration as being too lax, have allowed Boeing to purchase McDonnell Douglas if the deal had been the first major transaction to emerge in, say, 1995? One possible interpretation of FTC merger enforcement policy in the 1990s is that the successful challenges to the Staples and Cardinal transactions established the agency’s reputation for toughness. These litigation victories enabled the agency to say, when the Boeing merger and the petroleum deals came along, that it was willing to intervene when the facts required it, but was sufficiently discerning to stand down when in its view the transactions appeared benign. The queuing effect might have other consequences for an agency’s enforcement choices. Consider again the Boeing/McDonnell Douglas example. Many commentators criticized the FTC’s decision to take no action in relation to that transaction. Some observers accused the FTC of rolling over in the face of US political pressure applied by the White House to advance Boeing’s efforts to cope with the increasingly formidable commercial aircraft program of Airbus.11 Other critics said the FTC’s inaction clashed with Pitofsky’s scolding of the Reagan administration for allowing various mergers of large enterprises and his commitment to block deals that yielded especially high post-merger market shares. From a distance, the Boeing/McDonnell Douglas deal—a 3 to 2 transaction—seemed ripe for an FTC challenge. Roughly a year later, Heinz sought to acquire Beech-Nut’s baby food business, and the FTC ultimately succeeded in blocking the merger.12 On its face, the merger was a 3 to 2 transaction that would leave Gerber (the market leader with a 70 per cent share), Heinz (with roughly 28 per cent) and a fringe. Heinz argued, without success, that this was not a 3 to 2 merger but instead a 1 to 2 transaction—that the deal would enabled the merged enterprise to mount an effective challenge to the well-established industry leader, Gerber, whose dominance had persisted for over 50 years. There is some evidence that the Heinz/Beech-Nut efficiency and competitive effects story had merit, and Gerber’s share of the North America baby food market has climbed to over

11 In the months before the FTC’s decision not to intervene, Vice-President Al Gore visited one of Boeing’s production facilities outside Seattle and told an assembly of Boeing workers that moves by the European Commission to block the acquisition of McDonnell Douglas could lead to a trade war. I suspect the awareness of this preference was not the reason the FTC stood aside, but I also understand why many observers would conclude otherwise. This episode is recounted in William E Kovacic, ‘Transatlantic Turbulence: The Boeing-McDonnell Douglas Merger and International Competition Policy’, 68 Antitrust Law Journal 805 (2001). 12 FTC v HJ Heinz, 246 F3d 708 (DC Cir 2001). I was not a neutral observer in this case. I wrote a brief supporting the merger in the appeal to the DC Circuit.

404 William E Kovacic 80 per cent.13 Maybe the FTC stopped a 3 to 2 merger it should have allowed (Heinz) and should have imposed conditions on a 3 to 2 merger that it waived through (Boeing). The FTC’s decision in the baby food merger may have illustrated the queuing effect. The Heinz transaction arose a relatively short time after the Pitofsky FTC had decided not to oppose Boeing’s acquisition of McDonnell Douglas. Heinz/Beech-Nut also followed the FTC’s high-profile decision to permit Exxon to buy Mobil (a deal valued at over $100 billion and resolved with roughly $4 billion of divestitures). To allow Heinz to buy Beech-Nut (and to endorse a 3 to 2 deal on the basis of efficiency arguments that seldom rescue mergers in the highest atmosphere of concentration) would again have prompted aggressive questioning about the Commission’s merger program and Pitofsky’s commitment to tough merger enforcement. If the FTC had approved Heinz/Beech-Nut, Pitofsky’s legacy might have changed in unwelcome ways. He might have been remembered more as an official with a blind spot for big 3 to 2 deals (Boeing and Heinz) and petroleum industry consolidation rather than the man who spearheaded the Staples case.

3.7.4 The persistence of presumptions favoring intervention This discussion does not assume that activity levels are a suitable measure of agency quality. To say that an agency is bringing lots of cases, or collecting substantial fines, does not establish that its program is improving economic performance—the genuine test of effectiveness. Activity is not a synonym for accomplishment. If one acknowledges the hazards of simply counting cases and fines, why is activity so often taken to be an important signal for quality? One reason is that the assessment of outcomes of enforcement decisions (to intervene or not to intervene) can pose difficult methodological challenges and can be expensive. Given the choice between bringing the next case (with its immediate credit-claiming opportunities) and performing an ex post evaluation, agency leadership might well prefer the first course of action. To the extent that ex post assessments might discredit an agency’s existing presumptions about whether to intervene, the agency might prefer to assert the correctness of its policies on the basis of a priori reasoning and forego empirical testing. Another reason for the popularity of activity measures of effectiveness are the professional interests of numerous observers who shape opinions about the agency’s performance. The livelihoods of antitrust lawyers depend significantly upon the level of public agency enforcement activity.

13 See Viola Chen, ‘The Evolution of the Baby Food Industry, 2000–2008’, 6 Journal of Competition Law and Economics 423 (2010).

Creating a Respected Brand 405 The s ignificance of this activity grows where private rights of action do not exist or are feeble. Even the most ardent members of the defense bar require the fact or threat of intervention to obtain clients. The same is true for economic consultants. To say that there is economic self-interest among these professional groups does not discredit their endeavors, or their preference for higher (rather than lower) rates of enforcement. A physician who prescribes medical treatment as an alternative to bloodletting to cure sickness can be said to be advancing his own economic interests, yet his efforts to promote medical treatment generally coincide with the interests of prospective patients and society as a whole. There comes a point, however, at which a physician might prescribe a treatment regime that is excessive in the sense that it does not advance the patient’s well-being but instead results in an intensity of care-giving that serves mainly to raise the physician’s fees. The element of self-interest among professional advisors (economists and lawyers) in competition law resides not in their preference for certain levels of activity, but rather in the dosage they prescribe. Law firms and economic consultancies might well be correct in suggesting that antitrust agencies should exercise their enforcement power, but they might not be trusted entirely when they suggest how much. The same bias for action can be observed in the field of academics, especially where scholars also engage in professional consulting. An antitrust program with no enforcement would likely diminish the demand for courses in competition law and industrial organization economics, and would reduce the perceived value of papers on competition policy. In some respects, the ideal economic or legal model is one that creates a presumption favoring intervention in some area of competition law. The presumption towards intervention enables the academic/professional advisor to act for agencies or private firms that seek to intervene, but—just as importantly—it allows the academic/advisor to work on behalf of those who oppose intervention. In the latter case, the academic/advisor can say that, although he and his model are sympathetic to intervention generally, the facts of this case fail to satisfy the demanding conditions of the model. The ideal model creates opportunities for representation on behalf of plaintiffs and defendants, with the academic/ advisor serving as the knowledgeable expert who can discern when intervention or forbearance is appropriate. Legislators might also have a preference for activity that results in intervention, or the threat of intervention, by the competition agency. This is especially true in a system in which affected commercial interests are a crucial source of electoral resources (eg campaign contributions or the mobilization of voters) that legislators require to gain and hold office. The possibility of agency intervention creates two categories of parties which can provide such resources—firms that desire intervention and those that oppose it. If the agency intervenes, a legislator whose constituents benefit from the action can claim credit for it. At the same time, a legislator whose constituents and

406 William E Kovacic usiness patrons oppose intervention can seek to apply his influence to foreb stall or modulate the agency’s action. For both groups (legislators who desire agency intervention and those who oppose it), a decision-making presumption that favors agency activity serves their electoral interests. A propensity to intervene is also compatible with the preferences of media organizations which cover enforcement agencies. Journalists prefer to write about action rather than inaction. The prosecution of cases—most importantly, matters involving well-known enterprises—produces visible, ongoing events that demand reporting. Agencies that do not bring cases do not generate news coverage events that media organizations value highly. Journalists are more likely to write favorable accounts of agencies that create ‘news’ and to criticize, as ‘dormant’ or ‘moribund’, agencies that do not.

3.8 Quality control processes An agency can enhance its brand by creating and applying institutional processes that press the agency towards improving the quality of its work. In modern discussions about agency effectiveness, these quality control mechanisms are often analysed under the rubric of ‘procedural fairness’ or ‘procedural due process’. Procedural fairness or due process can be seen as having two dimensions. The first ingredient is the establishment of mechanisms that test evidence rigorously and thereby increase the agency’s capacity to diagnose behavior correctly and to impose appropriate cures for apparent competitive ills. Among other results, a regime of rigorous evidentiary testing overcomes the confirmation bias that can arise when an agency formulates an initial hypothesis about likely competitive effects and then refuses to abandon the hypothesis in the face of theories or evidence that dictate a rethink. The second dimension of due process or fairness is to create the appearance of truth seeking as well as to accomplish this result in fact. At a conceptual level, the appearance of fairness is achieved through the reliance (visible to agency outsiders) on mechanisms that test evidence rigorously and operate as safeguards against confirmation bias. Providing clarity about enforcement intentions and allowing affected parties the opportunity to contest agency theories and portrayals of fact are key means to these ends. The pursuit of procedural fairness, for the sake of sound substantive analysis and the appearance of fairness, has important implications for agency design. This is especially true for the decision to bundle or unbundle major decision-making tasks—the decision to investigate, the decision to prosecute, the determination of guilt or innocence, and the imposition of sanctions. As a general matter, the more a system moves towards complete integration of these functions in a single body, the more it is likely to confront greater difficulty in testing evidence rigorously and being seen as a fair decision maker.

Creating a Respected Brand 407 This is the challenge facing agencies such as the US FTC and the European Commission. Each institution has chosen different mechanisms to increase confidence in its process and to signal system quality to outsiders. In competition matters, the FTC board receives recommendations from three distinct entities in nearly each matter it addresses—the Bureau of Competition, the Bureau of Economics and the General Counsel. Administrative adjudication takes place before a hearing officer with substantial independence (and protection from dismissal by an unhappy board), testimony is taken under oath and witnesses are subject to cross-examination. Nonetheless, the same body that decides to prosecute a case (the board) is also the final body for appeals within the Commission on questions of liability and remedy. In a rough sense, the same entity that decided to bring the case gets to decide if its initial intuition about whether the respondent infringed the law was correct. The famous trilogy of merger setbacks for the European Commission before the Court of First Instance (now the General Court) in 2002 stimulated major procedural reforms within the Commission. DG Comp revamped its internal procedures to test evidence more rigorously and to give the outward appearance of using sound process. It created the position of Chief Economist and established internal adversary panels to challenge positions staked out by case-handling teams. DG Comp has been engaged in a continuing search for refinements of its hearing process for the same purpose. An important audience for an agency’s establishment of quality control mechanisms is the judiciary. Competition systems usually subject decisions of the competition agency to judicial review. Regardless of the nominal standard of review, the intensity of judicial oversight probably depends upon the court’s perception of the quality of the agency’s procedures. The stronger the agency’s internal safeguards, the better its prospects of convincing courts to show the deference which the statute says the agency is entitled to receive. As the stakes in a case increase, a reviewing court may well exercise still greater scrutiny. Courts that are willing to defer when the fine at issue is 1 million euros may be more demanding of the agency when the sum in question increases to 1 billion. Another important form of quality control is the willingness to engage in ex post assessment of past enforcement decisions to intervene or not. A commitment to submit the agency’s work to ex post review can increase confidence that the agency is determined to improve the quality of its programs and make necessary adjustments.

3.9 Engagement with external bodies Agencies can strengthen their brands through interaction with external bodies. Means to this end include speeches, the issuance of guidelines and the convening of events at which the agency elicits the views of outsiders

408 William E Kovacic about what it has done and what it should do in the future. Each public pronouncement—a speech, a case, a report or a guideline—is an opportunity to establish and reinforce a brand. As suggested above, doing this effectively requires an agency to identify what it wishes its brand to be, to clearly state its priorities and to develop a common understanding within the agency about how to deliver its message. It is impossible to be ‘on message’ without identifying what the messages ought to be. An agency should envision itself as being in a constant conversation with important external bodies. The judiciary, again, provides a good example. Agencies speak to the courts in many ways—most obviously in the filing of formal papers, but less obviously in the publication of reports, the preparation of guidelines and the delivery of speeches. An agency that seeks deference from the courts should want to convey the view to judges that it is a thinking person’s institution: its work standards are demanding, its commitment to good policy results is unyielding, and its procedures strive to test evidence rigorously and overcome confirmation bias. In the eyes of judges who see an agency on a recurring basis, the agency trades on what amounts to a stock exchange, and its reputation ascends or declines based upon the quality of the work it presents to the judges. An agency with a good reputation enters the courtroom with a halo and enjoys the benefit of the doubt before it speaks a word. An agency with a poor reputation is two goals behind before the match begins. On some occasions, an agency delivers messages about its programs directly to judges. After the Airtours/Schneider Electric/Tetra Laval trilogy in 2002, I observed several conferences at which DG Comp officials and judges from the CFI were speakers. DG Comp used these events to give speeches in which the Commissioner for Competition (at that time, Mario Monti) and the Director General (at that time, Phillip Lowe) carefully set out plans for internal reforms. With the judges in the audience, DG Comp, in effect, said that it had received the message delivered by the adverse merger decisions and was responding with improvements. On one occasion, a DG Comp official spoke first in a keynote address and was followed in a later keynote address by a CFI jurist (Bo Vesterdorf), who essentially said, ‘We are aware of and welcome these initiatives’.

3.10 Selection of leadership No group is more important to an agency’s branding efforts than its top leadership. This is most evident in the role that top leadership places as the public face of the agency in public appearances such as speeches and testimony before government bodies. The capacity of top leaders to set out a vision, to engage in debate and to demonstrate mastery of the substantive

Creating a Respected Brand 409 field shapes impressions about the agency and its quality. More than any other individuals, the agency leaders are most responsible for defining the agency’s ends and its strategy to attain them. This is particularly true for newer agencies in the first decades of their development. In many instances, the performance of newer agencies has been extraordinarily sensitive to the quality of leadership, and regime changes can produce dramatic variations in performance. Yet the branding role of top leadership is important in older agencies as well. The image of even an older agency rises or falls if it appears that the leader is out of his depth. International audiences are particularly sensitive to the apparent skill of an agency head. Top leaders are often political appointees. These largely come in two groups. In one group are individuals who are simply happy to have the job. In many cases they are indifferent to the agency’s longer-term interests and derive utility chiefly from the benefits associated with high office. In the other group are individuals who take the job to do something with it. These leaders tend to have deep expertise in their institution and a well-defined positive agenda for applying its powers.

3.11 Norms: aligning individual goals with agency needs A number of the measures that build a good brand (eg investments in nonlitigation activities that build knowledge or increase human capital) resemble long-term capital expenditures. This may not mesh well with the incentives that guide short-term political appointees. A political appointee may focus heavily on activities that yield immediate, appropriable gains. Case-centric performance criteria can accentuate attention to short-term individual needs rather than long-term agency requirements. A leader who desires credit-claiming events during his tenure might place a special emphasis on the initiation of new cases rather than, for example, the commencement of a study that will improve the agency’s understanding of a specific sector. It may be difficult in many institutions and political systems to develop a norm that induces incumbent leaders to make investments whose benefits emerge over the longer term and occur mainly (or entirely) after the person who launched the investment has left office. A propensity for short-term credit claiming can also cause a leader in the current period to denigrate the work of predecessors. The leader may engage in a self-serving form of framing which depicts earlier periods of work as severely deficient (‘It was terrible when I got here’) as a way of accentuating the accomplishments of current leadership (‘Through my heroic efforts, I have rescued the agency’). This behavior can be most intense at the time of a political regime change, but the simple force of personal pride and reputation

410 William E Kovacic polishing explains it as well. For political or reputational reasons, an incumbent political appointee may choose to magnify her own achievements by portraying predecessors as inadequate. The behavior can continue after an individual leader has left. Upon leaving office, the narrative then describes how the leader’s arrival saved a bad situation, but that one’s successors simply have not measured up, especially if they come from a different political party. This form of credit-claiming behavior may serve the interests of individual leaders, current or past, but it tarnishes the agency’s brand. Who, after all, would have confidence in an institution that seems so prone to wild variations in performance with each change of leadership, especially in a political regime change? Two good, rough signs of whether an agency has a culture that promotes improvements over time in the agency’s brand are whether and how often (i) an incumbent agency leader refers favorably to the work of immediate predecessors and (ii) former officials speak well of leaders who followed them. There is a different conception of policy making in which leaders treat public service as a relay race and not an individual event. From this perspective, success is the result of team production over a series of individual management episodes. Each individual leader in the succession of leaders envisions herself as a co-producer of a common product (the agency brand) and each seeks to ensure that, with each segment of the race, the position of the team has improved.

4. Conclusion A brand is an important capital asset for a commercial enterprise, and for a public agency. To an important degree, the effectiveness of competition agencies depends on the establishment and maintenance of a brand that signals quality in the institution’s process and substantive work. Good branding, in turn, requires effective systems to select a strategy, choose priorities and select programs to carry out the agency’s aims. The design of competition institutions can influence their ability to create good brands. Branding is also a function of the incentives that motivate public officials, especially top agency leadership. Deeper consideration of these design and incentive features provides a useful way to think about what makes for a good competition agency.

PART III Effectiveness and Legitimacy in International Enforcement Cooperation Chair: Philip Lowe Speakers: Jochen Burrichter Daniel Crane John Davies Alberto Heimler

Hwang Lee Paul Lugard Julie Soloway

Philip Lowe: All right, we have looked at effectiveness separately, we have looked at legitimacy separately and now we are going to look at effectiveness and legitimacy in an international context. As we said earlier, we are weak at quantification of some of the benefits of what we are doing, but if my arithmetic is correct we have got six speakers, and we are short on time. I would therefore ask the speakers to respect the time-frame. Paul will start us off. DD

Paul Lugard: Thank you. Our session is on enforcement and cooperation among antitrust agencies. I will focus on three possible ways to enhance cooperation between antitrust agencies when they would like to enforce their laws. Now, there is a general belief that effective cooperation between agencies is good, and we all know the reasons for that. One is that agencies learn from each other in various ways. They learn about best practices, and the way other agencies evaluate mergers and the way they detect anticompetitive conduct. And in this way the international community enlarges the transfer of knowledge, and that’s a good thing. More specifically, agencies work together in the actual enforcement of the law, and typically what we would see is that the lesser experienced agency would focus on knowledge transfer and the more experienced agencies would feel the need to work more closely together in the enforcement of the law and prosecution of specific cases. Now, the scope for, the practical need for and the frequency of international cooperation is increasing. There are more international cases, and one reason is of course that the business is becoming more global, more international. One source mentions that between 1990 and 2008 there have been 516 international DD

412 Part III: Effectiveness and Legitimacy in International Enforcement cases.1 So this is clearly one source of increased cooperation. Secondly, simply because of the fact that there are more agencies, the scope for agency cooperation increases. The incentives for cooperating, and the incentives of private parties to assist agencies when they would like to cooperate with each other, differ in merger cases. Private parties are often inclined to consent to waivers for the simple reason that the exchange of information may help the agencies to arrive more quickly at solutions. In contrast, in the cartel area, agencies are sometimes hindered because companies do not agree to provide waivers that would allow agencies to exchange confidential business information. This latter point is one of the three points that I would like to develop. The basis for international cooperation is manifold. There is, of course, a lot of multilateral work being done particularly in the ICN and the OECD, and that international work facilitates the cooperation between the agencies. Very generally, it would be fair to state that the ICN is based on the idea that agencies work voluntarily together, so I think the ICN is more an organisation that facilitates cooperation as opposed to prescribing precisely how that cooperation should be structured. The OECD is of another nature, and over time it has issued recommendations—including in particular the 1995 cooperation recommendation that tells OECD countries more clearly how international cooperation can be structured.2 Now, one significant obstacle to meaningful international enforcement cooperation is, as I just mentioned, the exchange of confidential business information. This is also recognised in the 1995 OECD recommendation, and there are now voices that suggest that this recommendation could be amended.3 And before I say something about it, I should note that there is an academic problem here because we don’t know precisely what inefficiencies result from the inability to cooperate. There has been a questionnaire—a combined ICN and OECD questionnaire—maybe John Davies will say something about it—and the questionnaire suggests that the inability to exchange confidential business information is a major source of inefficiencies. But we don’t know precisely how large the inefficiencies are. We don’t know how many more cases would have been prosecuted successfully if agencies were generally free to exchange confidential business information. How many more cartels would have been prosecuted? Having said

1 See Jay Pil Choi and Heiko Gerlach, ‘Global Cartels, Leniency Programs and International Antitrust Cooperation’, CESifo Working Paper 3005/2010, http://papers.ssrn.com/ sol3/papers.cfm?abstract_id=1587968. 2 Revised Recommendation of the Council Concerning Co-operation between M ember Countries on Anticompetitive Practices Affecting International Trade, OECD Doc C(95)130/FINAL (1995). 3 Following the Workshop, the OECD Council did adopt, on 16 September 2014, a new recommendation on international cooperation. The text is available at http://www.oecd. org/daf/competition/2014-rec-internat-coop-competition.pdf.

Part III: Effectiveness and Legitimacy in International Enforcement 413 that, there is significant scope to work on the amendments of the 1995 OECD recommendation, and there is an issue related to the definition of confidential information which could be improved. Another issue is the exchange of confidential agency information. Typically, confidential information of that nature is not shared because agencies feel it is strategic information, and sharing it might endanger their enforcement activities. But perhaps there is room to work on a better definition that would allow more meaningful cooperation based on confidential agency information. That would be distinct from business information because my sense is that it would be very difficult to make meaningful progress in that area. It is very difficult to create the right incentives for companies to come forward and agree to a waiver that allows agencies to exchange that information. It’s worth considering in this context how we can do something to improve the effective enforcement in cartels, in particular with regard to leniency applications. There are currently 50+ jurisdictions where leniency programmes are important, but the conditions to apply for leniency, and to apply for a marker, differ significantly. More uniformity might encourage companies to use leniency policies more effectively, and this would have a number of advantages: it would make life easier for companies because it would be easier for them to apply for leniency, and that is clearly desirable; and it would also allow lesser-developed agencies to develop their enforcement practice. The mechanism I have in mind is a one-stop shop for markers. Perhaps establishing a uniform and global system of leniency as such would be a step too far, particularly outside of Europe. Nevertheless, meaningful work could be done in the area of a uniform one-stop shop for markers. This would at least reserve a company’s place in line, and that might compensate for the lack of interagency exchange of confidential information. A second suggestion I would put forward would be built upon a system of allocation of cases internationally. Obviously we have in Europe a system in the ECN that is built around the concept of the ‘best-placed’ agency, or a ‘well-placed’ agency, and it would be interesting to see if that basic concept could be transposed to the international community. This would not be an initiative free of problems, since the reason why the ECN regime exists is because there is a system of parallel competences to apply EU law, whereas that legal foundation is simply not there internationally. But it might be an interesting experiment to see whether we could come up with a workable international system, perhaps a type of opt-in system that would allow a better allocation of cases internationally. That’s where I’ll stop. Julie Soloway: I’ll be making my remarks on effectiveness and legitimacy in international enforcement cooperation through the lens of Canada and the practice of the Canadian Competition Bureau. My overall point today is that, for the Canadian Bureau, international cooperation has been vital to its enforcement success. It is only through international cooperation DD

414 Part III: Effectiveness and Legitimacy in International Enforcement e nforcement that it has been able to satisfy the objective criteria for success along the lines of what Bill Kovacic described in his talk today. I’ll begin with the general remark that, historically and still today, the Bureau is highly motivated to enhance its cooperation efforts with its counterparts in other countries. It has much to gain from its formal relationships, on the basis of treaties, and from informal relationships, where they just pick up the telephone. In its own self-interest, the Bureau has relied on cooperation to benefit from the work of other, larger agencies around the world who are doing the lion’s share of investigation and prosecution. In some cases, they actually free ride of the efforts of others. This is a bit of a provocative statement, meant to generate some discussion among the agencies here—but you may see a bit of a free rider problem, since the US and the EU are doing all the hard work and the Bureau comes in at the last minute and is able to claim victory along with its US and EU counterparts. Is the overall principle of effectiveness served by international cooperation? Absolutely. In the merger field, you get a more streamlined and more efficient review; it reduces the costs of compliance in an era of reduced budgets. In the cartel field, cooperation has been crucial for the Bureau in the identification and prosecution of criminal conduct, and I would even go so far as to say their big wins and big fines have been predominantly dependent on the large international cases, which have again been led by the US and the EU. One can query what implications there may be for legitimacy, and whether it raises questions regarding the Bureau’s approach. Another issue I want to talk about today has to do with the Bureau’s method of information sharing, which again poses issues relating to the legitimacy of the enforcement process. Stepping back a minute, if you look at Canada, it has, in all fairness, a unique enforcement challenge. It is a relatively small country in terms of population, and it has a relatively small economy. It is highly integrated in the global economy, but Canadian competition concerns generally do not drive international transactions. When you speak to business people who are contemplating large global mergers, Canada is usually an afterthought. It is usually a small part of the transaction, and thoughts of Canada arise only after the parties think hard about how their transaction might be assessed in the US, the EU and, increasingly, China. But since Canada is so dependent on external trade, international issues of competition come up frequently, and international agency cooperation is really critical to the effectiveness of Canada’s competition regime. Our Commissioner in Canada, John Pecman, recognises this, and he has stated that his goal is to create a Bureau without borders, where its enforcement reach is not strictly limited by resources or constrained by jurisdictional boundaries. So you see these types of statements by our Commissioner, and it seems that the Bureau is even more focused on international cooperation than a lot of other agencies. With regard to legitimacy, I was very interested in Christine Parker’s comments earlier this morning when she described the compliance trap. In a situation where there is no widespread

Part III: Effectiveness and Legitimacy in International Enforcement 415 support for enforcement, you see a lack of legitimacy, a lack of compliance, and you get a see-sawing between a vigorous, formal, punitive approach and a softer, facilitative approach. This is exactly what is happening in Canada. In the past few years, under the previous Commissioner, you saw a really hard line approach. You even saw what some described as a capricious or arbitrary approach to competition enforcement. But the new Commissioner and the new administration are trying to build bridges with Canadian business. They are reaching out to foreign counterparts, and they are working hard to rebuild transparency and to reclaim a sense of trust in the regime. So how is the Bureau enhancing its efforts to collaborate internationally? What are the means of cooperation? First, the Bureau acts in accordance with formal cooperation agreements. It has formal cooperation agreements with 11 other competition agencies, and on the basis of those agreements they coordinate enforcement activities, exchange information, including sometimes confidential information, and they meet informally too. At the formal level, there are mutual legal assistant treaties (MLATs), which create a formal process to obtain evidence for criminal investigations, and of course the Bureau also has informal relationships with other agencies around the world. In many of the speeches of the previous and now the current Commissioner, they really brag about their pick-up-the-phone relationship with the other agencies. I haven’t heard the other agencies brag about their pick-up-the-phone relationship with Canada, so it seems the telephone calls are all outgoing calls by the Bureau. But one other area where I think legitimacy is compromised by the Bureau is in its treatment of confidential information, especially with respect to information obtained from parties in the context of mergers. There is a general confidentiality protection provision in the Competition Act, but the Bureau is allowed to share information with other enforcement agencies if it is for the purpose of administration and enforcement of the Act. So what this in effect means is that the Bureau can share all the confidential information that it receives from a party with any agency around the world if it feels that it is necessary for the administration and enforcement of the Act. They can do so without notice, without consent and without a waiver. As we know, some countries may treat confidential information with higher standards than other countries, so this has been a very serious concern for the competition bar in Canada and for many businesses. This unsatisfactory state of affairs is a major area that undermines the legitimacy of international cooperation and enforcement. Staying with the context of merger control, how does the Bureau coordinate its efforts with those of other agencies? Well, it has a principled approach to international coordination. What it says in its Bulletin is that global remedies are given due consideration, but it’ll seek to build a ‘Made in Canada’ tailored remedy where there are Canada-specific issues. In practice, it really takes three different approaches. First, if the remedy is satisfactorily addressed by a foreign authority, it’ll simply adopt it in Canada, and it can

416 Part III: Effectiveness and Legitimacy in International Enforcement do this by way of a press release; or, if it can do it by way of an undertaking from the parties saying they’ll apply in Canada the same remedies obtained in the US or the EU or wherever, that may suffice. Or the Bureau will do a very brief consent agreement that has to be registered with the tribunal, and this touches on what I said earlier: this is sort of a free rider situation, and I was on a case once where a DOJ official actually complained when speaking to our US counsel about free riding by the Canadian Bureau. The official said, We’ve done all the work, we’ve been on this for months and months, and they come in later to get a big win and a big press release. They’ve done virtually no work and they have virtually no assets in Canada that are related to this transaction.

So that’s one approach. The other approach is the Bureau may mirror the remedy obtained elsewhere. This may occur where there are more serious Canadian issues, and the Bureau will implement a reciprocal remedy that may have some unique Canadian features but in general it mirrors what’s done in other jurisdictions. A third approach is where there are really serious substantive competition issues. In this situation, the Bureau will develop a completely unique remedy tailored to Canada. That remedy will not necessarily take into account whatever the global remedy is. In the area of criminal matters, and specifically cartels, as I said before, many of the Bureau’s successes have come about through international cooperation. Although the fines are much smaller than you see in the US or the EU, which is reflective of the lower volume of commerce in Canada, the fines were in fact quite high by Canadian standards in the Japanese auto parts conspiracy. The fines in that case are now up to 35 million Canadian dollars, and in the air cargo carrier case, 22 million Canadian dollars in fines have been imposed. So it is overall through international cooperation that the Bureau has been able to enforce the Competition Act. And that cooperation arises largely because of the unique position of Canada and how interdependent it is with other geographical parts of the global economy. Jochen Burrichter: I would like to concentrate on the possibilities which already exist for cooperation and coordination between competition agencies, and I’ll cover some practical problems. Let me start with the possibilities of cooperation for cartel cases and their exchange of information. Article 12 of Regulation 1/2003 provides for a broad scope of cooperation between the Commission and the competition authorities of the Member States, for the purposes of applying Articles 101 and 102, and the Commission and the competition authorities of the Member States have the power to provide one another with and use as evidence any matter of fact or law, including confidential information. This includes information not only between the NCAs and the Commission, but also between the NCAs themselves. The information must have been gathered, of course, in a legal manner by the transmitting national authority in accordance with the law applicable in that Member State. DD

Part III: Effectiveness and Legitimacy in International Enforcement 417 What is the situation with third countries? The agreements between the EU or between some of the Member States and third countries, such as the US, provide in general for a broad scope for case notifications, and also for information exchange. In principle, information which is relevant according to the competent competition authority is to be exchanged. However, there are some important exceptions to this principle concerning the protection of business secrets and important national interests. This means that in many cases the exchange of information requires a waiver from the undertakings concerned. Furthermore, the exchange of information in most cases is restricted to existing information, so other competition authorities are not forced to investigate on their own account for purposes of the other procedure. Some problems arise with the exchange of information in leniency cases. In the absence of a Europe-wide system of fully harmonised leniency programmes, the exchange of information concerning applications for leniency between the Commission and the NCAs creates special problems which are partly dealt with in the Commission’s ‘Network Notice’.4 Under Article 11 of Regulation 1/2003, the Commission and the NCAs have to inform each other where a case has been initiated as a result of a leniency application. Such information will not be used by the other members of the ECN as the basis for starting an investigation on their own behalf, whether under the competition rules of the EU or, in the case of NCAs, under their national competition law. H owever, this does not preclude the relevant authority from opening an investigation on the basis of information received from other sources, or provided pursuant to A rticle 12, which I mentioned already. Information which has been voluntarily submitted by a leniency applicant will only be transmitted to another member of the ECN with the consent of the applicant. This is also true for other information gathered in inspections which could not have been carried out except as a result of a leniency application. However, the C ommission and the NCAs normally will encourage leniency applicants to provide consent. The details of the procedure are given in the Network Notice. Beyond the exchange of information between the national authorities and the Commission, there seems to be no exchange of information about leniency applications filed with other competition authorities. It remains a matter of discretion of the undertakings concerned to decide to apply for leniency in several Member States simultaneously. The next point concerns the access of third parties to leniency documents, and this is a crucial issue. In order to safeguard the effectiveness of leniency programmes, cooperation and coordination between the competent competition authorities with regard to the extent of granting third parties access to leniency documents is indispensable in cases of multiple leniency applications;

4 Commission Notice on cooperation within the Network of Competition Authorities, [2004] OJ C101/43.

418 Part III: Effectiveness and Legitimacy in International Enforcement otherwise, the disclosure of identical leniency documents in one jurisdiction would make the more restrictive practice in another jurisdiction obsolete. I come now to cooperation and coordination in the investigation process. In the EU, Article 22 of Regulation 1/2003 provides that one NCA may ask another for assistance in order to collect information on its behalf. This includes all kinds of fact-finding measures, including dawn raids. The information collected by the assisting NCA can be exchanged in accordance with the rules contained in Article 12 of the Regulation. The NCAs have made use of this possibility in various cases already. In third countries, under the MLATs between a number of international competition authorities, in particular where the US is concerned, there is intensive coordination of activities against international cartels, including the organisation of international dawn raids at the same point of time. This approach was used, for example, by the US, European, Japanese and Canadian competition authorities in the auto parts, marine hose and air cargo cases. This coordination of the precise time action is taken constitutes an important means to maintain the element of surprise. The consequence is that the cartel members are unable, despite the time differences between the various jurisdictions, to conceal pieces of evidence or to immediately start leniency applications in the other countries. Such internationally coordinated dawn raids can be facilitated if a member of an international cartel applies for leniency in various countries at the same time. With regard to the coordination of sanctions, and this is an underdevelopedarea, conflicts may arise in parallel proceedings by the NCAs at the national level, either within the framework of EU competition law or under national laws. A recent example of the need to coordinate sanctions imposed by different competition authorities is the flour case. In this case, the NCAs in Belgium, France, Germany and the Netherlands initiated proceedings against flour mills and imposed fines partly on the same undertakings. There is a special need for coordination in these cases where the undertakings concerned are not able to pay all the fines imposed due to an inability to pay. In these cases, situations may have to be avoided where there is a race between the NCAs to adopt the first enforceable decision. This apparently happened in the case that I just described, in particular as between the Dutch and the German authorities. Regarding the situation in third countries, there is practically no coordination between competition authorities of third countries concerning the fines to be imposed in cases of international cartels. This is critical, as it may be argued that parallel sanctions within the EU, but also outside it, and fines by various international competition authorities for the same international cartel, violate the principle of ne bis in idem.5

5 With regard to third countries, the position taken thus far by the Court of Justice is that the European Commission is not obliged to take account of proceedings or penalties to

Part III: Effectiveness and Legitimacy in International Enforcement 419 Now let me briefly cover cooperation in merger cases. There is a need for a coordination of proceedings in merger cases when a merger triggers multiple filings; a coordination of the proceedings in that situation is essential. The undertakings concerned have an interest in avoiding contradicting decisions. In that regard, it is crucial to avoid situations where one national a uthority is forced to decide earlier than another authority without being able to exchange relevant information and arguments. The parties in these cases may choose to notify at different points of time where this is possible under the national merger regimes, and the authorities may use their procedural possibilities to delay national proceedings in order to gain time until the authorities in jurisdictions which are crucial for the merger are ready. In addition to coordinated timing, the exchange of information between competent authorities is essential in order to avoid decisions based on insufficient factual findings. International agreements allowing for the exchange of such information should therefore be used to their fullest extent in order to avoid problems with professional secrets. The parties to the merger may also carefully consider agreeing to waivers for the exchange of information which normally could not be exchanged. A further point concerns the coordination of remedies. In merger cases especially, the fine-tuning of remedies and the compatibility of remedies across jurisdictions are crucial. Competition authorities should be allowed to exchange the conclusions they reach on the basis of market testing, for example in the case of commitments. Let me finish with two remarks concerning legal protection in coordination and cooperation between competition authorities. In international enforcement, cooperation between agencies must not lead to any restriction of the rights of the undertakings concerned with regard to access to the file and the right to be heard with regard to all arguments and aspects gathered and exchanged between the authorities. The other remark is that international enforcement cooperation must not risk or restrict legal protection under the relevant national competition regimes. This also includes, if granted under the relevant regime, the possibility to appeal against measures taken in the pursuit of that cooperation. Thank you. Alberto Heimler: Thank you very much. My topic today is effectiveness of enforcement cooperation in developing countries, and I would like to start by saying there are two objectives in international antitrust. The first is convergence of the substantive aspects of the law—of enforcement, which of course enhances trade and allows firms to be subject to the same rules. DD

which an undertaking has been subject in non-Member States, whether on ne bis in idem or any other legal grounds. See Case C-231/14 P Innolux Corp v Commission, EU:C:2015:451, para 75, with case references.

420 Part III: Effectiveness and Legitimacy in International Enforcement The second is cooperation that makes enforcement more effective and reduces transaction costs for firms. These two objectives, convergence and cooperation, are interlinked. The European Union is a good example for this. The EU, of course, developed an intense system of cooperation, and at the same time it made enforcement more effective and led to a large degree of substantive convergence among the Member States, which generally have similar rules and a similar interpretation of those rules. Let me start with cooperation. Cooperation has been promoted quite widely in developing countries. This is where I will concentrate my remarks today. There have been quite a number of regional agreements signed over the years, and it may be important to stress that the objectives of these agreements were much less ambitious than those of the European Union. Some of them created a free trade area, and sometimes a competition rule or authority or provisions were attached to it. Sometimes the aim was to create an internal market among these countries, but the degree of ambition and the effect of the legal provisions have never reached a level comparable to that of the EU. There are many such agreements, and all have performed quite badly. C ertainly, in terms of cooperation, they were unable to make enforcement more effective, and they have been unable to reduce transaction costs for firms. Indeed, sometimes they increased transaction costs. A number of such agreements exist in the developing world, such as the Mercado Commun do Sul (Mercosur), the Andean Community, the Common Market for Eastern and Southern Africa (COMESA), the Caribbean Community (CARICOM), the Southern African Custom Union (SACU), the West African Economic and Monetary Union (WAEMU) and the South African Development Community (SADC), to name a few. The Andean Community and CARICOM are both agreements among small economies, and neither of them has any jurisdiction over mergers. Transaction costs could have been reduced if mergers had been made subject to a centralised review process. However, these agreements do not concern mergers at all. But another telling example concerns COMESA. The law leading to the creation of the COMESA competition authority was in 2005, the authority was created in 2006 and the procedure for the functioning of these authorities was just issued in 2013. So it took almost more than eight years to get the authority established, and the regime faces a lot of problems— problems of judicial review and problems of jurisdiction, especially with regard to mergers. Indeed, COMESA now has a competition authority at the COMESA level, where notification is mandatory only if a merger produces its effects in two countries of the organisation—and there is no indication of a minimum turnover. Therefore, irrelevant mergers will have to be notified. Originally, the merger notification fee was quite high, but now, after a lot of criticism, it has been reduced. This episode also illustrates the need of international organisations to have some funding available. One means to secure funding is to have high notification fees. And, of course, the effectiveness of

Part III: Effectiveness and Legitimacy in International Enforcement 421 such an organisation is minimal. There are other problems as well. It is not clear, for example, whether, if a merger is being considered at the COMESA level, it also needs to be considered at the domestic level. There are issues of jurisdiction that still need to be settled. We can therefore see how this attempt to achieve cooperation through regional agreements has failed in the developing world. There is only one organisation that has had tangible success, and that is the West African Economic and Monetary Union. The WAEMU system mimics the European Union, but the reason why WAEMU showed some signs of effectiveness was that it eliminated the jurisdictional problem and concentrated all the power to enforce the law at the level of the Commission. This eliminated all possibility on the part of domestic authorities to enforce the law, which is probably quite effective in terms of transaction costs, but it is certainly not effective in terms of the use of public resources because these authorities continue to exist and it’s not clear what they’re here for. So cooperation initiatives in the developing world have mostly failed in terms of their objectives, and in terms of effectiveness. I would also like to mention that there are ways to address the issues of cooperation, which are much less institutionally oriented. A very simple and very important experience has been developed in the past decade between Australia and New Zealand. Under their close economic relations and a trade agreement between the two countries, each country decided to have a Commissioner nominated and appointed to the authority of the other country. This ensures that if a merger takes place and the effects would be cross-border, then there would be full information on the part of each authority with regard to what the other authority would do, and remedies could be decided in a common, cooperative way. So it seems that effective cooperation can be achieved without any need to create a separate institution, separate bureaucracy and separate system—which is really quite expensive to do. In developing countries, since there are no resources even for domestic authorities, creating supranational authorities can lead to worse results because scarce funds are then divided between authorities at different levels. Instead, those resources could be concentrated on the domestic authority, and cooperation can be promoted in a much less costly way. Let us now turn to the issue of convergence. Convergence is very important. In international antitrust, in the mid-1990s, in connection with the WTO Agreement, there was a very ambitious effort, especially on the part of the European Commission, to create a system of law where authorities would be subject to international control. The Commission never said this officially, but there was a report in 1995 on competition policy in the New Trade Order which made clear that the idea of this international order was for the international community to step in and eventually challenge decisions by domestic antitrust authorities and domestic courts. Sometimes Commission officials made reference to that report, and this was one of the reasons why the WTO exercise was regarded with great suspicion—it called into question the independence of domestic court systems. Eventually, in 2001, you will recall that

422 Part III: Effectiveness and Legitimacy in International Enforcement at the Doha meeting, trade ministers provided a clarification to the objectives of the WTO trade and competition agreement, and this was followed by a series of discussions. However, eventually, in 2004, the agreement was consigned to limbo, where it still is. I suppose that we cannot exclude that it may one day come back. In September 2000, at the 10-year anniversary of the Merger Regulation, there was an agreement to create the ICN, and in October 2001, at the Fordham Conference on International Antitrust in New York City, a group of top antitrust law officials from 14 jurisdictions6 created the International Competition Network (ICN). Its objective was to become ‘an informal network of antitrust agencies from developed and developing countries that will address antitrust enforcement and policy issues of common interest and formulate proposals for procedural and substantive convergence through a results-oriented agenda and structure’.7 At the meeting at Fordham, it was said that the ICN would favour procedural and substantive convergence while following a result-oriented agenda. Now the ICN is indeed the organisation that could promote convergence, and it has all the potential for promoting convergence. It is a virtual organisation without a secretariat, and in principle it is a very democratic organisation. There is no secretariat, there is a consensus on proposals made by the ICN, and every agency has the possibility to object. However, without a secretariat, there is no doubt that the larger agencies take the lead in the ICN—the US Department of Justice, the Federal Trade Commission and the European Commission. They have more resources than other authorities, and therefore they can put more people to work on ICN issues. So even though the ICN is in principle a democratic system, in the end there is a tendency to go along with the larger agencies. At least at the beginning, however, since the ICN was created to ensure convergence in antitrust enforcement, it was thought that there would be sufficient incentives for the more established agencies to engage with the smaller ones, and to make sure that they actively participate. There was an effort, at least in the beginning, to make sure that the smaller, developing countries would participate in the steering committee of the ICN and there was really an effort to promote their participation. With time, however, this effort in my view has diminished, even though some very positive results have been achieved. The ICN has issued around 100 recommendations; some of them have been used, but sometimes the ICN has had no influence on enforcement. But in more recent years, the

6 These were the authorities of Australia, Canada, the European Union, France, ermany, Israel, Italy, Japan, Korea, Mexico, South Africa, the United Kingdom, the G United States and Zambia. 7 See the Memorandum on the Establishment and Operation of the International Competition Network at http://www.internationalcompetitionnetwork.org/index.php/en/ about-icn/operational-framework.

Part III: Effectiveness and Legitimacy in International Enforcement 423 ICN has slowly moved away from the recommendation approach, especially since it has addressed more difficult and more controversial issues. So it has shifted from what I call the recommended practices approach to the ‘encyclopaedic’ approach. And when you read the report of the ICN on unilateral conduct, all you see is: ‘You do this; I do that’. Developing countries are left with no indication of what is best, or what might be better for them. Furthermore, the participation of developing countries is minimal—not the participation at ICN meetings, this has been increasing. In light of that participation, there is some optimism that the ICN may indeed secure convergence in a ntitrust enforcement—on procedural and substantive matters. But if you look at the papers of the ICN, the percentage of countries—developing countries— contributing directly to the ICN’s work is modest. I therefore think that extra effort should be made, if we want to achieve convergence as an instrument of antitrust, to engage developing countries more in the ICN’s substantive, hands-on work. Steps should be taken to enable them to contribute, and that means, among other things, finding ways to reduce their costs. Because what you see, especially when reading the procedural guidelines of COMESA, is a complete misunderstanding of the recommendations of the ICN on merger control. I believe that the ICN can provide guidelines and become, as we had originally thought early in the year 2000, a standard-setting organisation. Hwang Lee: I’d like to thank the organisers, Mel and Philip and G iorgio, for inviting me to participate in these very interesting discussions. We are talking about international convergence today, but I would like to raise the possibility of a degree of fragmentation and deviation on the part of Asian countries. As you know, in Asia, the three ‘north-east’ countries—China, Korea and Japan—play a significant leading role, and their influence stems from their unique assets, which differentiate them from many other Asian countries. I would like to concentrate on the fairness doctrine. Many, if not all, of you have heard about Confucius, and many Asian countries have been under the heavy philosophical and political influence of Confucianism for more than two thousand years. Korea is a typical example. Confucianism traditionally emphasises the importance of respecting fairness in politics, as well as daily life. Fairness is best achieved when it is carried out in both the result and process, and that respect it is not so remote from the Kantian tradition. The notion of fairness is a concept barely defined in modern economics, especially compared with the concept of efficiency, which is one of our central themes here at the Workshop. A number of Asian jurisdictions, like Korea, Japan and Taiwan, have seen the adaption of Kantian laws as a combined set of competition laws and unfair trade practice laws that regulate unfair business behaviour. This means that these countries are interested in making use of competition laws as a tool to achieve fairness and justice in addition to economic efficiency. And that combines with the traditional philosophy DD

424 Part III: Effectiveness and Legitimacy in International Enforcement of Confucianism. Indeed, in some cases, the fairness doctrine has dominated efficiency concerns. Today, most experts in the US and Europe agree that the top priority of competition laws is economic efficiency. Furthermore, many experts seem to agree that international cooperation is crucial to achieving this goal. Yet, for the reasons noted previously, the fairness doctrine in Korea–China–Japan may sometimes subordinate the issue of whether a given practice is efficient. So the notion of fairness in Confucianism may require a sacrifice of efficiency or monetary gain. This spirit can be best found in the Confucian doctrine known as ‘Rectification of Names’. Confucius held that, in order to have a well-ordered society, the most important thing is to carry out what he called the rectification of names. That is, things in actual fact should be made to accord with the implication attached to them by their names. In other words, every name contains certain implications which constitute the essence of that class of things to which this name applies. Such things, therefore, should agree with this ideal essence.8 One must not disregard the practical and/or policy reasons that contribute to this difference between East and West. Developing countries in Asia live a different economic reality compared to countries of the Western world. From the viewpoint of a developing country in Asia, strong industrial and trade policies that aim at spurring economic growth in the long term may seem to conflict with competition laws that are solely based on concepts of economic efficiency in the short term. One must also consider the fact that developing countries typically do not have the human or material resources necessary for a rigorous analysis efficiency, and they may consider that detailed analysis with such scarce resources is virtually impractical. As I said, the notion of fairness is a concept that is barely defined in modern economics. But fairness has its virtues. It can be used as a master key to solve problems that cannot be addressed solely by applying competition policies. In many Asian countries, similar to other developing countries in the world, competition policy tends to be considered as parallel to industrial policy, and it is regarded as a branch of general economic policy. Competition policy is subject to limits just like industrial policy, and the adoption of a complementary principle of fairness may aid in closing certain loopholes in

8 The above description follows the typical account given at http://history.cultural-china. com/en/165History878.html. The following quote from Lunyu, 13:3, is telling: ‘If names be not correct (zheng), language could not be fluently used. If language be not fluently used, affairs could not be carried [out successfully …] ritual/propriety (li) and music could not [flourish …] punishments could not be properly made [… and] the people would not know how to behave’. It is therefore imperative that language and action as well as titles and duties (eg the duties of a father, or of a ruler) should be harmonised. If discrepancies persist—if, for example, a ruler fails to conform to his predetermined role—there is a risk that the social structure will unravel and the civilisation in question will lose its way.

Part III: Effectiveness and Legitimacy in International Enforcement 425 c ompetition policy. It is therefore popular with policy makers. As long as the fairness doctrine can contribute to solve practical problems, it will be considered a useful policy tool today in Korea. As you can appreciate, this approach is quite different from the one prevailing in the US, where industrial policy is not widely accepted as a valid measure to address problems of general concentration. And this difference of approach tends to generate misunderstanding and hostility between the two continents. To reach some kind of solution for international harmonisation/ standardisation, Asian competition enforcement should make better efforts to rely on economic efficiency and market competition. But in reality, the Asian region cannot and will not abandon the standard of unfairness in competition enforcement. Hence, the issue really boils down to how Asian jurisdictions will define and apply such standards of unfairness. It would therefore be useful to explore how the concepts of fairness and unfairness should be defined and applied on the basis of a sound economic understanding. John Davies: My presentation is about inconsistent decisions. What happens when international coordination doesn’t work, and where agencies reach different decisions on essentially the same matter? I am going to talk about why that might happen, and about what the consequences might be when it does happen. I wish I could answer the excellent question Paul Lugard raised when he started the session: what is the cost of increased complexity and inconsistency among such a large number of different jurisdictions likely to be? At the OECD, we are trying to look at the drivers of the international competition law world.9 We’re looking at what we know about the number of jurisdictions that have competition laws and also at what we know about international businesses, and we’ll try to get some sense of where it might be going. That might partially provide the answer, or at least put some numbers on how much this all matters. I can’t do that here, so the presentation is just a short note to set out some of the questions we are trying to answer. Consider a global merger, or at least an international merger in which more than one jurisdiction is concerned. We sort of wave our hands around and say another case might be quite similar. I don’t think I can really wave my hands that much, because if we think about it, we think of other cases—both cartels and especially abuse of dominance, they are very different indeed. In our OECD work, we’ll go into that—but I’ll confine these remarks to mergers. So what happens when agencies reach a different decision on the same international merger? We think that might happen for three reasons. First of all, the agencies might be applying a different substantive test, and that could result DD

9 See OECD, ‘Challenges of International Co-operation in Competition Law E nforcement’ (2014), http://www.oecd.org/daf/competition/Challenges-CompetitionInternat-Coop-2014.pdf.

426 Part III: Effectiveness and Legitimacy in International Enforcement from a formula literally in the law or it could be in their guidelines, or it could be, if you like, a difference in the attitudes of different competition authorities. For example, one authority might pay more attention to market shares than another, and of course we know about differences in the way vertical mergers are treated, and so on. So substantive differences would be one, and obviously convergence would help with that. The second reason for differing outcomes is that the conditions of competition might vary: the merger might have different effects in different jurisdictions. That is a perfectly legitimate reason for agencies to come to different answers in what might appear to be the same case. But that justifiable result can still have adverse consequences, as I’ll talk about in a moment. The third reason why agencies might come to different opinions on mergers is essentially because competition law is hard. Reasonable people faced with the same evidence, and trying to apply the same legal test, will often come to different conclusions about a merger. What is interesting here is that, if you compare the frequency of disagreements between agencies with the frequency of, so to speak, dissenting agreements within agencies, it is quite surprising how few diverging, GE/Honeywell-type decisions there actually are.10 In a sense, competition agencies seem to agree with one another slightly more than they agree with themselves. This suggests that—slightly under cover, if you like—a certain amount of taking account of one another’s position is indeed going on, although it’s not clear to me quite what the legal basis of that might be. So we see three possible reasons why competition authorities might disagree on a single merger. When does it matter? Well, what’s really important for whether it matters in the framework we’re adopting has nothing to do with whether the markets are international or not; it is rather to do with remedies. It’s about whether you create a national, effective remedy, whether that’s a structural or behavioural remedy—because if you can do that, it’s fine. If you can do that, it doesn’t matter that I’ve reached a different decision from Dan, because I can implement the remedy in my jurisdiction and it doesn’t affect his jurisdiction, and that’s fine. But we know that a lot of remedies will actually have international or even global effects. In particular, although not exclusively, structural remedies will often have effects on the entire business of the merging parties. And, of course, where a merger is prohibited, this essentially implies that the merger is prohibited everywhere. That creates an externality, to use the economic jargon. As you all know, an externality is what happens when I take a decision that affects both Dan and me but I don’t take his

10 This finding is consistent with the observations discussed in Philip Lowe and Mel Marquis, eds, European Competition Law Annual 2010: Merger Control in European and International Perspective, Hart Publishing, 2013. See in particular Thomas Deisenhofer, ‘International Cooperation in Merger Cases—An EU Practitioner’s Perspective’, in ibid, 227 et seq.

Part III: Effectiveness and Legitimacy in International Enforcement 427 interests into account when I take that decision, and that leads to all sorts of inefficiencies. The thing that’s worrying about an externality is that an externality, in general, cannot be solved simply by communicating. Information exchange between the different agencies does not solve the externality because it is a fundamental problem of different incentives—I care about my consumers, you care about your consumers, and if my decision affects your consumers and affects your economy, then we have inefficiency between us. There is a big difference between small and big countries here; the definition will depend very much on the case. As many of you know, I used to run a competition authority in a very small country, Mauritius, and when we were faced with global mergers that we could do absolutely nothing about, we did absolutely nothing about them. Generally, we just sort of pretended not to notice. We had a discussion about this at the OECD about a year ago, and one fairly small jurisdiction, Austria, said that they frequently adopted behavioural remedies not because they thought that the behavioural remedy was necessarily the best but it was the only one which they could do. Since they were dealing with international mergers, they had no real ability to put into effect a significant structural remedy. In fact, you don’t have to be very small: there is a case in the UK, Dräger Medical,11 where medical equipment was manufactured in two countries, Germany and the US. The merger had an significant effect on competition in the UK market, and the UK authority imposed a behavioural remedy, simply stating that it was doing so because it was not in a position to impose a structural remedy. So there are circumstances where authorities in small countries really can’t do anything. At the same time, authorities in large countries can arguably do too much, as they can block global mergers. How large do you have to be to block a global or international merger? It kind of depends on the size of the market and it kind of depends on the size of the companies. In 2013, the UK Competition C ommission required a divestiture remedy, effectively blocking a merger between Eurotunnel and SeaFrance,12 which the Autorité de la Concurrence had previously approved with some behavioural commitments.13 Since Eurotunnel by its physical nature must sell into the UK market, the UK Commission’s decision prevailed. So in that case you had a deal that affected two countries and it was blocked, in both countries, by just one authority. As we heard a while ago, the GE/Honeywell case was a prime example, indeed the most famous example, of an authority from

11 Competition Commission (UK), Dräger Medical AG/Air-Shields Hillenbrand Industries Inc, 2004, https://www.gov.uk/cma-cases/dr-ger-medical-ag-air-shieldsfrom-hillenbrandindustries-inc-merger-inquiry-cc. 12 Competition Commission (UK), Eurotunnel/SeaFrance, 2013, https://www.gov.uk/ cma-cases/eurotunnel-seafrance-merger-inquiry. 13 Autorité de la Concurrence, Decision 12-DCC-154 of 7 November 2012, http:// www.autoritedelaconcurrence.fr/pdf/avis/12DCC154decision_version_publication.pdf (in French).

428 Part III: Effectiveness and Legitimacy in International Enforcement a large jurisdiction, the EU, effectively blocking, at a global level, a merger that affected the whole world. Our concern is that there are more and more countries that are in the position to act like that. Back in the early 1990s, if you just look at the European Union, the United States and Japan, you’ve covered 75 per cent of world GDP. So as long as those three competition authorities had some reasonable cooperation going, you pretty much had the problem of international coordination solved. That is obviously no longer the case. Those same three jurisdictions together are down to less than 50 per cent of world GDP today, and dropping. So what are the implications? In a world of several very large jurisdictions, each of which is large enough to block a global merger, you actually have a significantly tougher regime against global mergers than perhaps was planned—tougher than the ones that apply to domestic mergers. First of all, the toughest regime decides. I can decide to block a global merger if I am a big country, but there can also be global effects when I approve a merger s ubject to a significant structural remedy. For example, Glencore and Xstrata are Switzerland-based mining companies with production sites in various countries and global sales. The US Department of Justice cleared their merger unconditionally in 2012. China’s MOFCOM in 2013 required (in addition to some behavioural remedies) divestment of the Las Bambas copper mine, now under development in Peru.14 It is not clear whether the conditions of competition differ materially between these jurisdictions. However, it appears that the merged entity’s market share of copper sales in China would be less than 18 per cent, ie below the level at which most competition authorities would consider even intervening. Copper is a global market, and it is hard to escape the conclusion that the DOJ and China’s MOFCOM considered essentially the same facts. But since they reached different conclusions, the stricter standard—MOFCOM’s—has prevailed. Now, the structural r emedy imposed by MOFCOM is not necessarily a big problem in the sense that it does not stop the merger from going ahead. But presumably, when the DOJ investigated the transaction and decided to clear it, the message was that it was a desirable combination of assets. We presume mergers have efficiencies; we only block them when we have good reason to do so. But the MOFCOM decision in a sense scuttles that presumption in this case. I am struck by the thought of people sitting around and discussing, for example, precisely how tough the US merger regime should be and whether it should be a little tougher, or whether it should become a little softer or whatever. Because in the biggest mergers of all, none of that particularly matters if it has to be cleared by other authorities that are applying much tougher standards.

14 MOFCOM, Decision of 16 April 2013, Public Announcement No 20 [2013], http:// fldj.mofcom.gov.cn/article/ztxx/201304/20130400091222.shtml (in Chinese).

Part III: Effectiveness and Legitimacy in International Enforcement 429 So, in a world with numerous competition authorities, we may see that it is the toughest regime on the planet deciding global cases. Secondly, even if everybody is applying exactly the same substantive test—even if nobody is any tougher than anybody else—it is inherently more difficult to secure, say, five clearances than one. If a merger is a real borderline case, and you’re advising your clients they have a 50–50 chance of success—well, if it has to go through five big jurisdictions, any one of them can block it. So it’s not 50–50 after all—there is actually a high likelihood that the deal will be either prohibited or at least reshaped by structural remedies. So the global merger system is becoming much tougher than perhaps it needs to be—potentially with a chilling effect on the number of global mergers that will be pursued. Does any of this matter in practice? There are fewer inconsistent cases than perhaps one might expect. What we want to look at is what’s happening to the internationalisation of the world economy. One point that is noteworthy to me, at least, is that globalisation has got a very long way to go. Despite what you read if you pick up one of those fat business books sold at a irports, the world is not flat. The world is distinctly unflat—it’s local. Business remains local. There are very few genuinely international businesses, and the world has a long way to go even if it is true that the world is integrating very fast. Can the current system of ‘muddling through’ with a bit of under-the-counter exchanges of information about cases and some nods and winks to avoid inconsistent decisions really suffice? Thank you. Lowe: Okay, thank you very much, John. I’m not sure that the toughest regimes are always the same ones. Thanks to our other speakers as well. We’ll open the floor now to general discussion. It looks like a couple of people would like to comment, including myself. All right, let’s begin with Dan. DD

Dan Crane: Thank you. One of my concerns has long been that, as you get more major regimes in a position to block a merger, it gives authorities a ‘hold-out’ position. An authority could theoretically extract local benefits just because it has a hold-out position. That causes a real collective action problem and so I don’t think we have seen a lot of this yet, as you say, but over time we could see a major jurisdiction saying, ‘I am holding out for concessions that will play well back home as the price of clearing the merger’. So there is a whole sort of political economy problem that could develop over time. I also have a comment on your idea that Confucian values lead to differences between Asian competition and, so to speak, ‘Western’ competition law. One of the things to think about is that some of those same concerns may arise in fields of law other than competition law, and so one question is: why is competition law the residual place for those kinds of values? In the US, for example, we have all kinds of distributive justice and fairness concerns DD

430 Part III: Effectiveness and Legitimacy in International Enforcement around dealer franchise terminations, but we don’t think about that as being part of competition law. In effect, we reduce the core of competition law to a different set of efficiencies. So, even to the extent that you see those kinds of concerns manifested, is it inevitable that they will manifest themselves in competition law as opposed to other parts of law? Lowe: I think John was right to say that we certainly have not reached maturity as far as the many processes of globalisation are concerned. On the other hand, there are several areas where, beyond what was achieved in ICN in the years 2003 to 2007 or 2008, it seems that there is a lack of ambition, or a lack of will to go further than that. We could collectively express concern about how things are progressing, at least with regard to exchange of information—if not on waivers or mergers, exchange of information on leniency, merger waivers, maybe a series of other things. Julie has said that there is a free rider effect, which you would expect to be shared by more countries than just Canada. And, of course, where there are multiple agencies that can exercise jurisdiction, global solutions may be the best way to handle the case. If it is possible for Japan, Korea, the US, the EU or whoever to negotiate a global remedy, or maybe the company proposes a global remedy, this is clearly better than creating a jigsaw puzzle of responses to the competition concerns in different countries. So we should be trying to encourage global solutions where that is possible. I think there is a certain dissatisfaction that we haven’t gone further than what we’ve done up to now. This may also be linked to that fact that we’ve got authorities, perhaps in China and Africa and elsewhere, where there are policy considerations other than pure competition issues that influence decisions. It might be that what is needed is more ambition in the ICN and the OECD to produce a parallel substantive advance, now that so many governments have competition agencies and merger regimes. On a related point, I remember a particular merger we were involved in just before the accession of Slovakia, and the EU approved this merger between two agro-businesses. But in Slovakia there were two sugar plants, one owned by one of the companies and one owned by the other. Now, on the thinking that we’ve just been discussing, you say: ‘Well, obviously this merger is globally desirable; it produces efficiencies, etc’. But there was a problem in Slovakia. As it turned out, we found a solution but because Slovakia belongs to the European Union, and if there isn’t any problem of getting sugar into the Slovakian market—indeed, if there isn’t really a Slovakian market anymore—then there is no problem. But under other circumstances there may be local issues that have to be dealt with locally. I’m not sure that the experience we’ve had with the FTC and the DOJ are entirely under the counter. In fact, in many cases it has been organised by the parties. It’s in their interest to make sure that there is adequate exchange of information between agencies so that you’re doing your best to find a multi-jurisdictional solution. DD

Part III: Effectiveness and Legitimacy in International Enforcement 431 But here, again, John has emphasised how much we still have a deficit here which we ought to try and do something about, whether it’s a type I error, a type II error, or indeed a type III error, which is providing the right answer to the wrong questions. Anyway, I would just like to give the floor to the speakers in reverse order and then we’ll wrap up for this evening. So let me come back to John. Davies: I won’t come back on what I was just talking about, but let me just briefly say that I agree with Dan. Also, with regard to what Paul was saying at the start about the mechanisms of international cooperation, I agree with everything he said, and the OECD has decided to completely rewrite the 1995 recommendation on cooperation.15 That won’t deal with the sort of issues I was talking about, which are a sort of longer-term kind of thing for national laws, because the problems I have outlined concern the incentives of competition authorities. DD

Lee: On Dan’s questions, first of all, I think that the reasons why they are viewed as part of the competition law in Korea and China is that people simply say they are. A second reason is that the Korea Fair Trade Commission is responsible for public enforcement, and other issues like franchising problems are not enforced by public authorities. The KFTC usually tries to inject a ‘competition spirit’ into that regulation. A third reason is that unfair trade practices provisions are incorporated into Korean legislation and, just like Japan has done for many years, the KFTC has placed a traditional regulation over the abuse of market dominance. Finally, there are measures to protect small competitors in the market to develop vigorous market competition. DD

Heimler: My brief remark is that, with 120 competition laws in the world, we need some convergence, some standards to refer to; otherwise it’s going to be a full house—not just for mergers, but in general, for restrictive practices and abuse of dominance. As time goes on, younger authorities will enforce the law more and more actively, so the collective action problem Dan mentioned can be expected to grow more serious. DD

Burrichter: As I tried to explain in what I said, I think even before we have a kind of ICN initiative or international consensus, the competition authorities are able to cooperate, or they are not legally hindered from cooperating, and maybe this is also something legal advisers can help to initiate. Advisors can work to convince the competent authorities to enter into a relevant dialogue with their counterparts. DD

See above n 3.

15

432 Part III: Effectiveness and Legitimacy in International Enforcement Soloway: I have two quick points. First, Dan Crane mentioned this risk of leverage where some authority is the last hold-out. As more agencies review a merger, one agency can hold out for a better deal that better suits their local constituents. Let me make two comments on that. One comment is that I think the ICN provides a good environment where peer pressure can influence agencies. So it may be embarrassing for them to do that. The other comment is that it’s not obvious to me why that doesn’t happen more already right now, if they have that sort of leverage. The second point I wanted to make is that John Davies talked about the costs of divergent or conflicting outcomes. I don’t know if the OECD is planning on this, but it would be interesting to empirically measure that question so that we know more about those costs and how they can be reduced. DD

Lowe: I was just reminded of a chief executive of a very large multinational corporation who late at night said: ‘Philip, why didn’t you prohibit our merger with X? We’ve lost billions on that deal.’ [Laughter] Now, you’ve all been very patient, particularly those who are going to enlighten us tomorrow on the judicial perspectives you have of this vision of a legitimate and effective enforcement of competition law. Someone told me this morning that if you have a vision you should probably go to the doctor. But I think that we’ve gained a lot of insights today about a number of things which can be improved: not just in due process, but also in information gathering, the formulation of remedies and the dimensioning of those remedies in relation to the legal tests which have to be applied. We also just now looked more intensively at what is inevitably a process which is going to be difficult to control unless we start controlling it, which is how you develop a coherent international competition law enforcement regime in a world where most emerging countries want to prove that they are actually in control of their own destiny, and that is very difficult—not just in competition, but in trade and in other areas. Thanks again to everyone; we look forward to seeing you at dinner. DD

Julie Soloway, Charles Layton and Eric Richmond* International Cooperation in Antitrust Enforcement: A Canadian Perspective

‘My vision is to see a Bureau without borders—one where our enforcement reach is not strictly limited by resources or constrained by international or domestic jurisdictional boundaries’ (Canada’s Commissioner of Competition, John Pecman)1

1. Introduction The globalisation of the economy coupled with the proliferation of competition laws has led to an increase in the frequency and intensity of cooperation between global antitrust authorities. As the remarks cited above by Commissioner John Pecman illustrate, Canada’s Competition Bureau (the Bureau) is no exception to this trend. Highly dependent on external trade and investment, Canada is not able to effectively enforce its competition laws in isolation. As a result, the Bureau’s engagement with international antitrust authorities has expanded over time as the Canadian economy has become increasingly integrated with the North American and global economy. Indeed, today it would be noteworthy if cooperation did not occur between the Bureau and foreign antitrust authorities on a significant international cartel investigation or multi-jurisdictional merger. This chapter will discuss the effectiveness and legitimacy of cooperation in international antitrust enforcement from a Canadian perspective. It begins by briefly reviewing Canada’s approach to international cooperation that has developed over the past 20 years. Section 3 addresses the legitimacy of international enforcement cooperation, particularly with respect to the sharing of information without obtaining the parties’ consent. Section 4 considers recent efforts to coordinate merger reviews, with a particular emphasis on merger

* Julie Soloway is a partner, Charles Layton is an associate and Eric Richmond is a student-at-law in the Competition, Antitrust & Foreign Investment Group at Blake, Cassels & Graydon LLP, based in Toronto, Canada. The views herein are those of the authors and not necessarily of Blakes or any of its clients. 1 John Pecman, ‘Remarks by John Pecman, Interim Commissioner of Competition’, remarks delivered at the CBA Spring Speech, Toronto, Ontario, 28 May 2013, http://www. competitionbureau.gc.ca/eic/site/cb-bc.nsf/eng/03571.html.

434 Julie Soloway, Charles Layton and Eric Richmond remedies. The chapter concludes with a discussion on international cooperation in criminal matters.

2. Canada’s approach to international cooperation in antitrust enforcement The Bureau has long recognised the need for cooperation between international enforcement agencies. Over the past 15 years, the Bureau has signed formal cooperation agreements with 11 such agencies around the world. The Bureau also participates in several international forums, such as the Organization for Economic Cooperation and Development (OECD), the International Competition Network (ICN) and the International Consumer Protection and Enforcement Network (ICPEN), to develop and promote coordinated competition laws and policies in the globalised marketplace.2 The cooperation agreements established between the Bureau and its major international partners facilitate information exchange and cooperation. Primarily, the Bureau operates closely with the Department of Justice (the DOJ) and the Federal Trade Commission (FTC) in the United States and the European Commission (EC). Cooperation agreements between Canada and the United States and the European Union came into force in 1995 and 1999, respectively. Some of the highlights of the agreements include: (i) the coordination of enforcement activities; (ii) exchanges of information; (iii) semi-annual meetings to discuss matters of mutual interest; and (iv) the preservation of confidential information. Each jurisdiction must notify the foreign enforcement agency with respect to activities that may affect the other agency’s ‘important interests’.3 ‘Important interests’ are defined as any anticompetitive activities, merger or acquisition involving a company under the other jurisdiction, or application of remedies that would require or prohibit conduct in the territory of the other jurisdiction. When requested, each party must assist the other in its enforcement activities ‘to the extent compatible with the assisting Party’s laws and important interests’.4 The agreements also contain a ‘positive comity’ provision. This provision enables a country affected by anticompetitive activity arising from the other jurisdiction to request that the authorities of

2 ‘International Efforts’, http://www.competitionbureau.gc.ca/eic/site/cb-bc.nsf/eng/h_ 00128.html. 3 Agreement between the Government of Canada and the European Communities Regarding the Application of Their Competition Laws, [1999] OJ L175/50. 4 Ibid.

International Cooperation in Antitrust Enforcement 435 that jurisdiction initiate an investigation of the activity and take appropriate action. In the criminal context, Canada has signed mutual legal assistance treaties (MLATs) with most major jurisdictions, including the US, the EU and Australia. MLATs, administered by Canada’s Department of Justice, are more formal treaties used to request and obtain evidence for criminal investigations and prosecutions. A more detailed discussion of MLATs is located in Section 5 below.

3. Legitimacy of information sharing One of the primary concerns regarding the legitimacy of the Bureau’s international cooperation efforts pertains to its treatment of confidential information. The main provision governing the treatment of confidential information in the Bureau’s possession, section 29, provides that no official of the Bureau shall communicate information that is provided to, or obtained by, the Bureau in the course of executing its mandate. However, the section contains an exception such that, if the communication of information is to ‘a Canadian law enforcement agency or for the purposes of administration or enforcement’ of the Act, the Bureau may communicate the information.5 The Bureau takes the position that the exception to the confidentiality protections under section 29 of the Competition Act allows it to provide information to other antitrust authorities for the purposes of the ‘administration and enforcement’ of the Act.6 As a result, the Bureau’s practice is not to accept waivers from parties on the grounds that such waivers are not required by law. Moreover, the Bureau typically will not request the consent of the parties (or give notice to parties) when information is shared with a foreign antitrust agency. The Bureau’s expansive interpretation of section 29, namely that it has an unfettered ability to share information with foreign enforcement agencies, has raised questions surrounding the legitimacy of such information exchanges

Competition Act, RSC 1985, c C-34, s 29. The main provision for the treatment of confidential information in the possession of the Bureau is contained in section 29 of the Competition Act. Section 29 provides that no official of the Bureau shall communicate information that is provided to, or obtained by, the Bureau in the course of executing its mandate. However, the section contains an exception such that if the communication of information is to ‘a Canadian law enforcement agency or for the purposes of administration or enforcement’ of the Act, the Bureau may communicate the information. 5

6

436 Julie Soloway, Charles Layton and Eric Richmond and, more generally, regarding the Bureau’s treatment of confidential information. According to one commentator, The reference to ‘a Canadian law enforcement agency’ (and not merely ‘a law enforcement agency’) in subsection 29(1) and the comprehensive scheme for mutual legal assistance in Part III [of the Act] suggest that Parliament did not intend to grant the Bureau an unfettered ability to share information with foreign competition law enforcers.7

While the Bulletin on the Communication of Confidential Information under the Competition Act states that ‘[t]he decision to communicate confidential information to foreign authorities is not taken lightly’, it was nonetheless drafted under the assumption (without any explicit legal analysis or reasoning) that the Bureau’s right to communicate confidential information for the purposes of administration or enforcement of the Act allows the Bureau to share confidential information without the parties’ consent when coordinating enforcement actions with foreign law enforcement authorities.8 When the Bulletin was released in 2007, it was rightfully criticised by the C anadian Bar Association on the grounds that ‘[i]mportant issues, such as whether the Bureau is entitled to disclose information to foreign authorities, are not addressed in any significant detail’.9 Since the scope of section 29 has never been judicially considered, the Bureau’s broad interpretation of the provision remains unchecked despite concerns regarding the legitimacy of the Bureau’s approach to information sharing. Until the scope of section 29 becomes the subject of a judicial challenge, the Bureau will continue to be relatively unconstrained with respect to its practices for sharing information with foreign authorities.10

7 Omar Wakil, ‘Legislative Comment on s 29’, The 2013 Annotated Competition Act, Carswell, 2013, 43. 8 Competition Bureau, ‘Communication of Confidential Information under the Competition Act’, Information Bulletin 10 October 2007, paras 4.2.1.2, 4.2.2.1, http://www. competitionbureau.gc.ca/eic/site/cb-bc.nsf/eng/01277.html. 9 The Canadian Bar Association, ‘Information Bulletin (Consultation Draft) on the Communication and Treatment of Information under the Competition Act’, CBA, 2006, 8, http://www.competitionbureau.gc.ca/eic/site/cb-bc.nsf/vwapj/Infobulletin_cti_under%20 Comp_Act.pdf/$file/Infobulletin_cti_under%20Comp_Act.pdf. 10 While the Bureau does not require the consent of the parties before sharing confidential information, to ensure that the information remains confidential, it will, before sharing such information, seek to maintain the confidentiality of the information through either formal international instruments or assurances from the foreign authority. The Bureau also requires that use of the confidential information by the foreign authority be limited to the specific purposes for which it was provided. See section 4.2.2.2 of the Communication of Confidential Information Bulletin.

International Cooperation in Antitrust Enforcement 437

4. International coordination of merger control and merger remedies In the area of merger review, the Bulletin on Merger Remedies, published by the Bureau, sets out non-binding guidelines pertaining to the Bureau’s approach to international coordination and cooperation among competition authorities.11 As the Bulletin makes clear, the Bureau has adopted a case-specific approach with respect to the coordination of merger remedies with foreign antitrust agencies. In an inquiry involving a multi-jurisdictional merger, where Canadian competition concerns are addressed by remedies imposed by another agency, the Bureau will carefully consider the remedies crafted by that other competition authority in order to avoid unnecessary duplication of enforcement measures. The Bureau recognises that coordinating merger reviews with other enforcement agencies allows it to effectively mitigate anticompetitive effects in an efficient manner. Each additional authority seeking to obtain separate remedies brings with it a greater cost of compliance and administration, uncertainty over legal treatment, and duplication of resources. As a result, in appropriate cases, the Bureau will seek coordinated remedies to promote more effective and efficient compliance with competition laws. Such coordination: (i) eliminates or reduces duplicative negotiations with parties, thereby lessening the burden on businesses; and (ii) encourages parties who have transnational issues to negotiate remedies with the Bureau and other authorities concurrently. Notwithstanding the benefits that are derived from coordination, the Bureau’s willingness to adopt a deferential approach with respect to remedies is not unqualified. In particular, the Bureau must be satisfied that the actions taken by foreign authorities are sufficient to resolve any competition issues in Canada. Thus, although global remedies are given due consideration, the Bureau is more likely to formalise negotiated remedies within Canada when the matter raises Canada-specific issues, when the Canadian impact of the merger is particularly significant, when the assets to be divested reside in Canada or when Canadian-specific remedies are critical to the enforcement of the terms of the settlement.12 As noted in the Bulletin on Merger Remedies: While consistent and coordinated remedies are desirable, each jurisdiction must retain the authority and ability to ensure remedies that are sufficient within its own 11 Competition Bureau, ‘Information Bulletin on Merger Remedies in Canada’, Information Bulletin, 22 September 2006, http://www.competitionbureau.gc.ca/eic/site/cb-bc. nsf/eng/02170.html. 12 Ibid, para 78.

438 Julie Soloway, Charles Layton and Eric Richmond borders … Therefore, within the framework of its own laws, and to the extent compatible with its own interests, the Bureau will generally take another jurisdiction’s interests and policies into account only to the extent that such interests and policies do not limit or prevent Canadian competition concerns from being remedied.13

Thus, although consistent enforcement is a valued goal, the Bureau’s mandate to carry out the enforcement of the Competition Act in Canada is the Bureau’s overarching priority, even if specific remedies in Canada come at the expense of efficiency or consistency in international merger control. Worth noting is that the 2009 amendments to the Competition Act, which implemented two-stage merger review in Canada and aligned the timing for merger reviews with the Hart-Scott-Rodino Act, have facilitated a higher level of cooperation between the Bureau and US antitrust agencies. For example, the initial 30 day waiting period applied to mergers in Canada is now aligned with the timeline in the US and, likewise, the design of the supplementary information request (SIR) process in Canada was modelled on the US s econd request process. For global transactions, SIRs in Canada will typically focus on Canadian-specific issues and, provided that the parties file pre-merger notifications in or around the same time period, the alignment of the statutory time periods enables more seamless coordination with respect to the synchronisation of merger remedies. Thus, in general, the convergence between Canada’s Competition Act and the Hart-Scott-Rodino Act has led to increasingly effective cooperation between Canada and the US over the past four years. Below, we survey three recent transactions which illustrate the spectrum of the Bureau’s approach to the coordination of merger remedies. As will be demonstrated, the likelihood of the Bureau pursuing an independent remedy is dependent on: (i) whether remedies in other jurisdictions address Canadianspecific issues; (ii) whether such remedies must be specifically enforceable in Canada; and (iii) the nature of the parties’ operations in Canada (ie whether the parties have significant assets or operations in Canada).

4.1 Remedy satisfactorily addressed by another authority The Bureau has declined to seek a remedy where a substantial lessening or prevention of competition in Canada has been addressed through remedies imposed by another authority. The recent United Technologies/Goodrich transaction represents one such example. Both Goodrich and United Technologies were engaged in the manufacture and sale of commercial aircraft components. Following its review of the

Ibid, para 80 (emphasis added).

13

International Cooperation in Antitrust Enforcement 439 transaction, the Bureau raised concerns that the transaction would result in a substantial lessening of competition with respect to the sale of certain aircraft parts, which would have a downstream effect on Canadian carriers which purchase aircrafts containing these products. Throughout its investigation, the Bureau cooperated closely with the European Commission and the DOJ. The extent of the cooperation between the agencies is illustrated by the fact that all three agencies announced the outcome of their reviews on the same day.14 The Commission and DOJ approved the merger subject to the divestiture of certain operations in Europe and the US. Concurrently, the Bureau issued a press release stating that the remedial orders issued in the US and Europe ‘sufficiently mitigate the potential of anticompetitive effects in Canada’. Specifically, the remedies obtained in the US and Europe would ensure that Canadian companies could source aircraft parts on a competitive basis in the future. As a result, no remedy was required by the Bureau in Canada.15 The United Technologies/Goodrich transaction is an example of how effective international cooperation and the synchronisation of merger remedies can impose consistent obligations on merging parties, while eliminating unnecessary duplication and minimising the costs associated with antitrust compliance.16 In that regard, it is worth noting that in this case the Bureau concluded that the aerospace sector was ‘generally global in nature and scope’ and that customers adopted global sourcing strategies. In addition, no Canadian assets were involved, and the ‘vast majority’ of aerospace manufacturing facilities and customers of United Technologies and Goodrich were located in the US and Europe. Accordingly, the Bureau was comforted that it need not require ‘independent memorialisation’ of the remedies obtained by foreign antitrust agencies.17

14 Joseph Wayland, ‘International Cooperation at the Antitrust Division’,remarks delivered at the International Bar Association’s 16th Annual Competition Conference, Florence, 14 September 2012, http://www.justice.gov/atr/public/speeches/286979.pdf. 15 Competition Bureau, ‘Competition Bureau Statement Regarding United Technology Corporation’s Acquisition of Goodrich Corporation’, Press Release, 26 July 2012, www. competitionbureau.gc.ca/eic/site/cb-bc.nsf/eng/03483.html. 16 Other recent notable examples include: Dow Chemicals/Rohm (2009), where the Bureau accepted divestiture commitments of certain IP rights obtained by the FTC; Thomson/Reuters (2008), where the Bureau concluded that divestitures and other behavioural commitments obtained by the US DOJ and the European Commission were sufficient to address Canadian concerns; and Alcan/Pechiney (2003), where the Bureau accepted that divestiture commitments relating to rolling facilities and other behavioural comments concerning refining and smelting technology were sufficient to preserve competitive options for Canadian customers. 17 See Press Release, above n 15.

440 Julie Soloway, Charles Layton and Eric Richmond

4.2 Mirroring the remedy obtained by a foreign authority Where there is cause for concern regarding whether a global antitrust r emedy may be enforceable in Canada, the Bureau may replicate the remedy obtained by a foreign authority while requiring it to be registered as a consent agreement with Canada’s Competition Tribunal. This occurred in 2010, when the CocaCola Company attempted to purchase its bottler, Coca-Cola E nterprises Inc, for $12.3 billion.18 The FTC and the Bureau were concerned that sensitive business information of rival Dr Pepper Snapple Group Inc, which was bottled by Coca-Cola Enterprises, would be shared with Coca-Cola executives. The FTC and the Bureau worked cooperatively ‘to coordinate parallel reviews of the proposed acquisition and to negotiate mutually acceptable resolutions to this matter’.19 The FTC and Bureau each obtained largely identical remedies from Coca-Cola, which required the company to establish a firewall to partition Dr Pepper Snapple Group’s sensitive and confidential information and prevent it from being accessed by Coca-Cola executives. To ensure compliance with the remedy, the Bureau appointed an independent monitor to verify that Coca-Cola was complying with the restrictions set out in the consent agreement.20 A significant factor distinguishing the Coca-Cola transaction from United Technologies/Goodrich was that the remedy obtained was behavioural in nature rather than structural. In United Technologies/Goodrich, the Bureau was comforted by the fact that the divestitures required by the DOJ would be completed and would, once and for all, mitigate the Bureau’s concerns in Canada. In contrast, the behavioural remedy required in the Coca-Cola transaction required ongoing monitoring; accordingly, in order to address its concerns on an ongoing basis, the Bureau determined that it must obtain a remedy specifically enforceable in Canada and could not rely on the remedies imposed in other jurisdictions.

4.3 Specific remedies to address canadian-specific concerns Despite the movement towards convergence, antitrust laws are still rooted in domestic legislation. Competitive dynamics often vary across national

18 Jeff Bliss, ‘Coca-Cola Agrees to US Conditions to Buy Bottler’, Bloomberg, 27 September 2010, http://www.bloomberg.com/news/2010-09-27/coca-cola-agrees-to-u-sconditions-to-buy-soft-drink-bottler.html. 19 Competition Bureau, ‘Competition Bureau Requires Remedy in Coca-Cola Acquisition’, Press Release, 27 September 2010, http://www.competitionbureau.gc.ca/eic/site/ cb-bc.nsf/eng/03290.html. 20 Ibid.

International Cooperation in Antitrust Enforcement 441 oundaries such that a transaction may have widely disparate competitive b effects in different countries. These varying dynamics are a reflection of the fact that, while some antitrust product markets are global, others can be defined more narrowly as national, regional or even local. Accordingly, although there is scope for international cooperation in many global transactions, other transactions which raise unique domestic issues require r emedies tailored to a particular jurisdiction. Such cases illustrate the inherent limitations of international cooperation. As noted in the Bureau’s Information Bulletin on Merger Remedies in Canada, ‘the greater the extent to which competition issues identified in Canada are similar to those in other jurisdictions, the greater the likelihood that coordinated remedies will be e ffective’.21 However, the converse is also true, as demonstrated by Novartis’s $28.1 billion acquisition of Alcon in 2010. In Novartis/Alcon, the Bureau obtained remedies unique to Canada in order to address Canadian-specific competition concerns. Over the course of its review, the Bureau ‘worked closely’ with the FTC and the European Commission.22 The Commission, the FTC, the Bureau and several other international competition agencies raised concerns that the transaction would give Novartis a dominant position in research and development and in the supply of certain ophthalmic products. To resolve its concerns, the FTC entered into a consent agreement with Novartis, requiring the divestiture of its injectable miotics product, Miochotol-E. The Bureau required divestitures in injectable miotics, as well as multi-purpose contact lens solutions and an ophthalmic anti-allergy agent, the latter two of which went beyond what was divested in the United States pursuant to Novartis’s consent agreement with the FTC. The European Commission objected to the transaction with respect to an even greater number of products and required divestitures in certain Member States.23 In contrast, Australia, mirroring the approach taken by the FTC, required divestitures only with respect to injectable miotics. The differing remedies obtained in multiple jurisdictions in Novartis/Alcon demonstrate that, notwithstanding efforts by antitrust authorities to coordinate merger remedies, certain transactions raise national or subnational issues

See Bulletin, above n 11, para 77. Competition Bureau, ‘Competition Bureau Secures Divestitures in Novartis’ Acquisition of Alcon’, Press Release, 9 August 2010, http://www.competitionbureau.gc.ca/eic/site/ cb-bc.nsf/eng/03274.html. 23 Products addressed by remedies in the EU included: ophthalmological anti-infectives, anti-inflammatory/anti-infective combinations, anti-allergics, decongestants, antiseptics, mydriatics and cycloplegics, diagnostic agents, non-steroidal anti-inflammatories, injectable miotics, anti-glaucoma products, artificial tears, and multipurpose solutions for contact lenses. 21 22

442 Julie Soloway, Charles Layton and Eric Richmond and are not conducive to consistent international antitrust enforcement. As noted in the Bulletin on Merger Remedies, each jurisdiction must retain the authority and ability to ensure remedies that are sufficient within its own borders … the Bureau will generally take another jurisdiction’s interests and policies into account only to the extent that such interests and policies do not limit or prevent Canadian competition concerns from being remedied.24

Thus, while cooperation between the leading international antitrust agencies has become the norm, Novartis/Alcon demonstrates that such cooperation is no guarantee of coordinated outcomes or consistent antitrust enforcement.

5. International cooperation on criminal matters 5.1 Overview In criminal investigations, the Bureau will cooperate with antitrust authorities in foreign jurisdictions to identify and uncover criminal conduct and to share evidence required for prosecution. Ensuring compliance with the criminal provisions of the Competition Act and, in particular, the provisions aimed at cartels and bid-rigging, is an enforcement priority for the Bureau. In that regard, the Bureau has declared that, in order to combat international cartels, it will use its ‘full powers to participate in what must be a concerted international effort’.25 International cooperation is facilitated through formal agreements, but also occurs informally on an ad hoc basis.26 Following the initiation of an investigation or prosecution, a party may ask for investigative assistance under a mutual legal assistance treaty (MLAT). The MLAT is an agreement between two countries that they will assist each other in ‘all matters relating to the investigation, prosecution and suppression of criminal offences’.27

See Bulletin, above n 11, para 77. Calvin Goldman, Navin Joneja and Cassandra Brown, ‘Enforcement of the Competition Act in an Era of Cooperation and Convergence’, paper delivered at the Canadian Bar Association Annual Fall Conference, Ottawa, Ontario, 17–19 September 2008. 26 Calvin Goldman, Robert Kwinter and Angie Morris, ‘The Investigative Powers of the Canadian Competition Bureau: Domestic and International Dimensions’, in Claus-Dieter Ehlermann and Isabela Atanasiu, eds, European Competition Law Annual 2006: Enforcement of Prohibition of Cartels, Hart Publishing, 2006, 191 et seq. 27 Treaty between the Government of Canada and the Government of the United States of America on Mutual Legal Assistance in Criminal Matters, 18 March 1985, CTS 1990 No 19, article II. 24 25

International Cooperation in Antitrust Enforcement 443 The p urpose of the MLAT is to facilitate the sharing of evidence located in another jurisdiction that is considered criminal in the requesting party’s jurisdiction.28

5.2 Recent examples of international cooperation and cartel enforcement The Bureau recognises that cooperation between enforcement agencies leads to a greater likelihood of detection and prosecution of international cartels. As noted by the Commissioner of Competition, companies engaging in anti-competitive activities face an increasing regulatory risk—not only in Canada but around the world … the likelihood of detection and prosecution are increasing. And with increasing cooperation and coordination between agencies, there are fewer and fewer places to hide.29

Recently, coordinated simultaneous investigations were made into an international bid-rigging conspiracy by Japanese auto parts suppliers. This was one of the largest criminal antitrust probes on record. The Bureau’s investigation into auto parts suppliers was coordinated with a number of jurisdictions, including the US, Japan, the EU, the UK and Australia.30 The investigation revealed that Japanese auto parts suppliers conspired to submit bids in response to requests for quotations to supply automobile manufacturers (Honda and Toyota) with motor vehicle parts. The auto parts suppliers have been fined over $1 billion in Japan and the US, and additional fines are expected to be levied in the EU.31 Canada recently fined Yazaki Corporation $30 million and Furukawa Electric Co $5 million for their involvement. These fines are two of the largest ever ordered by a Canadian court for a bid-rigging offence.32 None of the bid-rigging took place in Canada or involved Canadian personnel, but had an indirect effect in Canada through the import of auto

See Goldman et al, above n 26. John Pecman, ‘Remarks by John Pecman, Interim Commissioner of Competition’, remarks delivered at Blake, Cassels & Graydon LLP, Toronto, Ontario, 30 October 2012, http://www.competitionbureau.gc.ca/eic/site/cb-bc.nsf/eng/02834.html. 30 Competition Bureau, ‘$5M Fine for a Japanese Supplier of Motor Vehicle Components’, Press Release, 4 April 2013, http://www.competitionbureau.gc.ca/eic/site/cb-bc.nsf/ eng/03555.html. 31 Ma Jie, Masatsugu Horie and Aoife White, ‘Japanese Auto Parts Makers Bracing for EU Cartel Fines’, Bloomberg, 31 January 2013, http://www.bloomberg.com/news/2013-0131/japanese-auto-parts-makers-bracing-for-eu-cartel-fines.html. 32 Competition Bureau, ‘Record $30M Fine Obtained by Competition Bureau against Japanese Auto Parts Supplier’, Press Release, 18 April 2013, http://www.competition bureau.gc.ca/eic/site/cb-bc.nsf/eng/03560.html. 28 29

444 Julie Soloway, Charles Layton and Eric Richmond parts. For example, nearly all Honda Civics sold in Canada are produced in Canada and 50 per cent of the Toyota vehicles sold to Canadians are produced in Canada. The total volume of commerce affected by the bid-rigging conspiracy in Canada was approximately $300 million.33 In another recent example, Canada worked extensively with the US, the EU, Korea and Australia to investigate and prosecute air cargo carriers based on allegations that the carriers had engaged in a conspiracy to fix rates and surcharges for international air cargo between 2002 and 2006.34 In the US, 19 air carriers pleaded guilty, resulting in $1.8 billion in fines.35 In the EU, fines totalling approximately €799.4 million have been levied.36 To date, Canada’s air cargo investigation has resulted in seven convictions and fines totalling more than $22.6 million.37 Both of these examples show that global businesses are subject to increasing scrutiny by the Bureau, which, by virtue of its international cooperation efforts, has been successful in supplementing its own limited resources by leveraging the resources and tools of other global, and often better funded, antitrust agencies.

6. Conclusion Coordinating enforcement efforts with foreign antitrust agencies has allowed the Bureau to mitigate anticompetitive effects in a more effective and efficient manner. While perfect consistency in international enforcement will never occur, given the exigencies of domestic legislation and given that competitive dynamics are often unique to domestic markets, nonetheless, there are signs that the Bureau will continue to enhance its cooperation with international enforcement agencies. As noted recently by Canada’s Commissioner

Ibid. Shearman & Sterling LLP, ‘Air Cargo’, http://www.carteldigest.com/cartel-detailpage.cfm?itemID=19. 35 United States Department of Justice, ‘Former Air France Cargo Executives Indicted in Conspiracy to Fix Rates and Surcharges on Air Cargo Shipments’, Press Release, 26 April 2011, http://www.justice.gov/atr/public/press_releases/2008/234435.pdf. 36 European Commission, ‘Antitrust: Commission Fines 11 Air Cargo Carriers €799 Million in Price Fixing Cartel’, Press Release, 9 November 2010, http://europa.eu/rapid/ press-release_IP-10-1487_en.htm. 37 Competition Bureau, ‘Korean Air Pleads Guilty to Price-Fixing Conspiracy’, Press Release, 19 July 2012, http://www.competitionbureau.gc.ca/eic/site/cb-bc.nsf/eng/03482. html. 33 34

International Cooperation in Antitrust Enforcement 445 of Competition, the Bureau will strive to develop ‘pick up the phone relationships’ with other enforcement agencies ‘when files of mutual concern arise’.38 In the mergers context, the Bureau’s efforts to coordinate merger reviews can often serve to streamline the regulatory process, thereby reducing the costs of compliance and increasing certainty of legal treatment for businesses. With respect to criminal enforcement, the Bureau recognises that cooperation between enforcement agencies is crucial for the identification and prosecution of criminal conduct. The growing number of statements from the Bureau citing Canada’s international coordination is indicative of the Bureau’s commitment to cooperation. This commitment will likely only strengthen over time as the Bureau seeks to perform its mandate in the face of budget cutbacks, which will force it to operate in an increasingly streamlined and efficient manner.

38 John Pecman, ‘Remarks by John Pecman, Interim Commissioner of Competition’, remarks delivered at the Vancouver Roundtable Speech, 5 December 2012, http://www. competitionbureau.gc.ca/eic/site/cb-bc.nsf/eng/03514.html.

446

Alberto Heimler* Effectiveness of Enforcement Cooperation in Developing Countries: What Role Can Existing Institutions Play?

1. Introduction As trade liberalisation has progressed, the gap between the rapid internationalisation of markets and the fragmentation of the systems of domestic antitrust laws has meant that competition authorities, especially in developing countries, have experienced considerable difficulties in fighting the transnational anticompetitive practices which originate abroad but have a negative effect on their domestic markets or to adequately intervene against transnational transactions initiated in other countries but affecting their home markets. To try to address this problem, the international competition community has experimented with a number of possible solutions, ranging from a failed attempt to promote a multilateral cooperation system in the area of competition to voluntary bilateral cooperation agreements between developed and developing countries, and to the creation of an organisation meant to facilitate cooperation in the field of antitrust—the International Competition Network (ICN). So far, those solutions have been mostly unsuccessful (in terms of avoiding international restrictions of competition) and frustrating (since cooperation has proved to be much more difficult than expected) especially for developing countries. Regional agreements on cooperation in competition were considered to be more promising, largely because it was believed that the cooperating countries may have more of an incentive to cooperate than pairs of developed and developing countries unequal in size and with different stakes and unbalanced trade flows. However, the lack of sufficient financial resources and insurmountable jurisdictional problems have so far prevented the regional agreements in developing countries from achieving tangible results. Developing countries have to be more inventive and find less costly ways to cooperate. Many steps forward are possible, even without creating supranational authorities. The example of the agreement between Australia and New Zealand can be of some help. As for convergence on the procedural and substantive aspects of antitrust law, the OECD, UNCTAD and the ICN all play an important role, with the ICN specifically SNA, Rome.

*

448 Alberto Heimler designed to meet the challenges facing developing countries in adapting to the effects-based approach which the more established jurisdictions have developed in recent decades. However, in order to become more effective in achieving its objectives, the ICN should take steps to overcome a number of organisational challenges so as to better engage non-OECD countries in its activities and ensure that the identification of best practices actually leads to a reduction in the transaction costs of firms and the creation of a level playing field in terms of procedures and substantive prohibitions.

2. The WTO effort to incorporate competition in trade agreements Picking up on a proposal by the European Commission, the 1996 World Trade Organization (WTO) Singapore intergovernmental conference created a working group to study issues relating to the interaction between trade and competition policy raised by WTO Members. The mandate of the group was very open, but it is clear from a number of speeches of DG Competition representatives that the agenda was very ambitious. In a speech delivered in Rome in 1995, Jean-François Pons, the Deputy Director General of DG Competition at the time, referred to the possibility of negotiating a plurilateral agreement on competition within the WTO.1 Countries would commit themselves to adopt a minimum set of antitrust rules and be subject to a dispute settlement mechanism. Pons did not go further, but in his intervention in Rome he cited several times the 1995 report, ‘Competition Policy in the New Trade Order: Strengthening International Cooperation and Rules’, drafted by a group of three external experts2 and five Commission experts.3 It made clear reference to the possibility that the proposed plurilateral agreement on competition would also deal with controversies over the way a jurisdiction decided on specific antitrust cases. As a result, this extension of the dispute settlement mechanism, even though it never made it into official Commission documents, was considered fully part of the Commission agenda. Most of the concerns over WTO involvement in antitrust enforcement originated from this.

1 Jean-François Pons, ‘Règles, Institutions et Relations internationales: Politique de concurrence et développement des échanges: pour un renforcement significatif de la coopération’, speech, Rome, 20–21 November 1995. 2 Ulrich Immenga, Frédéric Jenny and Ernst-Ulrich Petersmann. 3 Claus-Dieter Ehlermann, Roderick Abbott, François Lamoureux, Jean-François Marchipont and Alexis Jacquemin.

Effectiveness of Enforcement Cooperation in Developing Countries 449 These were not only theoretical or ideological concerns; they immediately had a practical application, partly as a result of a US action. In a WTO controversy initiated in 1996 and concluded in 1998,4 the United States government accused Japan of impeding efficient access of Kodak to the Japanese film market in favour of its Japanese rival, Fuji. According to the US, Japan failed to enforce Japanese antitrust laws against Fuji and, by allowing exclusive distribution agreements between Fuji and most retail shops in Japan, contributed to the exclusion of Kodak from the Japanese market. The criticism went further, with the US government arguing that ‘Japan’s entire retail system puts foreign competition at an unfair disadvantage’.5 The panel ruled against the US, concluding that the Japanese rules (including Japanese antitrust law), since they applied equally to domestic and foreign firms, did not discriminate against foreign companies.6 The substance of the matter, ie whether a nation would violate the GATT/ WTO rules because it failed to enforce its antitrust laws against private practices that might foreclose its domestic market, was never decided in the Kodak–Fuji dispute. If the panel had addressed the antitrust issues, it would have had a number of difficulties because exclusive distribution agreements are restrictive only in fact-specific circumstances. In particular, the WTO panel would have had to prove that Japanese consumers were actually hurt by the disputed practice, which would have required a lot of technical expertise and access to a set of quite specific information on the Japanese market. Furthermore, such a decision of the dispute settlement panel could have been in contradiction with an analogous decision of competition authorities, and with court judgments. The debates stirred by the Kodak–Fuji case and the discussions within the WTO working group were not able to eliminate all concerns. Thus, in the WTO meeting in Doha in 2001, ministers provided further clarifications as to the mandate of the group: 23. … we agree that negotiations will take place after the Fifth Session of the Ministerial Conference on the basis of a decision to be taken, by explicit consensus, at that session on modalities of negotiations. 24. … we shall work in cooperation with other relevant intergovernmental organisations, including UNCTAD, and through appropriate regional and bilateral

See http://www.wto.org/english/tratop_e/dispu_e/cases_e/ds44_e.htm. Emily Nelson, ‘Presentation of Data in Kodak’s case against Japan and Fuji delayed by US’, The Wall Street Journal, 20 September 1996, section B, 4. 6 The US failed to show that Japan had adopted any ‘measure’ to achieve the result of exclusivity and could not be held responsible under the traditional legal basis for breach of obligations. As an alternative theory, the US attempted to seek a ruling under the separate basis of a ‘non-violation complaint’, but failed in this respect as well. A non-violation complaint requires the complainant to show that the non-action of the defending WTO member could not have been foreseen. The panel ruled that this condition had not been satisfied. 4 5

450 Alberto Heimler c hannels, to provide strengthened and adequately resourced assistance to respond to these needs. 25. In the period until the Fifth Session, further work in the Working Group on the Interaction between Trade and Competition Policy will focus on the clarification of: core principles, including transparency, non-discrimination and procedural fairness, and provisions on hardcore cartels; modalities for voluntary cooperation; and support for progressive reinforcement of competition institutions in developing countries through capacity building. Full account shall be taken of the needs of developing and least-developed country participants and appropriate flexibility provided to address them.

As Clarke and Evenett7 note, paragraph 25 of the Declaration shifts the focus away from market access issues and demands that the WTO’s Working Group on the Interaction between Trade and Competition Policy focus on hard-core cartels, an area almost ignored before. However, even this significant limitation of the objectives to be pursued in multilateral negotiations led to no support and the WTO General Council’s ‘July package’ of 2004 put the whole subject of trade and competition in limbo. As a result of the WTO General Council’s decision, the possibility of resuming this work following the conclusion of the Doha Round is left open. This is not a new development in the history of international trade negotiations. Already in 1948 the Havana Charter required nations to address transnational restrictive business practices by authorising the then-proposed International Trade Organization to ‘take every possible remedial action’8 against them. The Havana Charter was never adopted because of the lack of support by the US government. Very similar developments have characterised this most recent attempt to launch negotiations on antitrust matters. Fox argues that, although strongly supported by the EU, the antitrust proposal lacked support from the United States, which feared a transfer of powers to a global bureaucracy and a lowest-common-denominator law, and it was opposed by developing countries because they feared a Trojan horse that would open floodgates to imports and disarm them from protecting their nations’ interests.9

In reality, over the course of the years, most of the problems that the US had identified were solved. Indeed, the major criticism made by the US authorities against the original project of the EU was that antitrust laws were enforced by judges and it would have been contrary to any principle of law to make governments responsible should a judgment (not a legal provision)

7 The State Secretariat of Economic Affairs and Simon Evenett, eds, The Singapore Issues and the World Trading System: The Road to Cancun and Beyond, World Trade Institute, June 2003. 8 See Havana Charter for an International Trade Organization, chapter 13, http://www. worldtradelaw.net/misc/havana.pdf. 9 Eleanor Fox, ‘The WTO’s First Antitrust Case—Mexican telecom: A Sleeping Victory for Trade and Competition’, 9 Journal of International Economic Law 271 (2006).

Effectiveness of Enforcement Cooperation in Developing Countries 451 be c onsidered unsatisfactory by a foreign nation. The 2001 Doha declaration was limited, so that only hard-core cartels (notably excluding export cartels) would be subject to a mandatory prohibition under the national laws of WTO Members. What led to the defeat of the trade and competition dossier was the opposition of developing countries. They argued that any negotiation on antitrust would only help the developed countries and their multinationals. If the Commission (and the EU Member States as well) had started with a much less ambitious, but feasible, agenda, aimed at helping developing countries in their efforts to introduce a domestic competition law system (as the Doha declaration implicitly acknowledges), there would have been a greater chance for the trade and competition dossier to flourish.

3. The International Competition Network and the cooperation agenda The lack of consensus for a WTO-led negotiation on antitrust did not impede further developments in promoting cooperation in antitrust. In February 2000, the International Competition Policy Advisory Committee published its final report in the US. It set forth recommendations to competition agencies throughout the world, designed to enhance merger enforcement and improve cooperation to address private restraints of competition that impede market access. On the subject of the intersection of trade and competition policy, the US Department of Justice press release stated:10 [T]he [ICPAC] report recommends further development of bilateral agreements with ‘positive comity’ provisions (which allow a nation affected by anticompetitive practices to request that the nation in which the alleged conduct is occurring initiate an appropriate enforcement action) as well as the use of extraterritorial enforcement tools where necessary. Further, the report argues that new multilateral approaches are also needed, although it does not see the WTO as the natural home for all global competition policy initiatives. Instead, it proposes a new Global Competition Initiative for addressing the broad global competition agenda.

In particular, the report recommended that: the United States explore the scope for collaborations among interested governments and international organizations to create a new venue where government officials, as well as private firms, non-governmental organizations (NGOs), and others can consult on matters of competition law and policy.11

See http://www.usdoj.gov/atr/icpac/4272.htm. See ICPAC Report, 282, http://www.justice.gov/atr/final-report.

10 11

452 Alberto Heimler Speaking in Brussels at the tenth anniversary of the EC merger regulation in September 2000, Joel Klein, at that time the Assistant Attorney General for Antitrust, took the ICPAC recommendations forward, suggesting that ‘whatever happens on antitrust at the WTO … we should move in the direction of a Global Competition Initiative’. More specifically, he suggested that interested jurisdictions along with the international bodies already thinking about these issues e.g., the OECD, WTO, UNCTAD, World Bank, and others might establish a joint working group—first for exchanging information and views (e.g., about ongoing and planned activities, common challenges, approaches each [is] taking to support sound enforcement practices, areas that are most vexing, greatest opportunities for cooperation, etc.) and then for fully exploring a Global Competition Initiative along the lines laid out in the ICPAC report.12

Speaking at the same conference, Mario Monti, at the time EC Commissioner for Competition, endorsed Assistant Attorney General Klein’s suggestion to create a new forum addressing the international challenges of antitrust enforcement. Meeting at the Fordham Conference on International Antitrust in New York City in October 2001, a group of top antitrust law officials from 14 jurisdictions13 created the International Competition Network. Its objective was to become ‘an informal network of antitrust agencies from developed and developing countries that will address antitrust enforcement and policy issues of common interest and formulate proposals for procedural and substantive convergence through a results-oriented agenda and structure’.14 The ICN quickly became a world forum for convergence of antitrust law enforcement practices. The ICN has well over 100 members—a sign of its success, but also the result of very low requirements for participating in the network (the existence of a domestic antitrust law and of an agency responsible for enforcement). Within the ICN there are three working groups on substantive issues whose mandates are to identify best practices in, respectively, merger control, the fight against hard-core cartels and the identification of abusive conduct by dominant companies. Furthermore, there is a working group—the Agency Effectiveness Group—which aims to overcome the challenges developing countries face in building up an efficient antitrust authority and an effective enforcement practice, and an Advocacy Working

12 See Joel Klein, ‘Time for a Global Competition Initiative?’, speech, Brussels, 14 September 2000, http://www.usdoj.gov/atr/public/speeches/6486.htm. 13 These were the authorities of Australia, Canada, the European Union, France, Germany, Israel, Italy, Japan, Korea, Mexico, South Africa, the United Kingdom, the United States and Zambia. 14 See ‘Memorandum on the Establishment and Operation of the International Competition Network’, http://www.internationalcompetitionnetwork.org/index.php/en/about-icn/ operational-framework.

Effectiveness of Enforcement Cooperation in Developing Countries 453 Group aiming to help agencies in their activity as promoter of a less restrictive regulatory environment. The ICN is a virtual organisation that operates without a secretariat. Every report is written by members or by non-governmental advisors who work for the ICN on a voluntary basis. Discussions are held via conference calls, and every agency that so desires can participate. Decision-making is for members only (ie competition authorities) and is consensual. In this sense, the organisation is highly democratic. The larger agencies, like the US Department of Justice, the US Federal Trade Commission and the European Commission, have more resources than the others and therefore can put more people to work on ICN issues than the other agencies. However, the fact that decision-making is consensual represents a very important safeguard that allows even very small agencies to have a voice. And in many instances, the veto power of small agencies has been successfully used, especially in the early years of the ICN. As time has gone by, more technically difficult competition issues have been taken up by the ICN, raising the barriers for entry for small agencies and in particular for those of developing countries. In this sense, the risk of them being excluded from decision-making has increased. At least at the beginning, the fact that the ICN was created to ensure convergence in antitrust enforcement, especially in developing countries, was thought to be a sufficient condition for maintaining the right incentive on the part of the most developed agencies to engage with the smaller ones and to make sure that they actively participate. To some extent, the original objective has been forgotten. The lack of a secretariat has made the activity of the ICN much more difficult to challenge on the part of the smaller authorities. At the same time, the lack of independence of those writing the reports (they are all tied to their own authorities) has made them disinclined to criticise their own jurisdictions, reducing the incentives for the ICN to become a standardsetting organisation. Nonetheless, some positive results have been achieved. Since its establishment in 2001, the ICN has issued more than 100 recommended practices to members on issues like merger control, cartels, abuse of dominance, banking, telecommunications and advocacy. These recommendations are non-binding, and it is left to governments and agencies to implement them as appropriate. These recommendations have increasingly become the benchmark that agencies and the private sector use to evaluate the appropriateness of laws and policies—and in that respect, there can be pressure for agencies to adopt regimes that conform to them. Three recent examples highlight the influence of the recommended practices on merger procedure and review. However, they also show the lack of substantive influence of the ICN. The first example is Korea, which has made significant changes to its merger notification thresholds. In doing so, it indicated publicly its desire to bring national law into line with the Recommended Practices for Merger

454 Alberto Heimler otification and Review Procedures. Second, India revised its merger regime, N in the summer of 2007, in a way that was at odds with some of the most important merger procedure recommended practices: local nexus, timelines and information requirements. Immediately thereafter, there was widespread criticism from the private sector, both within and outside of India, with the majority using the ICN recommendations as a benchmark for their complaints. As a result, the Indian agency proposed implementing regulations that would bring the regime into greater conformity with the ICN recommendations. Third, in countries that are just adopting new merger control regimes, like China a few years ago, the ICN has been cited many times by outside bodies commenting on various proposals. There was considerable interest by the Chinese government in understanding and incorporating the ICN recommendations into their merger control regime. However, since China became active in antitrust, it has completely ignored the ICN and its recommendations. As these examples show, the influence of the ICN in terms of compliance with recommended practices has so far been limited and, more importantly, its impact has been limited to the way rules and regulations are drafted. There is no experience of applying ICN recommendations to actual cases or of promoting cooperation among authorities. Some of the material produced by the ICN is user-/practically oriented, and such material has been used for training purposes and for enforcement. Unfortunately, there is no collected evidence of this. However, in more recent years, by addressing more controversial issues, the ICN has progressively abandoned the recommendation approach and has instead chosen an encyclopaedic approach, where a list of what every jurisdiction does is presented with equal weight, without any effort being made to suggest what the best practice is or should be. This is particularly the case for the activity on unilateral conduct and on advocacy, where the ICN reports cannot be used as examples of best practices (since there are too many of them and they are sometimes mutually exclusive) and therefore cannot be used as a point of reference for newer and less expert institutions. In any case, the major problem of the ICN is the lack of active participation of developing countries in its working groups and in the development of the material it provides to the antitrust community worldwide.15 How much

15 The roundtable on enforcement cooperation, aimed at providing a lesson on what had been learnt about cooperation in its first 10 years of existence, was organised by the ICN on the basis of a questionnaire that was filled out by the following agencies: Belgian Competition Authority; Tribunal de Defensa de la Libre Competencia (Chile); European Commission; Irish Competition Authority; Japan Fair Trade Commission; Monopo lies and Price Commission of Kenya; Korea Fair Trade Commission; Mexican Federal Competition Commission; Norwegian Competition Authority; Competition Commis sion of Pakistan; Commission Nationale de la Concurrence du Senegal; Competition

Effectiveness of Enforcement Cooperation in Developing Countries 455 this weakens the function of the ICN as a body seeking convergence in antitrust enforcement in the world is difficult to say. The evidence is indeed mixed. The annual conference the ICN organises each year in different parts of the world has seen the growing participation of developing countries: in the 2013 Warsaw conference, 81 jurisdictions were present, 40 of which were not from the OECD. However, it cannot be concluded that, as a result of this enhanced participation, ICN products are used more by these agencies in the practice of enforcement. ICN activities should concentrate more on identifying recommended practices and on the degree to which such recommendations are adopted and respected by member jurisdictions. Every year, jurisdictions should review themselves with respect to the recommended practices of the ICN, providing reasons for non-compliance or for partial compliance, and suggesting recommendations or issues member jurisdictions would like to see developed in the coming years. In this way, the agenda and the work products of the ICN would better reflect the needs of members, while recommendations by the ICN would receive a continuous check and their promotion would be the main instrument for achieving convergence in antitrust enforcement around the world.

4. Major problems in antitrust enforcement in developing countries The major problems developing countries face in antitrust enforcement are directly or indirectly related to the lack of adequate funding. In most of the developing countries that have adopted a competition law, the authorities that have been created are small and understaffed. In some countries, for example Tanzania, the World Bank has financed the functioning of the authority, but this is only for the short run and, once the funding stops, the authority risks returning to its original weak status. Recognising the weakness of domestic authorities, many international organisations, in an effort to minimise enforcement costs and to promote more effective enforcement, sponsor the introduction of regional cooperation agreements, even though the cases that are dealt with in these countries are

ommission of S C ingapore; Swedish Competition Authority; South African Competition Commission; Swiss Competition Commission; Taiwan Fair Trade Commission; the Office of Fair Trading (United Kingdom); United States Federal Trade Commission; and the United States Department of Justice—Antitrust Division. The vast majority of the countries represented are members of the OECD.

456 Alberto Heimler seldom instances of conduct with cross-border effects. As the experience with regional cooperation shows, regional supranational bodies are complementary to, or partially overlap with, domestic ones; they do not replace them. As a result, regional agreements very often create institutions that compete for funding with domestic ones, with the net result that neither is appropriately funded. Sometimes the funding of such supranational bodies is so low that staff cannot be hired, or basic functions of an antitrust authority cannot be performed. Moreover, in some cases, the hierarchy of legal systems is not clearly spelled out and the end result is endless jurisdictional conflicts. In any case, the record of regional bodies as antitrust enforcers is very discouraging.

4.1 Regional competition agreements in the developing world16 Such agreements are entered into by neighbouring countries, often in the context of a regional free trade agreement. A number of such agreements exist in the developing world, such as, among others, the Mercado Común del Sur (Mercosur), the Andean Community, the Common Market for Eastern and Southern Africa (COMESA), the Caribbean Community (CARICOM), the Southern African Customs Union (SACU), the West African Economic and Monetary Union (WAEMU) and the South African Development Community (SADC). Such agreements also exist, although there are fewer of them, in the developed world. The best known are the European Union treaties, which have a much broader agenda than free trade and aim at creating an internal market characterised by the free movement of goods, services, capital and people. On the other hand, in the developing world, the agreements have mostly been limited to free trade and, as a result, the competition provisions that they contain are not part of a wider array of policy instruments aimed at promoting a barrier-free international market. As a result, while the creation of a free trade area needs an institutional structure only for settling disputes among countries, the enforcement of competition provisions requires a wellfunctioning system that goes far beyond the simple existence of the rules—an apparatus that has been mostly lacking in these agreements, since the adoption of supranational antitrust provisions is not sufficient to influence the institutional structure and its funding. Furthermore, most of these agreements ignore the jurisdictional conflict that may arise between domestic and supranational competition provisions

16 This section of the chapter originates from a paper jointly written with Frédéric Jenny: ‘Regional Agreements’, in David Lewis, ed, Building New Competition Law Regimes: Selected Essays, Edward Elgar, 2013, chapter 7.

Effectiveness of Enforcement Cooperation in Developing Countries 457 and the institutional bodies that are enforcing them. In particular, regional agreements which exclude the jurisdiction of national competition authorities on competition cases affecting only their territory are unlikely to be as effective as agreements which contain a flexible system of case allocation and encourage cooperation in the enforcement of competition law by all of the institutions concerned. In Europe, the application of the principle of subsidiarity is not only very effective for a more thorough enforcement of competition law, it also reduces rivalry between institutions and encourages cooperation. Finally, regional agreements should make it clear to all concerned which law they are subject to, and which authority is in charge. Uncertainty about procedural issues is likely to create problems, increase the amount of litigation, and block the functioning of the agreement.

4.2 Free trade area agreements The creation of a free trade area may be the first step to establishing regional cooperation in antitrust enforcement. Two examples of cooperation agreements in the context of the creation of a free trade area will be analysed: Mercosur and the SACU. Both of these agreements contain provisions promoting increased cooperation among the competition authorities of member states. However, without the introduction of a system where confidential information can be freely exchanged and some convergence on the substance of the antitrust provisions, especially on mergers, this enhanced cooperation is clearly not effective. With respect to mergers, the cost associated with multiple filings may be substantial. First of all, parties would have to gather information on whether and when they would have to notify a merger. Then there would be costs associated with the different time frames and requirements of the decision-making process. While soft convergence (as achieved, for example, by Mercosur) may be more than sufficient to solve these informational problems, there has to be a system of incentives in place to promote them and to extend them to the remedy side. As for cartels, while (as we have seen) a few operate at the international level across countries and across continents, most operate nationally or regionally. Without some sort of regional agreement that would integrate the two jurisdictions (either directly or by the creation of a supranational body), it may be very difficult to ensure that domestic authorities cooperate in the investigation of cross-border violations. Domestic laws that protect the exchange of confidential information may make it very difficult for authorities operating in different jurisdictions to cooperate in a substantive way. The same would happen with the administration of leniency programmes and the setting of fines. A ‘federal’ system can help in overcoming these difficulties. But it is not the only option. Finally, regarding unilateral conduct, while a number

458 Alberto Heimler of violations are domestic in nature, there are instances where the same behaviour affects more than one jurisdiction. In these instances, a regional authority may be much more effective in identifying a common remedy than a succession of domestic proceedings. Here, too, there are other options—for example, the case could be conducted by the best placed authority; but it should have the power to sanction companies for violations that also produce effects outside of the domestic market.

4.2.1 Mercado Común del Sur Competition policy was introduced in Mercosur (which is an agreement between Argentina, Brazil, Paraguay and Uruguay, to which Venezuela became an additional party in July 2012) by Decision No 21/94 and strengthened by Decision No 18/96, the Protocol for the Defence of Competition (Fortaleza Protocol), which has addressed issues related to institutional design, substantive standards and increased cooperation among national competition authorities. The Fortaleza Protocol prohibits anticompetitive agreements and abuse of dominance. Proceedings against the violation of these provisions are initiated before the competition authority of each member state; and only if the practice affects trade within Mercosur can the authority submit the case to the Mercosur Committee for the Defence of Competition for a second draft decision, which must then be confirmed by the Mercosur Trade Commission, a body specialised in trade disputes. While a trade dispute body is not the best-suited institution for deciding on antitrust cases, it is worth noting that Argentina, Brazil and Venezuela, the only Mercosur countries that have domestic antitrust provisions, have very few incentives to refer a case to Mercosur, considering the very low (around zero) added value that the supranational body could bring to the investigation.

4.2.2 The Southern African Custom Union The same institutional weakness characterises the SACU, an international agreement between Botswana, Lesotho, Namibia, South Africa and Swaziland. Currently, South Africa and Namibia are the only SACU m ember states with a competition law. Like Mercosur, the SACU agreement also contains antitrust provisions. Article 40 requires all member states to have a competition law so as to be able to cooperate, but only two countries comply. Article 41 requires the SACU to develop, in an annex to the agreement, ‘policy and instruments to address unfair trade practices between member States’. Six years after the establishment of the SACU, Article 40 is not enforced, and the annex mentioned in Article 41 has still not been developed. As a result, regional competition issues have been completely ignored.

Effectiveness of Enforcement Cooperation in Developing Countries 459

4.3 Agreements aimed at the creation of an ‘internal market’ Other regional initiatives among developing countries are much more structured, and their aim is not simply the creation of a free trade area, but the building of an integrated common market. In terms of the effects of these initiatives on antitrust enforcement, however, experience shows that it is not sufficient for a federal authority to be created. A number of other elements must also be present. For example, sometimes the federal law does not cover merger control, the area most likely to benefit from a regional approach (CARICOM and Andean Community). In other circumstances, the jurisdictional uncertainty over which law to apply and the lack of funding at the federal level has reduced the incentives for applying the federal law (COMESA). Finally, the lack of a subsidiarity principle has eliminated the possibility of applying domestic laws, probably weakening overall antitrust enforcement (WAEMU).

4.3.1 The Andean Community In 1969, Bolivia, Chile, Colombia, Ecuador and Peru joined together in the Cartagena Agreement in order to ‘jointly improve their peoples’ standard of living through integration and economic and social cooperation’. On 30 October 1976, Chile withdrew, while Venezuela was a member from 13 February 1976 until 22 April 2006. The Andean countries created a free trade area in 1993, and in 1997 decided to expand it into a common market, the Andean Community. In 1991, competition rules were introduced into the Andean free trade agreement by Decision 285, which established common rules to address distortions of competition. However, the competition rules were strengthened at the time of the creation of the Andean Community. In 2005, by Decision 608, the Andean Community Commission made it clear that the Andean Community has jurisdiction over a matter when there are cross-border effects. However, merger control, the area where the benefits of cooperation are the greatest, is fully left to national authorities. Should a merger produce its effects on a geographic relevant market beyond that of a single member state, the difficulty of coordinating an appropriate remedy, for example a divestment action in a different member state, may lead to very ineffective decision-making. Another area in which there may be a problem is antitrust exemptions that are decided at the national level rather than at Community level (contrary to what happens in the EU). As a result, the creation of the Andean Community has only led to confusion, and consequently potential litigation. The difficulty the Andean Community Commission finds in claiming jurisdiction over antitrust matters is probably why the Andean Community has never adopted an antitrust decision in all these years.

460 Alberto Heimler

4.3.2 The Common Market for Eastern and Southern Africa The Common Market for Eastern and Southern Africa was conceived in the mid-1960s, but it was only in 1978, at a meeting of Ministers of Trade, Finance and Planning in Lusaka, that the creation of a regional economic community was recommended, beginning with a preferential trade area, which over a 10-year period evolved into a common market. Four years later, on 30 September 1982, the Treaty establishing the preferential trade area entered into force. Finally, the Treaty establishing the Common M arket for Eastern and Southern Africa, COMESA, was signed on 5 November 1993 in Kampala, Uganda, and was ratified in Lilongwe, Malawi, on 8 D ecember 1994. Nineteen countries are members of COMESA: Burundi, Comoros, Congo, Djibouti, Egypt, Eritrea, Ethiopia, Kenya, Libya, Madagascar, Malawi, Mauritius, Rwanda, Seychelles, Sudan, Swaziland, Uganda, Zambia and Zimbabwe, covering a very wide area and a total population of over 400 million people. Although COMESA was intended to become a common market, the provisions of the Treaty are not designed primarily for the creation of common rules but, rather, are devoted to the elimination of trade barriers, to the enforcement of antidumping rules and to the development of COMESA-wide investment projects. In 2006, a full Competition C ommission, composed of nine Commissioners from COMESA member states, was created. The Commission has been fully in place and operational since December 2009. The COMESA Competition Commission is responsible for enforcing the rules against abuse of dominance and cartel behaviour, and it also has some power with respect to merger control. Although the institutional structure of COMESA, and in particular the existence of a COMESA court of justice, could rapidly lead to an effective system of antitrust enforcement, as of this writing, there still have been no cases handled at the COMESA level. Indeed, jurisdictional issues are enormous, partly because a number of regional groupings in Africa have recently been created with overlapping boundaries and membership. In particular, the 16-member East and Southern Africa (ESA) association, created for the purpose of negotiating an economic partnership agreement with the European Union, has member states which belong to both COMESA and the SACU. As a result, the COMESA Competition Commission will have the preliminary task of going through the SACU, ESA and COMESA configurations when deciding on any case. The burden in terms of litigation could be enormous. Furthermore, the COMESA treaty did not guarantee adequate funding for the Competition Commission. As a result, although the Commission has been in place for almost four years, it has not made any decisions, probably also because it lacks funding. Indeed, on 25 April 2013 the COMESA Competition Commission introduced a system of merger control with a very high notification fee (US $500,000 or 0.5 per cent of the parties’ combined turnover or combined assets in the COMESA region, whichever is greater)

Effectiveness of Enforcement Cooperation in Developing Countries 461 and no turnover requirement for notification (just the presence of two of the parties of the transaction in two or more member states of COMESA)17—not a very efficient system to become a reputable institution. Indeed, the 1 July Communication by the COMESA Competition Commission reduced the filing fee to 0.01 per cent of the combined annual turnover or combined value of assets in the Common Market, whichever is higher—a much more reasonable fee. There is still no turnover threshold for notification.

4.3.3 The Caribbean Community CARICOM was established in 1973, transforming a pre-existing free trade agreement (CARIFTA) into a common market agreement where labour and capital would be free to circulate, and the agricultural, industrial and foreign policies of member states would be coordinated. In contrast to the European Union, which started as a political project (in the sense that the internal market was created in order to avoid wars and conflicts among EU Member States), CARICOM evolved from a free trade agreement with no visionary political objectives, just economic and administrative efficiency. As a result of this lack of a political project, CARICOM has been characterised by widespread implementation problems, and issues of sovereignty have always impeded the achievement of a much-desired increase in the degree of economic liberalisation of the region. The original Treaty of Chaguaramas was revised in 2001 and its new Chapter Eight created a Competition Commission responsible for enforcing the Treaty antitrust provisions on cross-border business practices. Although the Treaty requires each member state to ‘establish and maintain a national competition authority’, since CARICOM member countries are in some cases very small economies, they may not have the resources to establish their own domestic antitrust agency. Indeed, almost 10 years after the implementation of the revised treaty, only Jamaica, Guyana, Saint Vincent, Trinidad and Barbados have adopted a domestic competition law. This means that in the other jurisdictions (Antigua, Bahamas, Belize, Dominica, Grenada, Haiti, Montserrat, Saint Kitts, Santa Lucia and Suriname) only the CARICOM treaty provisions are in place. Unlike the European Union, where the interface between the Commission and national authorities has been developed through case law and practice, the CARICOM Treaty explicitly defines the relationship between the C ARICOM Commission and domestic authorities, creating ex lege a working network

17 See John Davies, Thomas Janssens and Chantal Lavoie, ‘COMESA’s Competition Commission Commences Operations’ (2013), http://www.mondaq.com/x/235910/Antitrust+ Competition/COMESAs+Competition+Commission+Commences+Operations.

462 Alberto Heimler among them. The operation of this network is rigidly defined, limiting the enforcement possibilities of the CARICOM Commission. Indeed, according to Article 176 of the Treaty, which provides for the possibility of ex officio investigation by the Commission, if there are reasons to believe that a firm is behaving in an anticompetitive manner, the Commission ‘shall request the national competition authority to undertake a preliminary investigation’. Only if it is dissatisfied with the outcome of its request can the Commission ‘initiate its own preliminary investigation’. Furthermore, if domestic authorities differ with the Commission as to whether a particular conduct is anticompetitive, the Commission must stop the investigation and refer the issue to the CARICOM Council for Trade and Economic Development. Because the Council is a political body, not a court of law, this may create opportunities for politically based decision-making that might prevent prosecution of politically sensitive cases. Furthermore, since this protection of member state interests is introduced only in the case of ex officio investigations, the independence of the Commission is reduced in the sense that the adopted procedure makes it difficult for the Commission to investigate the more sensitive cases. The treaty introduces a system of dispute settlement that is apparently only available for controversies among states. It is not clear whether, and under what circumstances, a firm can appeal a decision by the CARICOM Competition Commission. Certainly, if a company fails to adopt a remedy that the Commission has adopted, the Commission can take the firm to the Caribbean Court of Justice, where the firm can defend itself. In other circumstances, the rules are not clear. A final observation is that the CARICOM Commission was not given any power with respect to merger control. Especially in the case of the Caribbean Community, characterised by small island states, a regional merger control system would be particularly effective since remedies are much more easily identified in a regional setting than with a piecemeal approach state by state.

4.3.4 The West African Economic and Monetary Union With a treaty signed in Dakar on 10 January 1994, seven West African countries (Benin, Burkina Faso, Ivory Coast, Mali, Niger, Senegal and Togo) established the WAEMU. Guinea-Bissau joined on 2 May 1997. All the countries involved shared a common currency, the CFA franc. The 1994 Dakar treaty shows the clear intention of these countries to evolve from a monetary union to full economic integration where the four liberties (free movement of persons, goods, services and capital) are fully guaranteed. The treaty established a set of institutions very similar to those of the European Union: Conference of Head of States, Council of Ministers, Commission, Court of

Effectiveness of Enforcement Cooperation in Developing Countries 463 Justice, Parliament and Central Bank. Furthermore, the treaty established the principle of direct applicability of community legislation and the primacy of community law over domestic laws. Finally, the WAEMU Commission was made financially independent, and Articles 54 and 55 of the Treaty ensure that the Commission is properly funded. All these features are very important for ensuring the effectiveness of the WAEMU competition policy. However, one distinctive feature of the European approach that is missing in the WAEMU is the coexistence of domestic and community jurisdictions. In the European Union, the notion of effet utile has allowed national authorities to apply domestic law, under the constraint that their decisions (in situations where EU law could also be applied) had to be in line with the law and the case law of the EU. More recently, the principle of subsidiarity has been strengthened in the EU, with the Commission, domestic authorities and judges all applying the same European rules to practices capable of affecting trade between Member States. As a result, in the EU system, the same law is applied by different bodies across the whole common market, with very flexible rules for case allocation and no possibility of litigation over jurisdictional issues. By contrast, the WAEMU Court of Justice has ruled that only the Commission is in charge of applying community law, without exactly defining what ‘competition in the Union’ means and without identifying a role for domestic competition authorities. The territory of the WAEMU is extremely large, and leaving all antitrust enforcement to the WAEMU Commission may contribute to a weakening of enforcement; indeed, there are insufficient incentives for consumers, competitors or customers to report their complaints in strictly national cases to a distant authority. According to the WAEMU Court of Justice, the only role for a national authority is to cooperate with the WAEMU Commission. However, an institution has an incentive to act when it receives recognition for its actions. Very rarely will an institution act efficiently when it is only feeding into the decision-making process of another one. Indeed, domestic authorities have almost no role in the WAEMU Commission enforcement record. Adopting a principle of subsidiarity similar to the one adopted in the EU may help domestic competition authorities find an identifiable role in the procedure, leading to a much more effective system of antitrust enforcement. This does not mean that the WAEMU has been ineffective. On the contrary, the WAEMU Commission has taken up a number of important antitrust and state aid cases in a variety of sectors, such as cement, paper bags for cement, broadcasting, flour, pipelines, oil distribution, postal services, mobile telephony, audiovisual and maritime transport. The WAEMU has been by far the most successful of all the regional competition agreements in terms of enforcement. The point is that a better treatment of subsidiarity would enhance the role of domestic authorities, with a multiplying effect on the number and the relevance of cases.

464 Alberto Heimler

5. The role of the SADC: some reflections on the experience of three African countries (Zambia, Zimbabwe and Tanzania) There are a lot of overlaps in international cooperation agreements, especially in Africa. Just as an example, Zambia, Zimbabwe and Tanzania, three neighbouring countries in the South-East of Africa, are all members of a number of regional agreements. Tanzania is a member of the East African Community (EAC) and the SADC, while Zambia and Zimbabwe are both members of the COMESA and the SADC. As a result of this double or triple membership, there are overwhelming jurisdictional problems to be solved before any of these regional groupings may claim jurisdiction over a given case. Furthermore, the multiplication of cooperation efforts associated with the lack of funding creates a web of organisations, domestic and regional, all weak. The SADC, the regional organisation to which all three jurisdictions belong, is a 15-country organisation which aims to further the socio-economic cooperation and integration of its member states. Contrary to COMESA, which is an enforcement authority, the SADC has simply created a framework of cooperation among the 15 countries. In particular, SADC competition policy is governed by the Declaration on Regional Cooperation in Competition and Consumer Policies, which is aimed at facilitating investigations on anticompetitive practices that have cross-border effects. The declaration actively promotes cooperation among SADC member states, establishing comity principles among them and setting up a dedicated institutional structure, the Competition and Consumer Policy and Law Committee (CCOPOLC). The CCOPOLC was established in 2008. It is a forum where the competition authorities of the member countries exchange information and share enforcement experiences. It meets once a year, although not regularly. The SADC has a flexible structure and, since it does not have enforcement powers, is not very costly and does not need a large secretariat, could change its approach slightly: instead of simply promoting cooperation, as it now does (which may be premature at best, considering the low number of enforcement cases in its member countries, South Africa notwithstanding, and the even lower number of cases that have a cross-border effect), it could become a centre for the promotion of best practices on substance, procedures and institutional structure. SADC documents comparing the performance of the competition authorities of the organisation’s member states could be very useful domestically to promote more effective enforcement practices. More generally, the SADC may strengthen the role of the CCOPOLC and become an organisation providing the right environment for the authorities of member jurisdictions (or a subset of them) to elaborate solutions to substantive and procedural problems of common interest, by organising meetings of head of agencies leading up to the issuing of recommended practices,

Effectiveness of Enforcement Cooperation in Developing Countries 465 along the lines of the OECD Competition Committee. In this way, the SADC could be capable of promoting best practice solutions for these countries, both on the substantive and the procedural aspects of antitrust law enforcement. By excluding enforcement, the funding of the SADC would be a much less relevant issue, and its influence could be far reaching. The same developments could be promoted in a number of regional agreements that, for example, could become a focal point for the discussion of antitrust enforcement best practices at the regional level. Creating enforcers that do not enforce is too costly to maintain and undermines the efforts we all made to promote respect for antitrust provisions worldwide. On the other hand, developing best practice standards that would be dedicated to the solutions of common problems of a given area could well be an attainable objective for existing regional cooperation organisations.

6. What kind of cooperation for cross-border cases? When the original EU merger control regime (Regulation 4064/89) was being revised back in 1997, the European Commission argued that international cooperation could be effective only up to the point where three jurisdictions were involved. Three jurisdictions could easily cooperate among themselves without any need for a supranational body. The same approach was taken a few years later, when Regulation 1/2003 entered into force.18 The Commission would be best placed when four or more jurisdictions were affected by a given practice and where each one would be actively involved in pursuing the same case. In such circumstances, transaction costs and duplicative efforts would be too great; the Commission would thus be best placed to handle such a case. The number three did not originate from an analysis of transaction costs or from statistical studies showing the relevance of cross-border cases in the EU with differing numbers of jurisdictions being affected. Three was simply considered to be higher than two and smaller than four! The interesting aspect of the debate that took place at the time was that, even in a very structured regional cooperation agreement like the then-EC Treaty, some room was left for bilateral cooperation by domestic authorities. Such a bilateral approach could also be adopted by developing countries, where regional authorities seldom exist and where those that do exist are weak, understaffed and underfunded. In such an environment, bilateral and even multilateral agreements, short of creating supranational institutions, could be much more effective and much less costly to sustain.

18 See Commission Notice on cooperation within the Network of Competition Authorities, [2004] OJ C101/43.

466 Alberto Heimler We should indeed distinguish between the need for cooperation and the objective of promoting convergence in enforcement. Convergence could well be achieved by existing regional authorities or promoted by international bodies like the OECD, UNCTAD or the ICN.19 Cooperation could be achieved without the need to create a regional organisational structure. With a best practice application of bilateral agreements, Australia and New Zealand agreed in 1983 to sign a trade agreement promoting closer economic relations between the two countries. ANZCERTA, as the agreement is named, led to a sevenfold increase in the trade flow between the two countries.20 In 2010, under the auspices of ANZCERTA, and in particular within the Single Economic Market Outcomes Framework launched in 2009, the two countries agreed that each jurisdiction would nominate an associate commissioner to join the board of the competition authority of the other jurisdiction. These associate commissioners are involved in the process of examining cross-border mergers, so that basic concepts in the analysis can be articulated in a coherent way and, more importantly, so that remedies can be identified in an environment of perfect information with regard to the worries the merger would raise in the other jurisdiction. Such a deep cooperation is very effective since most of the work could be carried out from the respective capitals, as it is sufficient for documents, not people, to travel. Furthermore, with this type of framework, a costly organisational structure is not required to achieve effective cooperation. The example of Australia and New Zealand shows how the benefits of international cooperation could be achieved without a supranational institution, and in a very cost-effective way. Developing countries, where the issue of funding is much more severe, but where genuine supranational issues are much less frequent, would greatly benefit from such a flexible approach, while international organisations like the OECD, UNCTAD or the ICN can maintain their functions of promoting convergence and best practices on substantive issues.

19 The OECD Council Recommendation concerning International Co-operation on Competition Investigations and Proceedings (16 September 2004) is an exception, promoting both convergence and cooperation. First of all, the Recommendation suggests that member countries should make all possible efforts to promote cooperation in the antitrust field by taking the processes of rule-making in other jurisdictions into account. Furthermore, the Recommendation suggests that competition authorities should exchange information as much as possible, considering that statutory limitations on the exchange of confidential information typically prevent the authorities from exchanging a very narrow subset of all available information they have in their files. 20 See the information page of the Australian government, http://www.dfat.gov.au/fta/ anzcerta/anzcerta_bilateral_role.html.

Effectiveness of Enforcement Cooperation in Developing Countries 467

7. Conclusion In the last decade, the number of regional agreements on competition among developing countries has increased greatly, notably linked with regional efforts to set up free trade zones or common markets. The success of these regional agreements has been quite limited, but useful lessons can be learned from the experiences gained so far. In particular, WAEMU is the regional cooperation agreement on competition in the developing world that is closest to the European system and is also the only one that has been able to achieve some success in terms of actual case law and enforcement experience. The WAEMU Commission was able to enforce the law because the WAEMU Treaty eliminated the possibility for domestic authorities to apply the law (domestic and regional) themselves, a move that is not very efficient by itself, since these domestic authorities continue to exist and the end result may have been underenforcement by the WAEMU, given the size of the Union. Indeed, for a system of competition enforcement to function properly: (i) complainants have to be geographically close to the institution in charge, which implies that the relationship between the centre and the periphery needs to be governed in such a way that the incentive of domestic authorities to act is not weakened; and (ii) clarity has to be provided as to who is in charge of an investigation—either the regional or the domestic authority—otherwise, the whole system could collapse because of litigation over whether the legislation to be applied to a specific case should be that of the Community or that of domestic jurisdictions. On both of these aspects, regional agreements have not been very effective (including the WAEMU) partly because of a lack of proper structural funding. The very high level of the merger notification fee COMESA introduced originally is a clear indication of the financial problems faced by supranational organisations. Raising the merger notification fee to a very high level is not the answer to the lack of funding, because many firms would have decided not to notify, and with limited resources it would have been highly unlikely that COMESA could credibly threaten to sanction the non-notifying firms. The fee that was later introduced is much more reasonable. However, since all mergers have to be notified (irrespective of the importance of the acquired firm), the fee may indeed continue to be relatively very high for low-value acquisitions. Regional bodies are not necessary for enforcing the law in the very few cross-border cases that would benefit from a single centre of enforcement. Bilateral agreements can be very effective in this respect, as the Australia–New Zealand experience shows. By making it possible for agencies to coordinate on cross-border merger cases in a substantive way, the problems associated with cross-border practices can be more easily addressed.

468 Alberto Heimler What is important (for both domestic and cross-border decisions) is convergence on what is prohibited and what is allowed. In this respect, existing regional institutions could be reorganised to provide best practice recommendations to member countries and could serve as a forum for discussion where issues of common interest could be analysed in a cooperative way. Furthermore, existing international organisations, like the OECD, UNCTAD and the ICN, can be very helpful in developing standards on best practices to be adopted worldwide. The ICN, the organisation dedicated to antitrust enforcement, is particularly well placed in this respect, but extra effort should be made to involve developing countries more in the elaboration of ICN work products. The active members of the ICN are mostly OECD countries. Furthermore, the ICN should abandon the encyclopaedic approach which it has increasingly adopted in recent years, and which is practically useless as a tool for achieving convergence. If there are many options, how can I know which one is best for me? In many areas, recommendations are difficult to write, but the ICN should not give up the challenge; it should provide antitrust authorities with clear indications of what is best to do. As for promoting the adoption of its recommendations by members, the ICN should start a process of country reviews. Each jurisdiction could perform a self-review with respect to adherence to the ICN recommendations, both on substance and on procedures, giving reasons for non-compliance. The exercise would give relevance to the work of the ICN and at the same time engage member countries, both in the analysis of the ICN recommendations and in the process of adapting them to their specific circumstances. Every year, a few jurisdictions could be subject to a deeper review.

Antonio Capobianco, John Davies and Sean Ennis* The Need for International Cooperation in Merger Enforcement

1. Introduction More jurisdictions than ever before are applying competition law.1 Although substantive elements of these laws vary to some extent, and procedures vary even more, it is remarkable how similar these different regimes are. Almost all have a fairly strong prohibition against cartels, some form of assessment of mergers based primarily or exclusively on their effects on competition and (although this third area varies the most) something enabling action to be taken against firms with market power behaving anticompetitively. It is also worth bearing in mind that this is all very new. A prediction in the 1980s that by 2010 the People’s Republic of China would have antitrust laws—enforceable through private actions in courts—would have seemed quite bizarre. This explosion of competition laws and agencies is a positive development overall: whether for the individual jurisdictions newly applying competition law, for businesses based in longer-established competition jurisdictions or for the world economy as a whole. However, the additional complexity that this causes creates a number of challenges for the effective and consistent enforcement of competition law. In this chapter we will explore these, particularly the difficulties created when multiple authorities investigate the same international or global merger. Discussion of the problems this causes often focuses on the duplication of administrative costs.2 Our focus is on the costs of disagreement when d ifferent * The authors work in the Competition Division of the Organisation for Economic Co-operation and Development (OECD) in Paris. The views expressed in this chapter are those of the authors alone and do not purport to represent those of the OECD, its Competition Committee or any member countries of the OECD. We are grateful to Eleanor Fox and Stephen Davies for helpful comments on an earlier draft of this chapter, and to the organisers and participants at the EUI Competition Workshop in July 2013, especially Philip Lowe, Mel Marquis and Stephen Calkins. 1 The International Competition Network (ICN) now has about 130 competition agencies as members, although some cover the same jurisdiction. 2 See, eg ICN, ‘Report on the Costs and Burdens of Multijurisdictional Merger Review’ (2004); PricewaterhouseCoopers, ‘A Tax on Mergers? Surveying the Costs to Business of Multi-jurisdictional Merger Review’ (June 2003), commissioned by the International Bar Association and American Bar Association.

470 Antonio Capobianco, John Davies and Sean Ennis jurisdictions reach different conclusions on essentially the same matter, taking mergers as the example for this chapter. Throughout, we imagine a merger between ‘global’ companies: two firms— whether headquartered in the same country or not—that sell into several jurisdictions, overlapping in at least two. Under the widely applied ‘effects doctrine’, such a merger would come under scrutiny in multiple jurisdictions. The authorities in those jurisdictions might consult one another to some degree3 (depending, for example, on the provisions of their laws, particularly as regards exchange of confidential information), both on the substantive question of whether there is likely to be a loss of competition and on remedies. However, even if they consult one another, in general, competition authorities will take their decisions independently and with regard to the interests of their home jurisdiction (usually, though not always, identified as the interest of consumers in that jurisdiction) because of the absence of a legal basis for comity between many jurisdictions. In short, regulatory approval for an international merger will be decided through a multiplicity of independent national decisions.4 This chapter rather informally discusses some of the effects of this situation.

2. The scope for disagreement on international mergers We want to consider cases in which authorities in two or more different jurisdictions reach different views on a global merger. This could happen for at least three reasons: 1. the authorities in the two different jurisdictions are applying different rules of substantive analysis in assessing the merger, either because their laws require them to do so or because of different ‘understandings’ of how competition works; 2. conditions of competition are materially different in the two jurisdictions; or 3. the two authorities and any other decision makers have simply come to different views, perhaps on a borderline case or perhaps because the evidence collected and its interpretation were different.

3 The OECD and ICN have recently conducted a joint survey on the scale and effectiveness of international cooperation—see http://www.oecd.org/competition/oecd-icninternational-cooperation-survey.htm. 4 By ‘national’ we mean pertaining to a single jurisdiction. This could include multinational groupings that operate as a single jurisdiction such as, most obviously, the European Union.

The Need for International Cooperation in Merger Enforcement 471

2.1 Reasons to disagree 1: different substantive rules Competition specialists have discussed the problems caused by differing substantive rules at least since the GE-Honeywell case,5 because the different approaches taken to the assessment of a merger between suppliers of complementary products in that case was seen as being an important reason for the conflicting decisions taken by the US and EU authorities. In a few cases, substantive differences might arise from the law itself. For example, differing values might be embodied in legislation—such as over employment concerns, or protection of small sellers against buyer power— and can or must be taken into account by the competition authorities, and this can cause a divergence between agencies’ decisions. However, many—perhaps most—conflicting decisions under this category probably stem from less well-defined differences, simply reflecting different precedents and practices despite similar legal standards. One agency might pay more attention to market shares than another, or be more concerned about vertical linkages. A recent study6 comparing US and EU approaches to merger control, for example, found that the EU is ‘tougher’ overall, but the US is ‘tougher’ on coordinated effects cases. We would include in such substantive differences, differences in legal and procedural matters, such as where the burden of proof lies (for example, regarding efficiencies). Some of the discussion of international cooperation seems to imply that convergence on such substantive standards—whether written or unwritten— would essentially solve the coordination problem resulting from different decisions. Convergence is certainly important, and a lot of effort is devoted to it (not least by the OECD). However, even complete substantive convergence would not solve the problem, because there are at least two more reasons for competition authorities in different jurisdictions to reach conflicting opinions on a merger affecting both jurisdictions.

2.2 Reasons to disagree 2: differing conditions of competition It is obviously right for competition authorities to reach different conclusions when the merger’s effect is different in their respective jurisdictions. Mergers 5 This very well known case has been discussed in many places, but see, eg Fox’s account in Eleanor Fox and Daniel Crane, eds, Antitrust Stories, Foundation Press, 2007. Note that the General Court overturned the finding relating to complementary products, but upheld the European Commission’s decision on other grounds. 6 Mats Bergman, Malcolm Coate, Maria Jakobsson and Shawn Ulrick, ‘Comparing Merger Policies in the European Union and the United States’, 36 Review of Industrial Organization 305 (2010).

472 Antonio Capobianco, John Davies and Sean Ennis do not have the same effect everywhere, so the assessment of their effects will not be the same everywhere. The number of third-party suppliers may be different, for example. Very commonly, the difference will arise from different market sizes. A merger that is a ‘five to four’ in a large jurisdiction could be a ‘two to one’ in a smaller jurisdiction, for example, but have no effect whatever in another smaller jurisdiction with a different ‘two’ out of the ‘five’. Regulation can also result in products that are substitutes in one jurisdiction not being able to act as substitutes in another. Customer behaviour may differ, again resulting in agencies quite legitimately reaching different views either on market definition7 or on the competitive effects of the merger. This is a natural—and obviously legitimate—reason to disagree. However, that does not mean that this situation is unproblematic. Even in this case, national enforcement could reduce global welfare if the remedies adopted appropriately in one jurisdiction to remedy the merger’s adverse effects also have consequences for another jurisdiction where there are no adverse effects. This will be discussed in the latter half of this chapter.

2.3 Reasons to disagree 3: because people do Finally, competition authorities might simply disagree on the effects of a particular merger, even if the substantive test is the same and the conditions of competition are the same in both jurisdictions. Why would this ever happen? The simplest answer is: because assessment of competitive effects is hard. Reasonable people often differ on the likely effects of particular mergers, and all practitioners will be familiar with borderline cases, in which the evidence seems very evenly balanced. In some jurisdictions there is even empirical evidence of this, as decision-makers can record dissenting or minority opinions.8 It is quite likely that if one five-member decision-making body splits 3–2 on a decision, an equivalent body looking at the same case in another jurisdiction could split 3–2 the other way. If anything, the prevalence of international disagreements seems to be lower than the prevalence of disagreements between decision-makers within

7 The UK and French authorities reached different views on a merger between salmon producers, in part because French consumers regard Scottish and Norwegian salmon as being rather different, while British (or at least English) consumers do not—so the market is narrower in France. There may be no fundamental reason for this, but it seems to be true. See Pan Fish ASA/Marine Harvest NV merger inquiry, UK Competition Commission, 2006. 8 And the ‘true’ level of disagreement is surely still greater, as most people prefer to agree than to disagree, so dissenting or minority opinions will normally represent a rather significant disagreement.

The Need for International Cooperation in Merger Enforcement 473 an agency: agencies, so to speak, seem to disagree with one another less often then they have internal disagreements. If true, this perhaps reflects a desire to avoid international disagreements, particularly when the case seems finely balanced.

3. The costs of international disagreements Disagreements will occur, for any of the three reasons outlined earlier. The authority in jurisdiction A opposes a merger and the authority in jurisdiction B approves it. Whether this matters or not depends on whether there are any efficiencies from the merger and whether efficient remedies can be implemented purely nationally. If they can, the effects of the merger can be remedied in jurisdiction A, while allowing the merger in jurisdiction B. For example, if the merging firms’ business activities are essentially national even though their parent companies are merging globally, a divestment that effectively blocks the merger in jurisdiction A might be possible, while allowing it in effect to proceed in jurisdiction B (and the rest of the world). However, in many cases the most effective remedy available in jurisdiction A will have effects on jurisdiction B. This will be true of most structural remedies, particularly those affecting the ‘upstream’ level of trade. Quite a common model for global businesses is for production to be concentrated in a few locations, with sales across the world. These do not easily allow a national structural solution (different local marketing companies will not solve the fundamental identity of interests of the merged producers). This will particularly be true, of course, if the most effective ‘remedy’ is to block the merger completely. This is a classic externality, in economic language. One agent—the competition authority in one jurisdiction—takes a decision that will affect the citizens of another jurisdiction, but does not take those citizens’ interests into account. Consequently, the decision will not maximise the welfare of both sets of citizens, taken jointly. If the merger results in general benefits (through efficiencies) but imposes costs in one jurisdiction, then blocking the merger denies the benefits to other jurisdictions. In practice, when the most effective remedy would impose such an externality, the authority in jurisdiction A could instead choose a less effective ‘local’ remedy. It might do so out of concern that blocking a global merger because of purely local concerns would be disproportionate,9 or because any such prohibition (or significant upstream divestment) could not be enforced because

9 Whether it would have a duty to do so or, on the contrary, be forbidden to do so will depend on the precise wording of the legislation under which it operates.

474 Antonio Capobianco, John Davies and Sean Ennis the assets are not within the authority’s jurisdiction. We are not, of course, seeking to imply that behavioural remedies are necessarily less efficient, or that structural remedies can never be imposed at a purely national level. Both sorts of remedies will often be chosen freely. The point is that sometimes smaller jurisdictions will be constrained to choose them. For example, the Austrian delegation at an OECD roundtable on remedies is quoted in the summary of discussion10 as noting that While in some cases behavioural remedies may have simply been more appropriate, in others, they may have been a necessity since, due to Austria’s relatively small size, the assets that would have been subject to a structural remedy were located outside the jurisdiction.

How large does a jurisdiction have to be to impose a structural remedy on an international merger? The UK Competition Commission (CC) imposed behavioural remedies on a merger between producers of medical e quipment,11 both of whom manufactured products only outside the UK and then imported them. In its report, and in a subsequent retrospective analysis,12 the CC noted two practical constraints: one on its ability to enforce a structural remedy overseas and the other reflecting a concern that the merged entity could simply withdraw from the UK. As the UK is the eighth largest economy in the world, it is clear that most authorities will face some practical constraints on their ability (and willingness) to block global mergers. Only a handful of very large jurisdictions can impose remedies and block global mergers for reasons relating solely to their own jurisdictions: the USA, the EU, and perhaps increasingly the major emerging economies—China, India and so on.13 This is a matter of bargaining power against companies that might threaten to withdraw from markets, rather than legal power. The UK was reluctant to require divestment by Dräger (a German company) in 2004, whereas China was prepared to require a divestment in Peru by Xstrata (a Swiss company) in 2013. Competition authorities in the smallest jurisdictions are well aware that they cannot effectively block international mergers at all. What matters is the size of the market in the jurisdiction seeking to block the merger, compared to the total size of the markets for the merging firms. Mergers with a smaller geographic scope could be blocked by the authorities of smaller jurisdictions. For example, in 2013 the UK Competition Commission required a divestiture remedy effectively blocking a merger between

OECD Best Practice in Competition Policy, ‘Remedies in Merger Cases’ (2011), 291. Dräger Medical AG/Air-Shields Hillenbrand Industries Inc merger inquiry, UK Competition Commission, 2004. 12 Competition Commission, ‘Understanding Past Merger Remedies: Report on Case Study Research’ (2012). 13 Of course, even medium-sized countries will be important markets for some specific products or regionally. 10 11

The Need for International Cooperation in Merger Enforcement 475 Eurotunnel and SeaFrance,14 which the Autorité de la Concurrence had previously approved with some behavioural commitments.15 As Eurotunnel by its physical nature must sell into the UK market, the UK Commission’s decision prevailed.

4. Does it matter? In one way, this is not a bad outcome. It may be better for global economic welfare that only the largest states can block global mergers for national considerations, while the smaller cannot (but are able to block those for which their own jurisdictions represent a large share of the market). However, no economy—however large—constitutes much more than 20 per cent16 of world GDP, so any national decision with global consequences represents a large potential externality, affecting 80 per cent of the world economy. Secondly, the decisions to block or to clear are not symmetrical. A large state can impose its decision to block a merger on the world; it cannot impose a decision to clear it. Consequently, the more states there are that can block mergers (because they have a competition law and are large), the more mergers will be blocked. A truly global merger will be contingent on five, six or more states independently reaching a decision not to challenge it. The effect is to chill merger activity, for two reasons: —— the most interventionist standard will prevail; and —— even if all apply the same standards, requiring multiple independent clearances makes overall clearance less likely.

4.1 Chilling effect 1: the strictest standard will prevail It is fairly obvious that if competition authorities independently impose remedies with global implications, the most interventionist standard will be the one that prevails. In 2001, the US authorities cleared the GE/Honeywell merger without remedies while the EU required remedies that caused the deal to be abandoned. The literature relating to the merger is vast, but most commentators agree that the EU was applying a tougher standard to a merger

Eurotunnel/SeaFrance merger inquiry, UK Competition Commission, 2013. Autorité de la Concurrence, Decision 12-DCC-154 of 7 November 2012. 16 OECD Economic Outlook 2013 reports USA GDP at $13.8 trillion, 22% of World GDP at $62.3 trillion. The EU is the second largest economy; China is the third. 14 15

476 Antonio Capobianco, John Davies and Sean Ennis between producers of complementary products than the US applied. The merger did not go ahead, so the EU’s standard prevailed in this case. The same applies to globally effective remedies that fall short of prohibition. For example, Glencore and Xstrata are Switzerland-based mining companies with production at sites in various countries and global sales. The US Department of Justice cleared their merger unconditionally in 2012.17 However, in 2013, China’s MOFCOM required divestment of the Las Bambas copper mine, now under development in Peru18 (in addition to some behavioural remedies). It is not clear whether the conditions of competition differ materially between these jurisdictions. However, it appears that the merged entity’s market share of copper sales in China would be less than 18 per cent—below the level at which most competition authorities would consider intervening. Copper is a global market, and it is hard to escape the conclusion that the US Department of Justice (DOJ) and China’s MOFCOM considered essentially the same facts and reached different conclusions, and that the stricter standard—MOFCOM’s—has prevailed. For completeness, we also note that the European Commission imposed a structural remedy relating to the market for zinc.19 It is less clear whether this represented a specifically EEA-wide problem or, again, a different approach to a similar issue. We do not seek to imply that MOFCOM was wrong and the DOJ right. The point is that the US (presumably) considered whether the merged e ntity’s position in the copper market created competition concerns that might require the sale of an asset like Las Bambas, and (presumably) concluded that it did not. Whether or not that analysis was right, it was rendered irrelevant by MOFCOM’s decision.

4.2 Chilling effect 2: it is hard to get clearances everywhere Even if authorities apply the same standards, independent investigation in multiple jurisdictions reduces the overall likelihood of mergers being approved, and therefore raises the bar, cutting off many efficiency-promoting mergers that would otherwise be proposed. Fairly obviously, if a merger requires unanimous approval by authorities in all large jurisdictions, then the more such authorities there are, the less likely it is that the merger will

17 See, eg ‘The Long March: The Final Hurdle is Cleared in Glencore’s Takeover of Xstrata as Chinese Regulators Give Merger the Go Ahead’, Eversheds, 17 March 2013. 18 See, eg ‘China Clears Glencore’s Acquisition of Xstrata Subject to Remedies’, WilmerHale Publications and News, 26 April 2013. 19 Commission Decision of 22 November 2012 in CASE COMP/M.6541—Glencore International plc / Xstrata plc.

The Need for International Cooperation in Merger Enforcement 477 proceed—and the less likely it is that the merger would be attempted in the first place. This can be illustrated with a simple (oversimplified and artificial) example. Suppose that mergers have the same effects everywhere and all authorities apply the same standard: namely, that they will approve mergers that are likely to raise welfare and block those that are not. There are many mergers for which the facts and complexity of analysis make decisions rather uncertain: they are neither obviously harmful nor obviously harmless. These ‘borderline’ mergers can be considered to have a probabilistic chance of approval by competition authorities. For example, a highly problematic merger might be considered to have a 20 per cent chance of success, as evaluated before the merger is proposed.20 Suppose that businesses will propose a merger only if its chances of approval are 50 per cent or better. Finally, suppose that each authority’s assessment is independent of the others. In those circumstances, if only a single authority makes the decision (as in a purely national case), businesses will propose mergers if the ‘true’ probability that they are welfare enhancing is 50 per cent or greater. But for an international merger, more authorities will need to approve it independently. Suppose there are five important jurisdictions that must approve a ‘global’ merger for it to go ahead. Then the probability that a perfectly ‘borderline’ case with a 50 per cent chance of being welfare enhancing will obtain all five approvals is 50 per cent5 = 3 per cent. So only 3 per cent of such global mergers would be able to proceed, even though 50 per cent of them would be welfare enhancing. To have a 50 per cent chance of clearance by all five authorities, a proposed merger would need to have a probability of 87 per cent of being approved by any one authority (because 0.875 = 0.5). So, if firms will only propose mergers that have a 50 per cent chance of being approved globally, then only those mergers which are relatively safe bets—with an 87 per cent chance or more of being regarded by any one competition authority as enhancing welfare—will be proposed. Global mergers with a probability of enhancing welfare of between 50 and 87 per cent will not be proposed, because the chances of getting five approvals is less than half, even though each individual competition authority is more likely than not to approve the merger. Such mergers are (by definition) welfare enhancing, and are regarded as such by the majority of competition authorities assessing them; but they are not proposed, so welfare will be reduced in relative terms.

20 These probabilities could be thought of, for example, as the expert advisors’ opinions on the likelihood of successful clearance to the potential merging partners’ boards in advance of the announcement. Of course, merger decisions are not really probabilistic, but it is quite common in advance of those decisions for people to express likelihoods of success in such terms.

478 Antonio Capobianco, John Davies and Sean Ennis The multiplicity of authorities, each with the ability to veto a global merger, has much the same effect as a decision significantly to tighten the standards for merger approval. Yet while practitioners have debated and refined the appropriate standard for intervention for decades, some being concerned about possible ‘chilling’ effects of over-enforcement, there has been very l ittle discussion of how the need for multiple approvals might have an identical chilling effect on the very largest, global mergers.21 This has simply happened, and is still happening, without much awareness that this development might make the minutiae of domestic debates about precisely where to site the standard for merger control somewhat irrelevant, at least for the very largest mergers (which presumably matter the most).

4.3 A gap in governance? Even if the outcomes resulting from the current fragmented global decisionmaking were judged at present not to be too bad, there is (in the authors’ view) something rather unsatisfactory about the process. People who work in competition enforcement tend to believe that competition rules should be based on reasonably objective criteria (so that intervention takes place only when it is likely to improve welfare, or improve something anyway) and be applied in a consistent, fair and transparent way. It is hard to see that these criteria are met in the way global mergers are decided. Different authorities impose different remedies; some are effectively able to block mergers while many are powerless, regardless of the merger’s effects. If a national system exhibited this sort of chaotic process of decision-making, it would be unlikely to pass a peer review at the OECD. This is not a good way to decide important matters that can affect the global economy. There is a gap in governance. Earlier, we mentioned the 2013 disagreement between the British and French competition authorities over the Eurotunnel merger. Following those decisions, France’s transport minister, Frédéric Cuvillier, announced that he would seek a meeting with his British counterpart to ‘arbitrate between the decisions of the two competition authorities’.22

21 Of course, others have been concerned about underenforcement, and might therefore be tempted to argue that this chilling effect is a good thing. No doubt there are some mergers that should not go ahead on competition grounds (or are simply inefficient and welfarereducing) but are approved. However, there is no reason to suppose that the mergers that are blocked because of the chilling effect arising from requiring multiple approvals will be those that are the least beneficial. Our point is that, in practice, international firms face a tougher merger regime than purely domestic ones, which seems completely arbitrary. 22 Reported in Global Competition Review, 11 June 2013.

The Need for International Cooperation in Merger Enforcement 479 As a spokesman for the UK Competition Commission pointed out,23 even if the ministers agreed at such a meeting, it would not change the decision, which is for the Commission alone to take. This is all to the good—politicians should not decide merger cases. But the point is not just that the ministers had no legal basis to meet and cut a deal— no one does. If, instead, decision-makers from the British and French competition authorities had sought to avoid reaching conflicting decisions, it does not appear to us that they would have had any legal basis for ‘cutting a deal’ to resolve the inconsistency.24 If a competition authority changed its decision explicitly to achieve consistency with the decision of an overseas authority, it could well face an appeal if any party with standing objected. The effect is therefore that the slightly arbitrary combination of effects of independent decision-making in different jurisdictions stands as the overall decision on a global merger, with no formal mechanism for negotiation or arbitration to resolve inconsistencies. This seems unsatisfactory.

5. Will the situation get better or worse? There are some trends that are likely to result in the problems outlined above becoming less severe in future years. In particular, it is reasonable to expect continued convergence of substantive and legal standards, continuing the successful convergence between longer established agencies that has been seen in the last 10–15 years. However, there are also reasons to believe the problem might become worse (in the absence of better methods of coordination between jurisdictions). Firstly, the newer competition authorities, such as those of India and China, are likely to become more willing to impose remedies with global consequences. For example, China’s MOFCOM has shown great willingness to impose behavioural remedies in its first years. This might reflect a desire to find remedies that do not have global consequences, or might reflect MOFCOM’s

23 Ibid. ‘The UK has an independent competition regime designed to exclude government involvement in decision-making—and we have a duty to take the necessary action to protect competition and the interests of customers. In that context, it’s difficult to see what the role or purpose of such a meeting would b.’ (UK CC spokesman, quoted by Global Competition Review). 24 They might be able to convince one another that the facts or the analysis in the other’s decision were wrong, of course, and might thus agree on the case after all. Competition authorities are understandably keen to avoid reaching conflicting opinions, and communication between them on factual and analytical matters can help. However, that is a different matter from ‘cutting a deal’ when each has decided (before or after publishing that decision) where its analysis of the case has led it.

480 Antonio Capobianco, John Davies and Sean Ennis preference for less invasive behavioural solutions over more invasive structural ones. Assuming that MOFCOM experiences the same difficulties with behavioural remedies as longer-established authorities have encountered, MOFCOM’s practice in this respect is likely to change. Secondly, as more authorities take an active enforcement role around the globe, the lack of effective coordination tools increases the risk that agencies will develop different substantive rules in their enforcement practice over time.25 Enforcement coordination is not simply a tool to reduce inconsistencies in enforcement decisions—a worthwhile goal in itself—but also allows agencies to share views, experiences and legal and economic cultures on specific cases, which in turn will further convergence on substantive and procedural standards as well. Thirdly, the process of economic globalisation is continuing. The degree to which the world economy is globalised is often overstated. There are few truly global businesses, and more economic activity is local—particularly national—than might be supposed from reading the business books on sale at airport kiosks. This means that the transformation of the world economy that we have seen so far is only beginning. There will be more global activity, so there will be more global mergers and perhaps other sorts of competition cases too. Even if all jurisdictions apply exactly the same system, multiple independent decisions are likely to result in significantly greater externalities being imposed on the majority of the global economy than we see today. We are presently carrying out some research on this, relating the need for international decision-making to projections of global merger deals and other indicators of increased global economic integration. The OECD Competition Committee is discussing possible policy solutions that would help address the increasing need for better enforcement cooperation between competition enforcers.

See the first reason for disagreement mentioned above, ‘Different substantive rules’.

25

PART IV Issues for Courts and Perspectives on the Judicial Role Chair: Bill Kovacic Speakers: Gerald Barling Stephen Calkins José Luis Da Cruz Vilaça Ian Forrester Damien Gerard Alberto Heimler Wolfgang Kirchhoff Hwang Lee

Ioannis Lianos Frédéric Louis Renato Nazzini Jacqueline Riffault-Silk Savvas Papasavvas Mario Siragusa James Venit

Bill Kovacic: Good morning, everyone. Today we turn to the perspectives of members of the judiciary. For all the variations that you see across systems internationally, you can state the proposition very firmly that no competition system became proficient over time without enormous contributions from the courts. To say it another way, if you don’t have an effective judicial system, you don’t have an effective competition system, and that is a great challenge for many of the newest authorities around the world today. That critical role is manifest in a number of different activities. In some instances, it is the interpretation of statutes and secondary regulation. It is the interpretation and application of doctrine, and part of the role of courts is that they make a crucial contribution to quality control. By contrast, consider China as an example. China’s Antimonopoly Law is five years old, and the merger control regime assigned to MOFCOM is not subject to any judicial review. There is no administrative court, there is no supreme court review; there is no court review at all. The decisions of MOFCOM are the decisions of China with respect to merger policy. Imagine how merger policy would have evolved in Europe, in the US, and in any number of jurisdictions had there been no judicial review of administrative decision-making. So although it is commonly said at various conferences that there is no ‘one size fits all’, I think one principle does fit all: if you don’t have good courts, and if you don’t have an effective judicial system, then you do not have an effective and legitimate competition policy regime over time. We have several contributions from our judges today. Our format is much like that of yesterday: we have DD

482 Part IV: Issues for Courts and Perspectives on the Judicial Role 10-minute presentations from our four judges, followed by 10-minute presentations from our three discussants. Then we’ll turn to a discussion of the group as a whole, and convene after Philip’s closing remarks a little bit before lunch. And Gerald, if you could get us started. Gerald Barling: First of all, allow me to thank Mel and Philip for all the preparatory work they’ve done. It is a huge pleasure to be here again. I would also say—and I hope this is not a part of my 10 minutes—congratulationsto Bill and to Philip, both of whom have been appointed to the board of the new Competition and Markets Authority in the UK, which replaces the OFT and the Competition Commission. One of the great benefits of the UK system is that we allow people to be trained elsewhere for years until they are at their peak ability and expertise, and then we snatch them and obtain the benefit of all that wonderful wisdom and learning. That’s exactly what’s happened with Bill and Philip. In my paper I’ve set out various issues and potential developments confronting us in the United Kingdom at the moment, both in public and private enforcement. Some of them are very welcome developments; some give us cause for concern. Mainly they are good, it has to be said, particularly in the field of private enforcement. We are always free here at the Workshop to be provocative, so I will ask you two or three questions, and they’re only questions, mainly relating to public enforcement. The first one relates to the system. When I was last at this meeting four years ago, quite a lot of the same people were here—and what were we then hotly debating? We debated the large and ever-increasing size of penalties, and we debated whether the intensity of review available in the courts of the Member States, and to some extent at the General Court, was adequate. Our discussions were partly inspired by all of the consequences of a finding of infringement, which is essentially penal in nature, at least from the perspective of the ECHR. And what were we discussing yesterday? Issues not so very different from what we discussed then. If there is a persistent need to discuss an issue, does that mean there is something that needs to be changed? What could be at the root of the problem? You could argue that the root of the problem is that some national courts have the wrong standard of review, and that if you gave them greater scope to reassess, for example, the facts after the agency has looked at them, then perhaps that would ameliorate this problem. But is that actually the real cause of the problem? Or is it time to consider whether the root of the problem is in fact the administrative investigation and decision system itself ? I would like to pose the question of whether, as far as at least penal, ex post infringement cases are concerned, the administrative decision system has had its day. Why do we insist on placing the power of investigation, of prosecution and penal decision-making in the hands of officials of an agency, which cannot conceivably be regarded as impartial and independent? DD

Part IV: Issues for Courts and Perspectives on the Judicial Role 483 First, is this type of system down to the agency itself ? No matter how hard it tries sincerely to appear and indeed to be impartial, no matter how many safeguards, Chinese walls and hearing officers are put in place, it will never be regarded as impartial and independent, and it will always be suspected of confirmation bias. Is it fair, or realistic, to expect the agency to combine two essentially irreconcilable roles—that of prosecutor and that of judge? Second, and even more importantly, is it right or fair to the defendant to subject him to a penal process at the hands of an administrative agency that cannot fulfil the role of an independent and impartial decision-maker? Is it fair to the defendant to subject him to a process where his first opportunity to come before an independent and impartial tribunal is when he is already a convicted person? How can the presumption of innocence be given any real effect in those circumstances? Has the time come to accept that the problem will only be resolved when we abandon the administrative model as unsuited to the twenty-first century, when cartel infractions are morally reprehensible and are visited with very severe, quasi-criminal sanctions? The United Kingdom came very close to doing just that—abandoning that model—when it was going through the process of replacing the OFT with the Competition and Markets Authority. It was discussed whether the administrative system should be abandoned. Many people feel that it would have been better if the UK had bitten the bullet and done so, and reverted to a prosecutorial model. The latter model also has some attraction for the enforcement agency, which can still determine policy by choosing the cases it wishes to pursue, and can exert pressure to settle by threatening to take the alleged wrongdoer perhaps more quickly before a court, and the enforcer is liberated to become an unequivocal prosecutor. However, the United Kingdom is now consulting on going down the opposite route of a dumbing down of the standard of review, which we already have of the administrative decision. The German model has elements of both systems, but, provided that one has a right to a full rehearing—a full prosecution-type hearing before an independent tribunal, which decides whether there has been an infringement or not—it may be questionable whether you need to have the full decisionmaking process at the administrative stage or whether there is an unnecessary duplication. So that’s my first semi-provocative question. My second question relates to the points that Simon Bishop made yesterday when he pointed to the indistinct boundary between vigorously competitive conduct and conduct which is unlawful under Article 102. As Simon correctly stated, because that boundary between the two is blurred, firms and even their legal advisers are often unclear as to whether they are dominant and, if so, whether they have committed an abuse. And the situation of legal uncertainty arguably brings the law into disrepute and has a chilling effect on competition itself. When I was at the bar, I well remember an administrative procedure which took about seven years, in the course of which we had successive statements of o bjection,

484 Part IV: Issues for Courts and Perspectives on the Judicial Role each one reanalysing the cost allocation model to produce a different kind of result, some of which showed margin squeeze and predation, some of which showed that there was none, and eventually the agency abandoned the case. But it created a huge uncertainty and huge cost for a long period, and, of course, there were people waiting to bring follow-on claims if there had been an infringement decision. The criteria for some unilateral conduct infringements are still far from clear, particularly in relation to pricing abuses. Can businesses really be expected to engage several economists and lawyers in order to reassure themselves that their pricing policy is unlikely to be held unlawful? And so it is certainly worth examining carefully Simon’s suggested solution of not imposing penalties for some types of conduct. But should we go further? After all, even without penalties there is the finding of unlawful conduct and the risk of follow-on actions. There was a famous European statesman who said about the Schleswig-HolsteinQuestion that it was such a complicated matter of international diplomacy that only three people had ever understood it. One of them was dead, one had gone completely mad, and he himself was the third person but he had forgotten the answer! Before the United Kingdom joined the European Economic Community we had a very idiosyncratic competition law, if you could call it competition law at all. It went by the name of the Restrictive Trade Practices Act, and it was technically extraordinarily complex, very much like the Schleswig-Holstein Question. It was so complex that it took years and years for someone to actually write an intelligible book to explain it, and by the time the book came out the Act had been repealed! It was very sad for the authors, but it was a very, very good book. But the main thing about the Restrictive Trade Practices Act is that it was prospective. If your conduct fell within very, very complicated parameters criteria of the Act, then the contract would be void unless it was justified—it could be justified under various heads, and the job of justifying it was on the person who wished to continue the agreement. It was usually the Office of Fair Trading who argued the other side, and the court decided whether it was justified, and if it was decided that it could not be justified, then, in many cases, the court was asked to provide a permanent injunction against behaviour of that kind for the future. I am not suggesting that we should revive the Restrictive Trade P ractices Act or anything like it because it is better dead and buried, but, given the evolutionary nature of the law of A rticle 102, given the blurred boundaries, would the interest of legal certainty and justice, and of competition itself, be better served if some categories of what are now infringements were subject to only pro spective prohibition and pro spective penalties? This would remove the legal certainty objections that we often hear about. It would remove the penal nature of this category of cases, and it would enable a full proper economic assessment to be central to the evaluation of conduct—without endangering legal certainty. Of course, if so desired, you could still retain categories of

Part IV: Issues for Courts and Perspectives on the Judicial Role 485 ex post prohibitions of unilateral conduct which are considered sufficiently clear and heinous. So those are my provocative questions. Kovacic: Thanks for getting us off to a great start, Gerald. These are wonderful topics for the morning’s discussion. Jacqueline is next. DD

Jacqueline Riffault-Silk: It is a great pleasure to be with all of you in this stimulating atmosphere of discussion and debate. My first point concerns the framework in France. I think that much has been done in France to guarantee that there is proper enforcement—a legitimate enforcement— of competition law. First, it must be recalled that the French Competition Authority is organised on the basis of a strict separation between the functions of investigation, prosecution and decision. That is already something. For the judicial side, it must be said that, from the start, jurisdiction has been given to the judiciary to control the decisions of sanction and also the decisions taken in negotiated procedures. In France there is a principle of specialisation of judges and, furthermore—and this is a bit different from the EU system—there is a principle of full review. This is shared with many other continental Member States. And this principle of full review by the Court of Appeal is, of course, on the facts and the law, and there is an obligation for the Court of Appeal in the case of annulment of reformation of the decision of the Competition Authority to redecide the case itself. There is a principle of the principe dispositif, which means that if the Court of Appeal has all the elements of the case, it has to decide on them. That might not apply, however, if there is a need for further investigation. The principle is that the Court of Appeal—the judiciary—is not given the powers of inquiry of the administrative body. So in that scenario it has to send the case back to the fact-finder. With regard to the extent of the control exercised by the Cour de Cassation: as you know, the Cour de Cassation has a twofold control. First, it checks whether there has been a violation of the law; and second, it reviews the legal basis or lack thereof. To uphold a finding of infringement within the meaning of Article 102, the infringement has to be justified by pertinent, relevant factual analysis, and the Cour de Cassation is extremely demanding in that regard. Another point to underline is that tensions can be perceived as regards the administrative format of a decision and its perceived nature. According to the Conseil constitutionnel, there is no theoretical problem in giving a public body powers, provided that they are framed by strict guarantees—the same guarantees that are given in criminal procedures. But there are many other aspects in the proceedings. The Court de Cassation in plenary session has said that the rules applicable to the proceedings of the competition authority, absent indications to the contrary, should be those found in the Commercial Code. So the rules are of a civil nature, but this solution was completely at DD

486 Part IV: Issues for Courts and Perspectives on the Judicial Role odds with what was possible in the investigations of the Authority. There was a problem that involved a tape recording by a claimant, and the claimant had produced this as evidence to the Authority; in a criminal procedure, that would have been allowed, but under the circumstances, it was forbidden. The opposite is the case law of the criminal chamber on the seizure of numerical data and the repeated rulings of the criminal chamber that a mailbox is a non-separable document, analogous to a hard paper agenda—and this is very dangerous for civil liberties, because there are confidential elements and there is a fierce debate at the moment over whether confidentiality is sufficiently protected, given the technical possibility of separating some parts of the mailbox. Now I would like to say a brief word about the calculation of fines. I will refer to the steel cartel in France, where the Court of Appeal applied the ceiling criterion, which is of a criminal nature. Here there is a black box—you refer to some considerations, and at the end of the day you say the penalty should be this level. But it is a bit more complicated in the French system, and the problem was that, by means of this black box, the Court of Appeal reduced the fine by a factor of maybe 10. Finally, it proved to be very positive because the problem was highlighted, there was a debate, there was a report that was requested from a panel of professional people—judges and economists—and finally the Authority adopted a Notice on the subject, which is a very satisfactory element for more transparency. Maybe it does not change much the level of the fines, and maybe the problem of the very high level of fines as alleged by the undertakings is still not solved; but nevertheless there was a lot of progress for transparency. Yesterday it was asked why an authority does not take into consideration the general damage to the economy? This is a legal criterion in France—it is a French specificity, so to speak. In the early case law of the commercial chamber, it was presumed that if the infringement was proven then it was logically deduced that it has damaged the general economy. Then there was a reversal of the case law, and the commercial chamber ruled that the damage had to be proven or justified. So how can damage to the economy be proven? What should be the criterion with regard to quantification of the damage, given that this might lead to extremely complex assessments? So far, the commercial chamber has ruled, repeatedly, that a qualitative justification is enough, but it is has to be justified concretely. The Cour de Cassation very often states in relation to economic regulation that the Court of Appeal must be concrete in its motivation. Then in France we saw the rise of negotiated procedures and commitments. In France, they have been used very often, such as in the following circumstances. First, they may put an end to anticompetitive behaviour when a decision for interim measures has failed, the case is on lengthy appeal and there is urgency in the eyes of the Competition Authority. The second case is where commitments complement a finding of infringement and financial penalties. In France there is now very often a mixture of both,

Part IV: Issues for Courts and Perspectives on the Judicial Role 487 and an example of this was observed in the banking sector. That put an end to almost all the possibilities of inter-bank fees. Third, commitments can be a vehicle for remedies in specific sectors. One example is energy. In France, 80 per cent of the energy consumed is produced by nuclear plants. And we have a large incumbent, of course; a huge undertaking, with very significant power. So how can newcomers enter the market? It is almost impossible. But EDF is forced to sell, at cost and with absolutely no margin, a significant part of its production to new entrants. Some time before the sector was reformed, maybe two years earlier, the competition authority had accepted commitments from EDF to do that, and so in fact the reform was just a confirmation of those remedies. Professor Siragusa made a very good point because we have rulings of the Court de Cassation saying that the guarantees are not of the same level as they are with criminal procedure requirements. The guarantees are definitely weaker. And in the decisions of the commercial chamber of the Cour de Cassation, it is outlined that that procedure is a negotiation, and so we are on the civil side. I would like to add some comments about themes that came up in discussion yesterday. With regard to the subject of amicus curiae, it is remarkable that the European Commission has not developed its role before the Court of Appeal and the Cour de Cassation in cases where principles of European competition law were obviously at stake. Two points might be made there. The first is that preliminary rulings maintain their interest and their very important role. And second, just recalling what was said yesterday, the Commission prioritises when deciding whether to intervene as amicus curiae. You will recall the Expedia case,1 which concerned the interpretation of the Commission’s de minimis Notice. The dispute arose from the creation of a joint venture between the French railway company SNCF and the travel company Expedia to set up a company in charge of selling travel tickets online. Since SCNF is a dominant player in the travel transport market in France, in the railway transport market, the Competition Authority investigated this agreement. The Authority found that there was a restriction of competition by object and by effect under Article 101(1), and it fined the companies. They objected, alleging that their share in this new market was very small, and so the issue was whether the Notice was applicable or not. The Cour de cassation requested a preliminary ruling from the Court of Justice, and the Court decided first of all that the Notice does not bind national courts or national authorities. Second, it held that a restriction that has an appreciable effect on competition cannot be covered by the de minimis Notice. And third, it considered that a restriction of competition by object has by its very nature an appreciable effect on competition. Subsequently, the commercial chamber dismissed the appeal that had been brought by the two companies. The Court decided that, given the restriction of competition, the Notice was not Case C-226/11 Expedia Inc v Autorité de la concurrence and Others, EU:C:2012:795.

1

488 Part IV: Issues for Courts and Perspectives on the Judicial Role applicable. More recently, as many of you know, the Commission launched a revision of its Notice to reflect the ruling of the ECJ.2 So this case illustrates the interplay of different national and European institutions. Maybe there are other things that can be improved, but this case shows that the system we have in Europe sometimes works well. Savvas Papasavvas: As a European judge, it is appropriate for me to f ollow that reference to the European system. I would like to begin by underlining that the opinions I express are, of course, personal. Let me frame the first issue by asking: are there problems with some of the competition law cases before the General Court? Yes, the duration of some of them has been a problem. Are these problems as important as some want to project? I don’t think so. Is it a problem of the system in general or only of the General Court, or both? I’ll try to give you some answers. Let me remind you that we are not policy-makers, we are judges. And as far as I am concerned, the system in general is not bad. Hopefully, with the ongoing measures that are being taken at the EU and national levels to enhance private enforcement, the last really problematic part of the equation will find some solutions. Now, is the EU system compatible with Article 6 of the European Convention of Human Rights? Of course it is. Is it a good idea to switch from an administrative system to a different type of system? I don’t think so. It’s not up to me to decide, but I think that the system should remain administrative. Could we improve the one we have? One can always try to develop new solutions and new ideas. With regard to the occasional delays at the General Court, we regret them but, of course, our capacity is not unlimited. Like all modern jurisdictions, we have considerations of productivity, efficiency and backlog. We have been trying our best to adapt the situation, and we’ve been changing our working methods. But the truth is that our workload is considerable, and, as you know, we very often deal with quite voluminous administrative files. There are tables circulating four times a year with the total delays of each judge, and in general this provides an effective incentive to be as productive as the docket will allow. We have asked for reinforcements and so far we have not been able to have, for example, extra référendaires, but we will see how that progresses. As for the Court of Justice, the Court now has three more Advocates General, accompanied by 12 référendaires. Another element not to be neglected is the stability of our composition. We have had 10 new judges from September 2012 to September 2013, and integrating a large number of judges requires some time. There is also another element that might not be known to everyone: whereas, at the ECJ, the President of the Court directly assigns cases to judges, at the DD

2 The revised Notice and a corresponding Commission Staff Working Document providing guidance on restrictions of competition by object—updated to reflect the Cartes Bancaires judgment—are available at http://ec.europa.eu/competition/antitrust/legislation/ deminimis.html.

Part IV: Issues for Courts and Perspectives on the Judicial Role 489 General Court our President assigns them to the chambers. It is then for the President of the Chamber, according to a tour de rôle, to distribute them among the judges. Our system is more democratic, but not necessarily that efficient. If you combine this with the volume of the competition cases, it does not leave much room for a reattribution of a case if it becomes problematic— let alone a series of cases, like a cartel, which might be cumbersome. We do reassign some of the cases to take the burden off some of our colleagues, but this is not that easy. Can we change that? Yes we can, but it’s easier said than done. I also think that we are a jurisdiction that has some unnecessarily luxurious conditions for the parties. There are two rounds of pleadings before the hearing, with no enforceable possibility to limit the pages submitted if the lawyers refuse to comply with the registrar’s indications. Sometimes the excessive use of confidentiality measures causes delay as well. Let me now address the famous question of the intensity of judicial control. The lawyers complain, but sometimes they forget their own role, and sometimes they do not necessarily use the tools they have to discredit Commission decisions. For example, we rarely have expert witnesses appearing before us, nor do we see counter-reports to rebut the arguments of the C ommission. My dear friend President Marc Jaeger has written a very interesting article on the standard of review in competition law cases involving complex economic assessments.3 That is highly recommended bedtime reading. President Jaeger reminds us, among other things, that the Court of Justice has held that, although the Commission has a margin of discretion with regard to economic matters, that does not mean that the Court must refrain from reviewing the Commission’s interpretation of information of an economic nature. Not only must the EU Courts establish, inter alia, whether the evidence relied on is factually accurate, reliable and consistent, but also whether that evidence contains all the information which must be taken into account in order to assess a complicated situation, and whether it is capable of substantiating the conclusions drawn from it. Such a review is all the more necessary in the case of a prospective analysis, such as when examining a planned merger with conglomerate effects. I don’t find that type of control particularly loose—on the contrary. Now, specialisation is not an option for us, because it would be too unfair for some and too easy for others. However, the General Court is still the right court to handle competition law cases. We have an excellent overall view of the economic law of the EU, second to none, including state aid, antidumping and other areas. We blend generalists and specialists, and I think that this is important for the effectiveness of the institution. Furthermore, the Court

3 Marc Jaeger, ‘The Standard of Review in Competition Cases Involving Complex conomic Assessments: Towards the Marginalisation of the Marginal Review?’, 2 Journal E of European Competition Law & Practice 295 (2011).

490 Part IV: Issues for Courts and Perspectives on the Judicial Role of Justice holds us to a high standard, and it does not hesitate to overrule us where it finds flaws in our reasoning. President Jaeger, in the article I mentioned to you, cites the examples of Impala4 and Alrosa.5 I also have to remind you, despite all that is being said and written, especially as far as cartels are concerned, that it is rare for the parties to contest the infringement itself, and that says something about what the Commission does. Parties might contest the duration, the application of the leniency criteria or other things of that nature, but they rarely contest the finding of infringement itself. I would add that, in spite of our delays, we still have the fast-track procedure. And we still have excellent results most of the time when we handle fast-track cases. The fact that, for example, we got the EDP merger dealt with in six months—and that was quite a complex merger case—this is a very positive reflection.6 So I graciously accept that we can make things better, and we are trying. The incoming cases are rising once more, this year there was a rare exception, but nevertheless we are getting back to normal on our speed. Last but not least, I would like to thank Ian, because the title he used for his paper reminded me of my PhD years, and in particular it reminded me of an article written by EUI Professor Cappelletti about 30 years ago on the legitimacy of constitutional judges.7 José Luis Da Cruz Vilaça: First of all, I would like to thank Philip and Mel and Giorgio for very kindly inviting me. Mel and I tried to make that happen a few years ago and circumstances made that impossible, but I am very pleased and honoured to be here for the first time. I would just like to start by going back to some points that were raised around the table yesterday morning. And I will be guided by my own experience at the Court of First Instance more than 20 years ago, which means I’ll be digressing a little bit through the past—I am an old man, I apologise for that. My first reflection will be very brief. Some points have been touched upon already by Savvas, such as the intensity of judicial review in the competition cases. My reflection lies somewhere in the crossroads between effective enforcement and the protection of legitimate rights. The current system set out in the T reaties does not allow the Courts to substitute their own policy preferences for those of the Commission. The role of the Courts under the Treaty is to review the legality of Commission decisions—in terms of both external legality and internal DD

4 Case C-413/06 P Bertelsmann AG and Sony Corporation of America v Independent Music Publishers and Labels Association (Impala) [2008] ECR I-4951. 5 Case T-170/06 Alrosa v Commission [2007] ECR II-2601; set aside: Case C-441/07 P, Commission v Alrosa [2010] ECR I-5949. 6 Case T-87/05 EDP—Energias de Portugal, SA v Commission [2005] ECR II-3745. 7 Mauro Cappelletti, ‘Repudiating Montesquieu? The Expansions and Legitimacy of “Constitutional Justice”’, 35 Catholic University Law Review 1 (1985).

Part IV: Issues for Courts and Perspectives on the Judicial Role 491 legality. I share the considerations made before me. When I was President of the CFI, we hired an economist with a view to assisting the judges so that we could better understand the most complex and stratospheric arguments put forward by the parties. But that economist didn’t stay for long; he was needed only rarely by the judges. Indeed, the parties provided the court with their respective versions of the facts, and they brought their own experts in to explain the economic theories applicable to the facts. In case of doubt, we had to, and we still have to, rely on rules like the burden of proof. So the experience of the CFI tends to show that the advice of an in-house economist does not really make a difference. Forgive me once again for going back into the past, but I think frankly that the CFI from the beginning set the tone for a balanced approach as regards its judicial discretion. With regard to the protection of the rights of defendants and complainants, I believe that the CFI—many years before the entry into force of the EU Charter of Fundamental Rights—pushed the limits much further than the Court of Justice had done up to that point. This was done step by step. I recall the right of access to file, the Polypropylene case,8 the Hercules case9 and others. The CFI had a more open attitude as regards admissibility of appeals, and a more favourable view of legal privilege for in-house lawyers in Commission investigations. But at the same time, the CFI did not hesitate to strengthen the Commission’s powers of enforcement. I’ll just mention the Commission’s prerogative of defining its enforcement priorities, or the possibility to find an abuse of a dominant position where the abuse occurs in another neighbouring market, or a wider interpretation of the Commission’s power to adopt interim measures. Indeed, that power was wider than the Commission thought it was at the time. So the CFI sought to achieve a balanced approach, and the EU Courts still try to achieve a balance, but, of course, as judges and as institutions we can always do better. A second point touched upon yesterday concerns fines. Once again, I agree that higher fines require more intense judicial scrutiny. I never thought that the 10 per cent threshold should be scrapped and replaced by a totally different threshold. But why not reflect on a more nuanced, more balanced threshold? In any case, the ECJ and the General Court both exercise unlimited jurisdiction with regard to fines. And yet their respective positions differ. According to settled case law, it is not for the ECJ, when ruling on questions of law in the context of an appeal, to substitute, on grounds of fairness, its own assessment for that of the General Court exercising its unlimited jurisdiction. It is true that assessing the proportionality of a fine involves a legal analysis. But if the decision at first instance is based on legally assigned definitions and standards, then application of that principle to the concrete circumstances is

Case C-199/92 P Hüls AG v Commission [1999] ECR I-4287. Case C-51/92 P Hercules Chemicals NV v Commission [1999] ECR I-4235.

8 9

492 Part IV: Issues for Courts and Perspectives on the Judicial Role a matter of fact, which escapes the ECJ’s control. The exception there is where there is manifest error or a manifest disregard for the facts, which might turn the issue into a point of law. This may help to explain why, during my first year in office at the ECJ, I only had to deal with two little competition cases. I spent most of my time hearing about asylum seekers, child abduction and suspected terrorists. So my experience as an ECJ judge applying EU competition law is so far quite limited. As a matter of fact, in 2012 only 13 out of a total of 500 cases decided by the ECJ were on competition matters. It is therefore clear that the bulk of judicial work in this field still lies with the General Court. As Savvas noted a few moments ago, this is a great challenge for the General Court, which has had to cope in recent years with an increasing workload and backlog of cases. Should we still move to a completely different system of reassessment, with the Courts to reassess de novo the decisions of the Commission? I don’t think it is a suitable move, at least for now. But for the future, any change of this kind must be compatible with the specific characteristics of the EU judiciary. Any such reform must be very carefully designed to ensure that the Courts are not flooded with an immense workload. In a new wave of cases before the ECJ, it is claimed that the General Court breached its duty to adjudicate competition cases within a reasonable time.10 The Court is now called upon to provide guidance on how to deal with these claims. First of all, how can it be determined whether the delay is excessive or not, and how much time is excessive? Quantifying a delay is certainly not an exact science. With regard to the duration of inaction at the General Court, some delays are to some extent incompressible, in particular those concerning the written procedure. Should the ECJ set any kind of indicative delays? I think that drawing a table of indicative delays would be a hazardous task. The Court will also have to adjudicate on the remedies available in the case of a failure to adjudicate within a reasonable time. Two approaches follow from the case law—Baustahlgewebe11 and Der Grüne Punkt.12 In Baustahlgewebe, the Court decided to grant a further reduction of the fine imposed by the Commission, as a means of reasonable and immediate satisfaction, on grounds of procedural economy. The drawbacks of this approach may appear quite obvious—it cannot be applied where a fine is not imposed, as demonstrated later in Der Grüne Punkt. Furthermore, this approach makes it necessary to define some criteria to decide what amount of fine reduction is appropriate. In Baustahlgewebe the reduction amounted to 1.67 per cent of the fine. More fundamentally, there is

10 See, eg Case C-58/12 P Groupe Gascogne SA v Commission EU:C:2013:770, paras 59–97. 11 Case C-185/95 P Baustahlgewebe GmbH v Commission [1998] ECR I-8417. 12 Case C-385/07 P Der Grüne Punkt—Duales System Deutschland GmbH v Commission [2009] ECR I-6155.

Part IV: Issues for Courts and Perspectives on the Judicial Role 493 no link between the seriousness of the anticompetitive conduct that led to the fine being imposed and the reduction of the fine for reasons of excessive delay. The problem of how to find a suitable legal basis to introduce a reduction of the fine as a means of compensation is also a serious problem. An alternative approach was taken in Der Grüne Punkt. In that case, the ECJ ruled that the failure of the General Court to adjudicate within a reasonable time could give rise to a claim for damages against the Union. Such an approach has clear advantages, but also some problems, which I will skip over in order to respect my time limit. But a question that arises in this regard is whether an action for damages in such cases is subject to the same rules as normal actions for extra-contractual liability against the Union, or whether this type of action has specific characteristics that justify an award of compensation without the applicant having to prove damages. Some even envisage a kind of combination of both approaches. But, given the different approaches in the case law, it will certainly be interesting to see how the Court handles this issue of excessive delay.13 Two other problems are, first, at what stage of the judicial proceedings should the issue of reasonable delay be addressed, and second, which of the Courts would adjudicate remedies? If the Baustahlgewebe test is followed, of course both the ECJ and the General Court may apply the test; but if the procedure to be followed is an action for remedies before the General Court, some have questioned the compatibility with Article 6 of the European Convention in the light of the requirements of impartiality and objectivity. I reject this point of view for several reasons I will not mention. I do not share the concern that this may be incompatible with Article 6 of the European Convention.14 I will conclude by making some practical reflections. First, despite the efforts made by the General Court, the evidence seems to suggest that we are facing a kind of structural

13 Following the Workshop, in Group Gascogne, above n 10, the Court of Justice indicated that the General Court had breached its duty to decide within a reasonable timeframe as required under Article 47 of the Charter of Fundamental Rights. However, this ground of appeal (like all the other grounds of appeal) was rejected because, as the Court stated in paragraph 82, ‘a claim for damages brought against the European Union pursuant to Article 268 TFEU and the second paragraph of Article 340 TFEU [non-contractual liability of the Union] constitutes an effective remedy of general application for asserting and penalising such a breach, since such a claim can cover all the situations where a reasonable period of time has been exceeded in proceedings’. It is thus for the aggrieved party to seek compensation in a separate action before the General Court, which must sit in a composition different from the one that heard the underlying dispute (ibid, para 90). If the notion that the General Court is both judge and party in such a case seems awkward from a fundamental rights perspective (nemo judex in causa sua), the party seeking compensation will at least have the option, eventually, of appealing the General Court’s judgment in the Article 340 action to the Court of Justice. 14 For the solution adopted by the Court of Justice a few months after the Workshop convened, see previous footnote.

494 Part IV: Issues for Courts and Perspectives on the Judicial Role problem within the EU judicial system. It’s not a problem of the General Court, it’s a structural problem affecting the EU judiciary. There have always been structural problems affecting the EU judicial system, and in fact the courts alone cannot solve these problems. In my view, legislative intervention and probably also Treaty amendments are required. Personally, I think it would be useful to reflect on a redesign of the judicial bodies deciding at first instance in competition and other cases. With regard to the ECJ, I fear that it is quickly approaching the limits of its capacity, and that ultimately some kind of filtrage of the cases will become inevitable—not only for appeals against decisions of the General Court, but any kind of cases, such as references for preliminary rulings and other direct actions. For the rest, I’ll just add some reflections from my own experience this last judicial year at the ECJ. First of all, compared to the old days, the working methods have changed according to the evolution of the circumstances, and this is linked mainly to the increased size of the Court. Increased size, an increase in the number of legal systems that are represented, and, of course, a multiplication of languages: these are all relevant in this regard. We meet in principle in every case to decide the final deliberation, and then I’m happy, finally, to decide on points of law. We have been working to achieve better timing and earlier treatment of the files by Advocates General and by judges, with the possibility for an early warning of the need to refer a case from a given chamber to the Grand Chamber. The intervention of the Advocates General is in my view still very useful. The involvement of the Advocate General causes a new delay: naturally, by the time we receive the draft Opinion, several weeks or months have elapsed. Nevertheless, and regardless of whether we agree with the Advocate General or not, his or her intervention actually speeds up the process of the Court taking its decision. Finally, I can assure you the ECJ is constantly reflecting on ways to improve its working methods. Indeed, we organise from time to time—every two months, for instance—a forum of the members where we discuss a subject for reflection. This can lead to new methods and new proposals for internal reform. The next forum of the members is on the questions for preliminary rulings—how to deal with those questions, how to cooperate effectively with the national judges, but also how to speed up the proceedings and how to limit our intervention to what is really important for the decision of the national court. My time is up, so I will stop here. Thank you very much. Kovacic: Thank you, Judge. To discuss these four extremely informative perspectives on the operation of courts that have played such a crucial role in competition law, we have three exceptional observers of the judicial process at the national level and EU level. Jim Venit, Ian Forrester and Mario Siragusa have extensive experience as litigants, observers, authors and commentators. Each of them will give us his own observations, beginning with Jim. DD

Part IV: Issues for Courts and Perspectives on the Judicial Role 495 Jim Venit: Thanks, Bill. Let’s start by reminding ourselves of what is really the importance of the Workshop as an institution: it brings together different perspectives from different organisations that have different functions. Every year, the Workshop’s participants include judges, enforcement agencies, economists, lawyers, policy-makers and academics. It is a remarkable forum, and the exchange of views that has occurred in this forum over the past few years has had an impact and led to change. I am hopeful that that will be the case again. My paper from our 2009 proceedings15 sums up my argument that the combination of an administrative system of enforcement with a system of light judicial review is not sustainable in the long term in terms of credibility, legitimacy or the quality of the decisions of the enforcement agency. If I had to revisit that subject now, other than noting some of the changes that Philip introduced in the Best Practices16 and guidance on state-of-play meetings,17 which helped to contribute some greater transparency to the administrative process, I wouldn’t add much more to it. The big picture is that we do have a problem, and I agree with Mario’s observation yesterday that it is a problem that raises questions of legitimacy. I also agree with Bill’s remark today that courts contribute substantially to the quality of the law. I also agree with Gerald’s observation that it would be useful to consider some radical alternatives. We’ll come to that later, but in short, yes, there is a problem and I do think we need to try to address it. Do we have full or light review at the level of the EU Courts? That is a fascinating question, and the reference to Judge Jaeger’s article,18 which everybody read with great interest, is a reminder to me that there have been periods in time where the CFI was very proactive—some may even say aggressive—in its review. The period in which then-Judge Vesterdorf was President of the CFI was a time when things happened that we weren’t expecting, particularly in merger cases. At times, the Commission was caught by surprise, and there were many in the Commission who felt that the CFI had gone too far. But the real problem is that, although we’ve had periods and judgments that suggest a standard of review that actually may be more intense than the standard of deference to the Commission in cases of complex economic assessment, there is no consistency. You get some judgments where you seem to see that DD

15 James Venit, ‘Human All Too Human: The Gathering and Assessment of Evidence and the Appropriate Standard of Proof and Judicial Review in Commission Enforcement Proceedings Applyng Articles 81 and 82’, in Claus-Dieter Ehlermann and Mel Marquis, eds, European Competition Law Annual 2009: The Evaluation of Evidence and its Judicial Review in Competition Cases, Hart Publishing, 2011, 191 et seq. 16 Among the Commission’s other Best Practices documents, see Notice on best practices for the conduct of proceedings concerning Articles 101 and 102 TFEU, [2011] OJ C308/6. 17 See ibid, paras 60–66, 110, 119, 121, 132 and 139. 18 See above n 3 and accompanying text.

496 Part IV: Issues for Courts and Perspectives on the Judicial Role appening, but then you get Tomra.19 So there doesn’t seem to be a firm line h of development. We’ve seen flashes of review that was much more detailed and rigorous, but it has not been a consistent practice. Additionally, I think it’s important to bear in mind that all of these issues of legitimacy and the inter-relationship between an enforcement agency that has the powers of both prosecution and judgment and a court that doesn’t seem to fully exercise its powers of full judicial review are issues that are not limited to conduct cases or to quasi-criminal cases like cartels. They affect merger cases as much as other types of cases. Frankly, while there may be reasons to object to the level of fines seen in some cases, I’m more interested in whether the law is being applied in the right way. Are we making the right economic distinctions? To provide a small example: in a merger case in Europe, if you have a remedy imposed, the applicant is never going to appeal that decision, and the simple reason is that if it does and if it wins, the Commission gets to take the decision again three or four years later, under economic conditions that are unimaginable or unforeseeable at the time that the proposed merger is filed. That’s not a very interesting proposition for a client, even if in principle there is a very strong case to bring an appeal. So what should we do about this? Gerald said maybe we should think about moving to prosecutorial enforcement and strict judicial control. If you think about it, that illustrates how bad things have gotten when you consider moving to a full-blown prosecutorial system. However, at least in that type of system everybody would know what the rules were. From the Commission’s perspective, such a change might even be a liberation, because it often behaves like a prosecutor. Perhaps it would be better off if it were able to do that freely and openly—but again, under strict judicial control. A second option is that we can change the standard of review or make it clear that we have a very proactive, full review standard, similar to the standard prevailing in several national jurisdictions, as we discussed in 2009. Alternatively, we can just hope for gradual improvement. Sometimes that happens, but the road seems to be fairly uneven; I’ll come to the importance of this in a second. One issue which would be very easy to deal with relates to the notion of access to the file and confidentiality. I’ve taken the position for a long time that the Commission and the Courts ought to get rid of the current system of confidentiality. There should be protective orders—either from the Commission or the Court—and everything would be accessible. It would take the enormous burden of having to supervise redaction (and often making mistakes in doing so) off the Commission’s shoulders. I have a case now where the Commission has had to recall the file because confidential information appears in the text when you read it on a computer, and it’s going to cause the Commission considerable

19 Case T-155/06 Tomra Systems ASA and Others v Commission [2010] ECR II-4361; appeal dismissed: Case C-549/10 P, EU:C:2012:221.

Part IV: Issues for Courts and Perspectives on the Judicial Role 497 embarrassment and delay. If you had confidentiality agreements with disclosure orders, you would be rid of all of that. Neither the Commission nor the Courts would ever have to deal with the issue again, so I don’t understand why that is not being done. If I were a DG Comp employee, it’s the first thing I would do. It’s a small change, but it would make a big difference. My last point is that the importance of these issues goes beyond the EU. As Bill noted, the Chinese simply do not have judicial review of merger decisions. I would submit that the European institutions have a responsibility to be, if you will, globally responsible and to set the example by holding themselves to high standards of best practice. We all have a serious responsibility and serious tasks ahead of us that are quite important. If you consider the tendency of the Commission to sometimes over-regulate, and then look at the performance of the European economy as a report card on Europe’s tolerance for over-regulation, it’s disturbing. If you put that in a global context, it becomes even more disturbing. I’ll stop there. Thank you. Ian Forrester: This Workshop is a lovely opportunity to reflect on fundamental questions, and the old codgers who have white hair are thrilled by the active participation of young codgers who have bright ideas. The fact that the old codgers are repeating themselves does not mean that they are wrong. The acceptability and legitimacy of a system depends not just on the democratic conferral of power to apply the rules. The European Commission has that power with respect to competition, and that power is clearly assigned to the Commission by the Treaty of Rome. It has not been given that power nearly as clearly for telecoms or for food, or for a long series of other special responsibilities. But competition is undeniably a core competence of the Commission. Now, that doesn’t necessarily make what the C ommission does in competition matters legitimate. Legitimacy is an ongoing duty that requires continued effort. It requires procedures that are g enuine, consistent and visible, and at a time of economic and political crisis, we really ought to be fretting about whether Europe is going in the right direction in that regard. So what Jim just said about the example we should set for the rest of the world is especially relevant. My contention is that the basic procedures by which Commission competition decisions are taken are imperfect. They burden an excellent and distinguished institution with miserable processes—unfair processes. I am sorry to say that a client who has had a case in Germany or a case in London will feel, after a procedure with those authorities, far more satisfied than it will feel after an experience in Brussels. That’s not because there is anything wrong with the people in Brussels. They’re intelligent, talented, competent, polite, and they know what they’re doing. Some of them have even worked for me. Yet there is a problem. The problem is that the impression of the consumer of the decision is that the consumer is not being heard fairly and impartially. So that’s the problem in Brussels, and then the problem in Luxembourg is that DD

498 Part IV: Issues for Courts and Perspectives on the Judicial Role the judicial review—again, I’m echoing and foretelling what my colleagues will say and have said—judicial review in Luxembourg has not been consistent. At times, it has been tough and exacting—you could think of Woodpulp20 or Airtours21 as examples of that. But most of the time it has been ‘light’, or deferential. The check has merely been, or at least has appeared to be: ‘Well, this is technical, or economic; it’s not obvious that the Commission has gone totally off the rails, so we leave it in the C ommission’s hands’. A related problem is the inconsistency between how things are done in Brussels and the requirements of the European Rights Convention. According to Article 6 of the ECHR, as interpreted in dozens of cases by the E uropean Court of Human Rights, it’s okay to punish someone at first instance in a criminal matter, ‘criminal’ being broadly defined, provided that on appeal there is a full judicial review of the facts and the law. Now, the review carried out by the EU Courts under the Treaty is not a check of the facts as such, but of the ‘legality’ of the Commission’s decision. So the question—and I am being a little unfair, but not very unfair, that’s what lawyers are meant to do— the question is not whether the Commission was right, but rather whether its decision is marred by certain categories of error. Now, in a competition case, the facts are absolutely crucial to deciding whether or not there was an infringement. You can’t determine a competition matter without examining the facts. So, while ‘legality’ is an easy test for the Commission to satisfy, correctness is a difficult test. Regrettably, the European Courts have often felt obliged only to hold the Commission to the simpler, easier test. Let us remind ourselves of how the Commission system works. It begins normally with a suspicion or a complaint, and then there is an investigation by the case team. The case team will take months, probably years, to conduct extensive inquiries and then they decide, ‘Ah, there is something here’. Then, after many internal procedures and complex internal obstacles, they make an accusation called a Statement of Objections. Then there is access to the file and a hearing, and the hearing may last several days. But by this stage, the process might be 85 per cent complete, so the officials are already committed to condemning the company that is being investigated. Just recently, we had a hearing where we thought, ‘Gosh, that was really good, we have absolutely, clearly demonstrated why the accusations are all fundamentally mistaken’. But then we’re told: ‘It was a very nice hearing but we’re going ahead anyway; we’re not changing our views in any respect’. Now maybe we’re wrong, and maybe the Commission is completely right. But the process by which the Commission reaches its conclusion that the company is in the

20 Joined Cases C-89, 104, 114, 116, 117 and 125 to 129/85 A Ahlström Osakeyhtiö and others v Commission [1993] ECR I-1307. 21 Case T-342/99 Airtours plc v Commission [2002] ECR II-2585.

Part IV: Issues for Courts and Perspectives on the Judicial Role 499 wrong is miserable. I repeat: the officials are excellent and diligent—all the things you would want. But their minds seem to have been made up. So, in reality, the function of the hearing is to persuade them that maybe in one or other little respects they’ve got it wrong. Then the Commissioner announces a condemnation in public, normally with a fine and with the stigma that goes along with it. The decision is hundreds of pages long, and then that has to be appealed. And in matters of fining there is unlimited jurisdiction, but in matters of merit there is limited jurisdiction. Now, the Court of Justice has produced an interesting judgment in the Schindler case.22 Paraphrasing, the General Court had said that the Commission enjoyed a broad margin of assessment when conducting a complex assessment. In view of the margin of assessment, it is only where the Commission manifestly goes beyond the bounds of that margin that it may be criticised. So that was the test used by the General Court, and the Court of Justice said: ‘We’re not going to interfere’. The Court, too, said ‘We’re just looking for manifest error’, even though a closer look reveals that the Court was quite careful and thorough. It looked at all of the evidence and all of the facts. I find it difficult to reconcile that level of rigour with the jurisprudence of the European Court of Human Rights, as expressed particularly in the Menarini judgment.23 I observe in passing that the EU Courts in many cases—and this is the fun part of writing this year’s Workshop paper—have caused immense trouble for the Member States, creating new law in fields of free movement of goods and persons, nationalities, gender discrimination, food and dozens of other areas. So there are numerous contexts in which the Court of Justice has not hesitated to go far beyond what the Member States expected. Putting it in legal terms, the Court has applied to the Member States a leash of proportionality and discretion which in free movement cases is very short. The leash that binds the Commission in competition cases is, by contrast, a very loose-fitting one. So maybe it’s provocative, but I don’t think it’s untrue, that the EU Courts have disappointed the Member States much more than they have disappointed the Commission. And it’s no harm to criticise a sister institution from time to time. I also observe, and this is written up in my paper, that the Court’s judgments in preliminary reference cases are quite different from its judgments in cases deciding appeals against decisions of the Commission. The appellate cases cling too tightly to all-too familiar doctrine. But preliminary reference cases are sometimes startling. They can be quite reassuring, quite convincing and driven by common sense. Now in the domestic criminal context, we are told that there is a presumption of innocence. If there is

Case C-501/11 P Schindler Holding Ltd and Others v Commission, EU:C:2013:522. Menarini Diagnostics SRL v Italy, No 43509/08, 27 September 2011.

22 23

500 Part IV: Issues for Courts and Perspectives on the Judicial Role a presumption of innocence in EU competition cases in Brussels, the light shines, very faintly. I note that the French Conseil constitutionnel just on 5 July said, in a telecoms matter, that it was unacceptable for the enquiry and the decision to be fused within a single organisation.24 That brings me to some suggestions for the Commission and for the Court. I would suggest the following to the Commission: read what the Conseil constitutionnel said and split the case team in two—one team to do the investigation, one team to write the decision. Secondly, impose a page limit on your own decisions. Does anyone benefit from a 500-page decision? If we want effective judicial review, boil down the accusation and boil down the relevant facts. Everyone will benefit from a better and more efficient appeal. I would further suggest that you expand the powers of the Hearing Officer to include the power to decide contested facts. If the question is whether Mr Smith was there in the hotel bar on 13 January or not, and this person says he was and that person says he wasn’t, bring them together and let the Hearing Officer decide which story is true. Above all, I would suggest: reform now, when you are a world class, premier competition agency. You should show leadership by holding yourself to the highest standards of best practice. To the Courts, I say: please shorten your judgments. We don’t want hundreds of paragraphs; people with white beards prefer short stories to read. Shorten them, save paper, save the planet. Secondly, don’t be shy of criticising a sister institution. It is perfectly possible for the Commission to get it wrong. Treat the Commission as just another litigant: often you’ll agree with it, but not always. Then seek advice from the bar about accelerating the handling of cases. We realise perfectly well that you are cursed by your own procedures—they are what they are, you can’t change them; but you can live within them, and you might discover that members of the bar are much more sympathetic to helping you do your job than you might fear. Thank you. DD

Kovacic: That was fascinating, Ian. Thank you very much.

Mario Siragusa: I agree with everything that Jim and Ian have said. But my approach will be slightly different. I will try to be prospective in my remarks. I am now entering my forty-first year practising antitrust law, and since I intend to practice antitrust law for another 40 years at least, I will take a positive view. I will pose the question: how can we live within the system we have, and how can we feasibly improve it? Let us start from the point of all the criticism that has been aired. Today, we are talking about all those issues, including the improvement of judicial DD

Conseil Constitutionnel—Decision No 2013-331 QPC, 5 July 2013.

24

Part IV: Issues for Courts and Perspectives on the Judicial Role 501 review and improvement of the Commission’s procedures. These have become very prevalent topics. We have the problem of the Convention of Human Rights and we have case law developing at the national level. So there is no doubt that these issues will be at the centre of the debate. And all observers seem to understand that we need to perfect and improve the system. I am very impressed by the fact that the EU Courts are starting more and more to try and face the situation. As we have heard, in some cases, the review has been better. At the national level, clearly in some of our Member States, judicial review is very strong, and the principle of full review is fully recognised by law. What I regret is that, notwithstanding the genuine efforts of the EU Courts, they still seem to be prisoners to some of the old language, some of the old schemes. So, on the one hand, there is a trend towards a more substantive review; but on the other hand, the judges are prisoners of this language developed over 50 years of case law. So I come back to the comments of Hwang Lee about Confucianism: here in Europe I fully agree with the Confucian policy of the rectification of names, and the Court should adopt a new language. If, indeed, this trend is developing—if, indeed, we are now moving towards a more full exercise of judicial review—then let’s take the advice of Confucius, let’s change the language to reflect that. This constant referral to the Commission’s discretion and to the manifest error doctrine is no longer appropriate. A related point concerns the question of whether the situation is so bad that we need to change it completely. Ideally, as a practising lawyer, I would love to have a system in which the prosecutor is only a prosecutor and must prove his case in court to obtain a conviction. But it is probably not realistic today, in Europe, to think that we could change the system radically in that way. Having said that, there may be things that can be done, because, as Sir Gerald said, it is striking that the defendant goes before an independent body only after it has been condemned, only after the sanction has been imposed, and very often only after it either pays that sanction or posts a guarantee with interest. Why don’t we change that heavy-handed approach? One can envisage a mechanism whereby the Commission calculates the fine but the fine does not become effective unless and until the decision is confirmed by an independent decision-maker, the General Court. This would also restore some balance to the system. To oblige a company to pay this huge fine, and to put—I want to stress this, it is very important—to put the liability of the fine in the company’s books, when no independent authority has yet heard the case, that is objectionable to say the least. Even if, at the end of the day, the company wins the appeals, it has in the meantime suffered a blow to its accounts. This is something barely comprehensible and unacceptable to a great many companies. I would like now to make a third remark. This one touches on a discussion we had yesterday. I always thought that the commitment procedure under Article 9 of Regulation 1/2003 was precisely for the cases in which fines are inappropriate. It was supposed to be for a novel question of law, or a very

502 Part IV: Issues for Courts and Perspectives on the Judicial Role complicated problem or a new issue which had not arisen in the past. In those circumstances, it seems appropriate for the Commission to open a dialogue with the company and to come up with a negotiated solution by way of commitments. That’s what the procedure was supposed to be for in the beginning. But Article 9 has been completely corrupted; it is now being used as a pressure point for companies to come up with absurd commitments which often I don’t think are very useful. And this mechanism, this diabolical mechanism of threatening to impose a fine and thereby obtaining commitments from the company, is absolutely improper. The commitment procedure was thought to be—exactly in line with what Judge Barling was saying—for cases in which the proper solution was prospective because of the novelty of the case. The defendant could then say, ‘Okay, I will change my way of doing things and of course you will not punish me for that because I will change my behaviour for the future’. And there you would have the solution for the Article 102 issue that you were asking for, because in that case it is not a question of threatening with fines but simply of determining, prospectively, a change of behaviour. I think that a system of that kind would work, but it requires a complete change in the approach to the commitments, which, I repeat, are destroying our system! I have fought a lot with the Italian Competition Authority, which was enthusiastic with the commitments. Under the previous President, it was using the commitment procedure more than almost any other competition authority in Europe. Finally, after fighting for years, the new President has understood that this system is corrupting the entire system, and the Authority is now trying to refrain from commitment decisions. It is resorting to them much less often than it did in the past. The next point is that we have been focusing mainly on public enforcement, but I think that an efficient and legitimate implementation of competition rules requires a healthy possibility to apply those rules in private enforcement, and I think that the connection between the two things is very important— between public and private enforcement. The movement towards a more full judicial review is also key to the success of private enforcement. As you know, Regulation 1/2003 specifically says that the Commission decisions are binding on national courts, in the sense that the national courts cannot take a decision contrary to the decision taken by the Commission. And the new very important piece of proposed legislation—the Directive of the Commission, the proposal for damage actions under national law—the proposed Directive published by the Commission proposes to extend this binding effect also to the final decisions of national competition authorities.25 Well, this c annot 25 Article 9 the final version of Directive 2014/104/EU is somewhat toned down compared to the draft in that it only requires a Member State to attribute preclusive effects to a final infringement decision of a competition authority or review court of that p articular Member State. The final infringement decision of a competition authority of a sister Member State must be recognised as at least prima facie evidence that an infringement has occurred.

Part IV: Issues for Courts and Perspectives on the Judicial Role 503 be done if those decisions are not subject to full judicial review. So then it is absolutely clear, no judge—certainly no judge in Italy—is going to feel obliged to follow a Commission decision or a national authority decision if that decision is not subject to full judicial review on the merits. It would be against our Constitution! It is absolutely impossible to impose that obligation on a judge. So, for me, this link shows why judicial review is so important: it is the centrepiece of the entire system. Now as I said, the Directive is very important, so it is a very welcome proposal by the Commission. The most important thing that the Directive tries to do is to open up the rules of disclosure—to give a minimum of rules of disclosure in litigation before national judges in actions for damages. In that regard, it gives a central role to the judge. It is the judge who has to order disclosure under the system proposed by the Directive, and the parties, of course, have to demonstrate that the information they seek is relevant to sustain their claim. They should specify the pieces of the evidence or the category of evidence desired. This is a very important issue, because in some Member States there are already similar provisions. A judge can order disclosure if the plaintiff identifies exactly the document or piece of evidence or the category of evidence which is needed. But unfortunately those rules have been interpreted very strictly by judges and so, again, this is going to be key. I wish the Directive were a little bit more radical in permitting judges to order production of evidence even if the identification of the piece or the categorisation of the evidence is not so clear. This is very important if we are serious in trying to encourage stand-alone litigation. Another important thing the Directive tries to do is to regulate access to the file of the Commission or the relevant competition authority. And it is very interesting that access to most of the file is possible except for corporate leniency statements and settlement submissions. And if a party seeks access to a company’s reply to the questions of the authority, it has to wait for the termination of the proceeding before the disclosure of that information is possible. So there are conditions for some types of evidence, but in general it can be done—even in cases terminated without a finding of infringement, and even in cases revolved by proposal and acceptance of commitments. Finally, I’ll address the issue of the kind of orders that judges and authorities can issue. We have very different practices and different positions in the various Member States with regard to interim measures and final orders. At the Commission level, following the IMS Health case,26 we never had any other interim decisions in the competition field. That also makes

26 See Commission Decision 2002/165/EC of 3 July 2001—NDC Health/IMS Health: Interim Measures [2002] OJ L59/18, withdrawn by Commission Decision 2003/741/EC of 13 August 2003—NDC Health/IMS Health: Interim Measures [2003] OJ L268/69.

504 Part IV: Issues for Courts and Perspectives on the Judicial Role sense—judges are better placed to consider applications for interim measures. It is too much to entrust the power to adopt interim measures to an authority which is already prosecutor, judge and executioner. The Commission should go to the General Court to ask for interim measures and national authorities should go to judges to ask for interim measures. You can obtain interim measures in two weeks in this country, notwithstanding all the defects of our judiciary. And, as we know, there are varying practices across national authorities. Our friend the Chairman of the French Competition Authority is a very enthusiastic user of interim measures, while other Member State authorities do not use interim measures at all; again, this is an area in which some form of coordination and cooperation would be very important. And finally, final orders! We know that Regulation 1/2003 gives to the Commission the possibility to impose behavioural and structural commitments. I suggest that Article 5 of the Regulation should also be interpreted as granting these rights to national authorities, which is very important. Most of all, I think judges should also be able to adopt positive orders, because very often a plaintiff, even if he succeeds on the merits, is left without an effective remedy if the judge is not in a position to order the conclusion of a new contract. So, as you see, there are a lot of things which are happening. The system is evolving. As my colleagues have said, the Workshop has played a role in the evolution of this system. I remember the discussions we had here on the modernisation of Community competition law. This was an important forum for discussion of that great project, and I’m sure we will be able to do that for the topics we have discussed today. Kovacic: Thank you very much, Mario. I suggest that we go to the discussion, and I’ll make sure that our judges have time to respond to the comments they’ve heard during the discussion. Fred, you’re first. DD

Fred Louis: Thank you. Let me begin by saying that there is not a single word in the interventions by Ian, Jim and Mario with which I disagree, which can be taken as a telling sign of a consensus. We really all feel this way. As soon as you are a user of the system, you very quickly to realise there is something wrong. I’m very grateful to Mario for mentioning the proposed Directive on damages actions; two years ago, when we discussed private enforcement here at the Workshop, not many of us would have expected that the Commission would come with such a balanced and careful text. Certainly, we all see things that could be improved, but it is a phenomenal proposal. If you read A rticle 5(4) of the proposed Directive, you will see that the Commission is actually suggesting that Member States should introduce protective orders at the national level. It says: DD

Member States shall ensure that national courts have at their disposal effective measures to protect confidential information from improper use to the greatest

Part IV: Issues for Courts and Perspectives on the Judicial Role 505 extent possible whilst also ensuring that relevant evidence containing such information is available in the action for damages.27

That’s not the case in Commission procedures right now! I agree that some aspects of the appellate procedures in Luxembourg put significant constraints on the General Court. One idea that would alleviate its work would be to get rid of the Commission’s right to have the final word; get rid of the final rejoinder. You will have the decision or appeal, the Commission’s defence and a reply—two sets of documents from each side. That is actually the French system in cases before the Court d’appel de Paris, and it works perfectly. The Commission is already allowed, without any restraint on the number of pages and without any time limits, to set out everything it thinks about the cartel or the abuse. Why is it allowed to have the final word in writing? So you could get rid of the rejoinder. If you look at the Statute of the Court of Justice, the word ‘rejoinder’ doesn’t appear. It says there must be an application, a defence and a reply; that’s all. I was very much struck by Judge Da Cruz Vilaça reminding us of the first years of the CFI because that is when I started working, and my first day in Luxembourg was in front of him in the Post Bank interim measures case.28 And yes, the CFI in the very first year was an incredibly exciting place—it built very solid foundations upon which, unfortunately, nothing was built. Another case I started was the Solvay case.29 In 1995, in its judgment, the CFI not only enshrined access to file as a fundamental right, it also justified it by introducing the concept of equality of arms. When you read the judgment it said the administrative procedure is already skewed enough so at some point you have to re-establish the balance between the enforcer and companies. Unfortunately, nothing was built on the foundations of that judgment. On the contrary, access to file was emasculated by the CFI in the Cement judgment—in 2009, when you reached the end stage of the Soda Ash case, which lasted 21 years in court, the General Court determined that the fact that the Commission had lost a part of its file, and was unable to say what it had lost, had no incidence.30 Now, the Court of Justice put an end to that, and I think really there is a lot of work that the Court of Justice can do and must do. Fundamental rights now need to be taken seriously in Luxembourg, contrary to the cosy situation that we had when the ECHR was merely a source of inspiration. At that time, the ECJ could state the principle but then

27 The original language was later softened to some extent, and now appears in rticle 5(3)(c) of the final Directive. The final version now states that national courts A ‘shall, in particular, consider [inter alia] whether the evidence the disclosure of which is sought contains confidential information, especially concerning any third parties, and what arrangements are in place for protecting such confidential information’. 28 Case T-353/94 R Postbank NV v Commission [1994] ECR II-1141. 29 Case T-30/91 Solvay SA v Commission [1995] ECR II-1775. 30 Case C-110/10 P Solvay SA v Commission [2011] ECR I-10439.

506 Part IV: Issues for Courts and Perspectives on the Judicial Role isregard the case law of the Strasbourg court. That is impossible now under d the Charter. So there is a lot that the Court of Justice can do, and I would suggest that the first opportunity to do this is precisely on the issue of undue delay. Because—what is undue delay? It is the violation of the fundamental right to be judged within a reasonable period of time by an independent judge. It is a fundamental right. It is ranked just as high, or perhaps higher, than the principle of competition. So when you are balancing the principle of effective competition with the fundamental right against undue delay, at the very least they have to be given the same weight. When you are thinking of what to do, who is the culprit here? The culprit is the European Union, which has not created a system where the judiciary has the means to decide quickly. The Court of Justice has recognised that, and it has asked for more means, but basically the Member States have answered: no, we’re not interested. So, we were told yesterday by Director-General Italianer that the fines go into the EU budget. Every year the EU budget gets at least a billion euros, and normally much more in fines, and we are told that the few million euros that it would take to have a proper system, no, the EU legislator will not give it to us. Here the Court can say: ‘Well, just like you are saying that to have effective respect of competition law you need five hundred million euros, when there is a violation of undue delay we are going to slash hundreds of millions of those fines’. Suddenly the Member States will see an interest in giving you the means to be able to judge within a reasonable time. Damien Gerard: Whenever I read Article 31 of Regulation 1/2003, I wonder why it is that we have this endless debate about the intensity of judicial review. Because the Article says the Court of Justice shall have unlimited jurisdiction to review decisions whereby the Commission has fixed a fine. So, it’s a matter of jurisdiction, not of review within the scope of legality review. This provision, Article 31, is derived from Article 261 of the Treaty on the Functioning of the European Union—so we are outside of legality review here. So nothing prevents the General Court from reforming Commission decisions imposing fines to substitute its own appreciation of the facts for that of the Commission.31 Of course, 30 years ago the Court of Justice decided to go another way. It decided to limit the meaning of jurisdiction to just the definition of the fines, and to the fact that it could increase or reduce the fine. But frankly, it doesn’t require Treaty change; it just requires a different interpretation of that provision and of Article 261 for the system to be changed and to address most of the concerns that were expressed today. DD

31 For further support of this view, see also Ian Forrester, ‘A Bush in Need of Pruning: the Luxuriant Growth of “Light Judicial Review”’, in Claus-Dieter Ehlermann and Mel Marquis, eds, European Competition Law Annual 2009: The Evaluation of Evidence and its Judicial Review in Competition Cases, Hart Publishing, 2011, 407 et seq, 423 and 447–449.

Part IV: Issues for Courts and Perspectives on the Judicial Role 507 It’s really a matter of a judge taking the responsibility to say the system has changed, we need to adapt our judicial review system. The time where the Court of Justice took the view that unlimited jurisdiction was only applicable to the amount of the fine, that was 35 years ago, before there were high fines and way before we started this system of negotiated enforcement. Today we have very high fines, which definitely justifies a review of the facts as well. We entered this brave new world of modernisation that brought us into what our American colleagues call the ‘regulatory model’ of competition enforcement, and which we call negotiated enforcement, and if we want to ensure the sustainability of that model we need to ensure that there is a real alternative to it. Parties need to be sure that, if they decide not to seek a negotiated solution, they defend themselves in an infringement procedure where there is a risk of fines but then if they are sanctioned they can have full appellate review before the EU Courts. So this is really about ensuring the sustainability of the whole enforcement system going forward. I really believe that it doesn’t require Treaty change; the tools are already there. Kovacic: We now have eight people in the queue, and so to make sure we give the panel a chance to respond I would request that all speakers limit themselves to two minutes, please. Next is Alberto. DD

Alberto Heimler: Thank you. I will be quick. The curriculum of continental European lawyers is usually very much enshrined in the law; there is not much economics. I just wondered if some of you judges sit on the bench with an economist, and I wondered what role economists play as judges. I think the role can be very important. If we look back at the record of the General Court, if an economist had been on the bench we would not have had Michelin II32 and we would not have had British Airways,33 at least not the way they were decided. So I just wondered whether there was scope to involve trained economists—not necessarily as advisers or experts, as Judge Da Cruz Vilaça was mentioning before in the context of the CFI, but as judges. Would that help to enhance the quality of the judgments? DD

Wolfgang Kirchhoff: I’d like to try to give an answer to Judge Barling’s first provocative question: whether we should move to a pure prosecutorial system. In view of the experiences in Germany over the decades, at least in Germany, we absolutely must maintain our administrative system. I think DD

32 Case T-203/01 Manufacture française des pneumatiques Michelin v Commission [2003] II-4071. 33 Case T-219/99 British Airways plc v Commission [2003] ECR II-5917; appeal dismissed: Case C-95/04 P, [2007] ECR I-2331.

508 Part IV: Issues for Courts and Perspectives on the Judicial Role that most Bundeskartellamt lawyers and the judges agree on that. We think the system is a more efficient way to put an end to an infringement. The decision bodies of the BKA are already quite independent, and as far as I know there are no relevant deficiencies as to the protection of fundamental rights. In view of these experiences, I would also be reluctant to apply Ian’s suggestion about splitting the Commission into an investigating body and a decision-making body. I am aware of the fact that stock-taking at the EU level may result in a different analysis, and that it might be possible. But at least in Germany I would not suggest such a change. Why put an end to a system which has proved to be efficient and fair for decades? I do not see that we have a need for harmonisation within the EU as to the organisation of decision bodies for competition infringement. We should leave that to the Member States; they have to decide which system they believe to be most appropriate, provided, of course, that it absolutely guarantees fundamental rights protection in full. Thank you. Stephen Calkins: I have heard lots of wonderful, interesting points. We were told that the remedy should be prospective only, we were told there should be much tougher judicial review and that we want clarity and certainty. A quick reaction to that would be that this sounds very fine to me, especially if there is meaningful private enforcement coming along. What you would not want is something whereby everybody is told: ‘You can go out and do whatever you want, get the benefits, and if you don’t get challenged, great. If you do get challenged then you’ll be told, five years later, okay now stop doing it.’ In terms of tougher judicial review, it’s really important to have legitimate judicial review. I will just remind you that there are costs. Finally, I hear the great call for clarity, and yesterday I heard the call for convergence. I nod my head and I say that’s all great, but I have two reactions of concern. One is that there is a book out there called Defending Your Brand,34 where the author, a business school professor, says: ‘Let’s be honest, firms are dominant, they know they’re dominant, they have an incredibly valuable product, and their goal when somebody comes along is to drive that firm into the ground. They want to steal all its ideas and put them out of business by whatever means, often foul.’ The book is written by my brother Tim Calkins. I recommend that you all go buy it and help put his children through college, but often the rules are very clear and it’s only when you are near the border that they become unclear. And when people say ‘Yes, let’s have clear rules, let’s have convergence’, I go back and think about the Leegin case,35 where DD

34 Tim Calkins, Defending Your Brand: How Smart Companies Use Defensive Strategies to Deal with Competitive Attacks, Palgrave Macmillan, 2012. 35 Leegin Creative Leather Products, Inc v PSKS, Inc, 551 US 877 (2007).

Part IV: Issues for Courts and Perspectives on the Judicial Role 509 the US Supreme Court departed from the nice, clear, per se illegality of resale price maintenance. The US actually messed up convergence there because it distanced itself from the European Union’s approach. Yet you didn’t hear any proponents of convergence and clarity saying, ‘Oh this is terrible, we have a difference of views now, we have all the uncertainty of the rule of reason, how terrible it is’. So I always take this call for clarity and convergence with a grain of salt. Philip Lowe: I’ve said it before, but, as a Director-General of the ommission, what I’m about to say does in fact engage me, but it doesn’t C necessarily reflect the views of my staff. I want to start with that disclaimer because Ian pointed out that there are some excellent EU officials. This is true, but they are not always right, and they have different views. Some of them seemingly believe that they are in a prosecutorial system, whereas in reality they are in an administrative system. So the first sociological remark is that there would be plenty of people inside the Commission who would actually favour some of the things I’ve heard the codgers say for the last four or five years, and I personally have some sympathy for that. I have some sympathy for it for another reason too, which is that we thought that through the views of the Courts in Luxembourg there was some rectification of names, because we pleaded for something like a rule of reason, for a more facts-based approach, particularly in mergers, and we established the concept of a theory of harm. And we improved internal processes to a significant extent, I think, with the establishment of the Chief Economist and his team, and a lot of other adversarial mechanisms inside the department. I’m not sure that going back to the idea of having a team for investigation and a team to decide the case would be a truly meaningful solution. We’ve looked at those issues, and internally we introduced devil’s advocate panels. There are a sufficient n umber of stages, in principle, within the institution for the Commission to make a risk assessment of the cases concerned and to anticipate not just the robustness of the decision for the first appeal, but for the second appeal as well, and to come to a political judgement of how to dispose of the case. But the role of the Hearing Officer can still be considered further. I think that Mario’s points on fines are interesting; at the moment, it is on an escrow account. Maybe for some of the companies concerned it doesn’t really make a difference whether they put the money into an escrow account now or later, but it would be a signal that, ultimately speaking, the review would look closely at the level of the fine and therefore there wasn’t an automatic presumption of guilt. So those ideas are interesting. I want to say something political, though. If you’re thinking that the European Union and its political institutions are about to envisage a major reform of the competition system, I’m sorry. We’re in the middle of an economic recession, we’ve got a banking crisis, we’ve got a lot things going on. You say two things: reform things while we’re ‘winning’—yes, that’s correct. DD

510 Part IV: Issues for Courts and Perspectives on the Judicial Role But in a political context, where there are a lot of things going on, this is not the first thing on the agenda of the C ommission. That being said, I come back to the issue of rectification of names. I find it very difficult to justify a system of law where indeed the defence of any case always has to go back to language from cases decided 20 or 30 years ago or more, when the doctrine and the economic theories that inform our evaluation of cases has changed. The Commission announces a policy change and we apply it, but the Courts may not; indeed, our own in-house legal agent does not defend the cases uniquely on the basis of what we’ve proposed, but insists that we actually include the old theology as well as the new theology. So you have to be both Catholic and Lutheran! It’s justification by works, but by faith as well! [Laughter] The Court is understandably confused by this, and often turns back to the basic doctrine of manifest error and control of legality. However, as was pointed out by the judges, there has been some movement here, particularly in relation to the complexity of economic theories of harm. That is to say, the basic discussion we had four or five years ago turned on the issue that the more complex your theory of harm, the stricter will be the review by the Court, certainly by the General Court. I think there is a structural disconnect which both the Commission and the Court are suffering from: a failure to adapt to the evolution of thinking which reflects best practice convergence in a multijurisdictional context. Often this means not leaping to per se judgments, while acknowledging in certain instances that per se illegality is the right way to go. I’ve been out of the case-related work for three or four years, but what I see is actually a fair degree of stalemate between the Commission and the Courts regarding the basis on which the judgment should be built—the theories of harm, the economics and the legal reasoning behind it. I think that the next Commission will have to look at this, perhaps in a better economic atmosphere. It might even be able to undertake a reform of the system. But let me clarify that I am not representing the Commission with these particular remarks. My speculations are what you might call the ‘futuristic’ view. Ioannis Lianos: I found the parallels made by Mario Siragusa between private enforcement and public enforcement were quite interesting, because I thought of fines, the equivalent of fines, I would say could be treble damages, in the sense that they are substitutable remedies because they put a price tag on an infringement and they have a cumulative dimension. From that perspective, if we look a little bit at the way the debate goes on in the US, there isn’t a legitimacy crisis there with regard to treble damages. If you compare that with the debate we have in Europe and the legitimacy crisis we have in the EU, that tells us something about how the US system may be perceived as more fair, more independent and more impartial than the administrative system. But unfortunately in the EU we are trapped in a bad equilibrium. We have this administrative system—and, judging by Philip’s remarks, that DD

Part IV: Issues for Courts and Perspectives on the Judicial Role 511 is unlikely to change. So we need to think of ways to play with the system. Now there are several possible options, which can be cumulative. The first one follows on from what Mario Siragusa was saying. The idea is to suspend some of the negative effects of the system. I heard your proposal about, for example, suspending the payment of the fines, and this is something we also advise in a recent paper.36 Or we could suspend the application of the Commission’s decision pending a final review process. In Germany, there is a procedure of administrative self-regulation where there is a mandatory review step that takes place before the aggrieved party files a court action, and maybe something like that could be used.37 The second possibility would be to change the inner structure of the Commission, or at least the incentives that it has. That can be done in a manner similar to the suggestions we have heard here: separating the investigation function from the decision-making function. We also recommend the adoption of a ‘European Administrative Procedure Act’,38 which would develop further the administrative process rights for individuals and companies. In particular, it would include possibilities for longer oral hearings, crossexamination and so on, which would create more legitimacy in the process. Finally, improvements may be made at the level of the judiciary. In the General Court’s rules of procedure there is the possibility to appoint experts, but this has almost never been used by the General Court; the Court of Justice has made use of it only a couple of times. So, there are actually possibilities there to advance the intensity of the review, but at the same time, of course, there is a problem with the control of legality in the sense that it stops short of a review of the merits. You can see that, for example, in the General Court’s Ryanair judgment.39 The court tried to engage with the technical evidence and econometrics in this particular case, but then they said ‘Well, we know the plaintiffs’ arguments concerning the particular methods used by the C ommission, and the Commission actually engaged with these arguments in its decision’. Therefore—deference, end of story. Now, what we expect is that if you move to a specialised tribunal with full merits review, the Court will engage with the internal validity of this econometric evidence, but having looked a little bit to the cases of the Competition Appeal Tribunal— I haven’t actually looked at all of them—but I cannot actually see where the CAT has also done this type of intensive review, engaging with the

36 See Ioannis Lianos and Arianna Andreangeli, ‘The European Union: The Competition Law System and the Union’s Norms’, in Eleanor Fox and Michael Trebilcock, eds, The Design of Competition Law Institutions: Global Norms, Local Choices, Oxford University Press, 2013, chapter 9, 414 (citing Anne Macgregor and Bogdan Gecic, ‘Due Process in EU Competition Cases Following the Introduction of the New Best Practices Guidelines on Antitrust Proceedings’, 3 Journal of European Competition Law and Practice 425 (2012)). 37 See Lianos and Andreangeli, ibid. 38 Ibid, 442, with references to other authors making similar suggestions. 39 Case T-342/07 Ryanair Holdings plc v Commission [2010] ECR II-3457.

512 Part IV: Issues for Courts and Perspectives on the Judicial Role internal–external validity of the econometric evidence presented to it. So maybe the problem is the capacity of judges to do this type of work on their own. Maybe it is better for them to use more frequently the possibility to appoint experts, or the more direct option raised by Alberto, the possibility of having economists appointed as judges.40 I’ve written a paper comparing decisions of the European Courts with some US cases, where you see a lot of engagement with the e conometric evidence, particularly the internal–external validity of the evidence.41 So courts can sometimes scrutinise closely that type of evidence; they are not structurally incapable of doing that. Giorgio Monti: Just one point on the suggestion that undertakings abusing a dominant position should not be fined: I think there is a risk that that would create a loophole tempting dominant firms to step over the line and engage more often in abusive conduct. Furthermore, Article 9 is not really a substitute for an infringement decision because, after all, what we see with Article 9, as Damien suggested today, is that competition authorities look for the remedy and don’t really look for a theory of harm, and there is very little judges can do to control the commitments, so even going down that route is problematic. What about the suggestion to tone down the concept of dominance? If you look at the Guidance Paper, there are two paragraphs that define dominance that are in conflict with one another: one defines it as substantial market power; the other defines it the old-fashioned way, based on La Roche.42 If you focused on the first paragraph, you’d actually eliminate a lot of cases from the docket. DD

40 Article 253 TFEU provides: ‘The Judges and Advocates-General of the Court of J ustice shall be chosen from persons whose independence is beyond doubt and who possess the qualifications required for appointment to the highest judicial offices in their respective countries or who are jurisconsults of recognised competence’. In the case of judges from many Member States, this would seem to exclude pure economists, but there may be Member States that allow a more diverse range of skills to assume the highest judicial office (if one thinks, for example, of the Cour de cassation of France). Under Article 32 of the original ECSC Treaty, it was not required for judges to have any qualifications in law. The appointment of pure economists was thus permitted. The only true economist (though one could hardly call him an industrial organisation specialist) that served as a member of the EU Courts was Jacques Rueff, the original French Judge of the Court of Justice (1952–1962). (Dutchman Petrus Serrarens, like Rueff, lacked formal education in law.) A small number of members of the EU Courts have had postgraduate degrees in economics (José Luis Da Cruz Vilaça, Niall Fennelly and Josef Azizi), but this is a far cry from the more adventurous idea of appointing professional economists to bring a very different perspective with the aim of enriching and perhaps enhancing the decision-making process. For a positive view of multi-disciplinary courts, see the comments of Gerald Barling below. 41 See Ioannis Lianos and Christos Genakos, ‘Econometric Evidence in EU Competition Law: An Empirical and Theoretical Analysis’, in Ioanis Lianos and Damien Geradin, eds, Handbook on European Competition Law: Enforcement and Procedure, Edward Elgar, 2013, chapter 1. 42 Case 85/76 Hoffmann-La Roche & Co AG v Commission [1979] ECR 461.

Part IV: Issues for Courts and Perspectives on the Judicial Role 513 Renato Nazzini: I have a suggestion which has benefits but doesn’t require any Treaty or legislative changes, or any extra expenditure. It relates to the function of the Court of Justice in ensuring the correct interpretation of the law within the European Union. In particular, when the Court of Justice does change its jurisprudence, and it does so from time to time even in the competition law field, although it always refers to very antiquated cases, the Court never or at least hardly ever tells us why it is changing the case law. And it never gives us reasons when it decides not to change the law, although there are very good arguments for doing so. I’ll just give two quick examples. First, in Post Denmark,43 the Court includes in the classical definition of abuse a new and very important phrase about detriment to the consumer. The Court doesn’t tell us why, and it’s anybody’s guess what implications that has for the test for determining when a dominant company infringes Article 102 or not. Is it just making explicit what was implicit before, and nothing has changed? Is it just a statement of objectives—which was already in Continental Can44 in the sense that the provision is concerned with both direct and indirect harm to consumers? Or is it a more fundamental change? We simply don’t know. The Court really should just tell us. And just as the Court should tell us why it changes the language and perhaps the law, it should also tell us why it chooses not to change its case law. The example here, very quickly, is Toshiba,45 the case I was referring to yesterday on ne bis in idem. In that case, the parties were arguing that you should not apply this test because now you, the Court of Justice, have case law in another area of the law which does not apply the test. You also have Article 4, Protocol 7 of the ECHR, which does not apply the test. And you have the Advocate General saying you should not apply this test any longer. But the Court of Justice restates the same test, without telling us why it decided not to change the jurisprudence. Maybe there are good reasons, but we wouldn’t have any way to know that. So until the Court changes its methods and begins to provide more transparency about its reasoning, I’m afraid it will remain deficient with regard to some of its primary functions: ensuring legal certainty, developing the law and making clear why certain decisions are made.

DD

43 Case C-209/10 Post Danmark A/S v Konkurrencerådet, EU:C:2012:172, para 24. (See also ibid, para 44 and the operative part of the judgment.) For recent observations underlining, as Nazzini does, that it is not clear from the judgment exactly how the law has changed, see also Victoria Daskalova, ‘Consumer Welfare in EU Competition Law: What Is It (Not) About?’, TILEC Discussion Paper DP 2015-011. The paper does not directly discuss the Court’s new gloss, in paragraph 24 of Post Danmark, on the classic La Roche formula. 44 Case 6-72 Europemballage Corporation and Continental Can Company Inc v Commission [1973] ECR 215. 45 Case C-17/10 Toshiba Corporation and Others v Úřad pro ochranu hospodářské soutěže, EU:C:2011:72.

514 Part IV: Issues for Courts and Perspectives on the Judicial Role Hwang Lee: I would note first of all that there is a race between the Korean Fair Trade Commission and the Japanese Fair Trade Commission to influence various Asian countries in their enforcement of competition law. We have also witnessed that kind of competition between the European Union and the United States, and the European Union seems to be winning, especially in the Asian region.46 My comment is that the issue of the nature of the judicial review of the European Commission’s decision is therefore relevant as a reference point for many Asian countries. It seems that in the Asian region, Korea may be the only country where the competition authority’s decision is subject to very tight judicial review. Many companies say to KFTC officials, ‘I will see you in court’, and there is high probability that the KFTC’s decision will be overturned. The situation is different in Japan and China. In many cases, the Japanese authorities, and especially the Chinese authorities, say they are not going to depend on a heavy judicial review system. The US and the EU are doing a very good job with active judicial review, and in Korea we are confident that we can do similar things. But, especially at this early stage of establishing competition law in China, where all the political and administrative power is concentrated in the administration and the Communist Party, if there is no meaningful process of checking and control by the judiciary, there is a significant danger that administrative decisions will be flawed and may in fact be anticompetitive. So, once again, at least to some extent, the example set in the EU may be in other parts of the world. Rigorous judicial review here in Europe may potentially lead to positive developments elsewhere, for the benefit of both effectiveness and legitimacy. DD

Kovacic: Many thanks to all of you for your comments. Before we nished, let us give any of our judges who would like to do so the chance to fi come in with a final word on our discussions. First, Judge Barling. DD

Barling: Thank you, and I can be very quick. As I said to Savvas y esterday, the judges of the General Court are in my humble view engaged in a heroic task, given the work and the workload that they have. In that regard, the comments of Judge Da Cruz Vilaça are very much in point. My comments or questions were really related exclusively to the system. In fact, my first question was whether or not the problems might be rooted not so much in the intensity of judicial review as in the administrative system itself. If you change the administrative system, arguably the intensity of review problem just goes away. But I was not in any way criticising any national administrative systems, such as the German administrative system, because the German system quite DD

46 For extended discussion of the global competition between the US and the EU competition law systems, see Mel Marquis, ‘Idea Merchants and Paradigm Peddlers in Global Antitrust’, 28 Pacific McGeorge Global Business and Development Law Journal 155 (2015).

Part IV: Issues for Courts and Perspectives on the Judicial Role 515 clearly incorporates all the safeguards for criminal cases because it has a full de novo review of whatever has been decided. Savvas made the very good point that the parties do not necessarily always do their job in the General Court. He mentioned that they rarely bring in experts, and I wonder why that is. Maybe the parties are not always as proactive as they might be, and maybe they don’t always take full advantage of the General Court’s procedures. I do remember cases before the CFI where evidence was heard, witnesses were cross-examined and expert witnesses participated. As for Alberto’s question about the use of economists, the short answer is yes, they are invaluable as members of a competition court—that’s my experience. I have sat in the CAT with a multi-disciplinary bench. The economist judges play a full role in the decision making: they are full judges and, of course, when economic issues are involved they have tremendous insight that would be very unlikely to occur to a non-economist. So I do recommend that. Finally, in reference to Philip’s remarks about fines: does it hurt anyone so much if the money is put in escrow? He may have a point, because I remember being told by a company which had to wait many years before the case was finally resolved that they had made more money out of the interest on the account than they made doing their business during the same period. So maybe there is some force in what you say! Papasavvas: First of all, with reference to Fred’s comments, I would not reject, in principle, abolishing the rejoinder, provided, of course, that the applicant respects the logic of the reply. As for transforming the European system of judicial review into a full merits review—it can’t be done. That would be completely outside the logic of an administrative system, and it wouldn’t work. Now, on the issue of economists on the bench, economists could be useful, but competition is not the only thing we do at the General Court. Now, if you can find a candidate that has perfect knowledge of French, perfect knowledge of EU law, good knowledge of French administrative law and on top of that is an economist too, well, fair enough. But such a beast doesn’t appear often, and even if he exists that does not necessarily mean that the Member State he originates from will propose him for the Courts. With regard to suspensory effect, this would multiply appeals to the Court and it would multiply the grounds for appeal. It’s an interesting idea, but I don’t think that it could work in practice. Finally, Ian Forrester referred to the n umber of pages that we allow or don’t allow. Let me be clear: we are sufficiently flexible, when necessary, to authorise parties to exceed those 50 pages, especially for complex competition law cases. But when a lawyer on appeal from a decision of the Civil Service Tribunal brings me a 100-page appeal and I’m obliged not only to live with it but to answer it, there I have a problem. That only creates congestion that blocks our daily functions and affects the DD

516 Part IV: Issues for Courts and Perspectives on the Judicial Role way we treat competition law cases as well. So this is why I insist on saying yes, there is flexibility when it is really needed and suited to the needs of the case, as it may well be in a competition matter. So, in brief, those are my observations. Thank you. Da Cruz Vilaça: I’ll make just a couple of very telegraphic remarks. First, I want to thank Fred for joining me in my historical reminiscences. All the founders of the CFI have left, although two of them remain at the ECJ.47 But we do have good mutual understanding between the two Courts. I would like to thank my colleagues at the General Court for carrying out a huge task of coping with the backlog that accumulated during a certain period, and they are making important progress catching up. This is, of course, a crucial point as regards the legitimacy of the administration of justice. As regards the ECJ, I feel comfortable because in fact we don’t have a backlog. We are keeping up with the cases lodged with the court, which means around 600 matters per year. Yesterday and today we have heard references to unlimited jurisdiction, and I will submit that unlimited jurisdiction is really not the same as full judicial review. Unlimited jurisdiction means that fines can be annulled, reduced or increased. Unlimited jurisdiction is supposed to be the equivalent of the French concept, compétence de pleine jurisdiction. Full judicial review is something we have in Portugal—it applies in a procedure which is half a dministrative, half criminal. Now, is the Portuguese system working better than the EU system? Frankly, I do not think so at all. The factors of efficiency are multiple—it is not just a question of whether or not there is full judicial review. Finally, let me touch once again on the question of economists. I had an Italian colleague who in a hearing once quoted an old maxim from Roman law: Iudex non calculat.48 We have a wide range of legal backgrounds in both Courts, which is really a very good thing for the efficiency and richness of our case law. However, I agree that it’s good for judges deciding competition cases to have a good background in economics. The Court of Justice in the beginning had a French judge who was mainly an economist.49 My own doctoral degree is in international economics and that background is very helpful. It has served me very well as Advocate General and as Judge, and so I would DD

A reference to Judge Lenaerts and to Judge Da Cruz Vilaça himself. According a Wikipedia, Iudex non calculat is ‘a legal term that loosely translates as “The judge does not calculate”. It originates from the Roman legal concept that obvious calculation errors in a court decision are not harmful to the decision itself and can be corrected at any time. Figuratively it also describes the concept that a judge does not “add up” the number of arguments but instead bases their decision on the strength of those arguments. Jokingly, it is often used to mean “judges (or jurists) are unable to do the math”, often by jurists themselves,’ 49 See above n 40. 47 48

Part IV: Issues for Courts and Perspectives on the Judicial Role 517 agree that multi-disciplinary training is desirable. But at the most essential level, our task is that of judging legal disputes. So it is helpful to have diverse skills and qualifications, but we must also remember that these are mainly courts of law. Thank you. Kovacic: Jim Venit said that good systems require forums for discussion like this and I think he is certainly right. Any system that doesn’t have a good mechanism for discussion and dialogue is unlikely to improve itself and achieve superior results. With that remark, let me thank everyone for making this event possible, and for contributing to this session. For the very final of the final remarks, I yield the floor to Philip. DD

Closing remarks Lowe: Many thanks to Bill for chairing this last session, and thanks to everyone who participated in this morning’s discussions. I will not summarise what we’ve been speaking about this morning, but I will sum up the rest of the two days—it will be useful, I think, for us to recall some points that came up yesterday. During our discussions, the Workshop has examined most of the main determinants of legitimate and effective competition law enforcement by both competition authorities and courts. In the first panel session it was quickly recognised that, beyond the adequacy of its legal power, a competition authority’s public legitimacy depends on its capacity to intervene in the market effectively in order to protect competition and promote consumer welfare. The concern to demonstrate this effectiveness to the public and at a political level can lead authorities into action such as high-profile investigations and high fines. This raises the risk that the activities of an authority may not respect the principles of proportionality and necessity, or they may infringe principles of due process. If this risk materialises, it undermines the legitimacy of the authority, and this will in turn lead to a curtailment of its activities and its powers—and ultimately its effectiveness. A number of participants questioned the proportionality of high fines and their legitimacy, and questioned their effectiveness in preventing or addressing further repeat offences. The argument was made that the effectiveness and legitimacy of the final fines imposed on individual cartel participants was dubious since they bear little relation to individual responsibility, which needs to be sanctioned. The legitimacy of any fines at all for abuse of dominance was also questioned except for repeat offences, on the ground that recourse to industry-wide practices such as loyalty rebates could not be regarded per se as anticompetitive. Private enforcement was also available as a further deterrent. However, not everyone DD

518 Part IV: Issues for Courts and Perspectives on the Judicial Role agreed that fines should be taken completely off the table in abuse of dominance cases. There were calls for stricter and deeper judicial review of fines, with some favouring full merits review, and it was also stressed that authorities should take account of efforts made by defendants to avoid infringements, including in particular in the form of compliance programmes. Overall, the interventions of the Workshop participants emphasised the danger of overintrusive and disproportionate enforcement, which again may potentially undermine the legitimacy of enforcement. At the same time, it was acknowledged that, given low detection rates, action against cartels and the need to respect due process requirements already implies a heavy and drawn-out process. Any requirement to prove market effects as a prerequisite to sanctioning patently anticompetitive behaviour would unhelpfully reduce effectiveness. On procedural effectiveness, there was general agreement that the lack of harmonisation of procedural rules across the national jurisdictions of the EU was a major obstacle to effective enforcement of Regulation 1/2003 by the European Commission and by the national competition authorities of the EU. This lack of harmonisation was a serious impediment for authorities, but it is also a burden for both complainants and defending parties. With excessive complexity, in particular as far as leniency applications are concerned, the lack of clarity on antitrust liability in the context of legal succession was also highlighted as preventing effective law enforcement, in particular in Germany. This could be addressed through proposals for further harmonisation of national processes across the EU.50 Fines were seen as only one remedy for a competition infringement. Beyond the requirement to bring an infringement to an end, behavioural and structural remedies may be imposed and enforced. At the level of the EU, this can be done either through an Article 7 decision, often accompanied by fines, or by way of an Article 9 commitment decision without fines. The Workshop noted the increased use of commitment decisions in non-cartel cases by the Commission, an option that has been equally preferred by many parties for reasons of procedural expediency and for the absence of any finding of an infringement. However, several participants warned of the dangers of parties being pressured to offer commitments that went beyond the legal commitment requirements, in tension with the principles of proportionality and necessity. At the same time, it was acknowledged that voluntary commitments proposed by the parties frequently resulted in significant sector-wide benefits for consumers. Genuinely constructive discussions between the parties and 50 The discussion on prospects for procedural and institutional harmonisation—but also for further soft convergence—in the competition sphere has, of course, progressed. Together with collateral staff documents, see Commission, ‘Ten Years of Antitrust Enforcement under Regulation 1/2003: Achievements and Future Perspectives’, COM(2014) 453 final of 9 July 2014. See also Alexander Italianer, ‘Completing Convergence’, speech, Rome, 10 October 2014, 5–7.

Part IV: Issues for Courts and Perspectives on the Judicial Role 519 the authorities on remedies can enhance the effectiveness of law enforcement, on the understanding that if no agreement is reached the case may revert to the normal Article 7 procedure, with due regard for the rights of defence. In response to concerns that antitrust enforcement may assume an illegitimate regulatory character, emphasis was placed on the need for competition authorities to take a holistic view as to the most effective remedy to a competition problem—whether an antitrust case, the action of a regulator or legislative action. An effective remedy must resolve the competition problem within a time period relevant to the economic harm which it has caused and is likely to cause in the future. In the second session, the Workshop examined the potential benefits of amicus curiae interventions. In the US, amicus briefs are a widely used tool both in public and private enforcement. In the EU, the Commission has developed an efficient amicus practice under Regulation 1/2003 to contribute to the assessment of cases initiated by national authorities and reviewed by national courts. However, given its limited resources, the Commission must inevitably establish priorities as to when and in what circumstances to intervene. If the proposed private enforcement regime is to be effective beyond follow-on actions, the capacity for the Commission to respond to amicus requests in privately initiated cases will become critical. The Workshop then addressed the issue of effective control of state intervention, in particular in the EU and in the US. Recourse to the application of Article 106 TFEU by the Commission has been limited, even though, in a prolonged period of economic recession, there may be a strong and generally unjustified tendency for state intervention to restrict competition. In the US, as well as the EU, following initial initiatives to liberalise utility sectors, there has been understandable deference to democratic processes, under the assumption that any restriction of competition should be duly and transparently justified ex ante and on judicial appeal. Scrutiny of non-competition public policy objectives may be ensured either through a ‘hard look’ process or through formal competition impact assessments, or through wider market investigations, such as under the UK competition regime or other similar regimes. In the third session we returned to the bread-and-butter job of competition authorities and the challenges of parallel enforcement of competition law, both within the structured framework of the European Competition Network and between jurisdictions throughout the world. We also addressed the issue of agency priorities and case allocation. As far as parallel enforcement by the Commission and by one or more national authorities in Europe was concerned, there was a recognition that cases would legitimately be distinguished according to the effects in different jurisdictions and in relation to the time period relevant to the related infringements. However, fragmentation of cases beyond these parameters and in relation to the same facts may infringe the ne bis in idem principle and should be avoided—in respect of

520 Part IV: Issues for Courts and Perspectives on the Judicial Role both infringement and acquittal decisions. This does not preclude the possibility that one authority could be identified as best placed to handle a case, including in its work an assessment of impacts in other jurisdictions which may agree to defer to that best-placed authority. It was nevertheless generally accepted that full compliance with the ne bis in idem principle was predicated on EU-wide procedural harmonisation, including in particular with regard to the rights of defence and the rights of complainants. In this context, the Workshop again underlined the urgent need for the EU to progress beyond the current Model Leniency Programme to achieve a streamlined EU-wide system for leniency applications. It was recognised that, as far as competition authorities themselves are concerned, their effectiveness and legitimacy are heavily dependent on their ability to define their enforcement objectives clearly and to establish measurable indicators of their performance. Some activities, in particular, review procedures such as merger control, which are subject to legal deadlines, allow limited room for prioritisation, while other areas in antitrust require a focused approach. Case selection and internal control processes also depend on a holistic examination of what is the most effective resolution of a competition problem, whether through antitrust action or a regulatory response. Authorities can lose effectiveness and legitimacy by concentrating on high-profile cases to the detriment of a wider range of cases which can strengthen their enforcement record and provide a wider base of political and public support for their action. Giving some profile to particular cases, however, can help to communicate the practical benefits of enforcement action for consumers. The Workshop then turned its attention to parallel law enforcement and cooperation between national authorities and across the world. There was general acceptance that waivers for exchange of information between authorities can increase the effectiveness and efficiency of investigation, in particular for mergers. For cartels, the interface for administrative and criminal regimes still constitutes an obstacle to the useful exchange of information. The importance of inter-jurisdictional remedies was stressed. Some remedies are very specific to problems in particular national markets, but often worldwide remedies make more sense to the parties and more sense for consumer welfare. There was some evidence of effective cooperation on a case-by-case basis between the major jurisdictions, both on substantive issues and on procedural and timing issues. However, there was general concern that, despite the substantial progress made by virtue of, among other things, the original OECD Recommendation, and despite the establishment and work of the International Competition Network, there are now more than 120 competition authorities in the world and the convergence of standards and procedures is decreasing, not increasing. Many new authorities, in particular in major new, emerging countries, are reinforcing their competition laws and enforcing them rigorously—frequently with a domestic and industrial policy-related focus. At the same time, markets continue to globalise. It is urgent to use the

Part IV: Issues for Courts and Perspectives on the Judicial Role 521 processes of the OECD and the ICN to re-establish the path towards convergence on both substantive and procedural issues, in particular for mergers. That’s a partial summary, and later I’ll circulate my conclusions, including a summary of this morning’s further discussions. Many thanks to all the participants who have taken part so actively and vigorously—even passionately, in some cases—over the last 24 hours. Thanks also once again to Giorgio and Mel for the organisation of this event as well as some of the thinking behind our programme. We certainly had a nice wide range of issues which brought everyone together in the process, and some very important lessons have come out of it. I won’t keep anyone from lunch, I just want to wish you every success until we meet again. Thank you and goodbye for now!

522

Gerald Barling* The UK Competition Regime: Developments and Further Proposals for Change

1. Introduction The tectonic plates of the UK competition world have been on the move throughout my incumbency at the Competition Appeal Tribunal (CAT), with a bewildering number of reviews, consultations and concrete proposals across a number of related areas. In recent times the shaking and quaking have reached the very highest readings on the Richter scale. But the earthquake analogy should not be pushed too far, because many of the proposals are actually very positive for the UK system, even if some are less obviously so. Although the core of this chapter will consist of some comments on the changes which are in train with a view to beefing up the private enforcement of the competition rules in the UK, I will also make some observations about possible changes in the public enforcement arena. It is not possible entirely to separate public and private enforcement. For the most part, private actions are, and will continue to be, so-called ‘follow-on’ claims, wholly or mainly dependent on an infringement decision of the national competition authorities or the European Commission. Yet, as the 2013 proposal by the Commission for a directive on private enforcement (and now the Directive itself) has highlighted, there is a tension between public and private enforcement: on the one hand, a claimant’s requirement is for as much documentary disclosure as possible in order to prove its case; on the other hand, it is desirable to provide appropriate protection for leniency regimes at both EU and national levels. The conflict is capable of resolution because without leniency schemes there would be many fewer cartel decisions, and correspondingly fewer follow-on actions.

* A Judge of the High Court of England & Wales, and President of the UK Competition Appeal Tribunal (CAT). The content of this chapter represents the personal view of the author and not that of the CAT, the High Court or anyone else. It is based in part on the David Vaughan CBE, QC/Clifford Chance Annual Lecture on Anti-Trust Litigation, given by the author on 19 June 2013. In the time since the chapter was completed, significant developments have occurred in the field of UK competition law enforcement, including not least the actual launch of the Competition and Markets Authority, and the adoption of the Consumer Rights Act 2015; such reforms are not herein discussed retrospectively.

524 Gerald Barling In connection with proposed changes in the area of public enforcement, I will mention briefly the current UK government consultation on potentially significant procedural and jurisdictional changes affecting regulatory and competition appeals in the CAT and other UK appeal bodies.1 Although this is largely concerned with appeals from ex ante decisions of sectoral regulators, appeals against infringement decisions under both the EU and UK domestic competition rules are included within its scope. I will do no more than touch on the major shake-up to the UK’s competition institutions, and in particular the forthcoming replacement of the Competition Commission and the Office of Fair Trading by the new Competition and Markets Authority (CMA). Finally, I will say a word about the position of the CAT, of which I have been privileged to be President for nearly six years, and in particular about the importance of safeguarding judicial independence in our competition and regulatory system amidst the current cascade of institutional and other developments and reviews.

2. Institutional changes In April 2014, the Competition Commission and the Office of Fair Trading will be replaced by the new CMA, which will combine many of the characteristics of its predecessor institutions. When the merger was still the subject of consultation, one issue debated was whether the present administrative procedure should be retained for antitrust investigations, or whether instead of taking the decision itself the new authority should bring the alleged infringement before the CAT, so that the decision would be a judicial rather than an administrative one, and the role of the CMA would be that of a prosecutor rather than decision-maker. There were cogent arguments on each side of the debate. The main one in favour of retention of the administrative system was that the UK had had more than 10 years’ experience of that system, and that it was also the longstanding approach in the EU itself. Against this argument, a prosecutorial system would avoid the unsatisfactory situation where the decision-maker is also the investigator and prosecutor, and would enable the authority to discharge unequivocally the role of a prosecutor. Further, the current position, in both its EU and national manifestations, is now frequently the object of critical comment, not least because of the very large penalties and other legal consequences which can result from a finding of infringement. Compliance

1 ‘Streamlining Regulatory and Competition Appeals: Consultation on Options for Reform’ 19 June 2013.

The UK Competition Regime 525 of the administrative system with the ECHR has also been called into question by commentators. In the event, the UK government has opted for retention of the current administrative approach.

3. Private enforcement/CAT jurisdiction By far the most significant and positive development in the field of private enforcement in the UK since the advent of the Competition Act 1998 and the creation of the CAT is the government’s decision this year to establish the CAT as (in the government’s words) ‘a major venue’ for private enforcement of the competition rules in the UK. These proposals, if implemented, will, I believe, have a real impact on the effectiveness of our system.

3.1 Damages actions in the CAT—current position Well intentioned though it was, the follow-on damages jurisdiction afforded to the CAT by section 47A of the Competition Act 1998 (introduced by the Enterprise Act 2002) has proved something of a disappointment. The anomalous restrictions to which the power of the CAT to determine these claims is subject were not, perhaps, fully appreciated when the statutory provisions were framed. Their effect, as explained in an array of CAT and Court of Appeal decisions, is that, in the context of a claim for damages, the specialist competition tribunal has no power to decide whether an infringement of the competition rules has taken place; it cannot make any findings beyond the four corners of the authority’s infringement decision, and its jurisdiction is restricted to determining issues of causation and quantum of loss. As Lloyd LJ observed in Enron Coal Services Ltd (in liquidation) v English Welsh & Scottish Railway Ltd: ‘It seems somewhat anomalous that the specialist tribunal is entrusted with the decision as to infringement or no on an appeal from a regulator, but is not allowed to touch that question in a claim for damages … That can have a significant limiting effect on the scope of proceedings under section 47A’.2

And, as Patten LJ made plain in an earlier judgment in the Enron saga: It is not open to a claimant … to seek to recover damages through the medium of section 47A simply by identifying findings of fact which could arguably amount to

[2011] EWCA Civ 2, [143].

2

526 Gerald Barling such an infringement. No right of action exists unless the regulator has actually decided that such conduct constitutes an infringement … The Tribunal ought therefore … to be astute to recognise and reject cases where there is no clearly identifiable finding of infringement and where they are in effect being asked to make their own judgment on that issue.3

The illogicality of this situation is underscored by the fact, to which Lloyd LJ drew attention in the passage I have quoted, that in the context of its role as a tribunal sitting on an appeal from an infringement, or indeed from a noninfringement, decision, the CAT has full jurisdiction to determine the question whether or not an infringement has taken place. This jurisdictional disability, together with other procedural differences between a follow-on claim in the CAT and a similar claim in the High Court, have meant that the CAT’s jurisdiction in private enforcement has not been used as much as I believe it would have been but for these constraints. Only three follow-on damages actions have proceeded to trial before the CAT (Enron,4 2 Travel5 and Albion Water).6 Of course, that is three more than have reached that stage in the High Court—and in two of the three CAT actions damages were awarded—an event which has yet to occur in the High Court. Other such claims have been settled and/or withdrawn (eg Vion Food Group7 and Moy Park),8 or are mired in procedural and jurisdictional challenges, mainly as a result of the statutory constraints to which I have referred (eg Emerson,9 Deutsche Bahn10 and BCL Old).11 Throughout my time as President, we have pressed for jurisdiction over stand-alone and hybrid claims to be given to the CAT. This is now bearing fruit. The landmark reforms proposed by the government in its response to a consultation12 will, as I have said, mark the single biggest reform to UK competition law in recent years. The government has now published the text of the Consumer Rights Bill,13 schedule 7 of which contains provisions which

3 English Welsh & Scottish Railway Ltd v Enron Coal Services Ltd [2009] EWCA Civ 647, [31]; see also per Carnwath LJ, 64. 4 Case No 1106/5/7/08, above n 2. 5 Case No 1178/5/7/11, 2 Travel Group PLC (in liquidation) v Cardiff City Transport Services Limited [2012] CAT 19. 6 Case No 1166/5/7/10, Albion Water Limited v Dŵr Cymru Cyfyngedig [2013] CAT 6. 7 Case No 1201/5/7/12, Vion Food Group Limited & Ors v Tessenderlo Chemie NV and Britphos Limited. 8 Case No 1147/5/7/09, Moy Park Limited & Ors v Evonik Degussa GmbH and Degussa Limited. 9 Case No 1077/5/7/07, Emerson Electric Co & Ors v Morgan Crucible Company PLC. 10 Case No 1173/5/7/10, Deutsche Bahn AG & Ors v Morgan Crucible Company PLC & Ors. 11 Case No 1098/5/7/08, BCL Old Co Limited & Ors v BASF SE & Ors. 12 Private Actions in Competition Law: A Consultation on Options for Reform— GovernmentResponse, January 2013, 5. 13 Published by the Department for Business, Innovation and Skills (BIS) on 12 June 2013 (BIS/13/925).

The UK Competition Regime 527 will rationalise and significantly enhance the CAT’s private enforcement role. If and when they become law,14 these measures will enable the CAT to entertain stand-alone claims on the same basis as the High Court and, more controversially, opt-out collective actions. By coincidence, only the day before its publication, the European Commission published its own long-awaited proposals on related topics. From the UK perspective, the proposed changes to national procedures and jurisdiction are of greater significance. Nevertheless, I will say a word about the EU Directive too.

3.1.1 Stand-alone/hybrid claims for damages The proposed extension of the CAT’s jurisdiction to hear stand-alone as well as follow-on actions is to be reflected in a substituted section 47A of the Competition Act 1998, covering actions founded on alleged and not just established infringements of the prohibitions, as at present. This is a most welcome reform, enabling the CAT at last to address the question of infringement in private claims and not just in public enforcement cases. It will also allow the CAT to make better use of the considerable expertise in its panels of Chairmen and Members. The government has recognised that this specialist expertise, combined with the Tribunal’s flexible procedures and efficient case management, make the Tribunal well equipped to take on this enhanced role. However, there remain a number of unanswered questions about the new jurisdiction. For example, it is not clear whether, and if so how, the CAT will be able to address issues or causes of action that are not based on competition law but are connected with, and ancillary to, a competition law action. The proposed legislation is silent on this issue, and it would be unfortunate if the existence of a subsidiary cause of action were to prevent the CAT determining the competition claim, or lead to some bifurcation and transfer to the High Court of all or part of the proceedings. It is important not to introduce an unintended limitation on the Tribunal’s jurisdiction that could affect its desirability as a venue for damages actions and frustrate the stated intention of the reforms. There are a number of possible solutions to this question, but it will need to be addressed.

14 This chapter was completed prior to the adoption of the Bill, which is now the Consumer Rights Act 2015. For the final version of the legislation, see http://www. legislation.gov.uk/ukpga/2015/15/contents/enacted.

528 Gerald Barling

3.1.2 Injunctions The CAT will also be given power to grant interim and final injunctions, both in stand-alone actions and in the new collective redress proceedings. In deciding whether to do so, the CAT is to be required to apply the principles which the High Court would apply.15 This remedy will be a very important—indeed, an essential—tool once the new jurisdiction to determine stand-alone claims commences. Harm resulting from a breach of the competition rules is not always capable of being fully compensated by an award of damages. Irreparable damage can occur, and the ability to enjoin an allegedly infringing undertaking at an early stage, before such damage has been sustained, is likely to be a crucial form of relief in the context of abuse of dominance cases, especially for small and medium enterprises (SMEs). The requirement that the CAT must apply the same principles as the High Court raises the question whether the CAT would be able to waive the normal condition for grant of an interim injunction in England & Wales, namely that a suitable cross-undertaking in damages be given by the applicant. Arguments had been made that this rule should be relaxed, particularly in cases brought by SMEs under the proposed fast-track procedure. On the other hand, vigorous objections have also been made to any such relaxation. It remains to be seen how this question will be taken forward as the draft measure goes through the legislative process. As far as the enforcement of injunctions is concerned, the Bill stops short of providing the Tribunal with contempt powers, and requires the Tribunal to ‘certify’ a breach of an injunction to the High Court. The High Court will then enquire into the matter before determining whether the person is in contempt. This process seems to envisage a double trial of the alleged breach of the CAT’s order, with evidence and argument being capable of deployment on the same issues in both the CAT and the High Court. Although the CAT has UK-wide jurisdiction, the government does not propose to afford the CAT power to grant equivalent injunctive relief in Scotland.16 The Government’s Response noted that no tribunals currently sitting in Scotland have that power, and referred to the importance of not undermining the primacy of the Scottish Court of Session. Therefore, injunctive relief will only be available when the CAT is sitting as a tribunal in England & Wales or Northern Ireland.

See the proposed new subsections 47A(3)(c) and 47D(2) of the Competition Act 1998. The remedy is known as an ‘interdict’ in that jurisdiction.

15 16

The UK Competition Regime 529

3.1.3 Fast-track regime As I mentioned, the draft legislation requires the introduction of a ‘fast track’ in the CAT, for what are referred to in the Government’s Response as ‘simpler’ competition claims. The fast track is to be constructed within Tribunal rules, which will specify the procedure and identify the factors relevant to a determination of whether a claim is suitable for this treatment. The Bill also stipulates that such claims be heard by a chairman sitting alone, rather than by a normal Tribunal panel of three judges. The Government’s Response stated that the procedure is ‘intended to be principally for the benefit of SMEs, and the CAT will seek to prioritise cases involving companies which would otherwise find it more difficult to obtain access to justice’.17 A few comments are appropriate: —— The establishment of a fast track is a worthy aspiration. Indeed, the Tribunal already invariably aims, through its flexible rules and active case management, to streamline every case to the extent practicable in the light of its circumstances. —— There is no attempt to identify what would constitute a ‘simple’ case, and in the antitrust field simplicity is not often encountered. One cannot assume, for example, that because the parties are of a relatively small size the issues raised by the claim will be simple, any more than the opposite is true where larger parties are involved. Very complex issues can arise in relation to claims brought by small companies: 2 Travel and Albion Water are cases in point. —— It is not clear why it should be thought that the expertise of the Tribunal’s panel of Ordinary Members18 is dispensable in a fast-track case. Although there is something to be said for giving the Tribunal more flexibility to sit with a Chairman alone, for example where a case raises only questions of law, there is no reason to believe that this is more likely to be appropriate in fast-track cases. The reverse may be true—where matters are to be resolved swiftly, there may be all the more need for relevant economic, accountancy or other expertise, which the CAT’s multi-disciplinaryconstitution provides. I very much hope that the composition of the Tribunal will not ultimately be prescribed by legislation in this way, but will be left to the discretion of the Tribunal in the light of the particular circumstances. —— In its Response, the government stated that the Tribunal will have power to limit the amount of evidence and expert witnesses produced by each

Government’s Response, above n 12, point 4.22. Whilst the Chairman is always a senior lawyer with expertise in competition law, the ‘Ordinary Members’ are more usually non-lawyers, and are mainly distinguished economists, accountants or persons with relevant experience in industry. 17 18

530 Gerald Barling side. In fact, the Tribunal already has this power, and regularly exercises it as part of its general case management. —— Finally, the government indicates that the fast track should focus on injunctive relief. In practice, as I have said, the CAT’s power to grant such relief is likely to be one of the most important reforms to benefit SMEs.

3.1.4 Collective redress Of the proposals made by the government in January this year, the introduction of a new system of collective redress for competition claims is the most controversial by some margin, but I believe it is one that should be cautiously welcomed. There is simply no point in giving citizens rights if there are no effective remedies and procedures enabling those rights to be enforced. The deficiencies of section 47B of the Competition Act 199819 and of the other options available in civil proceedings in the High Court are well known. I became President of the CAT in the aftermath of the Football Shirts litigation,20 where it was recognised that section 47B, while a useful addition to the CAT’s jurisdictional arsenal, would not prove effective in the context of mass claims for relatively small sums of money. Nor is the position much better in the High Court: the Group Litigation Order procedure is fine so far as it goes, but it, too, is unsuitable for mass claims of that kind, being, like section 47B, opt-in only in nature. Any hopes that CPR Rule 19(6) might fill the procedural gap were removed by the decision of the Court of Appeal in Emerald Supplies.21 In that decision, the Court clarified the criteria for a representative action under the Rule in such a way that it would be difficult to use the procedure if damages were an element of the cause of action. So, as things stand, in the UK there is no effective procedural mechanism to be used as a vehicle for small mass claims—in any area of law. As a result, there is the potential for tortfeasors to avoid liability to many victims, and private enforcement of competition law is not as effective as it should be. In its proposals for reform, the government has rightly taken a cautious approach, mindful no doubt of the risk of opening the way to abuses of the kind which are perceived to have arisen in some other jurisdictions. Thus, under the proposed legislation, the new collective actions regime will not apply across the board but only to claims based on infringement of the competition

19 Section 47B provides for a representative action for damages in the CAT. The procedure is opt-in only, follow-on only and consumer only. 20 Case Nos 1019–1022/1/1/03. 21 See Emerald Supplies Limited & Anr v British Airways plc [2010] EWCA Civ 1284; on appeal from [2009] EWHC 741 (Ch).

The UK Competition Regime 531 rules. Also, the CAT will have exclusive jurisdiction in claims under this new system. The system will cover both follow-on and stand-alone claims, and— significantly—opt-out as well as opt-in procedures. The government recognises the need for there to be a set of strong safeguards in place to ensure that the introduction of a regime which includes the possibility of an opt-out procedure does not lead to the inadvertent creation of a litigation culture. Foremost amongst these safeguards will be the requirement that the CAT should certify that claims are suitable to be brought as collective proceedings, and that it is appropriate to authorise the claimant to act as a representative in those proceedings. A fundamental condition of the suitability of claims is that they raise the same, similar or related issues of fact or law. In addition, the Tribunal must identify the class of persons whose claims are eligible for inclusion in the proceedings, and must specify whether the proceedings are to be opt-in or opt-out. I will just say a word or two more about some of the features of the proposed scheme.

3.2 Representatives According to the Bill, claims will not be able to be brought by legal firms, funders or special purpose vehicles. However, there will be no requirement for a list of suitable bodies to be established by statutory instrument, as at present under section 47B of the 1998 Act. The CAT will assess the suitability of the representative at the certification stage of each case, and, for persons other than class members, whether it is ‘just and reasonable’ for that person to act as a representative in the claim. This will be assessed pursuant to factors to be contained in the new Tribunal rules.22 This leaves open the question of who, other than actual victims, will, in practice, be bringing these representative claims. In its Response, the government referred to ‘genuinely representative bodies’, such as trade and consumer associations, and those certainly seem the most likely candidates.

3.3 Suitability Another knotty question at the centre of the certification process relates to the requirement that the representative body demonstrate that a collective action of the kind sought would be the most appropriate way of bringing

Enterprise Act 2002, proposed new subsection 15B(2)(c).

22

532 Gerald Barling the claim(s) of the class members. Here the devil will be in the detail, which, again, is to be contained in Tribunal rules specifying the relevant factors to take into account in deciding whether claims are suitable for collective proceedings, and, if so, whether by opt-in or opt-out.23 It is to be hoped that these factors will be accompanied by a reference to ‘all other relevant circumstances’, as it would be very difficult to anticipate and formulate every factor which might be pertinent in a particular instance. Where the claim is a mass consumer claim for relatively small individual amounts, it is likely that the alternative to an opt-out collective claim would be no claim at all, and in principle this factor would favour such a procedure (provided all other criteria were satisfied). Where, however, the class is relatively small, perhaps comprising businesses, or a mixture of businesses and consumers, the matter may be less clear cut.

3.4 Commonality Under the proposed legislation, claims can only be brought as collective proceedings if the Tribunal considers that they raise ‘the same, similar or related issues of fact and law’. In some cases, this may be a straightforward matter. But anyone who is inclined to underestimate the problems which are likely to arise in relation to the requirement that the claims raise issues common to all members of the class has only to read the US Supreme Court’s decision in Comcast Corp v Behrend of 27 March 2013 to be disabused.24 There, the Supreme Court, by a slim majority of 5:4, overturned the certification of a class by both the District Court and the Court of Appeals on the basis that the claimants’ economist’s damages model was not sufficiently specific to the certified theory of harm to satisfy the rule that questions of fact or law common to the class must predominate over individual issues. According to the majority, it did not allow the Court to be satisfied that damages were capable of being quantified on a class-wide basis. The majority criticised the lower courts for rejecting the appellants’ arguments to this effect on the basis that to address it would mean getting too deeply involved in a question on the merits of the case. In a scathing dissent, the minority25 state that it is almost universally recognised that individualised damages calculations do not preclude certification, and that the majority’s judgment disturbed factual findings made by two lower courts, contrary to the Supreme Court’s established practice. Moreover,

Enterprise Act 2002, proposed new subsection 15B(2)(b). 569 US __ (2013). 25 Justices Ginsburg and Breyer, joined by Sotomayor and Kagan JJ. 23 24

The UK Competition Regime 533 those findings by the lower courts were probably correct, and the inadequacies of the model identified by the majority judgment were fundamentally misconceived and contrary to the established facts. Such diametrically opposed positions, in a court and jurisdiction already very familiar with the issues raised by opt-out collective actions, serve as a timely warning to all concerned that if the proposal is implemented the CAT will need to approach the task allotted to it with considerable caution. On any view, the issues become very complex where a proposed class includes both direct and indirect purchasers of cartelised goods. Although both may have a common interest in establishing the infringement, the need to demonstrate that the class as a whole was harmed by the alleged infringement might put the direct and indirect purchasers at loggerheads; they will almost certainly not have a common interest at the causation and quantification stage. Of course, in some cases such difficulties may be capable of resolution by the certification of issue-based subclasses.

3.5 Preliminary merits assessment In its Response, the government stated that the CAT would be required, as part of the certification process, to engage in a preliminary merits assessment. Presumably it is intended that this should be one of the ‘suitability’ factors, to be fleshed out in Tribunal rules, since there appears to be no express reference to it in the Consumer Rights Bill itself. Whilst this is, no doubt, intended as a sensible safeguard to ensure that unmeritorious claims are not certified, it raises a number of questions. If a merits threshold is applied in every case, then there is likely to be a mini-trial along the lines of an application for summary judgment, adding to the overall costs of proceedings. There is also a question as to the standard the CAT should apply when considering the merits. Should it merely be satisfied that the claim is not manifestly hopeless or should there be a higher threshold, such as ‘reasonably arguable’ or ‘realistic prospect of success’?

3.6 Costs and funding As to the funding of collective actions, damages-based fee agreements (DBAs) will not be enforceable for opt-out collective actions,26 whilst conditional fee

Proposed new subsection 47C(7).

26

534 Gerald Barling agreements and after-the-event insurance will remain available as funding options for such claims. The prohibition of DBAs in this field is said to be necessary to avoid the creation of a litigation culture. However, such agreements are now permitted, within certain limits, in general civil litigation as part of the Jackson reforms.27

3.7 Damages Unsurprisingly, the government has decided that US-style treble damages have no place in UK competition law.28 Exemplary damages are also to be prohibited in collective actions, whether opt-in or opt-out.29 Whilst this would carve out an exception from the general law of damages in England & Wales, in the context of collective redress, and would preclude an assessment such as those made by the Tribunal in two cases decided in 2012–13,30 this prohibition is unlikely to have a dramatic impact on the effectiveness of the proposed procedure. Awards of exemplary damages are extremely rare. Further, under Court of Appeal case law,31 one of the preconditions to the availability of exemplary damages is that no fine has been imposed on the infringing undertakings, and it is perhaps unlikely that there will be many infringements suitable for an opt-out collective claim which have not attracted a fine.

3.8 Disposal of residual damages In the consultation exercise preceding the Government’s Response, a hotly disputed question was what should happen to any unclaimed damages paid out by a defendant in an opt-out action. The government has decided that such sums should be paid to the Access to Justice Foundation.32 From a judicial point of view, this decision is very welcome. A cy-près approach would probably lead to undesirable and costly satellite litigation

27 See section 58AA of the Courts and Legal Services Act 1990, as amended by section 45 of the Legal Aid, Sentencing and Punishment of Offenders Act 2012. 28 Notwithstanding the probable origin of the idea in the English Statute of Monopolies of 1624. 29 Proposed new subsection 47C(1). 30 2 Travel, above n 5; Albion Water, above n 6. 31 Devenish Nutrition Limited v Sanofi-Aventis SA (France) and others [2008] EWCA Civ 1086. 32 The Access to Justice Foundation receives and distributes financial resources to help provide pro bono legal assistance to those who need it most. The Foundation is an initiative of the entire legal profession, working with the not for profit and voluntary legal advice sector. It is a prescribed charity under section 194 of the Legal Services Act 2007 and aims to provide practical ‘access to justice’ to those unable to afford help.

The UK Competition Regime 535 in the Tribunal between various would-be recipients. The proposed solution seems more likely to promote simplicity and certainty.

3.9 Collective settlements The proposed provision for a collective settlement scheme is also to be welcomed. It would provide the UK with a procedure similar to the apparently very successful mass settlement arrangements which have been operating in the Netherlands since about 2005 (across all sectors, not just competition).

4. EU-level developments At about the same time as the publication of the Consumer Rights Bill in the UK, the European Commission published its draft Directive on private enforcement (later adopted by the EP and the Council in 2014). This initiative is intended to remove procedural obstacles to damages actions in all Member States, and to address the important issue of access to leniency documents and the broader interaction of public and private enforcement. On the same day, the Commission unveiled a Communication and Recommendation on a common framework for collective redress mechanisms in connection with violations of EU law rights, and DG Comp also published two further documents—a Communication and a staff working paper—containing an extremely helpful vade mecum on the quantification of damages in competition law claims.

4.1 Directive on private damages actions In the EU Directive, probably the most dramatic feature for many Member States is the introduction of documentary disclosure obligations in actions for damages. Such disclosure is by no means the norm across Europe; indeed, in some Member States it has been close to anathema. From the UK perspective, the aspect of the Directive that is more significant is that certain leniency documents are to be protected from disclosure. Corporate statements in support of a leniency application and settlement submissions are absolutely protected, whereas certain other categories of documents enjoy temporary protection until the Commission’s administrative process is completed. The Directive also limits a successful immunity recipient’s liability for damages, so that in the first instance his liability does not exceed the loss he has caused to his own direct or indirect customers, and he is in principle relieved

536 Gerald Barling of joint and several liability for damages resulting from transactions of the other cartel members. However, the provision enables other claimants to have recourse to the leniency applicant if they fail to collect in full from the other cartelists. This provision may well prove a fruitful source of protracted litigation. The Directive also establishes a rebuttable presumption that a cartel has caused harm, and a measure to render final infringement decisions of national competition authorities (NCAs) and of national review courts binding on the courts of the same Member State when the latter are seised of actions for damages based on the same agreements or conduct. Final infringement decisions of NCAs and of national review courts of other Member States must be treated as at least prima facie evidence of an infringement.

4.2 Communication and Recommendation on collective redress The EU debate in the area of private damages actions extends not only to competition claims, but also to other circumstances of ‘mass harm’ resulting from any violation of an EU law right. In particular, the Recommendation adopts opt-in collective actions and collective settlement regimes as the appropriate basic standard, capable of being applied by all 28 Member States. Understandably, it does not propose to ask all those states, with very differing levels of sophistication in their justice systems, to attempt to introduce opt-out procedures. However, the Commission expressly envisages that in some Member States an opt-out remedy may be ‘justified by reasons of sound administration of justice’. This seems a sensible and pragmatic approach.

5. Regulatory appeals consultation This important document was published recently by BIS.33 The key questions about which views are being sought include the following: —— Whether the standard of review and permitted grounds of appeal in challenges to infringement decisions under the Competition Act 1998 and some regulatory decisions should be scaled back to the more limited

33 ‘Streamlining Regulatory and Competition Appeals: Consultation on Options for Reform (19 June 2013), https://www.gov.uk/government/uploads/system/uploads/ attachment_data/file/229758/bis-13-876-regulatory-and-competition-appeals-revised.pdf.

The UK Competition Regime 537

——

——

——

——

‘judicial review’ standard34 or to more closely defined grounds of appeal than at present. Whether, on appeal before the CAT, the introduction of material and evidence which was not put to the competition authority or regulator during the administrative procedure should be subject to greater restriction than at present. Whether decisions of the NCA taken in the course of an investigation into a possible infringement of the EU or domestic competition rules, but which are currently not open to review in the CAT, should be brought within the CAT’s jurisdiction, instead of having to proceed in the High Court. Whether a consistent standard of review should be introduced in other regulated sectors, including aviation, energy, postal services, water and rail, and whether appeals from these bodies should be rationalised in other respects, including the appeal body. Whether there are other possible improvements to regulatory investigations and decision-making, and to the operation of the CAT itself.

There are some very positive aspects to the consultation. For example, few would argue that the current patchwork of appeal or judicial review routes in respect of ex ante regulatory decisions from the various utility regulators is anything other than a complete mess, and could not usefully be rationalised. Also to be welcomed are a number of proposals relating to the CAT. These include the introduction of a mechanism to enable Court of Session judges in Scotland and High Court judges in Northern Ireland also to sit in the CAT, and the suggestion that the CAT rather than the Administrative Court should hear interlocutory disputes which arise in the course of an antitrust violation investigation. However, some aspects of the paper are less obviously positive. I will mention only two.

5.1 Standard of review in competition appeals When the government decided not to introduce a prosecutorial system for these cases with the advent of the new CMA, but to stick with the current administrative system, it was on the basis that appeals against infringement decisions of the CMA would be subject to an appeal on the merits, as now. In its March 2012 response to a consultation paper, the government said this:

34 In English law, judicial review is concerned with the legality and not the correctness of an administrative decision: the reviewing court will not overturn the original decision simply because it would have decided the matter in a different way itself.

538 Gerald Barling ‘The Government accepts the strong consensus from the consultation that it would be wrong to reduce parties’ rights and therefore intends that full merits appeal would be maintained in any strengthened administrative system’. However, in the present consultation the government is revisiting this issue and consulting again on whether to constrain the grounds upon which a competition infringement decision can be challenged. This is a cause for concern. Leaving aside the issue of compliance with Article 6 ECHR, a finding of infringement of the competition rules has a number of very serious reputational and financial consequences for a company and for the executives involved. It can result in huge penalties, including an uplift in penalty in the event of any recidivism by the company. Further, such a finding can be used as a binding base for a follow-on damages action. It is therefore puzzling to say the least that these apparent second thoughts should have arisen so soon. Any change of the kind envisaged would be ironic, given the current lively debate in Europe about the adequacy of the General Court’s jurisdiction to review the Commission’s infringement decisions. Of course, this is a consultation, not a decision, and it is possible that no change with regard to the standard of review will ultimately be made in the UK.

5.2 New evidence Another cause for concern relates to the possibility of significantly restricting the introduction, on appeal, of what the consultation paper calls ‘new evidence’. The first point to note is that what is being spoken of as ‘new evidence’ is nothing of the kind. In the administrative procedure, evidence is not placed before an impartial court or tribunal: this first happens on appeal to the CAT. So it is somewhat misleading to confuse that with the Ladd v Marshall35 situation, where a matter has been heard and decided by a lower court and a party seeks to adduce new evidence on appeal to a higher court. In competition appeals, the CAT is, in practice, a court of first instance. Further, there is simply no evidence that material which could have been adduced at the administrative stage is somehow being withheld in order to be deployed on appeal. The CAT’s current rules are perfectly adequate to enable it to exclude or limit evidence where the interests of justice so require. If restrictions of the kind set out in the paper are imposed on the CAT, then, far from streamlining appeals, which is the ostensible object of this consultation exercise, it will almost certainly lead to additional and/or longer

[1954] EWCA Civ 1.

35

The UK Competition Regime 539 appeals both in the CAT and in the Court of Appeal, as the parties dispute the CAT’s admission or exclusion of material by reference to the proposed statutory criteria.

6. Judicial independence Finally, perhaps I might be allowed one or two reflections on the importance in this field, as in all areas of law, of safeguarding judicial independence. The CAT is situated in the midst of some very big beasts, in the form of powerful regulatory bodies and huge industry players. Equally, the S ecretary of State for Business, Innovation and Skills is also, on occasion, a party to judicial review proceedings before the CAT, as in the HBOS/Lloyds TSB merger case.36 Compared to these institutions and companies, the CAT is tiny in budgetary and personnel terms, comprising only one full-time judge (the President), the Registrar and a staff of about a dozen. In addition, there are currently six part-time Chairmen and 15 inaptly named ‘Ordinary Members’—as I have already mentioned, they are far from ordinary, and are in fact all very distinguished in their specific fields. The CAT also has the valuable ability to call on the judges of the Chancery Division of the High Court as additional Chairmen. The CAT’s annual budget is about £3.7 million, and a large part of that represents the rent of its premises. In comparison with the regulators and companies who form the bulk of our litigants, the CAT is a small ‘cottage industry’. On the other hand, the cases with which it deals, as well as being factually, technically and legally complex, are often extremely sensitive and of great importance to the immediate litigants and others—sometimes a whole sector may be affected. Sensitivity and importance also exist from the point of view of the regulatory decision-maker. The authority or regulator may have been investigating for a long time, and may feel, rightly or wrongly, that its credibility is to some extent at stake in an appeal from its decision. In some cases, a setback in the CAT may be much harder for the decision-maker to swallow than for a losing appellant, who may be more likely to adopt a philosophical attitude of ‘you win some, you lose some’, even where the commercial consequences of losing an appeal are significant. Such differences in reaction are understandable, given the very different interests at stake.

36 Case No 1107/4/10/08, Merger Action Group v Secretary of State for Business, Enterprise and Regulatory Reform.

540 Gerald Barling However, it would be troubling if the risk of a regulatory or enforcement decision being overturned on appeal were to lead to a desire to protect decisions from an appropriate level of scrutiny by an independent judicial body. In the present regulatory climate, where very significant and commercially intrusive powers are given to regulators acting in the public interest, it is all the more important that the exercise of those powers should be subject to proper judicial oversight. It is not now appropriate, if it ever was, to speak of judicial deference in this context, if by deference it is meant that there should be ‘no go’ areas where, in effect, a rubber stamp will be applied by a court. This would be to place a virtually untrammelled power of commercial life and death into the hands of an administrative agency. Few, if any, would argue for such a situation. But it is equally important that the powers of a tribunal such as the CAT properly to examine and adjudicate upon a reasonable ground of challenge to a regulatory decision should not be overly circumscribed by artificial and restrictive rules as to, for example, the evidence which may or may not be admitted and considered by the court, or other matters which are normally within the scope of a court’s discretion when seeking justly to resolve disputes which fall within its jurisdiction. And it would be of even greater concern if pressures for changes of that kind were to be seen as a response to judgments of a court. All courts can and do go wrong. The proper way of addressing perceived errors of adjudication is by appeal to the relevant appeal court, not by seeking to undermine the effectiveness of judicial oversight by spurious suggestions for reform of the appeal process. We must also be wary of the consequences which could arise, albeit unintentionally, from constant pressures for review and change. Judicial independence is vital to all our well-being. It is in the public interest that it should be jealously safeguarded at all times. Our democracy and our freedom depend upon it. This applies no less in the competition and regulatory field, where infringement decisions and regulatory initiatives can have very real and sometimes adverse consequences for individuals and companies. It is crucially important that courts, particularly small specialist ones, whose judicial personnel are few in number and well known to their users, should not have to expect that giving a judgment to this or that effect might well lead to intense lobbying for jurisdictional and procedural changes, with the aim of lessening the scrutiny to which certain decisions would be subject in the future. To achieve such an aim would do nothing at all to improve the quality of regulatory decision-making, and would create unwholesome pressure on the courts concerned. I emphasise that I am not here speaking of proposals for reform where the need for reform is properly and fully evidenced by examples of where things have gone awry, or where genuine procedural improvements to the system can be made.

The UK Competition Regime 541 In an address to the Commonwealth Law Conference, the present Lord Chief Justice of England & Wales referred to the need for ‘eternal vigilance’ in matters of judicial independence, and to the importance of avoiding ‘the first small, even tiny, steps’ that might lead to something unacceptable. By way of example, he described how, in the context of the Control Orders issued under the Prevention of Terrorism Act, a former Secretary of State for Home Affairs was moved publicly to criticise the ‘total refusal’ of the justices of the UK Supreme Court (then the Judicial Committee of the House of Lords) to discuss the issues of principle involved in these matters; the former Home Secretary suggested that it was time for the senior judiciary to engage in a serious and considered debate about how best legally to confront terrorism in modern circumstances, and for some ‘proper discussion’ between the then Home Secretary and the justices with a view to the latter, in effect, providing guidance about what kind of measure would not be liable to be struck down. Surely, he said, the idea that such discussions would compromise the independence of the members of the court was ‘risible’. In his speech, the Lord Chief Justice observed that, had they taken place, such discussions between the members of the court and the executive would have represented one of those tiny first steps of which we should eternally beware. None of us can afford to be complacent. A cautionary tale of where ‘first tiny steps’ can end up is that of Judge Maria Afiuni, a Venezuelan judge who granted bail to a banker connected with the political opposition. The banker jumped bail and fled the country. The late President Chavez then had Judge Afiuni jailed, announcing on TV that in another era she might have been brought before a firing squad. He did not say whether this would have been preceded by a trial!

542

James S Venit* What Is To Be Done?1

It was argued at the 2013 Workshop that the credibility of the EU competition enforcement regime is itself on trial. The argument, summarized simply, is that the combination of an inquisitional (as opposed to adversarial) administrative system which confers substantial enforcement powers on the European Commission with a system of limited judicial review is too one-sided and calls into question the legitimacy of the EU regime of antitrust enforcement. Rigorous judicial review is a sophisticated form of quality control, and ultimately benefits both the providers and consumers of antitrust enforcement services. This author agrees with this basic criticism. At that point, the question that arises is: what is to be done? First, we need to reach a consensus that there is a problem: that is, to acknowledge that the combination of an administrative system in which the Commission acts as both prosecutor and judge with a system of light judicial review—limited to ‘controlling’ the ‘legality’ of the Commission decision with annulment only for manifest error or procedural violations, and referral back to the Commission to correct the errors in its decision—is not credible (in terms of legitimacy) and therefore not sustainable, particularly in an era of internationalization of antitrust enforcement, where less experienced jurisdictions are looking for role models and best practices. Second, we need to recognize that this is not a procedural issue, nor one confined to quasi-criminal cartel or abuse cases in which the Commission imposes large fines. The procedural defects in the EU administrative system and the lack of effective judicial review affect the quality of decisions—at both the evidentiary/factual and theoretical/economic levels. In other words, the problem is as much a substantive as a procedural one because the procedural defects affect the substantive quality of decisions and not merely the rights of defense. As Bill Kovacic put it in his remarks, high-quality judicial review ensures the quality of agency decisions. Third, we need to accept that the problems are structural and in theory require the type of wide-ranging reform that, as we have been told by Philip Lowe, is not going to be a serious agenda item for the European institutions

Partner, Skadden, Arps, Slate, Meagher and Flom, LLP. What Is To Be Done? Burning Questions of Our Movement was written by VI Lenin in 1901, its title being inspired by the eponymous novel by Nikolai Chernyshevsky. The pamphlet, which had a certain urgency, is otherwise totally unrelated to this chapter. * 1

544 James S Venit or the Member States in the foreseeable future, given, inter alia¸ the current and prolonged economic crisis.2 Fourth, accepting the limitations identified in point three, we need to focus on (i) what short to medium-term improvements can be made and (ii) encouraging and increasing judicial activism. While the first proposal is a little like applying a band-aid to a hemorrhage, the second could go far to curing the underlying disease without constitutional change. Since the second element is the more important, I will discuss it first before turning to more modest proposed procedural reforms. In light of the activism that the EU courts have sometimes displayed, the academic debate as to the scope of the European Courts’ powers of judicial review may not be the best place to start—or end. But, having acknowledged that, and with a nod to thoroughness, let us briefly review this issue. Pursuant to Article 263 of the Treaty on the Functioning of the European Union (TFEU), the General Court has jurisdiction to annul a Commission decision on the following grounds: (i) lack of competence; (ii) infringement of an essential procedural requirement; (iii) infringement of the Treaty or any rule of law relating to its application; or (iv) misuse of powers. On its face, the scope of the court’s review appears more circumscribed than its power under Article 261 TFEU, which confers upon the General Court an unlimited jurisdiction to review fines imposed by the Commission and thus to cancel, reduce or increase them, thereby substituting its own judgment, in matters of fact and law, for that of the Commission as far as the calculation and imposition of fines is concerned. Nevertheless, it is not clear whether, in exercising its powers under Article 263, there is a difference in the scope of the court’s substantive and procedural review of a decision as compared to its review of fines under Article 261, and, if so, the extent of that difference.3

2 Although this is undoubtedly true, we should perhaps also ask ourselves whether overenforcement and an overly regulatory approach to competition law has not itself contributed to the crisis. In all likelihood it has contributed a lot, though in fairness probably not as much as the Member States’ own resistance to liberalization and structural reform. 3 Under Article 230 EC (Article 263 TFEU), the Courts enjoy ‘only a rather limited, supervisory jurisdiction over the Commission’s measures, restricted to the grounds of review listed in the Treaty. They cannot therefore substitute their own assessment at the economic and legal appraisal conducted by the Commission and contained in the impugned decision, nor can they a fortiori ‘replace’ the decision by a new one, but will have to confine their scrutiny to whether the Commission respected the limits of its discretion, whether it respected the procedural rules and did not commit any manifest error of law or of fact or misused its powers.’ 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?’, GCLC Working Paper 04/08, 15 (citing articles by, inter alia, David Bailey, ‘Scope of Judicial Review under Article 81 EC’, 41 Common Market Law Review 1327, 1332 (2004); and Virpi Tiili and Jan Vanhamme, ‘The “Power of Appraisal” (pouvoir d’appréciation) of the Commission of the European Communities vis-à-vis the Powers of Judicial Review of the Communities’ Court of Justice and Court of First Instance’, 22 Fordham International Law Journal 885 (1998–1999)). The Working Group Paper also

What Is To Be Done? 545 Certainly the Court has at times expressly defined its power of judicial review in broad terms, as in this formulation by Judge Vesterdorf: As regards matters of law, the Community courts exercise full jurisdictional control. Indeed it is for the Community courts to provide the definitive interpretation of Community law, be it in Treaty provisions or secondary legal provisions such as those contained in the Merger Regulation, and this goes for both procedural and substantive legal provisions. The Community courts interpret the law and then check whether the Commission has applied the correct legal principles in the case under examination. There is no margin of appreciation left to the Commission in this respect as to what are the legal criteria to apply.4

The Court’s powers seem equally broad as concerns factual assessments notwithstanding the deference that the Court is supposed to accord the Commission in respect of ‘complex economic assessments’. Thus, former Judge Nicholas Forwood has taken the view that the mere fact that an assessment is made which requires the consideration of (possibly esoteric) economic arguments and the examination of economic evidence … does not necessarily make an assessment subject to judicial review a ‘complex’ one, subject only to limited control, even though it may make the task of the judge extremely difficult or burdensome.5

This distinction—or at least the General Court’s ability to review complex economic facts without deferring to the Commission’s margin of discretion in respect of complex economic assessments—has been endorsed in Tetra/Laval and, most recently, in Ryanair, where the General Court observed that [w]hilst the Courts of the European Union recognise that the Commission has a margin of discretion with regard to economic matters, that does not mean that they

notes that ‘[t]he case law suggests that the scope of the powers enjoyed by the Community Courts in the area varies in relation to the nature of the grounds of review raised by the parties. On this specific point, the case law suggests that the General Court will exercise a “comprehensive review” of the impugned decision when an allegation of error of fact or of procedural impropriety is made (see eg Case T-41/96 Bayer AG v Commission [2000] ECR II-3383, paras 67, 69, 71; confirmed in Case C-2/01 BAI and Commission v Bayer [2004] ECR I-23, para 48; Case T-44/90 La Cinq SA v Commission [1992] ECR II-1, para 48). By contrast, when the soundness of the “complex economic assessment” of the evidence before the Commission is questioned, the Community Courts have held that it is not for the EU judiciary to carry out a comprehensive review of the Commission’s economic appraisal of the evidence and of the conclusions reached on the basis of the applicable competition rules.’ Working Group Paper, 15. 4 Bo Vesterdorf, ‘Standard of Proof in Merger Cases: Reflections in the Light of Recent Case Law of the Community Courts’, 1 European Competition Journal 1, 5 (2005). 5 Nicholas Forwood, ‘The Commission’s More Economic Approach—Implications for the Role of the EU Courts, the Treatment of Economic Evidence and the Scope of Judicial Review’, in Claus-Dieter Ehlermann and Mel Marquis, eds, European Competition Law Annual 2009: The Evaluation of Evidence and its Judicial Review in Competition Cases, Hart Publishing, 2011, 255 et seq, 265.

546 James S Venit must refrain from reviewing the Commission’s interpretation of information of an economic nature.6

Even commentators who take the view that differences do exist between the Court’s powers under Article 261 and 263 qualify their position by noting that the deferential standard of review is the one ‘generally’ adopted by the Community Courts. This implies—and judgments such as Airtours, Tetra Laval and Glaxo would seem to confirm—that this deferential standard is not always applied. This, in turn, suggests that the constraint inferred from the contrast between Articles 263 and 261 may be flexible rather than a stringently binding one. Indeed, there is a strong argument that Article 263 should be interpreted in conformity with fundamental rights, including those recognized by the ECHR.7 Under this approach, since Article 263 does not preclude robust and extensive review, and since respect for the appropriate degree of judicial protection requires it, such review should be seen as inherent and thus within the scope of Article 263.8 Moreover, as a practical matter, the Court’s ability to annul a Commission decision for (i) factual error, (ii) material error in the assessment of the facts, (iii) the drawing of incorrect legal conclusions from the facts, or (iv) a procedural error resulting in a material infringement of the rights of defense in assembling the facts is sufficiently broad to permit the reversal of a Commission decision that was not firmly based on the evidence in the file even if not expressly permitting the rewriting of such a decision to conform to the facts or draw the appropriate conclusions from them. Thus, although the General Court frequently accords great deference to the Commission’s discretion in matters involving complicated economic assessments, this does not always restrict the scope of the Court’s review. For example, in Glaxo the General Court described its role as follows: the Court hearing an application for annulment of a decision applying [Article 101(1) TFEU] must undertake a comprehensive review of the examination carried out by

6 Case T-342/07 Ryanair Holdings plc v Commission [2010] ECR II-3457, para 30. pplying his analysis, Judge Forwood would, for example, classify assessments concerning A the granting of an exemption under Article 101(3), which require ‘value judgments’ as a complex economic assessment, whereas assessments as to the applicability of Article 101(1) or whether an undertaking holds a dominant position would involve a difficult economic appraisal which may, nevertheless, be subjected to comprehensive judicial review. See Forwood, ibid. 7 Indeed, in the Menarini judgment (A Menarini Diagnostics SRL v Italy, judgment of 27 September 2011), the ECtHR considered that a cartel fine imposed under Italian law was of a criminal nature which complied with the guarantees of Article 6 ECHR, provided that the competition authority’s decision was subject to review by a court having ‘full jurisdiction’. Based on the facts at issue, the ECtHR held that the Italian appeal court went beyond a ‘simple legality control’ and exercised full judicial review that covered points of fact and law. See also Working Group Paper, above n 3, 30; Opinion of Advocate General Bot in Bolloré cited in Working Paper, 30–31. 8 This is also the conclusion of the Working Paper: ‘it is not clear that the existing constitutional framework excludes full jurisdictional review of competition decisions by the Community Courts’. Working Paper, 29.

What Is To Be Done? 547 the Commission … unless that examination entails a complex economic assessment, in which case review by the Court is confined to ascertaining that there has been no misuse of powers, that the rules on procedure and on the statement of reasons have been complied with, that the facts have been accurately stated and that there has been no manifest error of assessment of those facts …9

It should be noted that in Glaxo this deferential language did not prevent the General Court from departing from the jurisprudence normally applied in parallel trade cases by rejecting a virtual per se approach in respect of measures that restricted parallel trade in prescription medicines.10 Moreover, although the ECJ reversed this finding, as a matter of law, it did not suggest that the Court had exceeded its jurisdictional powers of review.11 Nor did the deferential language prevent the General Court from annulling, in the same case, the Commission’s decision refusing an exemption under Article 101(3). Notwithstanding the Commission’s broad discretion when applying Article 101(3) and the complexity of the economic assessments that are involved in an exemption analysis, the General Court annulled this part of the decision for what it, in effect, characterized as the Commission’s failure to take seriously the applicant’s arguments, which, in its view, were ‘relevant, reliable and credible’.12 This part of the Court’s judgment was upheld by the ECJ on appeal. Similarly, in the Ryanair case which involved a merger, the General Court articulated a more rigorous and less deferential approach to complex economic assessments: the basic provisions of the [Merger] regulation … confer on the Commission a c ertain discretion especially with respect to assessments of an economic nature, and … consequently, review by the [EU] Courts … must take account of the margin of discretion implicit in the provisions of an economic nature which form part of the rules on concentrations …

But [w]hilst the Courts of the European Union recognise that the Commission has a margin of discretion with regard to economic matters, that does not mean that they must refrain from reviewing the Commission’s interpretation of information of an economic nature. Not only must they establish, in particular, whether the evidence relied on is factually accurate, reliable and consistent but also whether that evidence contains all the information which must be taken into account in order to assess a complex situation and whether it is capable of substantiating the conclusions drawn from it …

Case T-168/01 GlaxoSmithKline v Commission [2006] ECR II-2969, paras 57–58. See, to that effect, ibid, paras 116–174. Joined Cases C-501/06 P, C-513/06 P and C-519/06 P GlaxoSmithKline Services and Others v Commission and Others [2009] ECR I-9291. 12 Glaxo, above n 9, paras 303, 304–307, 308 and 316. 9

10 11

548 James S Venit In addition … where the institutions have a power of appraisal, respect for the rights guaranteed by the legal order of the European Union in administrative procedures is of even more fundamental importance. Those guarantees include, in particular, the duty of the Commission to examine carefully and impartially all the relevant aspects of the individual case, the right of the person concerned to make his views known and also his right to have an adequately reasoned decision …13

Certainly in merger cases, which invariably involve complex economic assessments, the Court has at times been willing to cast a critical eye over such assessments. This has been most notable in vertical and conglomerate merger cases such as Tetra Laval14 and GE/Honeywell,15 but has also occurred in horizontal cases such as Kali und Salz,16 Airtours17 and Schneider/Legrand.18 In the Sony/BMG appeal, Advocate General Kokott described the scope of the Court’s review with respect to the accuracy, reliability and completeness of the facts and the ability of the cited facts to support the conclusions drawn from them quite broadly. According to the Advocate General, in such a review, the Court will, when applying the manifest error standard, undertake its own analysis of the facts and the evidence:19 Having regard to this standard of review, it would be an error to assume that the Commission’s margin of discretion precludes the Community Courts in any event from giving their own analysis of the facts and the evidence. On the contrary, it is essential for the Community Courts to undertake such an assessment of their own where they are assessing whether the factual material on which the Commission’s decision was based was accurate, reliable, consistent and complete, and whether this factual material was capable of substantiating the conclusions the Commission drew from it. Otherwise, the Community Courts could not sensibly assess whether the Commission had stayed within the limits of the margin of discretion allowed to it or had committed a manifest error of assessment.

This conclusion was adopted by the Court of Justice, which stated that it is clear from the case-law that while, in the context of the control of concentrations, a field in which the Commission has a margin of assessment with regard to

Ryanair, above n 6, paras 29–31. Case T-5/02 Tetra Laval BV v Commission [2002] ECR II-4381. 15 Case T-209/01 Honeywell v Commission [2005] ECR II-5527. 16 Joined Cases C-68/94 and 30/95 France v Commission (Kali und Salz) [1998] ECR I-1375. 17 Case T-342/99 Airtours plc v Commission [2002] ECR II-2585. 18 Case T-310/01 Schneider Electric SA v Commission [2002] ECR II-4071. 19 Opinion of Advocate General Kokott in Case C-413/06 P Bertelsmann and Sony Corporation v Impala [2008] ECR I-4951, para 239. See also Bo Vesterdorf, ‘The E uropean Court’s Case Law in Merger Control’, in Abel Mateus and Teresa Moreira, eds, Competition Law and Economics: Advances in Competition Policy and Antitrust Enforcement, Kluwer Law International, 2007, chapter 13, 246: ‘Control of primary facts by the CFI is intensive with no room for discretion on the part of the Commission. This is inherent in the nature of the control of the accuracy of fact. Either a fact is a correct fact or it is not.’ 13 14

What Is To Be Done? 549 economic matters, review by the Court of First Instance is limited to establishing whether the evidence relied on is factually accurate and to establishing the absence of a manifest error of assessment, it none the less remains the case that the correctness, completeness and reliability of the facts on which a decision is based may be the subject of judicial review … Indeed, that is one of the ways in which the Community judicature can verify whether the factual and legal elements upon which the exercise of the power of assessment depends were present …20

However, the Advocate General did go on to note that such rigorous review does not entitle the Court to substitute its judgment for that of the Commission. The Advocate General articulated what she perceived as the only limitation on the Court’s powers in the following manner: The Court of First Instance exceeds the limits of judicial review of a Commission decision in the context of merger control only where the factual and evidential position reasonably allows different assessments, the Commission adopts one of them, and the Court of First Instance none the less substitutes its own different assessment for that of the Commission.21

In theory, the distinction between the Court’s rigorous assessment of the correctness, completeness and reliability of the facts and the appropriateness of the conclusions drawn from them, on the one hand, and the requirement that it not substitute its assessment for that of the Commission, on the other hand, is coherent in its articulation and can be neatly encapsulated in the notion that the Court can only annul where the Commission has committed a manifest error of assessment by drawing implausible or unwarranted conclusions from the facts. However, we are on a very slippery slope, and it is far from self-evidenthow, in practice, the Court can easily draw the line between thoroughly reviewing whether the facts are sufficient to support the conclusions the Commission has drawn from them while at the same time r especting the obligation not to substitute its judgment for that of the Commission.

Impala, ibid, para 69. Opinion of the Advocate General, above n 19, para 240. The different standard of judicial review with respect to the accuracy of facts, on the one hand, and their assessment by the Commission, in particular with respect to complex economic matters, on the other was also acknowledged by Advocate General Tizzano in Case C-12/03 P Commission v Tetra Laval [2005] ECR I-987, para 86: ‘With regard to the findings of fact, the review is clearly more intense, in that the issue is to verify objectively and materially the accuracy of certain facts and the correctness of the conclusions drawn in order to establish whether certain known facts make it possible to prove the existence of other facts to be ascertained. By contrast, with regard to the complex economic assessments made by the Commission, review by the Community judicature is necessarily more limited, since the latter has to respect the broad discretion inherent in that kind of assessment and may not substitute its own point of view for that of the body which is institutionally responsible for making those assessments.’ 20 21

550 James S Venit Respecting this distinction is likely to present serious problems in practice.22 This is because any determination as to whether the Commission’s conclusions are based on all the relevant facts and evidence and are sustainable in their light will invariably require the exercise of some independent judgment. And the need for such independent judgment increases when the Court seeks to determine whether the Commission’s conclusions are sustainable in light of the evidence. Review of the case law suggests that the degree of difficulty involved in conducting such a review without second guessing the Commission may vary substantially from cases to case. It further suggests that the burden of proof is the judicial tool that enables the Court to navigate between the Scylla of rigorous factual review and the Charybdis of having to defer to the Commission’s complex economic assessments. Airtours v Commission23 presents a good example of a relatively straightforward assessment, and, indeed, one which seems to verge on pure factual error. In Airtours, the Court fully re-examined the Commission’s projection of low market growth for package tours which, according to the Commission, made the relevant market conducive to coordinated effects. The C ommission’s conclusion was based on a single-page extract of an industry study which, the Court found, the Commission had misinterpreted by failing to take account of the actual wording of the document, which clearly indicated considerable market growth in the past decades. Against this background, the Court concluded: the Commission’s interpretation of the data available to it concerning growth demand was inaccurate in its disregard for the fact that the market had been marked by a clear tendency to considerable growth over the last decade in general, despite the volatile nature of demand from one year to another, and that the pace of demand growth has increased during recent years in particular. In that context of growth, and having failed to produce any more specific evidence establishing that the tendency to grow would be reversed in future years, the Commission was not entitled to conclude that market development was characterised by low growth, which was, in this instance, a factor conducive to the creation of a collective dominant position by the three remaining large tour operators.24

22 As the Court of Justice held in Case C-441/07 P Alrosa v Commission [2010] ECR I-5949, para 63: ‘The General Court could have held that the Commission had committed a manifest error of assessment only if it had found that the Commission’s conclusion was obviously unfounded, having regard to the facts established by it’. In Alrosa (not a merger case but a commitments case), the General Court had objected to the remedy imposed, concluding that a less invasive alternative existed. In reversing the General Court’s judgment annulling that aspect of the Commission’s decision, the Court of Justice drew a line between the appropriateness of the remedy, which was a matter to be left to the discretion of the Commission, and a manifest error of assessment, eg the need for a remedy, which could fall within the competence of the General Court. 23 Above n 17. 24 Ibid, para 133.

What Is To Be Done? 551 Tetra Laval v Commission25 provides perhaps the most striking example of a considerably more difficult case in which maintaining the distinction identified by Advocate General Kokott and endorsement by the Court of Justice appears to be considerably more difficult. In Tetra/Laval, the Court overturned the Commission’s assessment of the conglomerate effects of a merger which had been based on the Commission’s assessment of the likelihood of leveraging from aseptic carton markets, and the consequences of such leveraging, on the market for PET packaging. The Court based its conclusion on a careful examination of whether the evidence relied on by the Commission supported the decision’s findings with relation to complex factors such as (i) whether the likely levels of growth in the PET market would provide an incentive for the merging parties to leverage and (ii) the foreseeable consequences of leveraging on the various product markets.26 On the basis of a detailed factual analysis, the Court concluded that as regards the markets for asceptic and non-asceptic PET filling machines, the contested decision does not adduce evidence that suffices in law to show that the implementation of the modified merger would make it likely, that, as a result of leveraging practices essentially on Tetra’s current carton customers who wish to switch to PET for packaging sensitive products, a dominant position on those markets would be created in the future and at least by 2005.27

Three conclusions can be drawn from this example. First, the level of the Court’s review can be intense and, indeed, controversial, particularly when it involves the assessment of multiple factual elements underlying a single determination. Second, this intensive review can lead the Court to reject the Commission’s conclusions. Third, in doing so, the Court relies on the Commission’s failure to meet its burden of proof as the technical judicial tool in order to avoid overstepping the line and substituting its judgment for that of the Commission. As a legal matter, reliance on the Commission’s failure to meet its burden of proof is arguably non-controversial, since such an assessment would appear to be well within the scope of the Court’s powers of judicial review. More over, to the extent that such an assessment is based on the Court’s evaluation of the factual record, it also has the added benefit of insulating that assessment from subsequent judicial review, since the Court of Justice is limited to reviewing questions of law, but not of fact. That being said, the conclusion that there is insufficient factual support for the Commission’s conclusions on a key element in a merger decision involves a judgment that is far more complex than one concerning the accuracy or the completeness of the facts. There is therefore little wonder that the General Court’s judgment in

Above n 21. Ibid, paras 142–251. 27 Ibid, para 251. 25 26

552 James S Venit Tetra Laval provoked an appeal by the Commission to the Court of Justice. And, although the latter did uphold the General Court’s judgment, it did so in part by r efusing to reconsider issues of fact28 while disregarding the Advocate General’s conclusion that the General Court had exceeded its powers by substituting its assessment for that of the Commission.29 The particularly speculative nature of merger assessment further enhances the scope for broad judicial review. Thus, as the Court of Justice noted in Tetra Laval: A prospective analysis of the kind necessary in merger control must be carried out with great care since it does not entail the examination of past events—for which often many items of evidence are available which make it possible to understand the causes—or of current events, but rather a prediction of events which are more or less likely to occur in future if a decision prohibiting the planned concentration or laying down the conditions for it is not adopted.30

This consideration suggests—and the cases appear to support this conclusion—that: (i) the Court’s review may be considerably more independent and aggressive and the Commission’s reasoning and conclusions will be subject to increasingly intensive scrutiny, given the predictive and inevitably speculative nature of merger analysis; and (ii) the intensity of the Court’s analysis will be proportional to the speculativeness of the Commission’s theory of harm, with the intensity of judicial scrutiny being more intense in cases involving coordinated or conglomerate effects, as opposed to more straightforward horizontal effects. The General Court has confirmed as much by indicating that, where anticompetitive harm is not the normally expected outcome, the Commission’s need to base its decision on particularly convincing evidence, ie its burden of proof, rises considerably: The Commission’s analysis of a merger producing a conglomerate effect is conditioned by requirements similar to those defined by the Court with regard to the creation of a collective dominant position … Thus the Commission’s analysis of 28 See Tetra Laval, above n 21, para 47: ‘Amongst the other examples given by it, the Commission challenges the Court of First Instance’s finding, in paragraph 289 of the judgment under appeal, that “fresh milk is not a product for which the marketing advantages offered by PET have any particular importance” and its conclusions as to the cost of PET in comparison to that of carton, which are set out in paragraphs 288 and 328 of the judgment under appeal. It should be noted that these are findings of fact, which are not subject to review by the Court in appeal proceedings. It is therefore unnecessary to give a ruling on the merits of those findings by the Court of First Instance and it need be stated only that the Court of First Instance was able to base those findings on various items in the contested decision.’ 29 See Opinion of Advocate General Tizzano in Tetra Laval, above n 21, para 111: ‘As matters stand, it seems clear to me that, while the Court of First Instance could possibly have found that the Commission’s investigation was incomplete, or have criticised the logic, consistency and appropriateness of its reasoning, it definitely could not carry out its own assessment of the information in that institution’s possession in order to conclude that PET was “more expensive than carton”’. 30 Above n 21, para 42.

What Is To Be Done? 553 a merger transaction which is expected to have an anti-competitive conglomerate effect calls for a particularly close examination … As the Court has already held, where the Commission takes the view that a merger should be prohibited because it will create or strengthen a dominant position within a foreseeable period, it is incumbent upon it to produce convincing evidence thereof … Since the effects of a conglomerate-type merger are generally considered to be neutral, or even beneficial, for competition in the markets concerned, as is recognized in the present case … the proof of anti-competitive conglomerate effects of such a merger calls for a precise examination, supported by convincing evidence, of the circumstances which allegedly product those effects …31

In summary, these examples suggest that rigorous judicial review is a real possibility without a Treaty amendment. However, judicial review, no matter how rigorous, will never realize its full potentiality as long as the consequence of review is a remand to the Commission. In merger cases in which the Courts have annulled a prohibition decision, a remand may be less of a deterrent to appealing the Commission’s decision. However, in merger cases in which there is a conditional approval subject to a divestiture remedy, the provisions of the merger regulation which empower the Commission to take a new decision32 will effectively preclude an appeal, given the unknown consequences of a second merger review under unpredictable market circumstances, following a considerable lapse of time. Despite this major flaw (which could arguably be remedied by amendment of the Merger Regulation), thorough judicial review should be encouraged in order to ensure the quality of Commission decisions and to enable EU law to serve as a worthy role model in the international arena. As for more minor issues, other possible judicial reforms might include: 1. A correct interpretation of Article 263 TFEU, which would permit the Court to rule on Commission decisions without remanding. 2. Fully reasoned judgments explaining why precedent has been followed or overturned, as opposed to the blind recitation of precedent. 3. Increased transparency through the use of dissenting opinions. As for the band-aids, what follows is a short list of small measures, starting with the Hearing Officer.

1. The Hearing Officer The position of the Hearing Officer was created in 1982 in response to criticisms about the fairness of Commission proceedings and the lack of General Court in Tetra Laval, above n 14, para 155. See Article 10(5) of the Merger Regulation.

31 32

554 James S Venit eaningful judicial control over those proceedings during their pendency. The m Hearing Officer’s powers are set forth in the Commission decision establishing the Hearing Officer’s Mandate.33 These powers, which are largely but not exclusively procedural, include: (i) deciding claims by the defendant and third parties concerning confidentiality and file access;34 (ii) deciding on requests for extending the time limit for the response to the Statement of Objections;35 (iii) presiding over and running the oral hearing;36 and (iv) questioning the parties and the witnesses at the oral hearing.37 In addition, the Hearing Officer may present observations on ‘any matter’ arising out of ‘any Commission competition proceeding’.38 This provision is broad enough to permit the Hearing Officer to comment on the substance of the case. To date, this power has not been used extensively, and the Hearing Officer’s influence on substantive issues does not appear to have been significant in practice. In order to create some degree of independence, the Hearing Officer reports directly to the Commissioner for Competition rather than the Director General of DG Competition and has the power to adopt a report on the hearing (which provides some scope for criticism, where appropriate, of the case team’s handling of procedural—but not substantive or evidentiary—issues in any given case). Notwithstanding these two elements, the Hearing Officer remains a Commission official within DG Competition and for this reason the position is subject to the same inherent limitation as other attempts at institutional self-scrutiny. In this context, the direct reporting line to the Competition Commissioner does not automatically immunize the Hearing Officer from pressure to favor the case team in his or her rulings since the Competition Commissioner is no less likely than the Director General to have enforcement priorities. Experience suggests that the success of the Hearing Officer as a neutral decision-maker and arbiter of more than procedural issues has tended to depend on the independent character of the individual exercising that function and whether that individual envisioned a continuing career in DG Comp. Certainly the recent spate of complaints to the Ombudsman in competition cases39 reflects an increasing dissatisfaction as concerns the objectivity and independence of the Hearing Officer. Although the number of decisions

33 Decision of 13 October 2011 on the function and terms of reference of the hearing officer in certain competition proceedings, [2011] OJ L275/29 (the Hearing Officer’s Mandate). 34 Hearing Officer’s Mandate, Articles 14–15. 35 Hearing Officer’s Mandate, Article 9. 36 Article 14(1) of Regulation 773/2004 and Articles 10–13 of the Hearing Officer’s Mandate. 37 Article 14(7) of Regulation 773/2004. 38 Hearing Officer’s Mandate, Article 5. 39 For a comment on this case, see Bernard Amory and Yves Desmedt, ‘The European Ombudsman’s First Scrutiny of the EC Commission in Antitrust Matters’, 30 European Competition Law Review 205 (2009).

What Is To Be Done? 555 taken by the Ombudsman in competition cases to date is limited—with findings against the Commission being even more limited—the fact that parties are involving the Ombudsman does suggest a growing lack of confidence in the effectiveness and independence of the Hearing Officer. What can be done? Institutional independence, increased powers of factfinding, including the power to determine the facts, the power to investigate the Commission’s handling of the case, the introduction of adversarial procedures at the oral hearing and the power to challenge the Commission on substance are all possible improvements. The more the Hearing Officer resembles the Ombudsman, the less need there will be for recourse to the latter.

2. Other reforms Three additional suggestions may be made here: 1. Establish a system of complete file access based on protective orders which would eliminate the scandalously wasteful process of redaction, ensure fair file access and remove one of the procedural bases on which Commission decisions can be delayed and ultimately annulled. From the perspectives of the Commission, defendants and third parties, the preparation of non-confidential versions of file documents is probably the single most resource-intensive (and without doubt the most wasteful) element in current Commission procedures. It is also a potential source of serious procedural injustice in and/or annulment of Commission decisions. This reform has already been road tested on the basis of file access agreements between defendants and third parties, and would only be made more robust if incorporated in the Commission’s procedural toolbox. Prior to full implementation, file access agreements and abandonment of the practice of redaction and partial file access should become the rule rather than the exception. 2. Continue the development of a robust but critical system of private enforcement which frees the Commission to focus on its agenda and reduces the risk of agency over enforcement. Of course, for this to work the national courts will have to adopt a critical and soundly based economic approach. But one unanticipated benefit of private enforcement will be to divert cases from the European system where judicial control may not be adequate into national systems where it may be more robust. 3. Eliminate the ability of complainants to sue the Commission. The threat of such suits by rejected complainants actively promotes overenforcement. With an effective system of private enforcement, this right is no longer required to protect the interests of complainants.

556

Ian S Forrester* Quis custodiet ipsos custodes?1 Assessing the Judicial Role in a Lawful System of Competition Enforcement

1. Competition law and the challenge for the judge My topic is the judicial role in making European competition law both procedurally robust and substantively persuasive, at a time of unprecedented concern and uncertainty about the constitutional governance of the European Union. Advocates who appear in large and small judicial controversies frequently say that they have lost cases which they expected to win, or which they felt they deserved to win, and have won cases they expected to lose. As a practitioner who has argued a number of cases, I may properly be suspected of bias, or at least lack of detachment, when voicing a feeling that judicial review in Luxembourg in competition matters is not always appropriately rigorous. However, I believe that my sentiment is not unique, and indeed there is an abundance of literature on the topic, so I venture to assert that a fair number of practitioners are not comfortable with the current regime. Instead of proclaiming, like the Monday morning football fan, that the referee was no good, perhaps it is useful to reflect on the challenges confronting the Luxembourg Courts in competition matters. It is pointless to blame judges for doing the jobs ascribed to them, or to wish that the Courts enjoyed powers other than those with which they are currently endowed. The Treaty of Rome gave the Court the task of deciding appeals on the basis of an administrative law standard; legality, not correctness. Can the criticism directed at EU competition procedures be cured, given the Treaty’s constraints on judicial activity? * Queen’s Counsel at the Scots Bar; Honorary Professor, University of Glasgow. At the time of writing in 2013, Partner, White & Case, Brussels. I thank Deepa Varghese, Dr Sandra Keegan, Martin Möllman and Jérémie Jourdan for ideas and helpful contributions. The opinions expressed are wholly personal. 1 The quotation is usually translated as ‘Who will guard the guardians?’, but the entire text is more complicated: audio quid ueteres olim moneatis amici, ‘pone seram, cohibe.’ sed quis custodiet ipsos custodes? cauta est et ab illis incipit uxor: Juvenal, Satire VI, lines 347–348. Juvenal speculates that appointing watchers or guardians does not always solve a problem of spousal misbehaviour as the watching guardians may themselves fall into bad habits, so the reliability of the supervisors needs supervising.

558 Ian S Forrester I submit that standards of intensity of judicial review in competition matters vary to a considerable degree from jurisdiction to jurisdiction. The EU’s is not, I respectfully suggest, the most rigorous or the most predictable. Whereas in France and the UK enforcers are likely to be a little uneasy when confronting the uncertainties of litigation before their respective tribunals, enforcers who contemplate judicial review by the European Courts can be a little more confident, even if their professionalism makes them well prepared and ready for robust debate. Separately, the advocate who has defended a client at a competition law hearing in Brussels will usually be less satisfied than if the argument had occurred in London. The client will probably have even stronger feelings. And the advocate who appeals a decision to the EU Courts in Luxembourg has a considerable challenge to overcome. The Commission has an exceptional record of success in Luxembourg. Its last appellate defeat in a question of abuse of dominant position was some 30 years ago.2 In many cases, appellate scrutiny has consisted of ‘light touch’ judicial review.3 In any system of prosecution, it is natural that the prosecutor will be found right most of the time, so it would be wrong to conclude that there was a problem merely because appellants are usually unsuccessful. However, it is useful to consider whether the EU system appears to present problems and how these might be addressed. One elementary but fundamental challenge is the imprecision of the basic law. There is almost no proposition which is incapable of being advanced in a competition matter. Competition law lacks absolute unvarying principles. The basic prohibitions of EU competition law are contained in a few words in Articles 101 and 102 of the Treaty of Rome as amended. But, unlike other strict prohibitions in life (such as the Ten Commandments), competition law keeps changing. The rules are not absolute, rigid tax statutes; they are adaptable and plastic, reflecting evolution in policy. Common sense can take judicial analysis a long way, as well as some notion that agreements between competitors are likely to be sensitive and that monopolies should sometimes be restrained. However, there will be situations where the applicable legal rule will not be obvious, or where the policy needs of the factual situation are unclear. Rigour is essential, scepticism is healthy and argument brings clarity. On the merits of robust, clear, precise analysis, I quote the following from a book4 by Professor Gunther Bornkamm, the theologian father of Judge Joachim Bornkamm of the Bundesgerichthof: If the journey into this often misty country is to succeed, then the first requirement is the readiness for free and frank questioning, and the renunciation of an attitude

Case 322/81 Nederlandsche Banden Industrie Michelin v Commission [1983] ECR 3461. Ian Forrester, ‘A Bush in Need of Pruning: the Luxuriant Growth of “Light Judicial Review”’, in Claus-Dieter Ehlermann and Mel Marquis, eds, European Competition Law Annual 2009: Evaluation of Evidence and its Judicial Review in Competition Cases, Hart Publishing, 2011, 407 et seq. 4 Gunther Bornkamm, Jesus of Nazareth, Hodder & Stoughton, 1960, 10. 2 3

Quis custodiet ipsos custodes? 559 which simply seeks the confirmation of its own judgments arising from a background of belief or unbelief.

Courts are intrinsically unpredictable; that is the nature of ritual combat. The European Commission concluded in a survey that judges examining parallel patent disputes in different Member States reached different conclusions in 11 per cent of the cases.5 Thus, one in nine times, the German patent judge disagreed with the French judge on validity or infringement of the same patent and the same technology. In competition law, the role of the judge is particularly delicate since facts may be bitterly contested and can involve a direct conflict between individuals about technology, market size, nature and number of potential competitors, or the impact of new products, or between experts about how the market functions. In addition, competition law theories evolve, sometimes to an important degree. When doctrines have evolved, should a court rely on the earlier judgments of a competent court applying that doctrine or espouse the modern doctrine which departs from that old authority? Should the public authority in court have the task to educate and encourage the adoption of evolving policy? Or is its task to defend the line taken in the prosecution? Would judges do better if the lawyers acting for the public authority had to counsel the court rather than seeking to prevail judicially? And what happens when a court overrules a public authority: does this ruin the credibility and confidence of the agency? What level of scepticism, benign neutrality, respect or deference is it appropriate for the judges to deploy? Is it possible to assess how rigorous the level of review in Luxembourg is, and is that adequate under the European Convention on Human Rights (ECHR) and the Charter of Fundamental Rights? Is adequacy the relevant standard? At a time of convergence, it is worth considering the process by which one regulator’s opinion is favoured over another’s. Convergence between enforcement authorities as to the goals and the priorities is obviously desirable, but we do not have it in the European Union, within the United States or around the world. The International Competition Network does heroic work to promote transparency and enhance consistency, but large differences exist. Because the European Commission is the prime competition enforcer in the EU, and arguably the world leader in competition law creativity, its views are afforded especial weight. The Commission’s views are important because of its constitutional role in a devolved system and because it may be considered to have more expertise and experience. It will be worth studying carefully whether the Commission’s decision-making gives better results.

5 See Commission, ‘Pharmaceutical Sector Inquiry—Final Report’ (8 July 2009), http://ec.europa.eu/competition/sectors/pharmaceuticals/inquiry/staff_working_paper_ part1.pdf.

560 Ian S Forrester

2. Theories and fashions change The law (as applied in decisions by the College of Commissioners of the European Commission in 1962, 1991, 2004 or 2013) has evolved considerably. Europe is by no means the only continent where views or interpretations change or where enforcement choices are controversial. The arrangements of the United States for competition enforcement demonstrate that policy, trends and political tendencies have an effect on how antitrust decisions are taken. The Federal Trade Commission (FTC), created to remedy party political dissatisfaction over the vigour with which antitrust law was prosecuted, is an activist agency with a remit that is broader than pure antitrust. The FTC and the Department of Justice have had celebrated differences of opinion. They are not rivals, but they certainly are not identical twins. Yet the basic laws do not differ, whichever agency is interpreting them, though the two agencies may sometimes advocate different views of what antitrust law should provide. It is not the case that the American judiciary is confronted with a single fount of official antitrust wisdom. There is no doubt of the rigour with which US courts perform their task in antitrust matters. In the earliest days, the Sherman Act6 targeted large economic power, using populist rhetoric to justify action. The trusts were corporate equivalents of robber barons, so it is not surprising that their excesses were constrained by legislation. But there is little in common between those excesses and the infringing conduct which is today targeted by antitrust law. Even in the modern era theories have changed significantly. For nearly 50 years, the US and Canada were the only countries to ban cartels. Elsewhere, serious legal and economic opinion asserted that cartels were wholesome, desirable or even necessary to ensure national economic development. The Dutch construction sector was exceptionally articulate and robust in defending horizontal cooperation between competitors. Cementhandelaren7 is only one example of a case where horizontal cooperation is warmly endorsed by some parties. I am old enough to remember writing a memorandum for the revered Donald Holley summarising the different arguments in that case, and then listening to the debate between my betters about the merits of market allocations for consumers, employers and economic life generally. It is not the case that a rule of competition law will normally be perfectly clear and need only be applied to the established facts. Read Bill Kovacic, who bears the scars of many celebrated conflicts: The Post-Chicago School literature generally defines a broader zone for antitrust intervention. One body of Post-Chicago commentary describes how, in some

Sherman Antitrust Act, 2 July 1890. Case 8/72 Cementhandelaren v Commission [1972] ECR 977.

6 7

Quis custodiet ipsos custodes? 561 c ircumstances, exclusive dealing, tying, and other vertical restraints can facilitate the acquisition or maintenance of market power on grounds other than efficiency. Other Post-Chicago commentators have suggested how firms can use a mix of price and non-price strategies to diminish economic performance by deterring entry and expansion by rivals. Some Post-Chicago commentators accept the primacy of an efficiency framework, while others say that antitrust policy should serve distributional and other objectives. Post-Chicago observers generally express greater faith than do their Chicago School counterparts in the capacity of government institutions to make wise choices about when and how to intervene.8

I have been involved in several cases where the rules were freshly minted (as well as others where the challenge was to fit within an established line of cases). If we compare Volvo/Veng,9 Magill,10 IMS11 and Microsoft12 regarding compulsory licensing, it is evident that we have moved from a world where compulsion is barely conceivable to one where compulsory licensing is not fanciful in any circumstances. The law on vertical distribution used to be exceedingly prescriptive and ferociously penalised. The presence of a forbidden clause (an export ban) in one unenforced distribution agreement which had by accident not been remedied when similar distribution agreements were being cleaned up was deemed a serious infringement, and was fined in the case of Toshiba TEG.13 Another example of evolutionary policy can be seen in the first block exemption regulation for the distribution of motor cars, Regulation 123/85.14 It spelled out in elaborate detail the rules on security of tenure for dealers, the range of products to be carried, the situations in which cross-border purchases could be insisted upon by unwelcome consumers, and many other precise details. By contrast, the latest version, Regulation 461/2010,15 is much more tolerant and less prescriptive, leaving the parties to choose how to run their relationship. The block exemptions on distribution

8 See William Kovacic, ‘The Intellectual DNA of Modern US Competition Law for Dominant Firm Conduct: The Chicago/Harvard Double Helix’ [2007] Columbia Business Law Review 1. 9 Case 238/87 Volvo (UK) Ltd v Veng AB [1988] ECR 6211. 10 Commission Decision of 21 December 1988 in Case IV/31851—Magill TV Guide/ITP, BBC and RTE [1989] OJ L78/43, upheld on appeal: Case T-69/89 Radio Telefis Eireann v Commission [1991] ECR II-485, upheld on further appeal: Cases C-241/91 P and C-242/91 P Radio Telefis Eireann (RTE) and Independent Television Publications Ltd (ITP) v Commission [1995] ECR I-743. 11 Commission Decision of 3 July 2001 in Case COMP/38.044—NDC Health/IMS Health: Interim Measures [2002] OJ L59/18, withdrawn by Commission Decision of 13 August 2003 in Case COMP/38.044—NDC Health/IMS Health: Interim Measures [2003] OJ L268/69; Case C-418/01 IMS Health GmbH & Co OHG v NDC Health GmbH & Co KG [2004] ECR I-5039. 12 Case T-201/04 Microsoft Corp v Commission [2007] ECR II-3601. 13 Commission Decision of 5 June 1991 in Case IV/32.879—Viho/Toshiba [1991] OJ L287/39. 14 Commission Regulation (EEC) 123/85, [1985] OJ L15/24. 15 Commission Regulation (EU) 461/2010, [2010] OJ L129/52.

562 Ian S Forrester relationships16 and technology transfers17 are likewise more relaxed than their predecessors, in the sense that, instead of threatening dire punishment for contracts the terms of which do not fit the prescribed ‘safe’ template, they leave the parties the freedom to negotiate acceptable terms and to justify them if controversy arises. These evolutions in policy are further confirmations of my wider proposition—that competition law has few absolute rules. This fluidity makes judicial correction and supervision the more essential: they discipline an authority which could otherwise act in an arbitrary manner. It is, of course, relevant that the authority has a statutory mission and was in good faith, but those factors are not dispositive. The judicial task of performing quality control will also help to confer legitimacy on official action.

3. Differences between countries There are significant philosophical differences between nations’ competition laws: the US does not favour ‘free riding’ via parallel trade (GTE Sylvania18 is the classic case), whereas encouraging market integration via parallel trade used to be the most evident goal pursued by European competition law. A consideration of the number of early European cases which dealt with parallel trade would suggest that challenging private contractual obstacles to cross-border trade in consumer products was the prime enforcement objective of the Commission from 1970 to 1990. It was a feature distinguishing European law from other laws. Economic effects were not necessary, and the texts were enough to prove guilt. More fines were imposed for parallel trade infractions in that period than for any other category of conduct. In the United States, technology licences used to be governed by highly interventionist rules—the so-called ‘nine no-nos’,19 which have given way to

16 Commission Regulation (EU) 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. 17 Commission Regulation (EC) 772/2004 on the application of Article 81(3) of the Treaty to categories of technology transfer agreements, [2004] OJ L123/11. 18 Continental TV, Inc v GTE Sylvania Inc, 433 US 36 (1977). 19 See Bruce Wilson, ‘Know-How License Agreements: Field of Use, Territorial, Price and Quantity Restrictions’, remarks before the Fourth New England Antitrust Conference, 6 November 1970. The ‘nine no-nos’ were:

(1) (2) (3) (4) (5)

tying the purchase of unpatented materials as a condition of the license; requiring the licensee to assign back subsequent patents; restricting the right of the purchaser of the product in the resale of the product; restricting the licensee’s ability to deal in products outside the scope of the patent; a licensor’s agreement not to grant further licenses;

Quis custodiet ipsos custodes? 563 a less prescriptive, more tolerant and more realistic legal regime in which the merits and demerits of a practice are balanced. In this respect, European law has evolved in parallel: it used to be that almost everything in theory was forbidden but could be permitted with a theoretically available exemption which in reality usually was unobtainable. National courts coped with the practical absurdities of this regime quite successfully. The notification regime came to an end with the adoption of Regulation 1/2003.20

4. Differences in priorities Antitrust law evolves not because the rule changes like tax law, which is purely statutory, but because enforcement approaches and priorities change.21 I believe that, although regional cooperation in economic matters has taken firm root in Africa, South America and Asia, no other competition law regime has given priority to attacking contractual barriers to cross-border trade. But since, say, the year 2000, pursuing parallel trade cases seems to have been given a lower enforcement priority in Europe, although the topic has not entirely disappeared, as the various pharmaceutical cases about Greek and Spanish parallel trade problems demonstrate.22 Competition law has permeated deeply into the economic marrow of commercial life and business practice. Thus, more business leaders are conscious of the need to comply, and try to achieve compliance. But at the same time, what the law condemns has been evolving and changing. It is not like health (6) (7) (8) (9)

mandatory package licenses; royalty provisions not reasonably related to the licensee’s sales; restrictions on a licensee’s use of a product made by a patented process; and minimum resale price provisions for the licensed products.

20 Council Regulation (EC) 1/2003 on the implementation of the rules on competition laid down in Articles 81 and 82 of the Treaty, [2003] OJ L1/1. 21 See Damien Géradin and Nicolas Petit, ‘Judicial Review in European Union Competition Law: A Quantitative and Qualitative Assessment’, in Massimo Merola and Jacques Derenne, eds, The Role of the Court of Justice of the EU in Competition Law Cases, Bruylant, 2012, 21 et seq, 32: ‘While theoretical diversity may be a source of richness, and avoid the pitfalls of relying on one-sided theories, it may also be a source of confusion, and thus of errors, both in terms of competition policy-making, but also in terms of adjudication. Competition authorities will often be confronted with a patchwork of inconsistent theories concocted by clever complainants and defense counsel … This may lead to confused, and thus generally misguided, enforcement.’ 22 See Case C-53/03 Syfait and others v GlaxoSmithKline AEVE and GlaxoSmithKline plc [2005] ECR I-4609; Joined Cases C-468 to 478/06 Lelos kai Sia EE v GlaxoSmithKline [2008] ECR I-7139; Joined Cases C-501/06 P, C-513/06 P, C-515/06 P and C-519/06 P GlaxoSmithKline Services and Others v Commission and Others [2009] ECR I-9291; Commission v GlaxoSmithKline S ervices Unlimited; European Association of Euro Pharmaceutical Companies (EAEPC) v C ommission; Asociación de exportadores españoles de productos farmacéuticos (Aseprofar) v Commission [2009] ECR I-9291.

564 Ian S Forrester and safety rules, where the progression is linear in the direction of becoming stricter; competition law enforcement fluctuates in priorities, becoming stricter in some areas but less strict in others. And, as noted above, much can depend on the creativity of complainants, of talented lawyers and intelligible economists, who will contend that a new approach is only a trifle novel. In the previous sections I have pointed out that competition law is not a rigid and predictable set of rules but that, to the contrary, its enforcement varies according to times, fashions and geography. The law changes, getting more severe or less so. Enforcers make choices as to their targets and their accusations. They may choose no longer to challenge certain conduct, or to try alternative methods of addressing the supposed problem. At the same time, the law imposes penalties, is criminal for the purposes of human rights standards, and major decisions are announced with condemnatory relish. The quality of the enforcement policy is (wrongly, in my view) often deemed to be linked to the height of the fines imposed. For a number of years, fines levied in Europe for breaches of the competition rules were higher than penalties imposed for any offence anywhere on earth, and may indeed still be the highest. I have suggested that such levels lie above that which is necessary to achieve appropriate punishment and deterrence. In a legal world where the criteria and theories shift, fines should not be imposed for novel offences. It manifestly ought not to be the law that whatever DG Competition condemns is liable to a fine for that reason. Fines should be no higher than is necessary to achieve the appropriate level of punishment and deterrence. In contrast to the substantive law, which fluctuates, fines seem to evolve in only one direction—up. These characteristics make the judicial function even more important. Criminal law demands a high standard of procedural due process. Deterrence is a rational rationale only if the conduct is known to be unlawful in advance. That I have to record such an obvious proposition is regrettable. I further propose that a law which evolves significantly needs judicial oversight to guard against the dangers of arbitrariness.

5. Does ‘old’ law help when considering ‘new’ cases? The relevant authoritative texts will rarely offer a clear answer. Inevitably, the enforcement authority as well as the judge will need to consider whether a certain practice—viewed in its totality—is competitive or anticompetitive. National and European judges naturally give especial respect to the opinions of the prime European competition law legislator, drafter, enforcer and prosecutor.23 A separate question is how should the Court respond if the See Forrester, ‘A Bush in Need of Pruning’, above n 3.

23

Quis custodiet ipsos custodes? 565 Commission expresses itself inconsistently? Is the Court entitled to say to the prosecutor that certain conduct is no longer unacceptable in modern practice? Is the prosecutor entitled to rely on old judicial or administrative condemnations of a now-tolerated practice? Such challenges occur most often where the law has evolved from being prescriptive and even punitive, if certain e lements were present, into a modern law which considers the pro-competitive and anticompetitive factors in a neutral way, giving particular weight to the actual effects rather than the wording of the contracts. As one example, many practising lawyers and business people have difficulty in making sense of European law on discounts. The concern about discounts is that the dominant supplier may use them to exclude competitors; the merit of discounts is that they lower prices and intensify competition between suppliers. The European law on discounts is generally regarded as eccentric due to over-exuberant formulations of the principles in older cases. Modern theories on discounts by a dominant player are notably different from the doctrines enunciated by the ECJ in the classic early case of Hoffman-La Roche Vitamins,24 where a whistleblower, Stanley Adams, delivered to the Commission a mass of data showing that a Swiss company had established a system of rebates which depended on the customer’s loyalty, whether the customer was tiny or huge. Thus the customer would get an especially attractive price if it bought nearly 100 per cent of its needs of vitamins from the dominant supplier. In a sense, the customer was being ‘paid’ not to buy from a rival supplier. Viewed otherwise, the customer was given a good price and had a reliable supplier. The case attracted huge attention because of the misfortunes of Mr Adams, whose identity became evident to his former employer due to the line of questioning used by the Commission. He was subsequently arrested in Switzerland on charges of economic espionage, and his then-wife took her own life during his imprisonment. EC/Swiss relations were scarred by these events, and the accused company’s conduct was scrutinised with great scepticism. The prolonged, though ultimately successful, efforts of

Case 85/76 Hoffmann-La Roche & Co AG v Commission [1979] ECR 461, para 89:

24

An undertaking which is in a dominant position on a market and ties purchasers— even if it does so at their request—by an obligation or promise on their part to obtain all or most of their requirements exclusively from the said undertaking a buses its dominant position within the meaning of Article 86 of the Treaty, whether the obligation in question is stipulated without further qualification or whether it is undertaken in consideration of the grant of a rebate. The same applies if the said undertaking, without tying the purchasers by a formal obligation, applies, either under the terms of agreements concluded with these purchasers or unilaterally … discounts conditional on the customer’s obtaining all or most of its requirements—whether the quantity of its purchases be large or small—from the undertaking in a dominant position.

566 Ian S Forrester Mr Adams25 to get compensation kept the saga alive for years. His subsequent misfortunes have included a failed pig farm and a conviction for attempted murder. The outcome of the appeal was in the circumstances not surprising. Subsequent judicial analysis has focused not on whether there were actual economic effects of the pricing practices (even if not stipulated in writing) but on whether the practices might induce loyalty. In a succession of cases about discounts, the question has arisen of how to treat discount policies which were—not surprisingly—intended to attract and retain customers. Did it make any difference that the policies were successful or unsuccessful? Suppose the customer used one supplier’s discount to obtain a better price from the other supplier: was that a bad thing? Suppose customers regularly changed suppliers: was that irrelevant? Seeking to induce customer loyalty through discounts appears to be a kind of absolute offence. Competitors have successfully accused each other of abusive pricing (British Airways26—a very large enterprise; and Tomra27—a quite small enterprise). These were robustly defended on the grounds of commercial reasonableness. To what extent is it relevant that the discount or rebate did not hinder the success of the competitor? Is it relevant that there is no contractual duty on the customer which compels action to get the discount? Is offering a rebate for loyalty an offence in and of itself even if loyalty is not engendered? Does any reader think that a big supplier of goods or services should be punished for trying to keep customers loyal, indeed enthusiastic? What happens if the supposedly dominant player launches a programme of discounts which have no success in wooing customers but which can be said to have been liable, if they had encountered commercial success, to induce loyalty within the meaning of previous cases? Tomra (paragraph 70) provides an example: In the event that an undertaking in a dominant position makes use of a system of rebates, the Court has ruled that that undertaking abuses that position where, without tying the purchasers by a formal obligation, it applies, either under the terms of agreements concluded with these purchasers or unilaterally, a system of loyalty rebates, that is to say, discounts conditional on the customer’s obtaining—whether the quantity of its purchases is large or small—all or most of its requirements from the undertaking in a dominant position …

Thus, the old law appears to condemn discount practices which contain no contractual duty to refrain from buying from other suppliers but which induce loyalty. To constrain competitors’ attempts to undercut each other seems a rather improbable economic crime. The curiosities of European competition law on discounts are well recognised as distinctive and—I would

Case 145/83 Adams v Commission [1985] ECR 3539. Case C-95/04 P British Airways plc v Commission [2007] ECR I-2331. 27 Case T-155/06 Tomra Systems ASA and Others v Commission [2010] ECR II-4361; appeal dismissed: Case C-549/10 P, EU:C:2012:221. 25 26

Quis custodiet ipsos custodes? 567 argue—potentially irrational. National authorities have elected not to prosecute such discount policies: in the UK, the authorities have shown no appetite to develop a rule—even in accordance with modern doctrines—which may hinder competition on the merits and which is almost impossible to apply confidently. However, the European Courts have been broadly supportive of Commission challenges to discounts even where the two adversaries were evenly matched and competing ferociously. (Virgin and British Airways might both be thought well able to look after themselves.) Responsive to these criticisms, in its Guidance Paper on the Enforcement Priorities in Applying Article 82,28 the Commission eschewed the absolute offence and relied on effects: 23 … Vigorous price competition is generally beneficial to consumers. With a view to preventing anti-competitive foreclosure, the Commission will normally only intervene where the conduct concerned has already been or is capable of hampering competition from competitors which are considered to be as efficient as the dominant undertaking. 24 However, the Commission recognises that in certain circumstances a less efficient competitor may also exert a constraint which should be taken into account when considering whether particular price-based conduct leads to anti-competitiveforeclosure. The Commission will take a dynamic view of that constraint … 25 If the data clearly suggest that an equally efficient competitor can compete effectively with the pricing conduct of the dominant undertaking, the Commission will, in principle, infer that the dominant undertaking’s pricing conduct is not likely to have an adverse impact on effective competition, and thus on consumers, and will therefore be unlikely to intervene. If, on the contrary, the data suggest that the price charged by the dominant undertaking has the potential to foreclose equally efficient competitors, then the Commission will integrate this in the general assessment of anti-competitive foreclosure … taking into account other … evidence.

On this basis, the topic is complex, but everything will be carefully studied. Reality is indeed important. However, once a matter is before the judges in Luxembourg, the Commission has been seen to rely on the old law when rejecting the argument that it had not demonstrated actual effects on competition: there is in any event no requirement in the case-law to demonstrate actual foreclosure in order to prove an infringement of Article 102 of the Treaty … the Community Courts have established that ‘for the purposes of establishing an infringement of Article 82 EC, it is not necessary to demonstrate that the abuse in question had a concrete effect on the markets concerned. It is sufficient in that respect to demonstrate that the abusive conduct of the undertaking in a dominant position tends to

28 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 C45/02.

568 Ian S Forrester restrict competition or, in other words, that the conduct in question is capable of having or likely to have such an effect …’ with regard to conduct that constitutes granting fidelity discounts within the meaning of the Hoffmann La Roche case law, there is no requirement in the case-law even to demonstrate capability of foreclosure in order to prove an infringement of Article 102 of the Treaty … a violation of Article 102 TFEU may also result from the anticompetitive object of the practices pursued by a dominant undertaking. Indeed, the contested Decision found that establishing the potential foreclosure effects of … exclusivity rebates … was unnecessary for finding that these practices are in breach of Article 102 TFEU. The reason for this is because, as the Court explained, practices of this kind ‘will also be liable to have such an effect’.29

A discount that is unsuccessful but is intended to induce loyalty is on this basis hazardous. If the dominant player wishes to stay out of trouble, it will eschew discounts since it appears to make no difference whether the discount has an effect on customers, or whether it truly excludes other competing offerors from a realistic chance of making business in the face of the controversial discount. Thus, while the ‘advisory and hortatory department’ of the Commission acknowledges the merits of discounts (lower prices, intensity of competitive pressure, opportunities for the buyer to negotiate a better deal) and their dangers (foreclosure of meritorious potential competitors), the ‘litigation department’ may follow a different approach. Should the judge apply the modern doctrines? If there is a difference between the law as declared in the cases and the law as declared in the Guidance Paper, by what should the Court be guided? Is it possible that quasi-criminal law can be defined according to two inconsistent standards by the prosecutor, who can choose which version to advance? If such discrepancy exists, by which standard shall the Court be guided? I note that a court which has the benefit of a neutral expert has a real advantage over one that does not. In several countries the highest court had the advantage of receiving advice, as opposed to advocacy, from a person entrusted with functions comparable to that of the Advocate General. The amicus curiae helps the appellate policy function. The Advocate General is a valuable member of the Court of Justice of the European Union. Of course, no lawyer will wish to mislead the Court, but there is a functional difference between the impact of the lawyer who says to the Court ‘The contested action was valid and should be upheld’ and ‘The contested action presents the following points of principle’. Similarly, passionate arguments will no doubt ensue as to the currently topical issue of the settlement of pharmaceutical patent litigations. Are these I quote from a certain submission by the Commission to the Court.

29

Quis custodiet ipsos custodes? 569 to be regarded as ‘by-object’ offences, so inexcusably illegal that there is no need to enquire into their effects in the marketplace, like bribes to a competitor to leave the market, or are they susceptible to an analysis of the actual circumstances and whether they brought an end to litigation or legitimated the sale of a competing generic product? The FTC and the Commission have argued the former theory; the Supreme Court favoured a rule of reason enquiry (called by some the ‘sniff test’), not a blanket condemnation.30 This is one more illustration of the proposition that interpretations of basic texts and basic concepts evolve and vary considerably within and between jurisdictions.

6. The Court is robust and radical in some fields It is notable that, in the one area of Treaty competence where the Court of Justice has unlimited jurisdiction, namely competition cases involving penalties, there has been a tradition of deference, or what looked like deference.31 By contrast, with far less Treaty authority, the ECJ and its successors have revolutionised the constitutional status of European law (Van Gend en Loos32 through Les Verts)33, and damages for breach of the law (Factortame),34 and have more or less created modern law on freedom of movement of persons, residence, migration, access to social benefits, education, deportation and gender discrimination. The evolution of the law in these highly sensitive areas has been jerky, sometimes even explosive. The Court has been called activist, and its supposedly interventionist approach has been criticised with varying degrees of sharpness (Rasmussen35 was an early critic, and Bengoetxea36 is a modern defender of the Court’s record). I am not suggesting that we need to destabilise the pillars of the legal and administrative law temples. Constitutional doctrines occasionally move in jerks, but rarely come out of a totally cloudless sky. Van Gend en Loos37 was

FTC v Actavis, Inc, 570 US 756 (2013). See Ian Forrester, ‘A Challenge for Europe’s Judges: The Review of Fines in Competition Cases’, 36 European Law Review 185 (2011). 32 Case 26/62 NV Algemene Transport- en Expeditie Onderneming Van Gend en Loos v Netherlands Inland Revenue Administration [1963] ECR 3. 33 Case 294/82 Parti écologiste ‘Les Verts’ v European Parliament [1986] ECR 1339. 34 Joined Cases C-46 and 48/93 Brasserie du Pêcheur SA v Bundesrepublik Deutschland and the Queen v Secretary of State for Transport, ex parte: Factortame Ltd and others [1996] ECR I-1029. 35 Hjalte Rasmussen, On Law and Policy in the European Court of Justice: A Comparative Study in Judicial Policymaking, Martinus Nijhoff Publishers, 1986. 36 Joxerramón Bengoetxea, The Legal Reasoning of the European Court of Justice, Clarendon Press, 1993. 37 Van Gend en Loos, above n 32. 30 31

570 Ian S Forrester an interesting development, but the problem and possible approaches to the problem (national law and European law being out of step) had been extensively discussed before the case arrived in Luxembourg.38 It is not surprising that the Courts in Luxembourg regard the Commission’s policy appraisals as especially useful, authoritative and likely to be sound. By contrast, the Court of Justice has granted little weight to Member States’ policy appraisals in a succession of matters (deportation, food safety, environmental rules, technical regulations). Thus the Commission has enjoyed far more of a margin of discretion than the Member States. Putting it differently, the principle of proportionality has been invoked more frequently against Member States than against the Commission. It is noteworthy that, by contrast, the ECJ has been criticised repeatedly for being ‘activist’, for being too creative in the making of new law, for being an independent actor in the building of the European house and for stepping in where the Member States were unable to agree. Thus, in the fields of free movement of goods, nationality and residence, and free movement of persons, European law is often the fruit of the Court’s case law rather than legislation negotiated between the Member States.39 These phenomena prompt three remarks. First, the Court has shown itself able to disagree with public authorities in scores of cases. Second, it seems to leave more room to manoeuvre, more margin of discretion and perhaps a longer leash in the case of the Commission, a sister institution, than in the case of Member States. Third, European competition law has become more radical than American competition law as a result of the judicial successes of the Commission in major dominance cases.

7. The European Courts as locomotives of legal innovation and policy change It is not difficult to explain the potency of European law in an introductory lecture to law students. Such cases as Defrenne,40 Cassis de Dijon41 or

38 Indeed, the London–Leiden meeting (June 2013) of the British Institute of International and Comparative Law and the University of Leiden’s Europa Institute devoted an entire day to discussing those antecedents. 39 I would note, however, that disregarding national justifications for legislative measures had the effect of achieving market integration but did not guarantee that the rules of the integrated market correspond to international standards (food, environment, telecommunications and other areas are possible examples of international disparities). 40 Case 43/75 Defrenne v Société Anonyme Belge de Navigation Aérienne Sabena [1976] ECR 455. 41 Case 120/78 Rewe-Zentral AG v Bundesmonopolverwaltung für Branntwein [1979] ECR 649.

Quis custodiet ipsos custodes? 571 ravier v Liège42 each advanced the law constitutionally further than the G drafters of the Treaty of Rome had contemplated or where the Member States desired to travel. The topics of equality of employment terms for women, free movement of goods regulated differently in different Member States and the right to study abroad were, and remain, fundamental to the confirmation of the freedoms which Union citizens today enjoy. Each case arose as a judicial response to questions framed by a national court hearing a legal controversy with a European flavour.43 During the 1970s, a period of political stagnation, the Court (in the eyes of many) played a constitutional role as an independent actor in the construction of an integrated Europe. The Court has often been regarded as an independent force in the creation of a united Europe, by reference to its judicial activism in several areas of national law, often where the Member States could not agree on legislation. Horsley’s article on this topic44 comprehensively reviews the literature regarding the Court’s performance as a motor for integration and as an independent actor on the European stage. Sometimes the Court has had to consider the creative invocation of European law in order to challenge national rules which were controversial domestically but which were not intrinsically ‘European’ in orientation or intention. In the Sunday Trading45 cases the Court tried to produce a European rule but subsequently discontinued the effort and reversed itself. In Stoeckel46 there was a challenge to rules against night work by women in France, and in Johnston v Chief Constable47 there was a challenge based on European law to an essentially national rule, on the basis of principles of equality of opportunity for men and women. In cases like Reyners48 and Marks & Spencer49 (regarding, respectively, freedom of establishment of lawyers and the tax treatment of losses within corporate groups), the Court has found a way of short-circuiting stalled negotiations or a blocked legislative process. So there is no shortage of cases where the Court has been a lively, independent actor

Case 293/83 Gravier v City of Liège [1985] ECR 593. Ian Forrester, ‘The Judicial Function in European Law and Pleading in the European Courts’, 81 Tulane Law Review 647 (2007). 44 Thomas Horsley, ‘Reflections on the Role of the Court of Justice as the ‘Motor’ of European Integration: Legal Limits to Judicial Lawmaking’, 50 Common Market Law Review 1 (2013). 45 Case C-145/88 Torfaen Borough Council v B&Q plc [1989] ECR 3851. 46 Case C-345/89 Alfred Stoeckel [1991] ECR I-4047. 47 Case 222/84 Marguerite Johnston v Chief Constable of the Royal Ulster Constabulary [1986] ECR 1651. 48 Case C-2/74 Jean Reyners v Belgium [1974] ECR 631. 49 Case C-62/00 Marks & Spencer plc v Commissioners of Customs & Excise [2002] ECR 6325. 42 43

572 Ian S Forrester on the European integration front. It is not my purpose to praise or to criticise these decisions, or to regret them, as the Court was faced with a genuine dispute which required a determination in light of the Treaty, the legislation and previous jurisprudence. My purpose is instead to suggest that the Court has plenty of experience of making itself unpopular with public authorities. This brings us to the Court’s role in the shaping of European competition law in reference cases. In Grundig and Consten50 it acceded to the market integration goals proposed by the Commission. In Oscar Bronner51 it commended free enterprise and individual effort. In the compulsory licensing cases it produced a set of rational (though expanding) parameters in Volvo/Veng52 (reference), Magill 53 (appeal) and IMS 54 (reference after an appeal), but then seemed to abandon them in Microsoft 55 (appeal). One can note that when the Court responds to national requests for preliminary rulings, it is more ready to be expansive; in appeals it has on occasion been regrettably deferential. There have been appellate competition cases where the Luxembourg Courts were exceedingly interventionist. Woodpulp56 was one; Italian Flat Glass57 was another. Tetra Laval,58 Schneider Legrand59 and Airtours60 were three others, each in the field of mergers, where the Court’s overturning of three decisions within the space of about a year provoked a wholesome revolution in how merger cases were handled by the Commission. So it is not the case that the courts are unable to perform rigorous judicial review. They have sometimes done so, but not always. The impact of the European Courts’ appellate jurisdiction on the evolution of the law has been less than that of the US court hierarchy. I invited some friends to nominate their favourite ECJ competition case, and received a number of candidates: Magill (compulsory licensing of copyright); Bodson v Pompes Funèbres61 (access to public cemeteries); Italy v Commission62 (use of technology to divert traffic to cheaper service providers);

50 Joined Cases 56 and 58/64 Établissements Consten SARL and Grundig-Verkaufs-GmbH v Commission [1966] ECR 299. 51 Case C-7/97 Oscar Bronner GmbH & Co KG v Mediaprint Zeitungs- und Zeitschriftenverlag GmbH & Co KG and Others [1998] ECR I-7791. 52 Volvo v Veng, above n 9. 53 Magill, above n 10. 54 IMS Health, above n 11. 55 Microsoft, above n 12. 56 Joined Cases C-89, 104, 114, 116, 117 and 125 to 129/85 A Ahlström Osakeyhtiö and others v Commission [1993] ECR I-1307. 57 Joined Cases T-68, 77 and 78/89 Società Italiana Vetro SpA, Fabricca Pisana SpA and Vernante Pennitalia v Commission [1992] ECR II-1403. 58 Case T‑5/02 Tetra Laval v Commission [2002] ECR II‑4381. 59 Case T-310/01 Schneider Electric SA v Commission [2002] ECR II-4071. 60 Case T-342/99 Airtours plc v Commission [2002] ECR II-2585. 61 Case 30/87 Corinne Bodson v SA Pompes funèbres des régions libérées [1988] ECR 2479. 62 Case 41/83 Italy v Commission [1985] ECR 873.

Quis custodiet ipsos custodes? 573 Tournier63 (royalties); Night Services64 (economic reality); Port of Genoa65 (regulation of docking services); and Post Danmark66 (economic effects again). Giving a critique of these is beyond the scope of this article, but the range of the controversies addressed should suffice to demonstrate that the General Court and the Court of Justice of the European Union have demonstrated a capacity to produce judgments on competition law controversies which are robust and convincing. Disagreeing with the public authority is far from a disloyalty to a ‘sister institution’. It is a vital form of quality control. I respectfully submit that a court which is not afraid to displease the M ember States should not refrain from displeasing the Commission. If it is true that, in appellate matters, the Court’s performance has been more muted, this likely reflects in part the nature of judicial review prescribed in the Treaty of Rome as amended. Whereas in matters involving ‘penalties provided for’ in ‘Regulations adopted by the European Parliament and the Council’, the Court of Justice of the European Union has ‘unlimited jurisdiction’ in respect of other appellate matters, its jurisdiction is more modest according to paragraph 263 TFEU: The Court of Justice of the European Union shall review the legality of … acts of … the Commission … It shall for this purpose have jurisdiction on grounds of lack of competence, infringement of an essential procedural requirement, infringement of the Treaties or of any rule of law relating to their application, or misuse of powers.

I note in the bygoing that the Court’s judgments can at times be rather alarming. If one looks at Article 102, the merit of economic effects as a criterion can be recognised (as it was, for example, in Post Danmark),67 but one can see under Article 101 a mushrooming of new by-object infringements.68 Thus, Allianz Hungary,69 Expedia70 and the slightly different case of Pierre Fabre71 seem to endorse extremely aggressive and prescriptive competition

Case 395/87 Ministère Public v Jean-Louis Tournier [1989] ECR 2521. Joined Cases T-374/94, T-375, T-384/94 and T-388/94 European Night Services Ltd v Commission [1998] ECR II-3141. 65 Case C-179/90 Merci Convenzionali Porto di Genova SpA v Siderurgica Gabrielli SpA [1991] ECR I-5889. 66 Case C-209/10 Post Danmark A/S v Konkurrencerådet, EU:C:2012:172. 67 Ibid. 68 In the time since this chapter was completed, the Court of Justice appears to have circumscribed this expansive trend to a certain extent. See Case C-67/13 P Groupement des cartes bancaires (CB) v Commission, EU:C:2014:2204. 69 Case C-32/11 Allianz Hungária Biztosító Zrt and Others, EU:C:2013:160. 70 Case C-226/11 Expedia Inc v Autorité de la concurrence and Others, EU:C:2012:795. 71 Case C-439/09 Pierre Fabre Dermo-Cosmétique SAS v Président de l’Autorité de la concurrence [2011] ECR I-9419. 63 64

574 Ian S Forrester law principles under Article 101. In Allianz, the question was whether it was anticompetitive for a car repair shop to be accorded a higher hourly rate for carrying out repairs covered by insurance if the repair shop had sold insurance policies issued by the same insurer. The merits in competition law were evidently unsure: the Hungarian court was sufficiently unsure that it made a reference. The Advocate General found there to be no breach of the competition rules; the European Commission felt the contrary. The Allianz judgment confidently condemned the arrangements as a ‘by-object’ infringement, on the basis that, among other things, it was a means for the insurers to increase market share. In Pierre Fabre, a not-very-large skin-care company was told that it could not forbid its resellers to make internet sales. In both cases, it is not evident why such severe, indeed absolute, new infringements are being created. One can argue that, as a matter of consumer welfare, a car repairer should be paid the same price for replacing a dented bumper, whether or not the repairer sells insurance policies for the insurer who is going to pay the bill. One can argue that internet selling is the modern way, and that it is oldfashionedto prohibit resellers from using that method. But I would challenge the proposition that in either case there is a gross infringement of the competition rules, so profound that it is not necessary to check whether there is any effect in the marketplace. I question whether the Court was wise to follow the Commission’s encouragement to be so bold, whether the explanation is deference or exuberance. Now, carefully criticising these is for another publication (can it be that there is no need to show the effect of a contractual provision which seems to fit the de minimis exception?), but for present purposes I will say only that these judgments show that the Court of Justice is not afraid to take strong positions on matters of competition, and that encouragement by the European Commission plays a role therein. In the next section, I will review the numerous occasions where both Courts have taken weak appellate positions on matters of competition, and I will highlight the relevance of this state of affairs from the perspective of the ECHR.

8. Where do we stand as to the ECHR? What should and what will be the impact of the ECHR and the Charter of Fundamental Rights on how competition cases are handled by the European Union? Although the Courts in Luxembourg have always enjoyed unlimited jurisdiction in fining matters, in past years they have frequently declined to do more than make a bare review of legality, and have often refused on constitutional grounds to consider whether a fine was proportionate. In recent years, the General Court has often not engaged in significant proportionality

Quis custodiet ipsos custodes? 575 review of the fines.72 The ECJ, in turn, has generally declined to review the proportionality of fines with a view to overruling positions taken by the General Court: while … the Court cannot substitute, on grounds of fairness, its assessment for that of the Court of First Instance giving judgment in the exercise of its unlimited jurisdiction as to the amount of the fines imposed on undertakings by reason of their infringement of Community law …73

In Vitamins, the then-Court of First Instance (CFI) went so far as to suggest that its unlimited jurisdiction is only activated when there is illegality: It is possible for the Court to exercise its unlimited jurisdiction under Article 229 EC and Article 17 of Regulation No 17 only where it has made a finding of illegality affecting the decision, of which the undertaking concerned has complained in its action, and in order to remedy the consequences which that illegality has for determination of the amount of the fine imposed, by annulling or adjusting that fine if necessary.74

This went beyond deference into error, I suggest. Similar language was found in other cases: ‘the Commission nevertheless has a wide discretion in assessing the quality and usefulness of the cooperation provided by the various members of a cartel, and only a manifest abuse of that discretion can be censured’.75 The suggestion here that only a ‘manifest abuse’ of the Commission’s discretion is capable of being censured by the Courts seems very doubtful. The proper exercise of the Court’s unlimited jurisdiction requires the Court to correct any abuse or error that it detects in the Commission’s reasoning, manifest or otherwise. While it is true that in the Vitamins case the debate related to the value of a confession to the Commission, while the prosecution is best placed to assess the value of a guilty party’s confession and while it may be difficult for a court to reach a conclusion about the relative values of

72 See, eg Joined Cases T-71, 74, 87 and 91/03 Tokai Carbon and Others v Commission [2005] ECR II-10; Case T-38/02 Groupe Danone v Commission [2005] ECR II-4407; Joined Cases T-217 and 245/03 FNCBV and Others v Commission [2006] ECR II-4987; Case T-15/02 BASF AG v Commission [2006] ECR II-497; Case T-279/02 Degussa v Commission [2006] ECR II-897, where proportionality pleas were dismissed. 73 Case C-125/07 P Erste Group Bank AG v Commission [2009] ECR I-8681, para 187. See also Case C-320/95 P Società Finanziaria Siderurgica Finsider v Commission [1994] ECR I-5697, paras 45–46; Case C-310/93 P BPB Industries and British Gypsum v Commission [1995] ECR I-865, para 34; Joined Cases C-238/99 P, C-244/99 P, C-245/99 P, C-247/99 P, C-250/99 P to C-252/99 P and C-254/99 P Limburgse Vinyl Maatschappij and Others v Commission [2002] ECR I-8375, para 617. 74 BASF, above n 72, para 582. This BASF judgment is probably wrong, as it contradicts the notion of unlimited jurisdiction. 75 Joined Cases T-236, 239, 244 to 246, 251 and 252/01 Tokai Carbon and Others v Commission [2004] ECR II-1181, para 371; upheld on appeal: Case C-328/05 P SGL Carbon v Commission [2007] ECR I-3921, para 88.

576 Ian S Forrester c ontrasting pieces of testimony, it might be argued that the information thus tendered and the value thereof ought to be capable of objective assessment by a court. One aspect of the Commission’s decision seems no less susceptible in principle to an intense standard of review than any other. It seems legally wrong or, at least, institutionally inappropriate to have eschewed ‘unlimited jurisdiction’. In Wieland-Werke the Court stated: in areas such as determination of the amount of a fine imposed pursuant to Article 15(2) of Regulation No 17, where the Commission has a discretion, for example, as regards the amount of increase for the purposes of deterrence, review of the legality of those assessments is limited to determining the absence of manifest error of assessment.76

So the Luxembourg Courts have occasionally been reproached for applying inconsistently intense levels of judicial review in competition cases.77 The judgment of the Court of Justice in Chalkor78 may have changed matters. In Chalkor and KME,79 the appellants complained that the CFI had used its familiar deferential language from Wieland-Werke to limit the scope of its review as to fines, whereas with respect to fines there was no scope for deference in light of the Court’s unlimited jurisdiction as well as the requirements of the Convention. The ECJ said that, although the CFI had (inappropriately) used the language of deference, it had in fact exercised non-deferential review of the arguments raised. So the appeals were unsuccessful. However, the standard of judicial review in matters where the Court has unlimited jurisdiction was redefined for the future. Such review should henceforth include a verification of whether the factual assessments were free of manifest error. In future, the appellate judge should examine thoroughly the facts while at the same time being cautious about second-guessing Commission assessments in complex technical or economic matters. This might be called ‘legality plus’ or ‘merits minus’. Does the Chalkor judgment remedy the problem? I think not.

76 Case T-116/04 Wieland-Werke v Commission [2009] ECR II-1087, paras 32–33. See also Case T-241/01 Scandinavian Airlines System v Commission [2005] ECR II-2917, para 79, cited in Wieland-Werke (‘It next has to be examined whether the Commission’s assessment of the seriousness of the infringements, having regard to the three factors of their nature, the extent of the geographic market concerned and their actual impact on the market, is vitiated by obvious error’). 77 There is considerable literature on this topic. See, eg Merola and Derenne, eds, The Role of the Court of Justice, above n 21; Forrester, ‘A Bush in Need of Pruning’, above n 3; Forrester, ‘A Challenge for Europe’s Judges’, above n 31; Wouter Wils, ‘The Combination of the Investigative and Prosecutorial Function and the Adjudicative function’, 27 World Competition 202, 216 (2004); Fernando Castillo de la Torre, ‘The 2006 Guidelines on Fines: Reflections on the Commission’s Practice’, 33 World Competition 359 (2010). 78 Case C-386/10 P Chalkor AE Epexergasias Metallon v Commission [2011] ECR I-13085. 79 Case C-272/09 P KME Germany AG v Commission [2011] ECR I-12789.

Quis custodiet ipsos custodes? 577

9. The ECHR and competition decision-making If we (the community of teachers, enforcers, judges and practitioners) were asked to describe a perfect basis for enforcing the European competition rules in a satisfactory, efficient, robust manner, we would assuredly not opt for what exists today. There are several overlapping problems with the current system of enforcement which can for present purposes be summarised as administrative procedures and which seem not to offer an accused company the chance of an independent determination of its guilt or innocence after a fair examination of evidence. There is no hearing by a decision-maker and no confrontation of the accused with witnesses against him. The decision-making and the investigating function are not separated in the case team handling the reaching of a decision. There are truly rigorous internal checks and balances, but these are not visible to the outsider; indeed, Commission enforcers would justifiably complain that the procedural constraints on officials are much too burdensome, that the process is too slow and that concluding cartel cases clogs up the institution and prevents more interesting cases being advanced. The structural problems of Commission procedures are then exacerbated by the fact that judicial review appears to be unpredictable, and also uses the language of deference, notably with respect to ‘complex’ ‘economic’ or ‘technical’ ‘assessments’.80 The question for present purposes is not whether a better system could be devised, but whether the present system is satisfactory when viewed under the Charter and the Convention. It is to the Convention, and specifically its Articles 6(1) and 6(2), that we now turn. In order to assess the acceptability of current arrangements, we need to step back in the process of examining whether competition law procedures match modern standards of fairness in light of the ECHR. There have been many Strasbourg challenges to the imposition by the public authority on a citizen of some disadvantage: the loss of sheltered housing, a traffic ticket, the imposition of fiscal penalties and the loss of other advantages. The ECHR routinely treats as ‘criminal’ matters which are not so labelled as a matter of domestic law. We can easily agree that the citizen in dispute with the public authority has more procedural rights if the matter is ‘criminal’ than if it is not. The Engel81 criteria mean that if a matter is labelled ‘criminal’ in domestic parlance, that will guarantee the applicability of Article 6 ECHR; but the domestic categorisation of the infringement is no more than a starting point, implying that the authority cannot be sure of avoiding the inconvenience of Article 6 by calling the offence ‘administrative’. Putting it differently,

Microsoft, above n 12. Case No 5100/71, Engel v Netherlands, 8 June 1976.

80 81

578 Ian S Forrester almost no deference or credit is given to how the state defines the offence. After Engel, the scope of what was eligible for the protections of Article 6 ECHR was expanded steadily to cover a range of controversies with the state. It is, I believe, beyond argument that competition cases would be deemed ‘criminal’. But that is not at all the end of the debate. The difficult question is whether the availability of judicial review of fines in Luxembourg remedies the lack in Brussels of a public hearing by an impartial tribunal, where witnesses can confront the accused. In appropriate cases, the absence of a tribunal at first instance can be cured by the availability of an impartial tribunal on appeal. Assuming (which is not a trivial assumption) that competition cases fall into this category, the question arises of whether the constrained nature of the appeal presents a problem for purposes of the ECHR. Does a review of legality and not of correctness fall short of what the ECHR demands? Opinion is not unanimous.

10. The problem of penalties under the ECHR The rule of the ECHR should be that fines and other penalties are subject to appropriately intense appellate review. The application of the case law of the European Court of Human Rights (ECtHR) to ‘administrative’ penalties is difficult and confusing. In Jussila,82 the Court considered the imposition of a €300 fiscal surcharge by an administrative authority. It made a distinction between what the literature calls hard-core criminal offences and others: it is self-evident that there are criminal cases which do not carry any significant degree of stigma. There are clearly ‘criminal charges’ of differing weight. Tax surcharges differ from the hard core of criminal law; consequently, the criminal-head guarantees will not necessarily apply with their full stringency …83

Jussila was a decision of a Grand Chamber and was therefore highly authoritative, but dissents by Judges Zupančič and Spielmann (now President) made it less so. The Court also said ‘there must be at first instance a tribunal which fully meets the requirements of Article 6’. I note in the bygoing that I am assuming it is no longer plausibly contestable that the Union and the Commission are bound by the Convention in general, and its Article 6 in particular, in respect of the enforcement of competition law. Following Jussila, it seemed to me that the Court had declared unacceptable regimes under which an administrative agency imposed penalties in matters which did involve a ‘significant degree of stigma’.84

Case No 73053/01, Jussila v Finland, 23 November 2006. Ibid, para 43. 84 See Forrester, ‘A Challenge for Europe’s Judges’, above n 31. 82 83

Quis custodiet ipsos custodes? 579 Now, the public stigma of being condemned for competition infringement is much greater than losing an argument with the tax authorities. Citizens who have to pay tax penalties, or civil penalties for putting out rubbish on the wrong day or for customs irregularities, are not exposed to public disgrace. In tax cases, the facts in dispute are usually narrowly limited and the dispute turns on how to characterise them. The frequency and circumstances under which tax investigations are conducted differ substantially from the way the Commission conducts its competition investigations. Nor does the tax authority encourage private lawsuits against the taxpayer who paid too little. Thus the condemnation of a company found guilty of a competition law offence is a major disaster for the enterprise. Through their conduct employees are at risk of criminal charges, fines may be huge and years of civil claims in several countries may ensue. However, the ECtHR seems to have partially reversed Jussila by its judgment in Menarini.85 Perhaps ‘reversed’ is too strong. Maybe ‘departed from’ or ‘elected not to follow’ would be more accurate. In Menarini, the Court found that the imposition of a fine by the administrative agency in a competition case was acceptable in that the appellate courts had jurisdiction over questions of fact and law on appeal, could review the evidence and could review how the authority had exercised its discretion in imposing the fine. If Menarini (not a Grand Chamber decision, and weakened by the dissent of Judge Pinto de Albuquerque) is reliable, then punishment by an administrative agency may be made acceptable by the intensity of the available judicial review.86 This would mean that the Court must carefully enquire into the factual circumstances on which guilt is said to repose. Numerous questions are presented by the Commission’s ferocious fines in competition cases. The principle of proportionality has rarely been used to reduce fines by the Luxembourg Courts, even though they had unlimited jurisdiction. It seems curious that gigantic fines should be imposed for breaches of a law which is so prone to new interpretations. To the extent that arbitrariness is a risk, the institutional set-up in Brussels is not reassuring: the same officials study the complaint, decide whether to investigate, decide whether to accuse, decide if the accusation is well founded and decide on the penalty. By the end of the case, when the hearing occurs and levels of fine are discussed, the case team members must look like prosecutors even if they see

85 Menarini Diagnostics SRL v Italy, No 43509/08, 27 September 2011. For discussion, see Marco Bronkers and Anne Vallery, ‘Fair and Effective Competition Policy in the EU: Which Role for Authorities and Which Role for the Courts after Menarini?’, 8 European Competition Journal 283 (2012). 86 See further Renato Nazzini, ‘Administrative Enforcement, Judicial Review and Fundamental Rights in EU Competition Law: A Comparative Contextual-Functionalist Perspective’, 49 Common Market Law Review 971 (2012).

580 Ian S Forrester themselves as neutral investigators; if the presumption of innocence is meant to be present, the light it shines is quite feeble. I note the very keen approach taken by the French Conseil Constitutionnel in examining the law according to which the French telecoms regulator could impose sanctions on undertakings for breaches of telecom regulations.87 The Conseil ruled that the lack of separation of powers in investigating (‘les fonctions de poursuite et d’instruction’) and punishing (‘les fonctions de jugement’) exercised by the regulator would breach the principle of impartiality guaranteed by Article 16 of the Déclaration des Droits de l’Homme et du Citoyen de 1789. Thus, it is by no means obvious that as a matter of due process the mechanisms by which fines are imposed are satisfactory. Indeed, I respectfully submit that they are plainly unsatisfactory. It is also interesting to note that in the case of Ziegler,88 the Court of Justice had to consider a claim that it was improper for the Commission both to impose a penalty on a cartel and at the same time to claim damages for having had to pay too much to members of the cartel. Paragraphs 159–61 are particularly confident that there is no problem whatever: the court has already held that Commission decisions may be subject to review by the European Union judicature and that European Union law lays down a system enabling the courts to review Commission decisions, including decisions relating to procedures under Article [101] EC, which provides the guarantees required by Article 47 of the Charter … The Commission cannot, therefore, in any event be regarded as both the victim of an infringement and the judge responsible for imposing penalties for the infringement. In light of the foregoing, the General Court was justified in taking the view that the Commission had not failed in its duty of impartiality. It did not, therefore, err in law in rejecting Ziegler’s plea alleging infringement of the right to fair legal process and the general principle of good administration. Moreover … it is for the General Court alone to assess the value which should be attached to the evidence produced to it, save where the clear sense of the evidence has been distorted. 87 Conseil Constitutionnel—Décision No 2013-331 QPC du 05 juillet 2013, Société Numéricâble SAS et autre, para 12:

Considérant que, selon le premier alinéa de l’article L. 132 du code des postes et des communications électroniques, les services de l’Autorité de régulation des communications électroniques et des postes sont placés sous l’autorité du président de l’Autorité; que, selon l’article D. 292 du même code, le directeur général est nommé par le président de l’Autorité, est placé sous son autorité et assiste aux délibérations de l’Autorité; que, par suite et alors même que la décision de mise en demeure relève du directeur général, les dispositions des douze premiers alinéas de l’article L. 36-11 du code des postes et des communications électroniques, qui n’assurent pas la séparation au sein de l’Autorité e ntre, d’une part, les fonctions de poursuite et d’instruction des éventuels manquements et, d’autre part, les fonctions de jugement des mêmes manquements, méconnaissent le principe d’impartialité; que celles de ces dispositions qui sont de nature législative doivent être déclarées contraires à la Constitution. Case C-439/11 P Ziegler SA v Commission, EU:C:2013:513.

88

Quis custodiet ipsos custodes? 581 Arguably, the very last words, totally confident and allowing of no doubt or hesitation, seem to confirm the limited nature of the factual reappraisal which is often conducted by the EU Courts. This seems to take us back to manifest error again, and might seem to indicate that they endorse a limited factual check. The question is whether they have found a formula to satisfy the Charter and the ECHR by the judgment in Chalkor: I have doubts, though I concede that Menarini is read by some as endorsing the validity of how competition cases are decided by the EU.89

11. The approach of the European Free Trade Association (EFTA) Court to the problem One way of approaching the matter is to conclude that there is no single test by which the requirements of Article 6 can be assessed. On this reasoning, the totality of what happens at first instance may be relevant to the adequacy of the treatment accorded to the citizen. This involves discarding Jussila’s convincing distinction between categories of controversy, with a special status for the ‘hard core’ of offences. Something like this line was taken, not without some head-scratching, in Norge Post,90 where the EFTA Court stated: The criminal head guarantees of Article 6 are applied in a differentiated manner, depending on the nature of the issue and the degree of stigma carried by certain criminal cases on the one hand and, on the necessity of the guarantee in question for the requirements of a fair trial on the other. Thus, to what degree these guarantees apply in a given case must be determined with regard to the weight of the criminal charge at issue … the present case cannot be considered to concern a criminal charge of minor weight. The amount of the charge in this case is substantial and, moreover, the stigma attached to being held accountable for an abuse of a dominant position is not n egligible …91 keeping in mind the guarantees provided by Article 6(2) ECHR, it follows from the principle of the presumption of innocence that the undertaking … must be given the benefit of the doubt …92

This confirmed the notion that the presumption of innocence is relevant; it is a welcome concession, and not one that fits well in a regime where the

89 See also Segame SA v France, Application No 4837/06, 7 June 2012; Silvester’s Horeca’s Service v Belgium, Application No 47650/99, 4 March 2004. 90 Case E-15 Posten Norge AS v EFTA Surveillance Authority, judgment of 15 April 2012. 91 Ibid, paras 89–90. 92 Ibid, paras 93–94.

582 Ian S Forrester authority investigates, decides and punishes.93 However, the next problem was the extent of judicial review when the Treaty requires a review of legality yet common sense requires that a criminal charge be substantiated convincingly: as far as past events involving complex economic features are concerned, a situation may arise in which the Court, while still considering ESA’s reasoning to be capable of substantiating the conclusions drawn from the economic evidence, may come to a different assessment of a complex economic situation. However, the fact that the Court is restricted to a review of legality precludes it from annulling the contested decision if there can be no legal objection to the assessment of ESA, even if it is not the one which the Court would consider to be preferable … This does not, however, mean that the Court must refrain from reviewing ESA’s interpretation of information of an economic nature. Not only must the Court establish, among other things, whether the evidence relied on is factually accurate, reliable and consistent, but also whether that evidence contains all the information which must be taken into account in order to assess a complex situation and whether it is capable of substantiating the conclusions drawn from it …94

This language presents well the dilemma for a court when it thinks the agency came to a wrong conclusion but did not go wildly off the rails. The EFTA Court concludes that it should not interfere in such a case. As to fines, it is properly confident and on solid ground: when imposing fines for infringement of the competition rules, ESA cannot be regarded to have any margin of discretion in the assessment of complex economic matters which goes beyond the leeway that necessarily flows from the limitations inherent in the system of legality review.95

This formulation represents an attempt to square the circle of performing a legality review which also involves more than a quick look at the facts.96 It is a worthy effort, but it does not remove valid concerns. I therefore submit that the case law of the ECtHR in Strasbourg is not dispositive as yet, and that the case law of the European Courts in Luxembourg is inadequate to satisfy the requirements of Article 6 of the Convention in all cases; in consequence, the EU regime for taking competition decisions and judicially reviewing them is imperfect. The Luxembourg Courts, following Chalkor and Menarini, have tried to crack the problem of rendering acceptable a competition regime which imposes criminal sanctions without a hearing by an independent and impartial tribunal (and without according much of a presumption of innocence) by stating that the General Court will perform

93 See the discussions in Bronkers and Vallery, ‘Which Role For the Courts After Menarini?’, above n 85. 94 Posten Norge, above n 90, para 98. 95 Ibid, para 100. 96 Forrester, ‘A Bush in Need of Pruning’, above n 3.

Quis custodiet ipsos custodes? 583 careful examination of whether the Commission’s factual assessment was vitiated by manifest error. This is no more than a step in the right direction. I thus respectfully conclude that Menarini and Chalkor have not resolved the controversy over the adequacy of the regime by which the European competition rules are enforced.97 I cannot refrain from voicing regret that, although it has been known for years that the Union would be adhering to the Convention, we are still a long way from completing the process. Such delay conveys, even if there are institutional excuses, a sense of low priority which is dismaying.

12. Purely factual assessments can determine guilt or innocence; the problem of brevity In this section I will consider two practical difficulties which undermine the supposed effectiveness of the available judicial review. First, a conscientious authority will very frequently have some basis for believing that a given factual circumstance is present. Was the medicine prescribed because it was one of a class of competing medicines, or did it enjoy a unique status? Would a computer operating system with an upgraded directory program which operated in certain respects identically to another manufacturer’s directory program be a new product? These determinations are crucial to the existence of dominance or abuse. Was Mr Dupont absent from the cartel meeting on January 13 (as he contended) or was he present (as contended by Mr Smith, the employee of the leniency-seeking competitor)? There will often be more than one way of viewing a set of market share figures. Do they show that one product was largely immune to competitive pressure, and was therefore in a dominant position, or was it one of several competing products? There will usually be some basis for each point of view advanced by the authority. It will commonly be the case that officials have a profound certainty that their analysis is correct. It is never easy to show that an authority committed a denaturing of the facts or a manifest error of assessment. So a key legal question may often turn on a question of factual assessment. If that assessment is not satisfactorily tested (satisfaction being in the eye of the alleged infringer), then judicial review that assumes that the assessment is likely to be right is fragile. To this I add a common lament from coffee shops where lawyers congregate: how is it possible to appeal a huge, tentacular decision of 500 pages or 700 paragraphs in a 50-page succinct application to the General Court

97 See Christopher Bellamy, ‘ECHR and Competition Law Post Menarini: An Overview of EU and National Case Law’, e-Competitions No 47946, 5 July 2012.

584 Ian S Forrester in Luxembourg identifying the legal errors? Good decisions are necessarily voluminous. They cannot easily be summarised. The advocate does not know whether to pare the challenge to three big problems in 50 pages and risk failing to give sufficient detail or to write a longer pleading which helpfully describes the crucial legal pleas but risks being rejected as too prolix. Oral advocacy will not fill the gaps. Shortness in the initial oral presentations (20–30 minutes) is inescapable (the free-flowing question tradition of the General Court is quite generous, however). Thus, the Courts do not want lengthy submissions, but lengthy submissions are the only way of explaining the voluminous context of three or four legal problems. The two problems mentioned in this section are real, small manifestations of a bigger difficulty: adequacy of review. One standard asks if the Commission acted illegally by making a manifest error of appreciation of the facts. The other asks if the Commission acted correctly in assessing the facts. I submit that the former standard is no longer adequate to satisfy the standards established by the Convention. There is a big difference between saying that the authority correctly determined the controversial point and saying that the authority did not commit a manifest error of appreciation in its determination of the point. The judges’ duty is not to consider if the agency acted correctly, but whether it acted lawfully. Since an agency will act lawlessly in only very rare cases, the factual obstacle for the appellants to overcome is very high. The Commission is not a reckless or imprudent entity, and will not often commit truly gross errors of factual assessment; and its decisions are—very properly—written to resist judicial interference. Putting it differently, if the Chalkor standard is carefully applied, few appeals to the EU Courts are likely to succeed on factual grounds alone. These questions are forcefully reviewed in a magisterial opinion by Advocate General Wathelet in the case of Telefónica SA,98 in which there was a question of whether the level of the fine imposed on Telefónica was consistent with fines imposed on others. The Advocate General reviews the not-easyto-reconcile past decisions of the Commission and judgments of the Courts. Some cases suggest that the Commission was not obliged to be strictly consistent and could impose heavier fines if it felt enforcement so required. Other cases stated that while the Commission could usefully indicate how the fine was calculated, it need not feel constrained to follow strict arithmetical formulae. The Advocate General wonders whether the Commission could be

98 Opinion of Advocate General Wathelet in Case C-295/12 P Telefónica SA, Telefónica de España SAU v Commission, EU:C:2013:619. Following the completion of this chapter, the ECJ declined to follow the suggestions of the Advocate General. See Case C-295/12 P, EU:C:2014:2062.

Quis custodiet ipsos custodes? 585 silent both on how the fine is calculated and as to why the fine seemed higher than previous fines in comparable cases. He then considers how these cases and the decision-making practices of the Commission fit within the full jurisdiction of the Courts when reviewing penalties. Embarrassingly, the Advocate General observes that despite the judgment of the ECJ in Chalkor (regarding the necessity of full judicial review, not a mere check on legality), and despite the Opinions in the ECtHR case of Menarini, there had been cases in which the General Court had persisted in checking whether the Commission had had the power to impose the fine, not whether it had reached a correct conclusion. He concludes that the duty of the General Court is to exercise thoroughly and independently its power to review fines; and then in two trenchant pages concludes that the General Court failed to do its duty. Whether or not the Court of Justice finally agrees with the Advocate General, the Opinion suggests that there are grounds for concern as to the adequacy, consistency and persuasiveness of judicial review in competition matters. These concerns are amplified by the high levels of fines, the absence of a hearing by a person empowered to decide contested facts and the inconsistent functions attributed to those in charge of the administrative enquiry.

13. Competence and legitimacy I end with a few thoughts about the role and duty of the institutions in competition matters. There is a political debate about the legitimacy of the actions of the Union in a number of important areas. The triumphs of helping to end armed conflict and of achieving the reunification of East and West were massive, but they lie in the past. In matters of the European currency, early success has given way to apparently interminable crisis which looks likely to be ended only by a degree of political, fiscal and economic integration beyond the current appetite of many citizens. Separately, it is not certain that the Union’s 30-odd specialist regulatory agencies enjoy legitimacy because they perform well or because their output shows their skills, nor is their democratic mandate to regulate evident.99 By contrast, the drafters of the Treaty of Rome attributed to the Commission, as one of its core competences, the application of the competition rules, including the power to grant exemptions. In 2004,

99 The erratic and inconsistent performance of the Commission’s efforts to achieve good governance and to learn from successes or failures are elegantly documented by Sandra Keegan, ‘Legitimacy in the EU Single Market: The Role of Normative R egulatory Governance’, University of Edinburgh (2013), https://www.era.lib.ed.ac.uk/bitstream/ handle/1842/7808/Keegan2013.pdf ?sequence=2&isAllowed=y.

586 Ian S Forrester this power was shared in part with the Member States.100 We should reflect on whether the Commission deserves to be respected because of its constitutional role or because of the excellence of its output.101 One can argue over the rightness of individual competition decisions, but there is no doubt that the Commission is the prime advocate for, and enforcer of, competition law as set forth in the Treaty. That said, a necessary element of the legitimacy of the constitutional competence of the Commission depends upon the Courts in Luxembourg. It is correct that powers were granted by the Treaty to apply the competition rules, a very authoritative conferring of legitimacy. But legitimacy is a continuing obligation. The availability of rigorous, consistent and effective judicial review is not something ‘desirable’ and ‘worthy’, but an indispensable element in the ongoing acceptability of the whole system. We can observe that fines have risen from tens of thousands to hundreds of millions over 40 years, an evolution which has been largely endorsed without much judicial interference. The proportionality margin of discretion of the Commission has been checked only as to details, never as to the fundamental question of whether huge fines go beyond the level necessary to achieve compliance and deterrence, and on many occasions the judicial check was a quick look at the existence of powers, not how well they were exercised. That was a serious failing. I respectfully submit that a standard of review of ‘manifest error’ and ‘legality’ would undermine the legitimacy of the enforcement of the competition rules by the Commission. In light of the evolution of Union law and in light of the public law evolutions applicable to the Union after the Lisbon Treaty, it is fair to ask that the Luxembourg Courts’ standards of review satisfy a ‘correctness’ or ‘merits’ standard, and that the processes of reaching a decision in Brussels be adapted to remedy some of the weaknesses.

100 See Ian Forrester, ‘Modernisation: An Extension of the Powers of the Commission?’, in Damien Geradin, ed, Modernisation and Enlargement: Two Major Challenges for EC Competition Law, Intersentia, 2004, chapter 3. 101 Another useful form of quality control is for the authority to examine the beneficial, wasteful, unexpected or wholesome consequences of its actions ex post facto. What can we learn about the desirability of ‘technological tying’ (the inclusion of new features as standard in a technical product like a software program, as in Microsoft), or the practice whereby fielding a minimum number of home-grown players could be demanded of teams in the first division of professional sport leagues (as in Bosman)? In cases where mergers were permitted or where the merger was prohibited, have subsequent experiences revealed the wisdom or unwisdom of the official action taken? These questions deserve to be asked by a well-run and well-supervised public authority, yet at the moment they seem not to be examined consistently, rigorously and officially. The world’s pre-eminent enforcer of competition law is under a duty to be no less rigorous in looking at its past conduct than in examining the past conduct of enterprises under investigation.

Mario Siragusa* Interaction between Public and Private Enforcement of Competition Law

A system of effective enforcement of competition law needs harmonious development of both public and private enforcement. However, while public enforcement has enjoyed healthy development in Europe, private enforcement in general has been lacking. The adoption by the European Parliament and the Council of the Directive on damages actions1 under national law represents a significant effort to foster the development of private enforcement. Many of the key issues regarding effective enforcement have already been discussed in the other chapters of this book. I will focus on what I consider to be the most relevant issues, in particular for the effective interaction between public and private enforcement.

1. Full judicial review of decisions of the European Commission and of national competition authorities Full judicial review is a necessary precondition for the effective and legitimate enforcement of competition law based on the harmonious development of public and private enforcement. Pursuant to Article 16 of Regulation 1/2003, Commission decisions are binding on national courts. Courts therefore cannot take decisions running counter to the decision adopted by the Commission. Article 9 of the Directive assigns the same probative effect to a ‘final decision of a national competition authority or a review court’, while a final decision given in another Member State may be presented as at least prima facie evidence that an infringement of competition law has occurred. As long as European Commission decisions and national competition authorities decisions are not subject to full judicial review, their probative effect will be highly questionable under the European Convention of Human Rights and the constitutional laws of some of the EU Member States. Partner, Cleary Gottlieb Steen Hamilton LLP, Rome. Directive 2014/104/EU of the European Parliament and of the Council of 26 N ovember 2014 on certain rules governing actions for damages under national law for infringements of the competition law provisions of the Member States and of the European Union, [2014] OJ L349/1. * 1

588 Mario Siragusa

2. Disclosure of evidence The absence of disclosure rules is the key obstacle to the development of effective private enforcement of competition law. The Directive is a welcome albeit timid effort to introduce a minimum standard of disclosure in our Member States. It gives a central role to the judge, as disclosure can only be ordered by the judge and is subject to strict judicial control with regard to its necessity, scope and proportionality. In particular, according to Article 5(2), national courts are able to order the disclosure of specified pieces of evidence or relevant categories of evidence circumscribed as precisely and as narrowly as possible on the basis of reasonably available facts in the reasoned justification presented by the claimant. The need to specify the pieces or the categories of evidence as precisely and as narrowly as possible on the basis of reasonably available facts could be a serious obstacle in stand-alone cases. Similar requirements already exist in some of the national laws of Member States, and in general they have been interpreted quite strictly, making it very difficult for the requesting party to obtain access to evidence. Article 6(6) excludes leniency corporate statements and settlement submissions from disclosure. Furthermore, information that has been drawn up and sent to the parties by a competition authority in the course of its proceedings (such as a Statement of Objections), or by a party to these proceedings (such as replies to a request for information), as well as settlement submissions that have been withdrawn, are disclosable only after the competition authority has closed its proceedings by adopting a decision or otherwise. Therefore, this evidence should also be disclosable in cases in which the administrative proceedings have been closed without a finding of infringement and in cases of commitment decisions (which, in the case of the Commission, are adopted under Article 9 of Regulation 1/2003). Evidence that is obtained by a party through access to the file of a competition authority can be used in an action for damages by that party or by his successor, including any party that has acquired his claim (see Article 7(3) of the Directive).

3. Interim measures Interim measures are a very important tool for effective enforcement. Experience shows that this tool is better left to judges than to administrative authorities. Article 8 of Regulation 1/2003 empowers the Commission to adopt

Interaction Between Public and Private Enforcement of Competition Law 589 interim measures, although the Court of Justice had previously stated that the Commission was already implicitly entrusted with these powers. It is worth noting that Regulation 1/2003 only provides for interim measures adopted by the Commission ‘on its own initiative’, thus excluding decisions adopted on the basis of complaints. This very important limitation distinguishes the powers of the Commission from those of national competition authorities, which, pursuant to Article 5 of Regulation 1/2003, are authorised to order interim measures ‘on their own initiative or on a complaint’. Nevertheless, the adoption of interim measures by the Commission and/ or national competition authorities is subject to the same requirements: there must be a prima facie finding of infringement and some urgency due to the risk that competition will be seriously and irreparably damaged absent the requested measures. These standard elements are required both at the European level and by our own national laws. Moreover, the Commission and national competition authorities are also bound by the principle of proportionality, which has been strictly applied by the Court of Justice. The Commission cannot issue interim measures inaudita altera parte because Article 27 of Regulation 1/2003 provides that, before issuing interim measures, ‘the Commission shall give the undertakings or associations of undertakings which are the subject of the proceedings conducted by the Commission the opportunity of being heard’. Pursuant to Article 8(2), the measures have to be limited in time, although they can be renewed insofar as necessary and appropriate. The Commission may impose a very substantial fine of 10 per cent of the turnover against companies that violate a decision ordering interim measures (Article 23(2)(b)). At the Member States’ level, Article 5 of the Regulation gives national competition authorities the power to issue interim measures on their own initiative or on the basis of a complaint. For example, in Italy the Authority in 2006 was explicitly empowered to issue interim measures and to impose fines of up to 3 per cent of the turnover in cases of violation. The Italian Competition Authority may also issue interim measures inaudita altera parte. As we can see from its decision-making practice, the Commission has rarely adopted interim measures—de facto, it stopped issuing interim measures some years ago. Indeed, I found eight interim decisions made at the Commission level after the Camera Care case in 1980: three pursuant to Article 101 TFEU and five pursuant to Article 102 TFEU. I could find no cases of interim relief granted by the Commission since the famous IMS Health case of 2001. This can probably be explained by the fact that the IMS Health case was decided prior to the adoption of Regulation 1/2003. Therefore, one could argue that the European legislator has to a certain extent limited the powers of the Commission to issue interim measures by only allowing the Commission to act on its own initiative and by forcing it to consider only irreparable damages to competition, not to individual competitors.

590 Mario Siragusa The situation is different at the level of the Member States, as demonstrated in Italy and in France. In Italy, after a period of time during which the I talian Authority even issued inaudita altera parte interim orders, the Authority has practically ceased to issue any interim orders at all. However, there are several cases in which claimants sought injunctive relief before civil judges, and in some cases such relief was granted. Because this part of the Italian judicial system is quick and efficient, an interim order may be granted within a few weeks. In France, on the other hand, President Lasserre of the French Competition Authority was an enthusiastic supporter of interim measures, and thus the French Authority has used this tool often. In general, it seems judges are better equipped than administrative authorities to issue interim measures. Because administrative authorities are already prosecutors and judges on the merits, the power to issue interim measures should probably be concentrated in the judiciary. Ideally, administrative authorities would also be required to seek the issuance of an injunctive relief before a judge.

4. Final decisions and remedial powers Regarding final decisions, Article 7 of Regulation 1/2003 gives substantial powers to the Commission. Pursuant to this Article, not only can the Commission order an end to infringement, it can also impose behavioural or structural remedies, although they must be proportionate to the infringement committed by the investigated party and necessary to effectively bring the infringement to an end. At the EU level, there is a preference for behavioural remedies (as opposed to structural remedies), although structural remedies may be imposed—as has happened in at least two cases—where there is no equally effective behavioural remedy or where the behavioural remedy is more burdensome for the company than the structural remedy. The Commission has used its power to order specific remedies in a number of final decisions. However, this power cannot be used in proceedings closed pursuant to Article 9 of Regulation 1/2003, since in those cases it is up to the investigated party to propose commitments. Article 5 of the Regulation provides that national competition authorities are also empowered to order an investigated party to bring the infringements of Article 101 and/or Article 102 TFEU to an end, but the Article does not expressly mention the power to impose behavioural or structural remedies. However, since Article 7 of Regulation 1/2003 has granted this power to the Commission for the purpose of bringing infringements to an end, I think it is reasonable to believe that, when acting with the same purpose, national competition authorities are also implicitly granted the power to impose behavioural

Interaction Between Public and Private Enforcement of Competition Law 591 or structural remedies. Indeed, Recital 12 of the Regulation expressly states that it is intended to ‘make explicit provision for the Commission’s power to impose any remedy, whether behavioural or structural’, thereby clarifying, in my view, that this power is considered to be implicit in the power to order that infringements be brought to an end. It is probably fair to conclude that, in this respect, national competition authorities have the same powers as the Commission. This conclusion also makes sense in light of the fact that the EU legal framework is based on the parallel application of competition law by the Commission and the national competition authorities. It would therefore be very odd if national competition authorities did not have the same powers (at least in this respect) as the Commission. Effective enforcement would require that, in addition to declarations of nullity and damages, national judges must also be empowered to issue both negative and positive orders in their final judgments. In some Member States, including Italy, negative orders have generally been granted, but there is great reluctance to grant positive orders—typically on the ground that a judge ordering companies to enter into new contracts would be an intolerable violation of an enterprise’s constitutional freedom.

5. Limitation periods and joint and several liability The following are the most relevant provisions of the Directive on limitation periods and joint and several liability. Pursuant to Article 10(4), when a competition authority takes action for the purpose of the investigating or initiating proceedings with respect to an infringement, the limitation period is suspended (or, depending on national law, interrupted) until at least one year after the infringement decision has become final or the proceedings are otherwise terminated. The rule makes sense to the extent that authorities officially terminate the proceedings. However, as we all well know, the status of a proceeding may often remain uncertain for a very long period of time. In cases of such uncertainty, an interruption of the proceeding, rather than a suspension, may be more appropriate. Pursuant to Articles 11(4) and 11(5), an immunity recipient’s liability, and the contribution owed to co-infringers under joint and several liability, are limited to the harm the recipient has caused to its own direct or indirect purchasers or (in the case of a buying cartel) to its own direct or indirect providers. The immunity recipient remains fully liable as a last-resort debtor if the injured parties are unable to obtain full compensation from the other infringers.

592 Mario Siragusa

6. Quantification of harm Article 17(2) of the Directive introduces a presumption that cartel infringements cause harm. The infringer shall have the right to rebut this presumption. Recital 14 of the Directive also includes the definition of a ‘cartel’ as the: agreement or concerted practice between two or more competitors aimed at coordinating their competitive behaviour on the market or influencing the relevant parameters of competition, through practices such as, but not limited to, the fixing or coordination of purchase or selling prices or other trading conditions, including in relation to intellectual property rights, the allocation of production or sales quotas, the sharing of markets and customers, including bid-rigging, restrictions of imports or exports or anti-competitive actions against other competitors.

Philip Lowe Conclusions

During the two days of discussion, the Workshop examined some of the main determinants of legitimate and effective competition law enforcement by both competition authorities and the Courts. In the first panel session, it was quickly recognised that, beyond the adequacy of its legal powers, a competition authority’s public legitimacy depends on its capacity to intervene in the market effectively in order to protect competition and promote consumer welfare. However, the concern to demonstrate this effectiveness to the public and at a political level can lead authorities into such actions as high-profile investigations and high fines which do not respect the principles of proportionality and necessity, and flout due process. This, in time, can undermine the legitimacy of the authority, leading to a curtailment of its activities and its powers, and ultimately its effectiveness. A number of participants questioned the proportionality and legitimacy of high fines in relation to the effects of anticompetitive conduct and agreements, and their effectiveness in sanctioning and deterring them. As a result of complex immunity and leniency arrangements, the final fines imposed on individual cartel participants bore little relation to the extent of individual responsibility and infringement which needed to be sanctioned. The legitimacy of any fines for abuse of market power was also questioned, except for repeat offences, on the grounds that recourse to industry-wide practices, such as loyalty rebates, could not be regarded per se as sanctionable. Private enforcement was also available as a deterrent. There were additional calls for stricter and deeper judicial review of fines, especially by the European Courts, and the need to take account of measures taken by defendants to avoid infringements, such as compliance programmes. Overall, these interventions emphasised the danger of over-intensive and disproportionate enforcement, which weakened the legitimacy of competition authorities. At the same time, it was acknowledged that, given low detection rates, action against cartels and the due process associated with it was already a heavy and long-drawn-out process. The requirement to prove effects for patently anticompetitive behaviour could lead to a lack of effective results. Effects-based calculation of fines may also not necessarily lead to lower fines. On procedural effectiveness, there was general agreement that the lack of harmonisation of procedural rules across the national jurisdictions of the EU was a major obstacle to effective enforcement of Regulation 1/2003 by the Commission and national authorities. This was not only a serious impedi-

594 Philip Lowe ment to the effectiveness of the competition authorities, but also a major burden to plaintiffs and defending parties, with excessive complexity in particular as far as leniency applications were concerned. This call for further procedural harmonisation (in a Regulation 2?) was also echoed in a later discussion of parallel enforcement by national authorities and its lack of coherence with the principle of ne bis in idem. The lack of clarity on antitrust liability in the context of legal succession was also highlighted as preventing effective law enforcement in Germany. This could also be addressed through proposals for further harmonisation of national processes across the EU. Fines were seen as only one form of remedy to a competition infringement. Beyond requirements to bring an infringement to an end, behavioural and structural remedies could be enforced. Within the EU, this can be done either through an Article 7 decision (accompanied normally by fines) or through an Article 9 commitment decision without fines. The Workshop noted the increased use of commitments decisions by the European Commission, an option which was regularly preferred by many parties for reasons of procedural expediency and absence of a formal recognition of an infringement. However, several participants warned of the dangers of parties being put under pressure to offer commitments which went beyond the legal requirements and infringed the principles of proportionality and necessity. At the same time, it was acknowledged that commitment decisions which resulted from remedies proposed by the parties frequently resulted in significant sector-wide benefits to consumers. Genuinely constructive discussions between the parties and authorities on remedies could enhance effectiveness of law enforcement on the understanding that absence of agreement implied a return to an Article 7 decision, with due regard to the rights of defence. In response to concerns that antitrust enforcement could develop an illegitimate role in regulating markets, some emphasis was placed on the need for competition authorities to take a holistic view on the most effective remedy to a competition problem (an antitrust case, the action of a sectoral regulator or new, repealed or amended legislation). An effective remedy must resolve the competition problem in a time period relevant to the economic harm which it has caused and is likely to cause in the future. In the second session, the Workshop examined the potential benefits of amicus curiae briefs. In the US, amicus briefs were a widely used tool in both public and private enforcement. In the EU, the Commission had developed an efficient amicus practice under Regulation 1/2003 to contribute to the assessment of cases initiated by national authorities and reviewed by national courts. However, given its limited resources, the Commission inevitably has to establish priorities in responding to amicus requests. If the proposed private enforcement regime is to be effective beyond follow-on actions, the capacity of the Commission to respond to amicus requests in privately initiated cases will become critical.

Conclusions 595 The Workshop then addressed the issue of effective control of state intervention, in particular with reference to the EU and US jurisdictions. Recourse to application of Article 106 TFEU in the EU had recently been limited, although, in a prolonged period of economic recession, there was a strong and generally unjustified tendency for state intervention to restrict competition. In the US as well as the EU, initial initiatives to liberalise some utility sectors had been followed by an understandable deference by competition authorities to the democratic process of development of legislation. The underlying assumption here was that any restriction of competition would be duly and transparently justified ex ante, and on judicial appeal in relation to non-competition public policy objectives. This could be done through a ‘hard look’ process, formal impact assessments or market investigations, as under current UK competition law. However, given the lack of standing of complainants under Article 106, the Commission’s stance was an obstacle to effective enforcement. In the third session, we returned to the bread-and-butter job of competition authorities in the challenge of parallel enforcement of completion law, both within the structured framework of the European Competition Network and between jurisdictions throughout the world. We also addressed the issue of agency priorities and case allocation. As far as parallel enforcement by the European Commission and by one or more national authorities is concerned, there was recognition that cases could legitimately be distinguished according to the effects in different jurisdictions and in relation to the time period relevant to the related infringements. However, fragmentation of cases beyond these parameters, and in relation to the same facts, infringed the ne bis in idem principle and should be avoided, with respect to both infringement decisions and ‘acquittal’ decisions. This did not preclude the possibility that one authority could, in the terms of Regulation 1/2003, be identified as ‘best placed’ to handle a case, including in its work the assessment of impacts in other jurisdictions which may agree to defer to its decisions. It was nevertheless generally accepted that full compliance with the ne bis in idem principle was predicated on EU-wide procedural harmonisation, including in particular respect of the rights of defence and the rights of complainants. In this context, the Workshop underlined again the urgent need for the EU to progress beyond the Model Leniency Notice to a streamlined and EU-wide system for leniency applications. As far as competition authorities themselves are concerned, we recognised that their effectiveness and legitimacy were heavily dependent on their ability to define their objectives and enforcement priorities clearly, and to establish measurable indicators of their performance. Some activities—in particular, review procedures such as merger control, which are subject to legal deadlines—allowed limited room for prioritisation, but other areas in antitrust required a focused approach. Case selection and internal control

596 Philip Lowe rocesses also depended on a holistic examination of what was the most p effective solution for the competition problem concerned, whether through antitrust action or a regulatory response. Authorities frequently lost effectiveness and legitimacy by concentrating on high-profile cases to the detriment of a wider range of cases which could strengthen their enforcement record and provide a wider basis of political and public support for their action. Giving some profile to particular cases can, however, help to communicate the practical benefits of enforcement action for consumers. The Workshop then turned its attention to parallel law enforcement and cooperation between national authorities around the world. There was general acceptance that waivers for exchange of information between authorities could increase the effectiveness and efficiency of investigations, in particular for mergers. For cartels, the interface between administrative and criminal regimes still constituted an obstacle to information exchange. The importance of inter-jurisdictional cooperation on remedies was stressed. Some remedies were very specific to problems in particular national markets, but often worldwide remedies made more sense for the parties and for consumer welfare in all jurisdictions. There was some evidence of effective cooperation on a case-by-case basis between the major jurisdictions, both on substantive issues and on procedural and timing issues. However, there was general concern that, despite the substantial progress made through the original OECD Recommendation and the establishment of and work within the International Competition Network (ICN), there were now more than 120 national competition authorities in the world and convergence between standards and procedures is decreasing, not increasing. Many new authorities, in particular in major emerging countries, were enforcing their competition laws rigorously and frequently with a domestic and industrial policy-related focus. At the same time, markets continue to globalise. It was urgent to use the processes of the OECD and the ICN to re-establish the path towards convergence on both substantive and procedural issues, in particular for mergers. In our fourth and final session, the Workshop reviewed the role and the influence of the Courts in competition law enforcement. The issue of the scope of judicial review of decisions taken by competition authorities was extensively discussed. Many participants suggested that the prevalent administrative system of law enforcement in most European jurisdictions could only maintain its legitimacy if it was underpinned by a more comprehensive judicial review or a full merits appeal (of legality, facts and arguments). Otherwise the terms of the European Convention on Human Rights could not be respected. It was stressed in particular that the traditional tests applied to European Commission decisions (legality and manifest error) were insufficient to guarantee the necessity and proportionality of these decisions, especially on merger authorisations and prohibitions, as well as on antitrust infringements, fines and commitments.

Conclusions 597 In response to concerns about delays in the treatment of cases in the uxembourg Courts, attention was drawn to the speed with which the L General Court had already responded rapidly to merger appeals (to keep transactions alive) and to the need for the Courts to set priorities, given their limited resources. In addition, parties often lacked focus in their appeals and lengthened procedures unnecessarily. As to the scope of judicial review, there was deference to an administrative authority in the assessment of complex economic issues. But the more complex the theory of harm advanced in the decision of a competition authority, the higher the burden of proof had to be. At the same time, a number of participants also underlined the lack of consistency of the European Courts insofar as the judgments were based on old case law and legal theory, whereas the decisions under appeal were based on a more effects-based assessment of the facts of the alleged competition law infringements. Most interventions did not support the idea of introduction of prosecutorial proceedings in Europe, although the idea was advanced that the powers of the Commission should be limited to proposing only fines on which the Court would then decide. In general, there was a call for the European Commission and the Courts to strengthen the due process associated with the present administrative system, both through more rigorous and transparent processes within the Commission and through strengthened review of decisions by the Courts.

598

SELECTED REFERENCES

A Ackermann, Thomas, ‘Kartellgeldbußen als Instrument der Wirtschaftsaufsicht’, [2012] Zeitschrift für Wettbewerbsrecht 3 Adam, Michael and Frank Maier-Rigaud, ‘The Law and Economics of Article 82 EC and the Commission Guidance Paper on Exclusionary Conduct’, [2009] Zeitschrift für Wettbewerbsrecht 131 Anderson, Lloyd, ‘Mocking the Public Interest: Congress Restores Meaningful Judicial Review of Government Antitrust Consent Decrees’, 31 Vermont Law Review 594 (2007) Andreangeli, Arianna, ‘Between Economic Freedom and Effective Competition Enforcement: The Impact of the Antitrust Remedies Provided by the Modernisation Regulation on Investigated Parties’ Freedom to Contract and to Enjoy Property’, 6(2) Competition Law Review 225 (2010) Ashton, David and David Henry, eds, Competition Damages Actions in the EU: Law and Practice, Edward Elgar, 2013 Ayres, Ian and John Braithwaite, Responsive Regulation: Transcending the Deregulation Debate, Oxford University Press, 1992

B Bailey, David, ‘Scope of judicial review under Article 81 EC’, 41 Common Market Law Review 1327 (2004) Balaguer, Victoria, ‘Beef Industry Development Society: A Warning on Industry Restructuring’, 33 European Competition Law Review 296 (2012) Baumol, William and Daniel Swanson, ‘The New Economy and Ubiquitous Competitive Price Discrimination: Identifying Defensible Criteria of Market Power’, 70 Antitrust Law Journal 661 (2003) Beaton-Wells, Caron and Ariel Ezrachi, eds, Criminalising Cartels: Critical Studies of an International Regulatory Movement, Hart Publishing, 2011 Beaton-Wells, Caron and Brent Fisse, Australian Cartel Regulation: Law, Policy and Practice in an International Context, Cambridge University Press, 2011 Beaton-Wells, Caron and Christine Parker, ‘Justifying Criminal Sanctions for Cartel Conduct: A Hard Case’, 1 Journal of Antitrust Enforcement 198 (2013) Becker, Gary, ‘Crime and Punishment: An Economic Approach’, 76 Journal of Political Economy 169 (1968) Bengoetxea, Joxerramón, The Legal Reasoning of the European Court of Justice, Clarendon Press, 1993

600 Selected References Bergman, Mats, Malcolm Coate, Maria Jakobsson and Shawn Ulrick, ‘Comparing Merger Policies in the European Union and the United States’, 36 Review of Industrial Organization 305 (2010) Bernitz, Ulf, Xavier Groussot and Felix Schulyok, eds, General Principles of EU Law and European Private Law, Kluwer Law International, 2013 Bernstein, David, Rehabilitating Lochner: Defending Individual Rights against Progressive Reform, University of Chicago Press, 2011 Bickel, Alexander, The Least Dangerous Branch, 2nd edn, Yale University Press, 1986 Black, Ryan and Ryan Owens, ‘Agenda Setting in the Supreme Court: The Collision of Policy and Jurisprudence’, 71 Journal of Politics 1062 (2009) Bourgeois, Jacques, and Denis Waelbroeck, eds, Ten Years of Effects-based Approach in EU Competition Law—State of Play and Perspective, Bruylant, 2013 Braithwaite, John, Restorative Justice and Responsive Regulation, Oxford University Press, 2002 Brannon, Leah and Douglas Ginsburg, ‘Antitrust Decisions of the US Supreme Court, 1967 to 2007’, 3(2) CPI Journal 1 (2007) Bronkers, Marco and Anne Vallery, ‘Fair and Effective Competition Policy in the EU: Which Role for Authorities and Which Role for the Courts after Menarini?’, 8 European Competition Journal 283 (2012) Brown, Adrian and Morris Schonberg, ‘Widening the Net: The General Court Extends the Principle of Successor Liability in EU Competition Law’, 34 European Competition Law Review 1 (2013) Buccirossi, Paolo, ed, Handbook of Antitrust Economics, MIT Press, 2008 Buendía Sierra, José Luis, Exclusive Rights and State Monopolies under the EC Law: Article 86 (Former Article 90) of the Treaty, Oxford University Press, 1999

C Caldeira, Gregory and John Wright, ‘Organized Interests and Agenda Setting in the US Supreme Court’, 82 American Political Science Review 1109 (1988) Calkins, Stephen, ‘The Antitrust Conversation’, 68 Antitrust Law Journal 625 (2001) Calkins, Tim, Defending Your Brand: How Smart Companies Use Defensive Strategies to Deal with Competitive Attacks, Palgrave Macmillan, 2012 Cappelletti, Mauro, ‘Repudiating Montesquieu? The Expansions and Legitimacy of “Constitutional Justice”’, 35 Catholic University Law Review 1 (1985) Castillo de la Torre, Fernando, ‘The 2006 Guidelines on Fines: Reflections on the Commission’s Practice’, 33 World Competition 359 (2010) Cengiz, Firat, ‘Multi-level Governance in Competition Policy: The European Competition Network’, 35 European Law Review 660 (2010) Cengiz, Firat, ‘Judicial Review and the Rule of Law in the EU Competition law Regime after Alrosa’, 7 European Competition Journal 127 (2011) Chalmers, Damien, Gareth Davies and Giorgio Monti, European Union Law: Cases and Materials, 2nd edn, Cambridge University Press, 2010 Collins, Paul and Lisa Solowiej, ‘Interest Group Participation, Competition, and Conflict in the US Supreme Court’, 32 Law and Social Inquiry 955 (2007)

Selected References 601 Connor, John, ‘Has the European Commission Become More Severe in Punishing Cartels? Effects of the 2006 Guidelines’, 32 European Competition Law Review 27 (2011) Connor, John and Robert Lande, ‘How High do Cartels Raise Prices: Implications for Optimal Cartel Fines’, 80 Tulane Law Review 513 (2005) Crane, Daniel, ‘Optimizing Private Antitrust Enforcement’, 62 Vanderbilt Law Review 675 (2010) Crane, Daniel, The Institutional Structure of Antitrust Enforcement, Oxford University Press, 2011 Crane, Daniel, ‘Enacted Legislative Findings and the Deference Problem’, 102 Georgia Law Journal 637 (2013) Crane, Daniel, ‘Judicial Review of Anticompetitive State Action: Two Models in Comparative Perspective’, 1 Journal of Antitrust Enforcement 418 (2013)

D Davies, Stephen and Bruce Lyons, Mergers and Merger Remedies in the EU: Assessing the Consequences for Competition, Edward Elgar, 2007 Dunne, Niamh, ‘Commitment Decisions in EU Competition Law’, 10 Journal of Competition Law and Economics 399 (2014)

E Easterbrook, Frank, ‘The Limits of Antitrust’, 63 Texas Law Review 1 (1984) Economides, Nicholas and Ioannis Lianos, ‘A Critical Appraisal of Remedies in the EU Antitrust Microsoft Cases’, 10 Columbia Business Law Review 346 (2010) Ehlermann, Claus-Dieter and Isabela Atanasiu, eds, European Competition Law Annual 2000: The Modernization of EC Antitrust Policy, Hart Publishing, 2001 Ehlermann, Claus-Dieter and Isabela Atanasiu, eds, European Competition Law Annual 2001: Effective Private Enforcement of EC Competition Law, Hart Publishing, 2003 Ehlermann, Claus-Dieter and Isabela Atanasiu, eds, European Competition Law Annual 2002: Constructing the EU Network of Competition Authorities, Hart Publishing, 2005 Ehlermann, Claus-Dieter and Isabela Atanasiu, eds, European Competition Law Annual 2006: Enforcement of Prohibition of Cartels, Hart Publishing, 2007 Ehlermann, Claus-Dieter and Mel Marquis, eds, European Competition Law Annual 2008: Antitrust Settlements under EC Competition Law, Hart Publishing, 2010 Ehlermann, Claus-Dieter and Mel Marquis, eds, European Competition Law Annual 2009: The Evaluation of Evidence and its Judicial Review in Competition Cases, Hart Publishing, 2011 Ehlermann, Claus-Dieter and Mel Marquis, eds, European Competition Law Annual 2012: Competition, Regulation and Public Policies, Hart Publishing, 2014

602 Selected References Ely, John Hart, Democracy and Distrust: A Theory of Judicial Review, Harvard University Press, 1980 Epstein, Lee, William Landes and Richard Posner, ‘How Business Fares in the Supreme Court’, 97 Minnesota Law Review 1431 (2013) Epstein, Richard and Michael Greve, Competition Laws in Conflict: Antitrust Jurisdiction in the Global Economy, AEI Press, 2004 Epstein, Richard, Antitrust Consent Decrees in Theory and Practice, AEI Press, 2007

F Fallon, Richard Jr , ‘Scholars’ Briefs and the Vocation of a Law Professor’, 4 Journal of Legal Analysis 223 (2012) Faull, Jonathan and Ali Nikpay, eds, The EC Law of Competition, 3rd edn, Oxford University Press, 2014 Forrester, Ian, ‘The Judicial Function in European Law and Pleading in the European Courts’, 81 Tulane Law Review 647 (2007) Forrester, Ian, ‘A Challenge for Europe’s Judges: The Review of Fines in Competition Cases’, 36 European Law Review 185 (2011) Fox, Eleanor, ‘The WTO’s First Antitrust Case—Mexican telecom: A Sleeping Victory for Trade and Competition’, 9 Journal of International Economic Law 271 (2006) Fox, Eleanor and Daniel Crane, eds, Antitrust Stories, Foundation Press, 2007 Fox, Eleanor and Michael Trebilcock, eds, The Design of Competition Law Institutions: Global Norms, Local Choices, Oxford University Press, 2013 Francioni, Francesco, ‘Access to Justice, Denial of Justice and International Investment Law’, 20 European Journal of International Law 729 (2009) Franklin, David, ‘Legislative Rules, Nonlegislative Rules, and the Perils of the Short Cut’, 120 Yale Law Journal 276 (2010) Freyer, Tony, Antitrust and Global Capitalism, 1930–2004, Cambridge University Press, 2006 Fuller, Lon and Kenneth Winston, ‘The Forms and Limits of Adjudication’, 92 Harvard Law Review 353 (1978) Fumagalli, Chiara and Massimo Motta, ‘Exclusive Dealing and Entry, When Buyers Compete’, 96 American Economic Review 785 (2006)

G Garland, Merrick, ‘Deregulation and Judicial Review’, 98 Harvard Law Review 505 (1985) Garland, Merrick, ‘Antitrust and State Action: Economic Efficiency and the Political Process’, 96 Yale Law Journal 486 (1987)

Selected References 603 Geradin, Damien, ed, Modernisation and Enlargement: Two Major Challenges for EC Competition Law, Intersentia, 2004 Geradin, Damien and David Henry, ‘The EC Fining Policy for Violations of Competition Law: An Empirical Review of the Commission Decisional Practice and the Community Courts’ Judgments’, 1 European Competition Journal 401 (2005) Gerard, Damien, ‘Breaking the EU Antitrust Enforcement Deadlock: Re-empowering the Courts?’, 36 European Law Review 457 (2011) Gerber, David, Law and Competition in Twentieth Century Europe, Oxford University Press, 1998 Ginsburg, Douglas, ‘Originalism and Economic Analysis: Two Case Studies of Consistency and Coherence in Supreme Court Decision Making’, 33 Harvard Journal of Law and Public Policy 217 (2010) Ginsburg, Douglas and Joshua Wright, ‘Antitrust Sanctions’, 6(2) CPI Journal 1 (2010) Gheur, Charles and Nicolas Petit, eds, Alternative Enforcement Techniques in EC Competition Law, Bruylant, 2009

H Hawk, Barry, ed, International Antitrust Law & Policy: Fordham Corporate Law Institute 2003, Juris Publishing, 2004 Hawk, Barry, ed, International Antitrust Law and Policy: Fordham Corporate Law 2004, Juris Publishing, 2005 Hawk, Barry, ed, International Law & Policy: Fordham Competition Law Institute 2007, Juris Publishing, 2008 Heinichen, Christian, Unternehmensbegriff und Haftungsnachfolge im Europäischen Kartellrecht, Nomos, 2011 Hellström, Per, Frank Maier-Rigaud and Friedrich Wenzel Bulst, ‘Remedies in European Antitrust Law’, 76 Antitrust Law Journal 43 (2009) Horsley, Thomas, ‘Reflections on the Role of the Court of Justice as the ‘Motor’ of European Integration: Legal Limits to Judicial Lawmaking’, 50 Common Market Law Review 1 (2013) Hylton, Keith, ed, Antitrust Law and Economics, Edward Elgar, 2010

I Ibañez Colomo, Pablo, ‘On the Application of Competition Law as Regulation: Elements for a Theory’, 29 Yearbook of European Law 261 (2010) Inman, Robert and Daniel Rubinfeld, ‘Making Sense of the State Action Doctrine: Balancing Political Participation and Economic Efficiency in Regulatory Federalism’, 75 Texas Law Review 1203 (1997)

604 Selected References

J Jaeger, Marc, ‘The Standard of Review in Competition Cases Involving Complex Economic Assessments: Towards the Marginalisation of the Marginal Review?’, 2 Journal of European Competition Law and Practice 295 (2011) Jones, Alison, ‘The Journey toward an Effects-based Approach under Article 101 TFEU—the Case of Hardcore Constraints’, 55 Antitrust Bulletin 783 (2010)

K Khan, Nicholas, Kerse & Khan on EU Antitrust Procedure, 6th edn, Sweet & Maxwell, 2012 Kochevar, Steven, ‘Amici Curiae in Civil Law Jurisdictions’, 122 Yale Law Journal 1653 (2013) Kokott, Julianne and Daniel Dittert, ‘Die Verantwortlichkeit von Muttergesellschaften für Kartellvergehen ihrer Tochtergesellschaften im Lichte der Rechtsprechung der Unionsgerichte’, [2012] Wirtschaft und Wettbewerb 670 Konkurrensverket, The Pros And Cons of Vertical Restraints, SCA, 2008 Kovacic, William, ‘Transatlantic Turbulence: The Boeing–McDonnell Douglas Merger and International Competition Policy’, 68 Antitrust Law Journal 805 (2001) Kovacic, William, ‘The Intellectual DNA of Modern US Competition Law for Dominant Firm Conduct: The Chicago–Harvard Double Helix’, [2007] Columbia Business Law Review 1 Kovacic, William and David A Hyman, ‘Competition Agency Design: What’s on the Menu?’, 8 European Competition Journal 527 (2012).

L Landes, William, ‘Optimal Sanctions for Antitrust Violations’, 50 University of Chicago Law Review 652 (1983) Lévêque, François and Howard Shelanski, eds, Merger Remedies in American and European Union Competition Law, Edward Elgar, 2003 Levine, Eugenia, ‘Amicus Curiae in International Investment Arbitration: The Implications of an Increase in Third-Party Participation’, 29 Berkeley Journal of International Law 200 (2011) Lewis, David, ed, Building New Competition Law Regimes: Selected Essays, Edward Elgar, 2013 Lianos, Ioannis and Damien Geradin, eds, Handbook of EU Competition Law: Enforcement and Procedure, Edward Elgar, 2013

Selected References 605 Lianos, Ioannis and D Daniel Sokol, eds, The Global Limits of Competition Law, Stanford University Press, 2012 Louis, Frédéric and Gabriele Accardo, ‘Ne Bis in Idem, Part “Bis”’, 34 World Competition 97 (2011) Lowe, Philip and Mel Marquis, eds, European Competition Law Annual 2010: Merger Control in European and International Perspective, Hart Publishing, 2013 Lynch, Kelly, ‘Best Friends?: Supreme Court Law Clerks on Effective Amicus Curiae Briefs’, 20 Journal of Law and Politics 33 (2004)

M Maduro, Miguel Poiares, We the Court: The European Court of Justice and the European Economic Constitution, Hart Publications, 1998 Marquis, Mel, ‘Abuse of Administrative Power to Restrict Competition in China: Four Reflections, Two Ideas and a Thought’, in Michael Faure and Xinzhu Zhang, eds, Regulation and Competition Policy in China: New Developments and Empirical Evidence, Edward Elgar, 2013, 73 et seq Marquis, Mel, ‘Idea Merchants and Paradigm Peddlers in Global Antitrust’, 28 Pacific McGeorge Global Business and Development Law Journal 155 (2015) Marty, François, ‘Une régulation du secteur de l’énergie au travers des procedures d’engagements? Réflexions sur le contentieux concurrentiel européen’, 26–27 économie publique 93 (2011) Mateus, Abel and Teresa Moreira, eds, Competition Law and Economics: Advances in Competition Policy and Antitrust Enforcement, Kluwer Law International, 2007 Meiklejohn, Alexander, Political Freedom: The Constitutional Powers of the People, Harper & Brothers, 1960 Melamed, A Douglas, ‘Antitrust: The New Regulation’, 10 Antitrust 13 (Fall 1995) Merola, Massimo and Jacques Derenne, eds, The Role of the Court of Justice of the EU in Competition Law Cases, Bruylant, 2012 Micklitz, Hans-Wolfgang and Bruno de Witte, eds, The ECJ and the Autonomy of Member States, Intersentia, 2012 Miles, Thomas and Cass Sunstein, ‘The Real World of Arbitrariness Review’, 75 University of Chicago Law Review 761 (2008) Monti, Giorgio, ‘Managing the Intersection of Utilities Regulation and EC Competition Law’, 4(2) Competition Law Review 123 (2008) Monti, Giorgio, ‘Article 82 EC: What Future for the Effects-Based Approach?’, 1 Journal of European Competition Law and Practice 2 (2010)

N Nazzini, Renato, The Foundations of European Union Competition Law. The Objective and Principles of Article 102, Oxford University Press, 2011

606 Selected References Nazzini, Renato, ‘Administrative Enforcement, Judicial Review and Fundamental Rights in EU Competition Law: A Comparative Contextual-Functionalist Perspective’, 49 Common Market Law Review 971 (2012)

O O’Donoghue, Robert and Jorge Padilla, The Law and Economics of Article 102 TFEU, 2nd edn, Hart Publishing, 2013

P Page, William and John Lopatka, ‘State Regulation in the Shadow of Antitrust: FTC v Ticor Title Insurance Co’, 3 Supreme Court Economic Review 189, 229 (1993) Parker, Christine, ‘The “Compliance” Trap: The Moral Message in Regulatory Enforcement’, 40 Law and Society Review 591 (2006) Parker, Christine, ‘Twenty Years of Responsive Regulation: An Appreciation and Appraisal’, 7 Regulation and Governance 2 (2013) Parker, Christine, ‘The War on Cartels and the Social Meaning of Deterrence’, 7 Regulation and Governance 174 (2013) Parker, Christine and Vibeke Lehmann Nielsen, Explaining Compliance: Business Responses to Regulation, Edward Elgar, 2011 Parker, Christine and Vibeke Lehmann Nielsen, ‘The Fels Effect: Responsive Regulation and the Impact of Business Opinions of the ACCC’, 20 Griffith Law Review 91 (2011) Pedersen, William, ‘Formal Records and Information Rulemaking’, 85 Yale Law Journal 38 (1975) Petit, Nicolas, ‘From Formalism to Effects?—The Commission’s Guidance on Article 82 EC’, 32 World Competition 485 (2009)

R Rasmusen, Eric, Mark Ramseyer and John Wiley Jr , ‘Naked Exclusion’, 81 American Economic Review 1137 (1991) Rasmussen, Hjalte, On Law and Policy in the European Court of Justice: A Comparative Study in Judicial Policymaking, Martinus Nijhoff Publishers, 1986 Riley, Alan, ‘The ECHR Implications of the Investigation Provisions of the Draft Competition Regulation’ 51 International and Comparative Law Quarterly 55 (2002) Riley, Alan, ‘EC Antitrust Modernisation: The Commission Does Very Nicely— Thank You! Part One: Regulation 1 and the Notification Burden’, 24 European Competition Law Review 604 (2003)

Selected References 607 Rustad, Michael and Thomas Koenig, ‘The Supreme Court and Junk Social Science: Selective Distortion in Amicus Briefs’, 72 North Carolina Law Review 91 (1993)

S Sandstrom Simard, Linda, ‘An Empirical Study of Amici Curiae in Federal Court: A Fine Balance of Access, Efficiency, and Adversarialism’, 27 Review of Litigation 669 (2008) Segal, Ilya and Michael Whinston, ‘Naked Exclusion: Comment’, 90 American Economic Review 296 (2000) Shapiro, Martin, ‘The Administrative Procedure Act: A Fortieth Anniversary Symposium: APA: Past, Present, Future’, 72 Virginia Law Review 447 (1986) Shelton, Dinah, ‘The Participation of Nongovernmental Organizations in International Judicial Proceedings’, 88 American Journal of International Law 611 (1994) Sofianatos, Gerasimos, Injonctions et Engagements en Droit de la Concurrence, LGDJ, 2009 Stack, Kevin, ‘The Constitutional Foundations of Chenery’, 116 Yale Law Journal 952 (2007) Stack, Kevin, ‘Interpreting Regulations’, 111 Michigan Law Review 355 (2012) Stewart, Richard, ‘The Reformation of American Administrative Law’, 88 Harvard Law Review 1667 (1975) Sunstein, Cass, ‘Constitutionalism after the New Deal’, 101 Harvard Law Review 421 (1987)

V Veljanovski, Cento, ‘Cartel Fines in Europe: Law, Practice and Deterrence’ 29 World Competition 1 (2007) Vesterdorf, Bo, ‘Standard of Proof in Merger Cases: Reflections in the Light of Recent Case Law of the Community Courts’, 1 European Competition Journal 1 (2005)

W Waelbroeck, Denis, ‘Le développement en droit européen de la concurrence des solutions négociées (engagements, clémence, non-contestation des faits et transactions): que va-t-il rester aux juges?’, Global Competition Law Center Working Paper 1/08 Wagner von Papp, Florian, ‘Best and Even Better Practices in the European Commitment Procedure after Alrosa: The Dangers of Abandoning the “Struggle for Competition Law”’, 49 Common Market Law Review 929 (2012)

608 Selected References Wang, Wei, ‘Structural Remedies in EU Antitrust and Merger Control’, 34 World Competition 571 (2011) Watts, Kathryn, ‘Proposing a Place for Politics in Arbitrary and Capricious Review’, 119 Yale Law Journal 2 (2009) Webb Yackee, Jason and Susan Webb Yackee, ‘Testing the Ossification Thesis: An Empirical Examination of Federal Regulatory Volume and Speed, 1950–1990’, 80 George Washington Law Review 1414 (2012) Weinrib, Earnest, Corrective Justice, Oxford University Press, 2012 Wesseling, Rein and Marc van der Woude, ‘The Lawfulness and Acceptability of Enforcement of European Cartel Law’, 35 World Competition 573 (2012) Whelan, Peter, ‘A Principled Argument for Personal Criminal Sanctions as Punishment under EC Cartel Law’, 4(1) Competition Law Review 7 (2007) Whish, Richard and David Bailey, Competition Law, 7th edn, Oxford University Press, 2012 Wiley, John Shepherd Jr , ‘A Capture Theory of Antitrust Federalism’, 99 Harvard Law Review 713 (1986) Wils, Wouter, Principles of European Antitrust Enforcement, Hart Publishing, 2005 Wils, Wouter, Efficiency and Justice in European Antitrust Enforcement, Hart Publishing, 2008 Wils, Wouter, ‘The Increased Level of EU Antitrust Fines, Judical Review, and the European Convention on Human Rights’, 33 World Competition 5 (2010) Wils, Wouter, ‘EU Antitrust Enforcement Powers and Procedural Rights and Guarantees: The Interplay between EU Law, National Law, the Charter of Fundamental Rights of the EU and the European Convention on Human Rights’, 34 World Competition 189 (2011) Wils, Wouter, ‘Recidivism in EU Antitrust Enforcement: A Legal and Economic Analysis’, 35 World Competition 5 (2012) Wind, Elena, ‘Remedies and Sanctions in Article 82 of the EC Treaty’, 26 European Competition Law Review 659 (2005)