Article 82 EC: Reflections on its Recent Evolution 9781472560605, 9781841132501

Citation preview

Preface The landscape of European competition law has seen significant changes in the past decade, in terms both of enforcement and substantive application. One of the last frontiers to be subjected to scrutiny and debate has been Article 82. A lively debate has developed in recent years regarding the introduction of an effects-based approach to Article 82, and the way such an approach may affect Article 82’s nature and scope. The debate was stimulated by the publication of three papers: the Commission Discussion Paper, the EAGCP Report, and the Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings. Moving from a formalistic approach to Article 82 to an economic-based one, we are crossing somewhat uncertain grounds. While emphasis on the actual or expected economic effect is welcome, as it aims to curtail only those activities which are harmful to competition, it is challenging to apply effects-based analysis in practice. There are questions relating to the relevant benchmarks to use for establishing abuse and to the principles that limit their applications. The European courts may have a significant role in this evolutionary process. Several landmark cases, including Microsoft, France Télécom (Wanadoo) and GlaxoSmithKline, provided the courts with opportunities to have their say on the boundaries of Article 82. These judgments are interesting not only for their contribution to the development of the case law, but also for their role in shaping the future and scope of the ‘effectsbased approach’. We stand in the midst of a transition period in which the Commission and courts grapple with complex questions as to the treatments of market power and the abuse of dominance. Interestingly, the debate in relation to the adequate level of intervention is not confined to Europe. This is notably demonstrated in the US by the rift between the Department of Justice and the Federal Trade Commission on the treatment of market power. Discussion in relation to the evolution of Article 82 has also dominated the agenda of The University of Oxford Centre for Competition Law and Policy (CCLP). Guest lectures and round table discussions have added to the debate in past years. Chapters in this book are primarily a reflection of this ongoing debate at the CCLP. The book groups together 10 contributions which explore recent developments in relation to Article 82 and consider the future of its enforcement.

viii

Preface

In chapter one, James Kavanagh, Neil Marshall and Gunnar Niels provide an economic perspective on Article 82 and the effects-based approach. They discuss the reasons for the need for reform and analyse whether recent reforms have succeeded in moving the law toward an effects-based approach. They also discuss the extent to which there may be convergence between European and US policy in relation to predation, taking into account the recent opinion of the Advocate-General in France Télécom. In chapter two, Ioannis Lianos examines the shortcomings of the current approach to Article 82 and stresses the need for a reconceptualisation of the current categorisation of abuses. The chapter analyses the recent Commission Guidance on its enforcement priorities in applying Article 82 EC and considers whether it constitutes a real effects-based approach. The possible implications for competition law enforcement are explored, with a particular focus on the notion of consumer sovereignty. In chapter three, Ariel Ezrachi identifies some of the inherent difficulties associated with the departure from a formalistic approach and the introduction of effect-based analysis. The author considers how the Commission’s limited mandate to advance a reform which departs from traditional case law, impacts on the boundaries of Article 82. Furthermore, it highlights how the Commission’s attempt to widen the analytical framework of Article 82 may trigger inconsistencies, in so much as ‘effects-based variants’ are being advanced outside the precedent system. In chapter four, Orit Dayagi-Epstein sets out the evolution of the notion of consumer interest under Article 82, from a formalistic to an effectsbased approach that emphasises consumer welfare. In this framework the discrepancies between the Discussion Paper and the Commission Guidance with respect to consumer welfare are discussed. The chapter concludes by arguing that despite the modernisation process, the evolution of consumer interest is yet to be supported by the Community courts. In chapter five, Steven Anderman asks what is left of the ‘exceptional circumstances’ test following the Microsoft ruling. He argues that the test may have been widened to cover conduct limiting the technical development of secondary markets, but stresses that the test is itself subject to a finding of ‘essential facilities’ dominance. He discusses the exceptional nature of this concept and the effect that this type of dominance has upon the concept of abuse. In chapter six, Dan Eklöf discusses the interface between intellectual property (IP) and competition law, and reflects on the legal standards applied in Microsoft. He goes on to explore the software environment and its legal foundation in the EC computer programs directive, and the impact this had in the proceedings. Thereafter, it is argued that Microsoft follows a pattern of challenges to questionable IP rights under Article 82. The internal copyright provision targeting interoperability between different makes of software is then discussed, and the chapter concludes by expressing qualified support for antitrust-based compulsory licensing.

Preface

ix

In chapter seven, Kathryn McMahon considers one aspect of the debate that the Commission has placed at the forefront of its approach—the special responsibility of dominant firms not to allow their conduct to impair genuine, undistorted competition. The chapter analyses what ‘special responsibility’ can tell us about the goals of Article 82 and its significance within the reformed approach to the abuse of dominance. This is contrasted with the US approach, and the impact of the concept on the wider goals of EC competition law is discussed. In chapter eight, Pranvera Këllezi analyses recent developments in European competition law with regard to tying and bundling, and assesses them in light of economic theory. The chapter discusses whether the case law and the new Commission Guidance is consistent with economic thinking and whether it is flexible enough to allow for economic learning to be taken into account. In chapter nine, Ariel Ezrachi and David Gilo consider an area yet untouched by reform. They explore and question the main grounds commonly used to justify non-intervention in cases of alleged excessive pricing. The discussion focuses on one of the grounds for non-intervention—the complexity of evaluating whether a price is excessive—and puts forward a proposal for a post-entry price-cut benchmark which may facilitate the finding of excessive pricing. In chapter ten, Ulf Bernitz considers the consequences of a finding of abuse under Article 82, focusing on the fact that, in contrast to Article 81, there is no provision for voidness. The issue of voidness under Article 82 is revisited in light of the recent Commission White Paper on damages. In the author’s view, there is a strong link between the right to damages and the issue of voidness. A number of people have contributed to the success of this project. In addition to the authors, whose compelling contributions and close cooperation are much appreciated, I would like to thank Emma Cochrane for her dedicated help in editing the book. I would also like to express gratitude to Jenny Dix, administrator of the IECL and CCLP, who was instrumental in organising various CCLP events. Last but not least, thanks are due to the CCLP speakers, fellows, guests and students who have contributed their time, talent and expertise to creating vibrant debates on competition law. Ariel Ezrachi Oxford, April 2009

List of Contributors Steven Anderman, Professor of Law, University of Essex. Ulf Bernitz, Professor of European Law, Stockholm University; Director, Oxford/Stockholm Wallenberg Venture in European Law, Associate Senior Fellow, Pembroke College, Oxford. Orit Dayagi-Epstein, Research Affiliate, The University of Oxford Centre for Competition Law and Policy. Dan Eklöf, Associate Professor, Stockholm University; Associate, Sandart & Partners, law firm. Ariel Ezrachi, Director, The University of Oxford Centre for Competition Law and Policy; Slaughter and May Lecturer in Competition Law, Oxford University; Fellow, Pembroke College, Oxford. David Gilo, Associate Professor, Buchmann Faculty of Law, Tel Aviv University. James Kavanagh, Senior Consultant, Oxera Consulting Ltd. Pranvera Këllezi, Legal adviser for competition law at European Broadcasting Union, Geneva, Switzerland. Ioannis Lianos, City Solicitors’ Education Trust Reader in European Union Law; Jevons Institute of Competition Law and Economics, Faculty of Laws, University College London; Emile Noel Fellow, New York University Law School. Neil Marshall, Managing Consultant, Oxera Consulting Ltd. Kathryn McMahon, Associate Professor, School of Law, University of Warwick. Gunnar Niels, Director, Oxera Consulting Ltd.

Table of Cases Court of First Instance of the European Communities Michelin II see Case T–203/01 Case T–30/89 Hilti v Commission [1991] ECR II–1439, [1992] 4 CMLR 16................................................................................... 68, 162 Case T–65/89 BPB Industries plc and British Gypsum Ltd v Commission (British Gypsum) [1993] ECR II–385 ..................................... 71, 72 Joined Cases T–69, T–70, T–76/89 RTE v Commission [1991] ECR II–485 ........ 73 Case T–83/91 Tetra Pak v Commission [1993] ECR II–755 ....................... 122, 125 Joined Cases T–24/93 to T–26/93 and T–28/93 Compagnie Maritime Belge Transports and Others v Commission [1996] ECR II–1201 .................... 59 Case T–395/94 Atlantic Container Line [2002] ECR II–875 ................................ 53 Case T–111/96 ITT Promedia v Commission [1998] ECR II–2937 .................................................................................... 24, 122, 134 Case T–5/97 Industrie des poudres sphériques SA v Commission [2000] ECR II–3755............................................................... 42, 46 Case T–228/97 Irish Sugar plc v Commission [1999] ECR II–2969 ........................................................................ 59, 69, 70, 122, 126 Case T–65/98 Van den Bergh Foods Ltd v Commission [2003] ECR II–4653 ...................................................................... 24, 25, 38, 48, 71, 72 Joined Cases T–191/98, T–212/98 and T–214/98 Atlantic Container Line (TACA), [2003] ECR II–3275 ..................................... 53 Case T–219/99 British Airways v Commission [2003] ECR II–5917 .......................................... 7, 70, 71, 79, 82, 83, 88, 134, 138, 165 Case T–168/01 GlaxoSmithKline Services v Commission [2006] ECR II–2969 .......................................................................... vii, 31, 35, 84, 130 Case T–203/01 Manufacture Française des Pneumatiques Michelin v EC Commission [2003] ECR II–4071; [2004] 4 CMLR 18 (‘Michelin II’) ............................... 7, 23, 30, 58, 69, 70, 84, 85, 122 Case T–210/01 General Electric v Commission [2005] ECR II–5575 ................. 127 Joined Cases T–213/01 and 214/01 Osterreichische Postsparkasse AG and Bank für Arbeit und Wirtschaft AG v EC Commission [2006] ECR II–1601 .................................... 84 Case T–193/02 Laurent Piau v Commission [2005] ECR II–209 .......................... 52 Case T–271/03 Deutsche Telekom AG v Commission [2008] ECR II–477 ............................................................................................ 2, 42, 46 Case T–340/03 France Télécom SA v EC Commission [2007] ECR II–107: [2007] 4 CMLR 21 ..................................... vii, 2, 4, 11, 14, 17, 30, 58, 82, 83, 134, 135, 136 Case T–99/04 AC–Treuhand AG v Commission [2008] ECR 00 .......................... 56 Case T–201/04R Microsoft Corp v Commission [2004] ECR II–4463 ....... 106, 107

xx

Table of Cases

Case T–201/04 Microsoft Corp v Commission [2007] ECR II–3601 ................ vii, viii, 2, 23, 27, 28, 38, 58, 82, 83, 84, 87, 91, 93, 94, 95, 96, 97, 99, 100, 102, 103, 104, 105, 108, 109, 110, 111, 114, 117, 118, 127, 133, 134, 137, 139, 147, 148, 149, 157, 159, 163, 167 Case T–167/08 Microsoft Appeal......................................................................... 99

Court of Justice of the European Communities Cewal see Joined Cases C–395/96P Magill TV Guide see Case 241/91P Michelin I see Case 322/81 Case 26/62 NV Algemene Transport- en Expeditie Onderneming van Gend & Loos v Netherlands Inland Revenue Administration [1963] ECR 1 ................ 183 Case 56/65 Société La Technique Minière (LTM) v Maschinenbau Ulm [1966] ECR 235 .............................................................. 191 Case 40/70 Sirena v Eda [1971] ECR 69 .................................................... 171, 175 Case 6/72 Europemballage Corp. and Continental Can Co Inc v Commission [1973] ECR 215 ................... 68, 69, 70, 83, 138, 139 Joined Cases C–6/73 and C–7/73 Institute Chemioterapico Italiano SpA and Commercial Solvents Corp v Commission [1974] ECR 223 ............. 72, 81, 85, 89, 95, 96, 97 Case 40/73 Suiker Unie v Commission [1975] ECR 1663 .................................. 139 Case 127/73 Belgische Radio en Televisie v SV SABAM and NV Fonior [1974] ECR 313 .................................................... 188, 189, 190 Case 26/75 General Motors v Commission [1975] ECR 1367 ........... 171, 175, 179 Case 26/76 Metro SB-Grossmärkte GmbH & Co KG v Commission [1977] ECR 1875 .......................................................... 47 Case 27/76 United Brands v Commission [1978] ECR 207............................................. 85, 88, 123, 129, 134, 171, 175, 176, 177 Case 85/76 Hoffmann-La Roche AG v Commission [1979] ECR I–461 .............................. 69, 71, 73, 83, 84, 85, 123, 124, 129, 132 Case 22/78 Hugin Kassaregister AB v Commission [1979] ECR I–1869.............................................................................................. 68, 129 Case 30/78 Distillers Co v Commission [1980] ECR 2229 ................................... 35 Case 32/81 Michelin v Commission [1983] ECR 3461 ........................................ 91 Case 262/81 Coditel II [1982] ECR 3381........................................................... 107 Case 322/81 NV Nederlandsche Banden-Industrie Michelin NV v Commission [1983] ECR 3461 (‘Michelin I’) ......... 3, 23, 69, 122 Case 226/84 British Leyland Plc v Commission [1986] ECR 3263; [1987] 1 CMLR 185 .................................................................... 179 Case 311/84 Centre Belge d’Etudes de Marché-Télémarketing v CLT [1985] ECR 3261 ...................................... 97, 162 Case C–325/85 Commission v Ireland [1987] ECR 5041 ..................................... 36 Case C–62/86 AKZO Chemie BV v Commission [1991] ECR I–3359; [1993] 5 CMLR 215 ............. 5, 15, 46, 70, 85, 124, 134, 135, 176

Table of Cases

xxi

Case 66/86 Ahmed Saeed Flugreisen and Silver Line Reisebüro GmbH v Zentrale zur Bekämpfung unlauteren Wettbewerbs eV [1989] ECR 803 ....................... 60, 188 Case 30/87 Corinne Bodson v SA Pompes Funèbres des Régions Libérées [1988] ECR 2479 ........................................... 178 Case 238/87 Volvo v Veng (UK) Ltd [1988] ECR 6211 ........................ 92, 101, 114 Case 395/87 Ministère Public v Jean-Louis Tournier [1989] ECR 2521 ............. 178 Case 110/88 Lucazeau v SACEM [1989] ECR 2811 .......................................... 177 Case C–202/88 France v Commission [1991] ECR I–1223 ................................ 141 Case C–234/89 Stergios Delimitis v Henninger Bräu AG [1991] ECR I–935 ........ 60 Case 241/91P Magill TV Guide/ITP, BBC and RTE [1989] OJ L78/43 [1989] 4 CMLR 757 ................... 73, 81, 84, 106, 107, 111, 114, 115 Joined Cases C–241/91P and C–242/91P RTE and ITV v Commission [1995] ECR I–743 ........................ 73, 87, 91, 92, 93, 94, 95, 104, 105, 133, 139 Case C–53/92P Hilti v Commission [1994] ECR I–667...................... 68, 70, 74, 83 Case C–333/94P Tetra Pak International SA v Commission [1996] ECR I–5951; [1997] 4 CMLR 662 (Tetra Pak II) ........ 5, 15, 70, 74, 83, 124, 162, 176 Case C–242/95 GT-Link v De Danske Statsbaner [1997] ECR I–4449 ............... 193 Case C–200/96 Metronome [1998] ECR I–1953........................................ 111, 112 Joined Cases C–395/96P and C–396/96P Compagnie Maritime Belge and others v Commission [2000] ECR I–1365 (‘Cewal’) ................ 4, 6, 125, 126 Case C–7/97 Oscar Bronner GmbH & Co KG v Mediaprint [1998] ECR I–7791.................................................. 41, 42, 43, 88, 90, 91, 93, 115, 139 Case C–61/97 Laserdisken [1998] ECR I–5171 ................................................. 111 Case C–453/99 Courage Ltd v Bernard Crehan [2001] ECR I–6297.................................................................... 183, 188, 192, 193, 194 Case C–497/99 Irish Sugar plc v EC Commission [2001] ECR I–5333...................................................................................................... 69 Case C–2/01 and C–3/01P Bundesverband der Arzneimittel-Importeure eV & Commission v Bayer [2004] ECR I–23.................................................... 22 Case C–418/01 IMS Health GmbH & Co OHG v NDC Health GmbH & Co KG [2004] ECR I–5039 ......................... 73, 84, 87, 92, 93, 94, 95, 104, 105, 106, 107, 114, 133 Joined Cases 189/02P, C–202/02P, C–205/02P to C–208/02P and C–213/02P Dansk Rorindustri and Others v Commission [2005] ECR I–5425................... 55 Case C–552/03 Unilever Bestfoods Ireland v Commission [2006] ECR I–9091.......................................................................................... 25, 38, 48 Case C–95/04P British Airways v Commission [2007] ECR I–2331 ...................................................... 2, 7, 17, 23, 25, 28, 56, 57, 58, 71, 82, 83, 84, 134, 138, 140, 141 Case C–167/04P JCB Service v Commission of the European Communities [2006] ECR I–8935 .................................................................... 55 Joined Cases C–295/04 to C–298/04 Vincenzo Manfredi v Lloyd Adriatico Assicurazioni SpA et al [2006] ECR I–6619 ............ 190, 192, 193, 194 Joined Cases C–468/06 to C–478/06 Sot Lelos kai Sia v GlaxoSmithKline [2008] ECR 00 ........................... 30, 31, 32, 33, 34, 57, 59, 85 Cases C–501/06P, C–513/06P, C–515/06P, C–519/06P GlaxoSmithKline appeal ............................................................................. 31, 35

xxii Table of Cases Case C–49/07 Motosykletistiki Omospondia Ellados NPID (MOTOE) v Greece [2008] 5 CMLR 11......................................................... 141 Case C–202/07P France Télécom SA v Commission [2008] (decided 2 April 2009) ................................................................... viii, 2, 15, 134 Case C–8/08 T–Mobile Netherlands BV and Others [2009] ECR 00 ............. 19, 33

European Commission Decisions AOL/Time Warner, (2001/718/EC) (Case IV/M1845) [2001] OJ L268/28 ........... 74 BBI/Boosey & Hawkes: Interim Measures, [1987] OJ L286/36, [1988] 4 CMLR 67......................................................................................... 123 British Post Office v Deutsche Post AG, Case COMP/C–1/36.915, [2001] OJ L331/40; [2002] 4 CMLR 17......................................... 171, 175, 178 Deutsche Post AG–Interception of cross-border mail, [2001] OJ L331/40, [2002] 4 CMLR 598....................................................... 125 GB-Inno-BM/Fedetab, (78/670) [1978] OJ L224/29............................................. 24 General Electric, Case COMP/M.2220............................................................... 127 IBM ................................................................................................................... 109 Microsoft 2007/53/EC, Case COMP/C–3.37.792, [2007] OJ L32/23 .......................................... 58, 73, 84, 98, 124, 125, 147, 153, 158, 159, 160, 161, 163, 167 Napier Brown/British Sugar, (88/518/EEC) [1988] OJ L284/41 ........................... 42 Port of Rodby, [1994] OJ L55/52 ......................................................................... 90 Prokent Tomra, Case COMP/E–1/38.113 [2008] OJ C219/12 ............................. 55 Sea Containers Ltd/Stena Sealink, [1994] OJ L15/8 ............................................. 90 United Brands, [1976] OJ L95/1 ........................................................................ 171 Virgin/British Airways, [2000] OJ L30/1 .................................... 134, 135, 138, 165 Wanadoo Espana v Telefónica, [2007] Case COMP/38.784 ........................... 42, 43 Wanadoo Interactive, Case COMP/38.233, [2005] 5 CMLR 120................................................................... 14, 88, 134, 136 Windows Media Player (Microsoft) ................................................. 40, 41, 42, 160

Australia Melway Publishing Pty Ltd v Robert Hicks Pty Ltd (2001) 205 CLR 1...................................................................................................... 124 Queensland Wire v BHP (1989) 167 CLR 177, 202 ........................................... 124

Sweden Scandinavian Airlines System v Swedish Civil Aviation Administration (unreported) 27 April 2001, Case No T 33/00 (Arlanda Case) (Gota Court of Appeal) .............. 189, 191, 193

Table of Cases

xxiii

United Kingdom Aberdeen Journals Ltd v Director-General of Fair Trading [2002] Competition Appeals Tribunal 4;[2002] Competition Appeals Reports 167........................ 5, 6 Albion Water Limited v Water Services Regulation Authority [2006] CAT 36 ............................................................................... 177 At the Races Limited v The British Horse Racing Limits and others [2007] EWCA Civ 38................................................................................................... 62 British Horseracing Board v Office of Fair Trading [2005] CAT 29...................... 27 British Horseracing Board v Victor Chandler International [2005] EWHC 1074 (Ch)................................................................................... 176, 177 Inntrepreneur Pub Company (CPC) and others v Crehan [2006] UKHL 38.......................................................................................................... 61 Napp Pharmaceutical Holdings Ltd v Director-General of Fair Trading CA98/2/2001; [2002]CAT 1; Comp Appeal Reports 13 ................. 125, 173, 178

United States of America Berkey Photo v Eastman Kodak Co, 603 F 2d 263, 274 (2nd Cir, 1979) (US) ........................................................ 122, 126, 170, 172 Blue Cross & Blue Shield United v Marshfield Clinic, 65 F 3d 1406, 1413 (7th Cir 1995) (US)......................................................... 170 Brooke Group Ltd v Brown & Williamson Tobacco Corp, 509 US 209 (1993) ....................................................................... 4, 16, 122, 143 Brown Shoe Co v United States, 370 US 294 (1962) .......................................... 129 California Dental Association v FTC, 128 F 3d 720 (9th Cir 1997), rev’d 119 S Ct 1604 (1999).............................. 144 Chicago Professional Sports Ltd Partnership v NBA, 95 F 3d 593, 597 (7th Cir 1996) (US)............................................................. 170 Covad Communications Company v Bell Atlantic Corp, 398 F 3d 666 (DC Cir 2005) (US) .................................................................... 42 Data General Corp v Grumman Sys Support Corp, 36 F 3d 1147 (1st Cir 1994) ........................................................................... 100 Eastman Kodak Co v Image Technical Services Inc, 504 US 451 (1992) ................................................................................. 123, 144 Linkline Communications Inc v Pacific Bell Telephone Co, AT&T, 503 F 3d 876 (9th cir 2007) ......................................................... 42, 128 Olympia Equipment Leasing v Western Union Telegraph, 797 F 2d 370 (7th Cir 1986) .................................................................. 126, 127 Pacific Bell Telephone Co v Linkline Communications Inc, (No. 07–512) 503 F 3d 876 US (2009) ....................................... 43, 48, 128, 143 Standard Oil Co (Cal) v United States, 337 US 293, 305–306 (1949) ................ 147 United States v Aluminium Co of America (Alcoa), 148 F 2d 416, 444 (1945)....................................................... 122, 125, 129, 171 United States v American Can Co, 230 F 859 .................................................... 170 United States v Colgate & Co, 250 US 300 (1919)............................................... 22

xxiv

Table of Cases

United States v Microsoft Corp, 87 F Supp 2d 30 (DDC); 253 F 3d 34 (DC Cir 2001); 2001 US App LEXIS 14324, HN36 ........... 144, 148 United States v Terminal Railroad Assn of St Louis, 224 US 383 (1912) .............. 89 United States v Trans-Missouri Freight Assn, 166 US 290 (1897) ...................... 175 US v Swift & Co, 286 US 106 ............................................................................ 122 Verizon Communications v Trinko, 124 S Ct 872 (US)............................................... 48, 89, 110, 127, 143, 170, 174 Virgin Atlantic Airways Ltd v British Airways plc, 257 F 3d 256 (2d Cir 2001)........................................................................ 8, 142 Weyerhaeuser Co v Ross-Simmons Hardwood Lumber Co, 127 S Ct 1069 (2007) ......................................................................... 4, 122, 143

Table of Legislation European Union EC Treaty ..................................................................... 22, 33, 60, 71, 77, 140, 141 Art 3(1)(g) .............................................................................................. 130, 139 Art 3(g)........................................................................................................... 142 Art 10 ............................................................................................................. 189 Art 81 ........................................... ix, 1, 9, 19, 22, 23, 24, 25, 27, 28, 29, 31, 33, 35, 36, 45, 48, 49, 54, 56, 60, 61, 63, 68, 69, 74, 83, 84, 107, 108, 109, 112, 182, 187, 189, 192, 195 Art 81(1) .................................................................. 25, 26, 28, 33, 56, 188, 192 Art 81(2) ................................................................................................ 188, 195 Art 81(3) .................. 13, 25, 26, 27, 28, 29, 30, 33, 52, 54, 60, 74, 84, 130, 165 Art 82 ................................ vii, viii, ix, 1, 2, 3, 7, 8, 9, 13, 17, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 38, 39, 41, 42, 43, 44, 45, 47, 48, 49, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 67, 68, 69, 70, 71, 72, 74, 75, 77, 79, 82, 83, 84, 85, 86, 87, 88, 89, 91, 92, 94, 95, 96, 97, 98, 99, 100, 101, 102, 104, 105, 106, 107, 108, 109, 110, 111, 112, 114, 115, 117, 118, 119, 121, 122, 123, 124, 126, 127, 129, 131, 133, 134, 136, 137, 138, 139, 140, 141, 142, 143, 147, 148, 149, 157, 165, 166, 169, 171, 175, 182, 184, 187, 188, 189, 192, 193, 195, 196 Art 82(a)..................................................................................... 31, 68, 171, 175 Art 82(b) ................................................ 31, 42, 68, 87, 93, 94, 95, 96, 139, 148 Art 82(c) ........................................................................................... 31, 138, 139 Art 82(d) .......................................................................... 31, 147, 148, 159, 160 Art 85(1) .......................................................................................................... 56 Art 86 ............................................................................................................. 141 Art 234 ............................................................................................... 57, 62, 141 Art 249 ............................................................................................................. 82 Lisbon Reform Treaty ........................................................................................ 144 Directive 91/250/EEC (Software) ....................... 100, 108, 109, 110, 111, 117, 118 Preamble......................................................................................................... 109 Arts 6–7.......................................................................................................... 116 Art 9 ............................................................................................................... 116 Directive 92/100/EEC (Rental Rights and Lending Rights)................................. 111 Directive 96/9/EC (Legal protection of databases), [1996] OJ L77/20 ........ 108, 109 Preamble, Recital 47 ....................................................................................... 109 Directive 2001/29/EC (Harmonisation of certain aspects of copyright and related rights in the information society), Preamble, Recital 50 ............... 116 Directive 2006/115/EC (Rental Rights and Lending Rights) ....................... 111, 112

xxvi

Table of Legislation

Regulation (EC) No 1/2003 (Implementation of the rules on competition laid down in Articles 81 and 82 of the Treaty), [2003] OJ L1/1........................ 25 Recital 5 ........................................................................................................... 26 Art 1(3) .......................................................................................................... 188 Art 2 ................................................................................................................. 26 Art 7(1) .......................................................................................................... 182 Regulation (EC) No 139/2004 (Control of concentrations between undertakings (Merger Regulation)), [2004] OJ L24/1 Art 2 ................................................................................................................. 23 Art 16 ............................................................................................................... 61

European Commission Guidelines and Notices Guidance on the Commission’s Enforcement Priorities in applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings (Brussels, 3 December 2008).................................. vii, viii, ix, 2, 8, 9, 11, 12, 13, 14, 17, 19, 21, 25, 26, 44, 45, 46, 49, 51, 53, 54, 55, 56, 57, 60, 61, 62, 64, 67, 68, 74, 76, 77, 79, 80, 81, 82, 85, 86, 105, 121, 122, 123, 126, 131, 132, 139, 142, 149, 150, 154, 155, 156, 157, 158, 161, 162, 163, 164, 166, 167, 169, 176 Pt D .................................................................................................................. 43 Section III.D ....................................................................................................... 9 Section IV.C ...................................................................................................... 11 para 1 ............................................................................................. 122, 123, 162 para 3 ....................................................................................................... 55, 122 para 5 ......................................................................................... 48, 77, 121, 131 para 6 ............................................................................................. 8, 14, 79, 131 para 7 ....................................................................................................... 76, 121 para 9 ..................................................................................................... 122, 123 para 10 ............................................................................................................. 24 para 11 ................................................................................................. 24, 75, 77 para 14 ..................................................................................................... 14, 124 para 15 ........................................................................................................... 126 para 17 ........................................................................................................... 124 para 19 ..................................................................... 10, 76, 77, 79, 81, 131, 157 paras 19ff ....................................................................................................... 105 para 20 ................................................................................. 8, 79, 126, 161, 162 paras 20ff ....................................................................................................... 107 para 21 ............................................................................................................. 80 para 22 ............................................................................. 8, 33, 45, 46, 132, 166 para 23 ............................................................................................... 12, 39, 132 para 24 ....................................................................................................... 39, 45 para 25 ........................................................................................................... 132 para 26 ............................................................................................................. 39 paras 27ff ....................................................................................................... 105 para 29 ............................................................................................. 13, 141, 166

Table of Legislation

xxvii

para 30 ......................................................................................... 26, 27, 54, 166 para 31 ........................................................................................................... 166 para 36 ........................................................................................................... 132 para 37 ............................................................................................................. 54 para 38 ....................................................................................................... 10, 54 para 39 ............................................................................................................. 54 para 40 ............................................................................................... 11, 54, 141 para 41 ....................................................................................................... 11, 54 para 42 ..................................................................................................... 54, 142 paras 42ff ......................................................................................................... 53 para 43 ............................................................................................. 54, 136, 142 paras 44–46 ...................................................................................................... 54 para 49 ..................................................................................................... 38, 166 para 50 ........................................................................................................... 159 para 52 ............................................................................................................. 48 para 53 ........................................................................................... 155, 163, 164 para 54 ........................................................................................................... 155 para 56 ........................................................................................................... 164 para 57 ........................................................................................................... 155 para 59 ............................................................................................. 39, 163, 164 paras 59ff ......................................................................................................... 53 para 60 ..................................................................................................... 39, 164 para 62 ........................................................................................... 132, 136, 166 paras 63–65 .................................................................................................... 132 para 67 ........................................................................................................... 144 para 69 ....................................................................................................... 80, 83 para 73 ..................................................................................................... 13, 136 para 74 ................................................................................................... 107, 108 paras 74ff ............................................................................................... 105, 108 para 77 ................................................................................................... 107, 108 para 79 ............................................................................................................. 47 para 80 ........................................................................................................... 133 para 81 ................................................................................................... 107, 108 para 82 ........................................................................................................... 139 para 83 ................................................................................................... 107, 108 para 84 ..................................................................................................... 52, 133 para 85 ............................................................................... 52, 79, 107, 108, 133 para 86 ................................................................................. 52, 77, 79, 107, 133 para 87 ............................................................................................... 52, 79, 133 paras 88–89 .............................................................................................. 52, 107 paras 90–92 ...................................................................................................... 52 para 118 ........................................................................................................... 11 Guidelines on the applicability of Article 81 of the EC Treaty to horizontal cooperation agreements [2001] OJ C3/2 para 36 ........................................................................................................... 165 para 71 ........................................................................................................... 165 paras 79–86 ...................................................................................................... 74 para 105 ......................................................................................................... 165

xxviii Table of Legislation para 134 ......................................................................................................... 165 para 155 ......................................................................................................... 165 Guidelines on the Application of Article 81 of the EC Treaty to Technology Transfer Agreements [2004] OJ C101/2 ................................ 74, 160 para 8 ............................................................................................................. 107 para 147 ......................................................................................................... 107 para 151 ......................................................................................................... 165 Guidelines on the Application of Article 81(3) of the Treaty [2004] OJ C101/97................................................................................ 69 para 11–06 ....................................................................................................... 53 para 21 ....................................................................................................... 74, 76 para 24 ....................................................................................................... 28, 74 para 46 ............................................................................................................. 76 para 51 ............................................................................................................. 28 para 84 ......................................................................................... 68, 74, 84, 130 para 85 ............................................................................................................. 26 para 96 ....................................................................................................... 69, 74 para 102 ..................................................................................................... 69, 74 para 106 ......................................................................................................... 165 Guidelines on the Assessment of Horizontal Mergers under the Council Regulation on the Control of Concentrations between Undertakings [2004] OJ C31/3, paras 24–38 ...................................... 23 Guidelines on the Assessment of Horizontal Mergers under the Council Regulation on the Control of Concentrations between Undertakings [2004] OJ C31/5, paras 79–80 ...................................... 69 Guidelines on Vertical Restraints [2000] OJ C291/1 para 104 ........................................................................................................... 45 para 135 ......................................................................................................... 165 para 153 ......................................................................................................... 165 para 216 ......................................................................................................... 160 paras 221–222 ................................................................................................ 165 Notice on the Application to Access Agreements in the Telecommunication Sector [1998] OJ C265/2 .................................................. 90

Discussion Paper on Article 82 Discussion Paper on the Application of Article 82 of the Treaty to Exclusionary Abuses (December 2005)................. vii, viii, ix, 1, 2, 7, 8, 9, 10, 11, 12, 17, 19, 20, 24, 38, 44, 45, 46, 52, 67, 68, 74, 76, 77, 78, 80, 81, 82, 85, 86, 101, 121, 133, 142, 149, 161, 162, 163, 166, 169 Section 6 ........................................................................................................... 12 Section 7 ........................................................................................................... 10 para 4 ................................................................................... 75, 77, 78, 121, 131 para 6 ............................................................................................................... 55 para 7 ............................................................................................................... 53 para 8 ............................................................................................................... 52

Table of Legislation

xxix

paras 23–24 ...................................................................................................... 24 para 54 ............................................................................................... 76, 78, 131 para 55 ................................................................................................. 68, 76, 78 para 56 ....................................................................................................... 76, 78 para 58 ................................................................................................. 10, 76, 78 para 63 ............................................................................................................. 46 para 67 ............................................................................................................. 39 para 73 ............................................................................................................. 45 para 80 ............................................................................................................. 32 paras 81–82 ...................................................................................................... 78 para 88 ............................................................................................................. 76 para 91 ........................................................................................................... 126 para 92 ............................................................................................................... 4 para 122 ..................................................................................................... 17, 80 para 129 ........................................................................................................... 12 para 133 ........................................................................................................... 12 para 143 ............................................................................................................. 9 para 153 ............................................................................................................. 9 para 154 ................................................................................................. 9, 10, 77 para 178 ......................................................................................................... 166 para 182 ............................................................................................. 37, 38, 163 para 188 ......................................................................................................... 162 para 189 ................................................................................................... 38, 163 para 196ff....................................................................................................... 162 para 199 ......................................................................................................... 162 para 213 ......................................................................................................... 107 para 230 ......................................................................................................... 107 para 235 ......................................................................................................... 107 para 241 ......................................................................................................... 107 para 252 ......................................................................................................... 101 para 264 ......................................................................................................... 101

Denmark Contract Act, s 36 .............................................................................................. 191

Finland Contract Act, s 36 .............................................................................................. 191

Germany Act against Restraints of Competition s 20................................................................................................................. 129 s 20(2) ............................................................................................................ 129

xxx Table of Legislation Iceland Contract Act, s 36 .............................................................................................. 191

Norway Contract Act, s 36 .............................................................................................. 191

Sweden Contract Act, s 36 .............................................................................................. 191

United Kingdom Competition Act 1998...................................................................................... 5, 61 s 60(2) .............................................................................................................. 61 s 60(3) .............................................................................................................. 61

United States of America Sherman Act............................................................................................... 127, 129 s 1.................................................................................................................... 22 s 2.......................................................................... 22, 29, 43, 89, 126, 127, 143

1 Reform of Article 82 EC—Can the Law and the Economics be Reconciled? JAMES KAVANAGH, NEIL MARSHALL AND GUNNAR NIELS*

I. INTRODUCTION

T

HE EUROPEAN COMMISSION’S review of its approach to Article 82 of the EC Treaty put abuse of dominance prominently on the agenda of policy issues in EU competition law in 2005. The review followed the substantial changes made in the preceding years to the other two pillars of competition law: merger control and the policy towards restrictive agreements under Article 81 of the EC Treaty. One outcome of these reforms is that EU policy on mergers and agreements has been brought more into line with current economic thinking. A similar objective was set for the Article 82 reform. European Union case law on abuse of dominance had for years been criticised as legalistic and interventionist, and the review was seen by many as an opportunity for change. Various commentators have stressed the desirability of moving towards an approach that emphasises the actual or expected economic effects of allegedly abusive behaviour by dominant firms, rather than the form of the behaviour.1 The Commission commenced public debate on the reform process by issuing its Discussion Paper on Article 82 in December 2005.2 This was followed by public hearings and much debate among economists, lawyers *

Oxera Consulting Ltd. The views expressed in this article are those of the authors. See, eg, J Vickers, ‘Abuse of Market Power’ (2005) 115 Economic Journal, F244–61; Competition Law Forum Article 82 Review Group, ‘The Reform of Article 82: Recommendations on Key Policy Objectives’, European Competition Journal, 1, 179–83, see also www.biicl.org/ files/53_the_reform_of_article_82_recommendations_on_key_policy_objectives.pdf. 2 DG Competition, Discussion Paper on the Application of Article 82 of the Treaty to Exclusionary Abuses (December 2005), http://europa.eu.int/comm/competition/antitrust/ others/discpaper2005.pdf. 1

2

James Kavanagh, Neil Marshall and Gunnar Niels

and policy-makers about the way forward. In the meantime, a number of important Article 82 judgments by the European Court of First Instance (CFI) influenced the debate (and did not always go in the same direction as the Commission Paper). These included judgments upholding the Commission’s finding against Microsoft for tying and refusal to supply,3 the predatory pricing decision against France Télécom4 and the margin squeeze case against Deutsche Telekom.5 In another case the European Court of Justice (ECJ) upheld the Commission’s decision concerning loyalty rebates offered by British Airways.6 In December 2008, the Commission published what it called Guidance on ‘enforcement priorities’ in applying Article 82.7 Although not the full set of guidelines that some had hoped for, the Guidance demonstrates the Commission’s further push towards an effectsbased approach. In this chapter we first set out some of the economic perspectives on why the policy towards abuse of dominance is in need of change. We then describe how the Commission’s 2005 Discussion Paper and 2008 Guidance on enforcement priorities have indeed sought to move Article 82 policy towards a more effects-based approach (although perhaps not as far, or as consistently, as we would have liked), and balance this with established case law. Lastly, we discuss to what extent there may be convergence between European and US policy in relation to predation, taking into account the recent (September 2008) Opinion of the Advocate-General in France Télécom, which in fact goes against the prior rulings of both the CFI and the Commission, and may therefore herald change in line with the Commission’s Article 82 reform.8

II. WHY DOES ARTICLE 82 POLICY NEED REFORM?

European Union case law on abuse of dominance has, for years, been criticised as legalistic and interventionist. Perhaps less attention has been paid to where it is exactly that the current approach fails. Many have argued that Article 82 protects competitors rather than competition—but how does this come about? We explain below that one of the fundamental shortcomings in the current case law is the virtual per se prohibition of certain practices once a

3

Case T-201/04 Microsoft Corp v Commission [2007] ECR II-3601. Case T-340/03 France Télécom SA (formerly Wanadoo Interactive SA) v EC Commission [2007] 4 CMLR 21. 5 Case T-271/03 Deutsche Telekom AG v Commission [2008] ECR II-477. 6 Case C-95/04 P British Airways v Commission [2007] ECR I-2331. 7 DG Competition, Communication from the Commission, Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings (Brussels, 3 December 2008). 8 Case C-202/07P France Télécom SA v Commission [2008] Opinion of A-G Mazák. 4

Reform of Article 82 EC—Can Law and Economics be Reconciled? 3 firm is deemed to be dominant; in other words, dominance is used as a shortcut to infer anti-competitive effects. This can be misleading, as the implied link between dominance and effects does not always hold. Article 82 policy in Europe has traditionally been influenced by the ‘ordo-liberal school’. In essence, this school of thought emphasises individual freedom as the primary objective of competition policy, and considers that the presence of dominant firms weakens the competitive process and reduces the economic freedom of other market participants. The notion that the mere existence of dominant firms is a threat to competition is still deeply embedded in EU law. A dominant firm is in effect regarded as the proverbial bull in a china shop—it must be restrained to prevent it from inflicting further damage to its already fragile surroundings.9 As formally established in Michelin I, a dominant firm has a ‘special responsibility not to allow its conduct to impair genuine undistorted competition on the common market’.10 This view of how competition works is somewhat outdated. Economic theory and practical experience over the past 30–40 years have shown that competitive dynamics can function well even if a market has some very large players. Indeed, large players and/or temporary positions of market power can improve competitive dynamics. The theory has also established that certain behaviour can have positive efficiency effects, even if practised by dominant firms. In other words, using the bull in a china shop analogy to describe dominant firms does not fit well with current thinking on how markets work. The shortcomings of the current approach to Article 82 arise from a combination of the following three policy aspects. First, EU case law has only one threshold for dominance, regardless of the type of practice in question. Secondly, this threshold is set relatively low. Thirdly, the threshold constitutes the basis for a form-based per se prohibition of certain behaviour. With regard to the first of these, it is important to bear in mind that market power is a matter of degree—Microsoft has market power, but so has a small corner shop (provided there are few other shops nearby). Dominance can be interpreted as a very high degree of market power—one that enables a firm to sustain prices above the competitive level in the long run without inducing customer switching or competitor entry. The assessment of dominance can be a useful intermediary step in the analysis of an alleged abuse of dominance. In particular, many types of behaviour are likely to be of little competitive concern if the accused firm is not actually dominant—in other words, dominance is a necessary 9 See Niels and Jenkins, ‘Reform of Article 82: Where the Link Between Dominance and Effects Breaks Down’ (2005) 26(11) European Competition Law Review 605. 10 Case 322/81 Nederlansche Banden-Industrie Michelin NV v Commission [1983] ECR 3461 (‘Michelin I’).

4

James Kavanagh, Neil Marshall and Gunnar Niels

condition, and can be used to filter out cases below the threshold. However, in EU law, dominance is also used the other way around—to apply per se prohibitions to firms above the threshold. This policy approach can be problematic. From an economic perspective, the competitive effects of any business practice will depend on, first, the type of practice in question, and secondly, the degree of market power of the firm under scrutiny. The current ‘one-size-fits-all’ dominance threshold, the traditional market share rule of thumb of 40 per cent (combined with some indications of entry barriers), still seems to be the norm. This is arguably set too low to allow for any strong inferences of anti-competitive effects, as discussed in this chapter. This is not to say that a per se prohibition above a certain dominance threshold can never be the correct policy approach. Such a prohibition might be justified if the threshold were set sufficiently high—some types of behaviour (such as refusal to deal or margin squeeze) are very likely to be anti-competitive if engaged in by a firm with a near monopoly. In this respect it is worth noting that some EU cases have referred to the concept of ‘super-dominance’, which might be a reasonable threshold above which the behaviour under investigation can be presumed to have anti-competitive effects.11 However, in our view, if the current, lower threshold for dominance is maintained, this can only be used as an intermediate step in the analysis of actual or likely anti-competitive effects, ie as a necessary but not sufficient condition. A further assessment of those effects is required. Take the examples of predatory pricing and targeted discounting—two practices where the link between dominance and effects may not hold. Antitrust law in the US has established the ‘recoupment test’ for predation cases, which places strong emphasis on whether market structure is such that predation is feasible and hence the initial losses from predation can be recouped through subsequent monopoly profits.12 While sometimes criticised as overly harsh on complainants, the recoupment test has consumer welfare at its heart, as it essentially allows any aggressive price cut as long as there is no realistic expectation of successful monopolisation of the market. In contrast, and subject to the interpretation of Advocate-General Mazák in his Opinion in France Télécom (see section IV below), EU law has explicitly rejected the recoupment test—it considers pricing below the acceptable 11 See Joined Cases C-395/96 P and C-396/96 P Compagnie Maritime Belge SA [2000] ECR I-1365, Opinion of A-G Fennelly, at para 137: super-dominance was defined as a ‘position of overwhelming dominance verging on monopoly’; the Discussion Paper (above n 2), para 92, seems to interpret this as a market share in excess of 75% and limited competition from the other firms in the market. 12 Brooke Group Ltd v Brown & Williamson Tobacco Corp, 509 US 209 (1993); the US Supreme Court reaffirmed this effects-based test in Weyerhaeuser Co v Ross-Simmons Hardwood Lumber Co Inc, Opinion of the Court, 05-381, 20 February 2007.

Reform of Article 82 EC—Can Law and Economics be Reconciled? 5 level (see below) by dominant firms as abusive, if there is a risk that competitors are eliminated. It therefore sees no need to establish the feasibility of predation.13 Another (more economic) justification for rejecting the recoupment test might be that the possession of dominance in itself means that predation is likely to be feasible. However, the conditions for recoupment are typically more stringent than those for dominance, as defined in EU law. For example, a very high market share is needed from the start, and greater emphasis is placed on the possibility of successfully maintaining monopoly prices in future.14 Instead, predatory pricing cases in the EU focus primarily on the relationship between prices and costs. In the AKZO judgment, the ECJ determined that predation can be presumed if a dominant firm sets prices below average variable cost.15 In this respect, this test is in line with basic economic theory, which also identifies marginal cost (or some variant) as a relevant price floor for predation cases. Yet the risk is that pricing below average variable cost by dominant firms is outlawed per se, without any consideration of whether such pricing: a)

actually has negative effects on competition, which is not always the case—for example, pricing below average variable cost during one month only (the basis for the UK Office of Fair Trading’s finding of predation in Aberdeen Journals16) is arguably not sufficient to infer an anti-competitive effect; b) may be justified on efficiency grounds—for example, if there are strong network effects, pricing below average variable cost may be required to gain critical mass, even in the absence of any competitors.17 The link between dominance and effects on competition may also be questionable in targeted discounting and ‘fighting brands’ cases. Targeted discounting arises where a dominant firm sets low prices only for those customers who are using, or are likely to switch to, a competitor, but keeps other prices unchanged (the AKZO case is an example). Fighting

13

Case C-333/94 P Tetra Pak v Commission [1996] ECR I-5951. See Niels and Ten Kate, ‘Predatory Pricing Standards: Is there a Growing International Consensus’ (2000) 45 Antitrust Bulletin 787. 15 Case C-62/86 AKZO Chemie v Commission [1991] ECR 1-3359, [1993] 5 CMLR 215. Prices in the range between average variable cost and average total cost are deemed predatory if the purpose of the conduct was to eliminate a competitor. 16 Aberdeen Journals Ltd v Director-General of Fair Trading [2002] Competition Appeals Tribunal 4, [2002] Competition Appeals Reports 167. Note that there were also legal reasons why the OFT challenge was limited to a one-month period, relating to the date of entry into force of the Competition Act 1998. 17 See Ten Kate and Niels, ‘Below-cost Pricing in the Presence of Network Externalities’ in Swedish Competition Authority, The Pros and Cons of Low Pricing (Stockholm, Swedish Competition Authority, 2003). 14

6

James Kavanagh, Neil Marshall and Gunnar Niels

brands—or fighting ships in Compagnie Maritime Belge and fighting titles in Aberdeen Journals—are also specifically targeted at competitors. They are priced low, while the price of the main brand remains unaffected. In ‘standard’ predation cases, economic theory establishes a link between market share and the likelihood of success of predation—the higher a predator’s market share, the quicker it can depress market price, and hence the greater the chance that rivals will leave the market soon.18 With targeted discounting and fighting brands, however, this link does not hold because the predator is not reducing the prices on all of its products in the market. Take the example in Figure 1: a dominant firm with 90 per cent of the market pitches a fighting brand (representing 10 per cent) against a competitor with a 10 per cent market share. The firm therefore sets prices below average variable cost only for the fighting brand, keeping the price of its main brand unaffected. In other words, the predator is not using the full weight of its dominance (ie, its 90 per cent share), and hence the possession of such dominance does not lead directly to actual or likely effects on competition. This is not to say that complaints about these forms of predation should be rejected outright. Indeed, the ability to use a fighting brand may make predation more likely, particularly in the context of differentiated products. However, the point is that the actual or likely effects of targeted discounting

Main brand 80% Pitched below average variable cost at rival Fighting brand 10%

Rival brand 10%

Figure 1 Predation through a fighting brand

18 To give an example, if market demand is inelastic (elasticity = –0.5), and predation requires the market price to be cut by 40 per cent, a predator with a 10 per cent market share would have to triple its output; a predator with 40 per cent still must increase output by half; while a predator with 90 per cent needs to increase output by only 20 per cent.

Reform of Article 82 EC—Can Law and Economics be Reconciled? 7 or fighting brands need to be assessed more carefully, and the association between the fighting brand and the company’s overall position examined to analyse the incentives that this association generates. The effects cannot simply be inferred from the position of dominance in the broader market as such, nor from the fact that the fighting brand is priced below average variable cost. Form-based rules have also been applied to practices that may result in vertical foreclosure by dominant firms, such as loyalty incentives for distributors. Under EU law, dominant firms are prevented from engaging in any such practice, as confirmed in British Airways.19 Again, such a per se approach may be overly intrusive and unrelated to the actual or likely effects on competition. Dominance is a necessary but not a sufficient condition for vertical foreclosure to occur. Whether a significant part of the distribution channel is indeed foreclosed must also be assessed. For example, if a dominant firm with 60 per cent of the market imposes exclusivity requirements on 10 per cent of all distributors, the foreclosure effect is likely to be limited. In this case, a prohibition seems less appropriate than if, say, 60 per cent of the distribution channel were foreclosed, particularly if such exclusivity generates certain efficiency benefits (which would also need to be assessed as part of the effects-based test). The treatment of loyalty rebates illustrates the shortcomings of the formbased approach in EU law. In two judgments in 2003—Michelin II and British Airways—the CFI confirmed the long-established EU policy that dominant firms are allowed to offer discounts that relate to cost savings, but cannot offer discounts to encourage loyalty.20 Any considerations of the actual or likely effects of such discounts on competition are deemed irrelevant. This view was upheld by the ECJ in its 2007 judgment in British Airways.21 Notably, this judgment was issued after the Commission had published the Discussion Paper (December 2005) in which it sought to move Article 82 policy towards an effects-based approach. In British Airways, the concern was that the loyalty incentives offered by the airline to travel agents were retrospective, ie paid on all ticket sales once the target threshold had been reached and not just on the incremental sales above that target. This could thus induce agents who were close to their sales target to promote British Airways rather than rival airlines. The form of this incentive scheme was considered an abuse, without much consideration of the effects. One such effect could have been the foreclosure of sales channels to Virgin Atlantic, a competitor and complainant in this case. However, Virgin had continued to gain market share throughout the 19

Case C-95/04 P British Airways v Commission [2007] ECR I-2331, above n 6. Case T-203/01 Michelin v Commission [2003] ECR II-4071; Case T-219/99 British Airways v Commission [2003] ECR II-5917. 21 Above n 6. 20

8

James Kavanagh, Neil Marshall and Gunnar Niels

relevant period, which indicates that the competitive effect of the loyalty schemes was probably limited. For the US court that reviewed the same facts, this was one reason (among others) to reject the complaint filed by Virgin in that jurisdiction.22 In contrast, the European Commission simply made the point that Virgin would have had even more success in the absence of the loyalty schemes. This may well have been the case, but should not automatically lead to a finding of abuse under Article 82. III. EMBRACING THE PRINCIPLES OF AN ‘EFFECTS-BASED’ APPROACH

In its Guidance on enforcement priorities the Commission explicitly endorses the principle that it is the economic effects of allegedly abusive behaviour on competition and consumers that count, rather than the form of the behaviour. In this it follows the approach first set out in the 2005 Discussion Paper. It states that ‘what really matters is to protect an effective competitive process and not simply protecting competitors’,23 thus seeking to address the criticism frequently made in the past that the Commission is more concerned with the fate of competitors than competition. The Commission then sets out principles for assessing exclusionary abuses that indeed open the door to a more effects-based approach. The Paper makes the following points clear: a) The degree of dominance matters:24 this makes sense from an economic perspective, as discussed above. Practices such as below-cost pricing and refusal to supply are more likely to have an anti-competitive effect if the perpetrator has a very high degree of dominance (or even super-dominance) than if it just passes the (currently low) threshold for dominance. b) The degree of foreclosure matters:25 the Guidance states that it is not just the nature or form of the conduct that matters, but also its incidence (ie the extent to which the dominant firm is engaging in that conduct). c) Only conduct that would exclude ‘as-efficient’ competitors is abusive:26 this principle seeks to draw the line between exclusionary conduct that simply reflects ‘competition on the merits’ and exclusionary conduct that is harmful to consumers.

22 23 24 25 26

Virgin Atlantic Airways Ltd v British Airways plc, 257 F 3d 256 (2d Cir 2001). Commission Guidance, above n 7, para 6. Ibid, para 20, 1st bullet. Ibid, para 20, 5th bullet. Ibid, para 22.

Reform of Article 82 EC—Can Law and Economics be Reconciled? 9 d)

Exclusionary behaviour can still be justified on grounds of objective necessity, meeting competition, or efficiency:27 the efficiency defence is a new element in Article 82 policy, and recognises the economic principle (also embodied in Article 81) that some restrictive practices may be required to achieve efficiencies.

However, a critical question remains: does the Commission consistently stick to the principles? Perhaps not surprisingly, given the complex nature of Article 82 policy and the established case law, this question must be answered in the negative. The 2005 Discussion Paper (but less so the 2008 Guidance) lends itself to different interpretations of the proposed new approach to Article 82, including one interpretation where little changes with respect to current case law.28 The inconsistencies become particularly clear where the 2005 Discussion Paper and the 2008 Guidance discuss specific practices (predatory pricing, exclusive dealing,29 tying and bundling, refusal to supply and margin squeeze), as the following points relating to the above four principles illustrate.

A. The degree of dominance matters An example of where this principle gets somewhat lost is where the 2005 Paper discusses rebates. A number of statements in this section are reminders of the ‘ordo-liberal’ view of a dominant firm as the ‘bull in a china shop’. For example, we read that for a good part of demand on the market there are no proper substitutes to the dominant supplier’s product, because for instance its brand is a ‘must stock item’ preferred by many final consumers or because the capacity constraints on the other suppliers are such that a good part of demand can only be provided for by the dominant supplier.30

In other words, there is an inelastic or ‘non-contestable’ portion of the demand of each customer31 for which competitors, which are ‘smaller’, cannot compete.32 All these statements about a dominant firm may well be true, but this cannot be assumed from the outset—indeed, they should form an integral part of the effects-based analysis. In the 2008 Guidance, the Commission 27

Ibid, section III.D. See Niels, ‘The Article 82 Discussion Paper: A Comment on the Economic Principles’, April 2006, available at www.oxera.com/cmsDocuments/Agenda_April%2006/The%20 Article%2082%20discussion%20paper.pdf. 29 In the Discussion Paper the Commission used the term ‘single-branding’, but in the Guidance it changed this to ‘exclusive dealing’, which is more conventional. 30 Discussion Paper, above n 2, para 143. 31 Ibid, para 153. 32 Ibid, para 154. 28

10

James Kavanagh, Neil Marshall and Gunnar Niels

acknowledges this point. In its discussion about conditional rebates in that document,33 the Commission stresses the importance of establishing whether there is actually a ‘non-contestable’ portion of demand for which rivals of the dominant firm compete, as opposed to assuming from the outset (as it did in the Discussion Paper) that rivals are hampered in that way.

B. The degree of foreclosure matters In the Discussion Paper, the Commission had proposed a definition of foreclosure that focused very much on the effect on individual competitors: ‘By foreclosure is meant that actual or potential competitors are completely or partially denied profitable access to a market’.34 In another significant development between the Discussion Paper and the Guidance document, the Commission has redefined potentially foreclosing behaviour as that which enables the dominant firm to ‘profitably increase prices to the detriment of consumers’.35 The reformulation contained in the Guidance recognises that, to be significant, the behaviour of the incumbent should affect most or all competitors, not just individual ones. However, there remain concerns in relation to the test to establish whether loyalty rebates could be considered abusive. In the 2005 Discussion Paper, a numerical example was included which led to the opposite result than that implied by the appropriate definition of foreclosure. In particular, the first part of the test for loyalty rebates established in Section 7 of the Discussion Paper resulted in quite counterintuitive logic: the numerical example presented in the box following paragraph 154 implied that the smaller the ‘commercially viable amount’ that a competitor would require to operate in the market, the more likely it is that an abuse would be found.36 In other words, if there is room for very small competitors in the market, intervention against rebates is more likely than where

33

Commission Guidance, above n 7, para 38. Discussion Paper, above n 2, para 58. 35 Commission Guidance, above n 7, para 19. 36 The rebate test in para 154 seeks to calculate the effective price for the last X number of units bought before the retroactive rebate is applied, where X is the commercially viable amount. In the example, a 2.5% rebate is applied to all sales once 1m units are sold (price falls from €100 to €97.5), and the commercially viable amount is assumed to be 50,000. The effective price over the last 50,000 units (units 950,001 to 1m) is then €50—the 950,000 units would have been sold at an average price of €100, generating revenue of €95m, while the new revenue is €97.5m, so the last 50,000 units generated extra revenue of €2.5m. This effective price of €50 is then compared with average total cost. This approach means that the smaller the commercially viable amount, the lower the effective price for this amount, and hence the more likely it is that the rebate will be considered capable of foreclosure. In terms of this example, if the commercially viable amount were smaller, say 25,000, the effective price for this amount would be €0, and hence likely to be below average total cost. 34

Reform of Article 82 EC—Can Law and Economics be Reconciled? 11 entry requires a large scale. This is inappropriate since a market in which small competitors can flourish is typically more difficult to monopolise. While the numerical example has been removed from the Guidance, similar concerns arise in relation to the qualitative description of the test to be applied when assessing rebates. The test will be applied in relation to a ‘relevant range’ to be determined according to how much demand could be or could have been switched.37 When assessing retrospective discounts, the smaller the relevant range, the greater the effective discount would be per unit of output, and therefore the greater likelihood that the effective price would fall below any particular threshold. This generates the same perverse outcome as the numerical example in the 2005 Discussion Paper. Another area where the Commission’s policy does not appear to stick to the principle that the degree of foreclosure matters is in relation to predatory pricing.38 As noted in section II above, a claim of predation is much more straightforward if the dominant firm’s price cuts apply to all of its products or customers. For selective price cuts (eg targeted discounting or ‘fighting brands’), the direct link between dominance and the likelihood of success of predation breaks down—the dominant firm is not using the full weight of its market power to exclude rivals, so the existence of dominance in itself would not be sufficient to conclude a high likelihood of success of the predatory strategy. Yet the Commission applies the opposite logic, looking with more suspicion on selective price cuts, stating that ‘a general price decrease applied to all the output of a dominant company is in general less likely to be part of a predatory strategy’.39 This approach may be valid, but it would be necessary to examine how the competitive position of the dominant firm’s (non-discounted) activities is strengthened by its selective discounting. We return to the issue of predatory pricing below, when discussing the Advocate-General’s opinion in France Télécom.

C. Only conduct that would exclude ‘as-efficient’ competitors is abusive This principle is already reflected, for example, in the case law on predatory pricing, which states that predation can be presumed if a dominant firm sets prices below average variable cost (the Guidance document emphasises average avoidable cost; both are variations on marginal cost). If prices are above this level, equally efficient firms—those with similar marginal costs— can still compete in the market. Interestingly, the Guidance document uses average avoidable cost as the relevant benchmark for both predation and 37 38 39

Commission Guidance, above n 7, paras 40–41. Ibid, Section IV.C. Ibid, para 118.

12

James Kavanagh, Neil Marshall and Gunnar Niels

loyalty rebates (in the latter, the cost benchmark is compared to the price over the ‘relevant range’ as discussed above), while the 2005 Discussion Paper used average total cost for loyalty rebates. However, the as-efficient competitor test may pose practical problems in markets where an incumbent firm faces competition from new entrants which do not enjoy the same scale advantages, and are therefore not as efficient, or at least not yet as efficient.40 The Guidance document sets out that, when network or learning effects exist, a competitor that is currently less efficient than the dominant firm may be anti-competitively foreclosed by ‘particular price-based conduct’.41 The implication of this statement is that, in these circumstances, as was expressed in the Discussion Paper, even prices above average total cost could be deemed predatory.42 Other competition authorities have also considered entrants’ costs rather than the incumbent’s cost, eg Ofcom’s investigation of a residential telephony service.43 However, this is risky territory from a public policy perspective, since it could lead to the protection of inefficient firms, which is of little benefit to consumers in the longer term.

D. The efficiency defence The Discussion Paper placed the burden of proof for the efficiency defence on the defendant. However, whether the efficiency defence should be a separate part at the end of the effects analysis is questionable from an economic perspective—some efficiencies are simply inherent in the conduct being assessed, and may therefore be overlooked in the separate efficiency defence. The clearest example of this is predatory pricing. At the end of Section 6 of the Discussion Paper, the Commission stated that: An efficiency defence can in general not be applied to predatory pricing. It is highly unlikely that clear efficiencies from predation can be shown and even when they exist that predation is the least restrictive way to achieve them. In addition, it is similarly unlikely that, in the case that such benefits arise, that in the longer run some of these benefits are passed on to the customers and that these benefits outweigh the loss of competition brought about by the predation.44

Nowhere in the 2005 Discussion Paper is it recognised that predation involves low prices for consumers, and that this is the most immediate form of consumer welfare gain one can imagine. 40

See Vickers, above n 1. Commission Guidance, above n 7, para 23. 42 Discussion Paper, above n 2, para 129. 43 Case CW/00760/03/04 Ofcom (2004), ‘Investigation against BT about Potential Anticompetitive Behaviour’, 12 July. 44 Discussion Paper, above n 2, para 133. 41

Reform of Article 82 EC—Can Law and Economics be Reconciled? 13 In the 2008 Guidance document, the Commission has softened its stance somewhat, indicating that, while it considers it unlikely that predation will create efficiencies,45 it will subject any such claims to the multi-part test that it sets out in paragraph 29 of the Guidance. That test echoes Article 81(3) EC and requires that: a) the efficiencies have been or will be realised; b) the conduct is indispensable to achieving the efficiencies; c) the efficiencies outweigh the costs of any negative effect on competition and consumer welfare; and d) the conduct does not eliminate effective competition. Various commentators have expressed the concern that an effects-based approach would create legal uncertainty. With the current form-based approach, so the argument goes, dominant firms know exactly what they can and cannot do. To a great extent this concern arises from a misunderstanding of what an effects-based test would, or should, be about. Such a test does not mean fully quantifying all the pro- and anti-competitive effects of the conduct in each case and then weighing these against each other—even though, admittedly, some economists may have created this impression. Rather, an effects-based test is about asking some further critical questions beyond the current ‘dominance-plus-form’ questions, such as: Does this conduct foreclose a significant part of the market? Or is it likely to exclude as-efficient competitors and distort competition? In other words, an effects-based test is about putting forward a plausible theory of harm to competition and consumers. This theory should be supported by the facts and, where feasible, by some empirical evidence. Such an approach would constitute a significant step beyond the current form-based approach. An effects-based approach does not have to result in legal uncertainty. Indeed, legal certainty could be significantly improved by adopting certain safe harbours in the Article 82 assessment. An example would be a market share threshold below which dominance will not be found—the ‘magic’ 40 per cent threshold could be used as a safe harbour, rather than as a presumption of dominance, although from an economic perspective, a higher threshold (say, 50 per cent) might also make sense. The Guidance document contributes to some extent to the legal uncertainty by avoiding the use of safe harbours. Many criteria that are established in the Guidance are followed by ‘exceptional circumstances’ where those criteria do not hold. Hence, market shares in excess of 50 per cent, held for some time, are ‘very likely’ to indicate dominance, but ‘also undertakings with market shares below 40 per cent could be considered to be in a

45

Commission Guidance, above n 7, para 73.

14

James Kavanagh, Neil Marshall and Gunnar Niels

dominant position’.46 Likewise, as noted above, the Commission’s Guidance on predation sets out that prices below average avoidable costs would be an indication of sacrifice. But a firm would also be deemed to have incurred a sacrifice if its conduct led in the short term to net revenues lower than could have been expected from a reasonable alternative conduct, ie whether the dominant undertaking incurred a loss that it could have avoided.47

This statement introduces uncertainty for any dominant firm. First, a stand-alone assessment of a particular pricing strategy may show it to be profitable. However, if the Commission can identify an alternative strategy that it purports to be more profitable, the former pricing strategy would be deemed loss-making. This compounds the uncertainty that a dominant firm now faces in relation to the precise cost floor it should adopt in assessing the risks of predating. In theory, it could be found to be predating not only when it is pricing above the level of its average avoidable costs, but also if it failed to identify a more profitable pricing strategy than the one it pursued. In the next section, we discuss issues relating to predation in more depth, drawing on the various judgments and opinions expressed since the Commission found France Télécom to have predated illegally on the French broadband market.

IV. THE FRANCE TÉLÉCOM JUDGMENT: HELP OR HINDRANCE TO ARTICLE 82 REFORM?

In January 2007, the CFI upheld a fine of €10.4m imposed by the European Commission for abuse of dominance in France’s high-speed Internet access market.48 France Télécom’s Internet arm, formerly Wanadoo Interactive, was found by the Commission to have engaged in predatory pricing in 2001–02, offering its services below cost in order to eliminate competitors from the market for Internet access.49

46

Ibid, para 14. Ibid, para 6. 48 Case T-304/03 France Télécom SA v Commission of the European Communities, above n 4; Court of First Instance press release, ‘Court Upholds Commission Decision Finding that France Télécom Abused its Dominant Position on the French Market for Internet Access’, 09/07, 30 January 2007. See also Niels, ‘Establishing Predation? France Telecom and Article 82 Reform’, February 2007, available at www.oxera.com/cmsDocuments/Agenda_Feb07/ Establishing%20predation.pdf. 49 European Commission (2003), ‘Commission Decision of 16 July 2003 Relating to a Proceeding under Article 82 of the EC Treaty (COMP/38.233—Wanadoo Interactive)’, 16 July. 47

Reform of Article 82 EC—Can Law and Economics be Reconciled? 15 The CFI decision to uphold the Commission’s ruling is controversial since it reiterates the form-based precedent in AKZO, relying purely on the rule that pricing below average variable cost is always predatory, while pricing between average variable cost and average total cost is always predatory if there is intent to exclude competitors.50 Although the AKZO test is well-established—and, as we noted in section II, is reasonably in line with economic principles—the issue is the exclusive, form-based focus on this rule alone. The CFI’s final decision states that: a)

prices below average variable costs give grounds for assuming that a pricing practice is eliminatory; prices below average total costs but above average variable costs must be regarded as abusive if they are determined as part of a plan for eliminating a competitor; it was not necessary to establish proof that Wanadoo had a realistic chance of recouping its losses; and Wanadoo could not rely on a right to align its prices with those of its competitors to justify its conduct.

b)

c) d)

The key element of the judgment is its reliance on form-based rules of conduct for predation, establishing that there is no need to show that actual harm to competition occurred as a result of below-cost pricing. However, a fundamental change in course may well be on its way now at the Luxembourg courts. Following an appeal against the CFI’s decision, the Advocate-General published an Opinion in September 2008 that, if carried through into the final judgment from the ECJ, would represent a significant development to EU case law on predation. In particular, Advocate-General Mazák’s Opinion states, in referring to the CFI’s reference to Tetra Pak II: I consider that the Court of First Instance’s, and for that matter the Commission’s, interpretation of the Court’s case law is incorrect. In my view that case law requires the possibility of recoupment to be proven … Unless there is a possibility of recoupment, the dominant undertaking is probably engaged in normal competition.51

Although a requirement to prove recoupment is not explicitly in the Commission’s guidelines, a shift towards consideration of recoupment can be inferred from the following wording: 69. Generally speaking, consumers are likely to be harmed if the dominant undertaking can reasonably expect its market power after the predatory conduct comes

50

Case C-62/86 AKZO Chemie BV v Commission [1991] ECR I-3359. Case C-202/07 P, France Télécom SA v Commission, above n 8, Opinion of A-G Mazák, paras 69 and 73. 51

16

James Kavanagh, Neil Marshall and Gunnar Niels

to an end to be greater than it would have been had the firm not engaged in that conduct in the first place, ie if the firm is likely to be in a position to benefit from the sacrifice. 70. This does not mean that the Commission will only intervene if the dominant firm would be likely to be able to increase its prices above the level persisting in the market before the conduct. It is sufficient, for instance, that the conduct would be likely to prevent or delay a decline in prices that would otherwise have occurred. Identifying consumer harm is not a mechanical calculation of profits and losses, and proof of overall profits is not required. Likely consumer harm may be demonstrated by assessing the likely foreclosure effect of the conduct, combined with consideration of other factors, such as entry barriers.

The European Commission’s revised approach appears to bring EU policy closer to that of the US, where recoupment is a key part of the predation test since Brooke Group, referred to in section II above, as also shown by the Department of Justice’s (DoJ) statement in its recent policy document on single-firm conduct: Without a dangerous probability that the investment in below-cost prices will be recouped through later, supracompetitive pricing, below-cost prices most likely reflect nothing more than intense price competition that is in the interest of consumers.52

Rather than discussing the recoupment test as a rigid mathematical exercise, the DoJ refers to it as ‘a valuable screening device to identify implausible predatory-pricing claims’.53 This may suggest that the DoJ sees the place of the recoupment test more as a reality check to weed out claims that have little or no merit, thus avoiding the need for detailed analysis of costs and revenues in such cases. In another recent development, the UK Competition Commission also confirmed that it does not consider EU case law as best practice for its market investigations. It states that for below-cost selling to be a predatory strategy, it is necessary to show that barriers to entry or re-entry into convenience store and specialist grocery retailing are high so that new convenience or specialist grocery stores could not open in response to a weakening of the retail offer by large grocery retailers and prevent recoupment of the losses incurred during the predation stage. (emphasis added)54

This same logic should be applied to predation (and rebates): the current price cuts—which are of immediate benefit to consumers—should be of concern to competition authorities only if there is a high likelihood that they

52 US Department of Justice, ‘Competition and Monopoly: Single-firm Conduct Under Section 2 of the Sherman Act’, September 2008, p ix. 53 Ibid, p 69. 54 Competition Commission, ‘Groceries Market Investigation, Final Report’, 30 April 2008, para 5.57.

Reform of Article 82 EC—Can Law and Economics be Reconciled? 17 will be offset by future price increases once predation has succeeded. This is effectively a recoupment test, as applied in the USA but rejected in the EU— both in existing case law, and in the Discussion Paper.55 Such a test should form an inherent part of any assessment of alleged predatory prices, and the burden of proof should not necessarily fall on the defendant alone. V. CONCLUSIONS

The Commission has embarked on the reform of Article 82, and with its Discussion Paper of December 2005 it paved the way for an effects-based test. The Guidance adopted in December 2008 confirms this change in policy. While many commentators (including these authors) have argued that the reform as set out in the Discussion Paper and the Guidance does not go far enough, it should of course be recognised that the Commission also has to take account of current case law on Article 82, which is still very much form-based (although arguably, this case law is, in part, of the Commission’s own making). Reconciling these two positions is far from straightforward. One problem is that the France Télécom judgment from the CFI, which came out after the Discussion Paper was published, does not follow the Commission’s move towards a more effects-based approach (much as the Commission’s Discussion Paper was criticised for not going far enough in that direction). The same holds for the British Airways judgment by the ECJ, which also came out after the Discussion Paper and upheld a previous (form-based) Commission decision and CFI ruling on loyalty rebates. Yet the European courts may well be the ultimate drivers of change. Of great significance is the Advocate-General’s Opinion in the France Télécom appeal currently before the ECJ, to the effect that recoupment (which itself is more in line with an effects-based approach to abuse of dominance) is indeed an important factor when assessing predatory pricing cases. This view would, if adopted in the final judgment, indicate that, at least in relation to predation, the Commission may have to consider in more detail the likelihood of recoupment as part of its assessment. Lastly, it will also be interesting to see how the Commission intends to deal with the other types of practice addressed by Article 82 but not covered in the Discussion Paper or in the Guidance, in particular excessive pricing, price discrimination and other exploitative abuses. The current emphasis on exclusionary abuses may have helped focus the discussions in the last few years. However, in-depth discussion and reform of the policy on exploitative abuses is also long overdue, and in this regard the Commission’s December 2008 Guidance on ‘enforcement priorities’ in fact has not fully set out what the priorities should be. 55

Discussion Paper, above n 2, para 122.

2 Categorical Thinking in Competition Law and the ‘Effects-based’ Approach in Article 82 EC IOANNIS LIANOS*

I. INTRODUCTION

T

HE DEBATE OVER the soul of Article 82 EC (hereafter ‘Article 82’) has been raging for some time now. Encouraged by the relatively anodyne and quick reform of Article 81 EC, as well as that of EC merger control, the European Commission initiated a process of review of Article 82 with the aim of introducing a more economics-orientated approach. The key results of this process are the European Commission’s Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings (hereafter the ‘Commission Guidance’)1 and the Discussion Paper on the Application of Article 82 of the Treaty to Exclusionary Abuses (hereafter the ‘Discussion

* City Solicitors’ Education Trust Reader in European Union Law; Jevons Institute of Competition Law and Economics, Faculty of Laws, University College London; Emile Noel Fellow, New York University Law School. Many thanks to my colleague Valentine Korah for very helpful comments, as always. Any errors or omissions are those of the author only. 1 DG Competition, Communication from the Commission, Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings (Brussels, 3 December 2008), http://ec.europa.eu/comm/ competition/antitrust/art82/guidance.pdf. Note that the Guidance on enforcement priorities is a softer law instrument than guidelines: it is complementary to the Commission’s specific enforcement decisions. The choice of the instrument of guidance on enforcement priorities offers to the Commission more leeway in presenting its approach for Article 82. The Commission could not have adopted guidelines contrary to the rulings of the European courts (see the most recent reminder by Case C-8/08 T-Mobile Netherlands BV and Others [2009] ECR 00. Opinion of A-G Kokkott, para 29). The Commission maintains the ability to reject a complaint when it considers that a case lacks priority for other reasons (eg lack of Community interest). The Commission Guidance will be published in the Official Journal of the European Communities (OJ).

20

Ioannis Lianos

Paper’).2 In addition, the Economic Advisory Group for Competition Policy (EAGCP) published a report for the Commission, entitled Report on an Economic Approach to Article 82 EC (hereafter the ‘EAGCP Report’).3 An essential component of the call for reform was the opposition to the current ‘form-based approach’ followed by the courts in their interpretation of Article 82. The proponents of the reform of Article 82 advocate an ‘effects-based approach’, which focuses ‘on the presence of anti-competitive effects that harm consumers, and is based on the examination of each specific case, based on sound economics and grounded on facts’.4 This is in opposition to the ‘form-based’ approach, which emphasises instead the nature of the conduct of the dominant firm. This is judged problematic by economists. According to the EAGCP Report: The discussion and management of Article 82 cases are often organised by categories of conduct, such as predatory pricing, discrimination, fidelity rebates or tying. However, such a form-based approach is problematic. In many instances alternative practices can serve the same purpose.5

In other words, the ‘effects-based’ approach does not rely on an evaluation in abstracto of the effect of a specific category of commercial practices but takes an agnostic perspective on the compatibility of a specific commercial practice with Article 82. Thus, when deciding whether the conduct of a dominant firm is compatible with competition law, the starting point should not be to classify the specific practice as a pre-existing behavioural category, but to perform a concrete analysis of the effects of the conduct on consumers. The meaning of ‘effects on consumers’ remains unclear. Effects may indicate empirical, observable findings on the worsening of the situation, in terms of price or quality, of specific groups of consumers, following the adoption of the anticompetitive practice (actual effects). It may also refer to situations where there are no observable findings of effects on consumers but there is ‘a consistent theory of consumer harm’ which is empirically validated, that is the theory of harm should be consistent with factual observations (ex ante validation) and ‘that the market outcomes should be consistent with the predictions of the theory’, probable or likely nonobservable effects (ex post validation).6

2 DG Competition, Discussion Paper on the Application of Article 82 of the Treaty to Exclusionary Abuses (December 2005), http://europa.eu.int/comm/competition/antitrust/ others/discpaper2005.pdf. The Discussion Paper is a consultation document, prepared by the staff of DG Competition. It has not been published in the OJ and therefore does not produce any legal effect. 3 EAGCP, Report on an Economic Approach to Article 82 EC (July 2005), http://europa. eu.int/comm/competition/publications/studies/eagcp_july_21_05.pdf, pp 2, 7. 4 Ibid, p 2. 5 Ibid, p 5. 6 Penelope Papandropoulos, ‘Implementing an effects-based approach under Article 82’ (2008) Concurrences 1–5, at 3.

Categorical Thinking in Competition Law 21 An effects-based approach does not reject the need for categorical thinking. The identification of a theory of consumer harm involves a process of categorisation of the practice as falling within the boundaries of a specific theory, eg raising rivals’ cost, leveraging, predation, maintenance of monopoly strategy, a two-sided market situation. Specific effects will follow from the classification of the practice as falling within a specific antitrust theory. There is an important difference, nonetheless, between the categorical thinking of the ‘form-based’ approach and that of the ‘effects-based’ approach, as this study will attempt to show. The study will briefly discuss the current practice of the classification of abuses under Article 82. It will then examine the shortcomings of the current approach and will stress the need for a reconceptualisation of the current categorisation of abuses in Article 82. It will finally analyse the Commission Guidance on its enforcement priorities in applying Article 82, and consider whether it constitutes a real effects-based approach and its possible implications for competition law enforcement.

II. CATEGORICAL THINKING IN ARTICLE 82: A CRITICAL ASSESSMENT

An ‘effects-based’ approach is compatible with the emergence of specific antitrust categories. A full ‘effects-based’ approach presents important shortcomings, such as decision costs (costs of information gathering and processing): it is impossible to perform a thorough antitrust analysis of all commercial practices. Decision makers employ a sequential informationgathering process in order to reduce information costs.7 The decision to acquire more information is a trade-off between two types of costs: ‘error costs on the one hand’, ie the decision maker may mistakenly identify a pro-competitive practice as being anticompetitive or the opposite; and ‘information costs on the other’.8 Antitrust categories constitute the necessary analytical shortcuts that will ensure that the factual investigation and logical analysis performed are well focused. The existence of information costs, therefore, sets limits to the expansion of a full effects-based approach in Article 82. Categorical thinking may also reflect legal presumptions, that is rules that dispense with any further factual investigation or, if these are absolute presumptions, any further logical analysis. This raises the issue of the compatibility of an effects-based approach with the possibility of per se prohibitions in Article 82.

7 C Frederick Beckner III and Steven C Salop, ‘Decision Theory and Antitrust Rules’ (1999) 67 Antitrust Law Journal 41, at 43. 8 Ibid, at 46.

22

Ioannis Lianos

A. The limits of a full effects-based approach in Article 82 Classification is of particular interest for competition law. For example, the competition provisions of the EC Treaty, as well as those of the Sherman Act in the United States, are based on the dichotomy between collusive and non-collusive practices. The effort of classification is not, however, always clear-cut. One could easily distinguish an agreement between competitors from the refusal of an undertaking to deal with its competitors. However, there are more difficult cases. Although a distribution practice/agreement may involve more than one undertaking, the combination might find its source in the unilateral exercise of the economic or bargaining power of one of the parties.9 Would the specific practice be classified in this case as unilateral conduct? What is the degree of constraint (and exercise of power) that is needed in order to characterise a practice as unilateral conduct? More importantly, what is the purpose of distinguishing unilateral from collusive practices, if the competition authority follows an effects-based approach? Does this distinction make any sense if the analytical steps required by Articles 81 and 82 are similar and, therefore, the classification does not have any instrumental objective?10 Judge Posner’s suggestion to repeal the antitrust laws other than Section 1 of the Sherman Act, or alternatively to repeal all antitrust laws except the provision of Section 2 that forbids monopolising, ‘if by monopolizing we were to mean simply conduct that unjustifiably promotes supra competitive pricing,’ illustrates the futility of the distinction if one adopts a full effects-based approach.11 The classification between unilateral and collusive behaviour may nevertheless serve the purpose of providing a safe harbour for those activities that it would be otherwise costly, in terms of risks of enforcement error, to subject to the scope of competition law. The Colgate exception in US antitrust law,12 or the Bayer rule in EC competition law,13 satisfy this objective. 9 For an analysis of the distinction between unilateral and collusive practices in a vertical context, see Ioannis Lianos, ‘Collusion in vertical relations under Article 81 EC’ (2008) 45 CML Rev 1027–77. 10 Jay M Feinman, ‘The Jurisprudence of Classification’ (1989) 41 Stanford Law Review 661, at 664, observes that ‘Classification presupposes objects that are to be classified, categories into which the objects are to be grouped, and some means of separating the objects into the categories’. There are two objectives to be served by classifying: first, an instrumental one, which ‘looks to the reasons that the categories were formulated in the first place’ and recognises that ‘behind the different categories lie distinct objectives, principles’; secondly, an analytical one, as the aim of classifying is also to create and maintain ‘a rational structure for doctrine’ that would be rigorous enough to fit in different factual contexts. 11 Richard Posner, Antitrust Law, 2nd edn (University of Chicago Press, Chicago, 2001), p 260. 12 United States v Colgate & Co, 250 US 300 (1919). 13 Case C-2/01 and C-3/01 P Bundesverband der Arzneimittel-Importeure eV & Commission v Bayer [2004] ECR I-23.

Categorical Thinking in Competition Law 23 Unilateral behaviour by non-dominant firms benefits from a forbearance regime, in other words per se legality, regardless of the effects of the commercial practice on consumers. This gap in Article 82 essentially constitutes a form-based approach. Unilateral practices of non-dominant firms are exempt from antitrust scrutiny, even if they may harm consumers. Merger control in the EC has recently moved to close this enforcement gap by integrating the analysis of unilateral effects, in the absence of dominance by a single firm.14 This has not yet been the case in Article 82, although it would be easier to identify consumer harm with greater precision ex post rather than ex ante. Error costs are more likely in the context of an ex ante control, which relies on prospective analysis, in comparison to an ex post control, which could be based on a finding of anticompetitive effects. This paradox highlights that the main justification for the ‘gap’ in Article 82 is essentially a value judgement: expansion of market power by merger is perceived as less meritorious than expansion by unilateral commercial practices. John Vickers critically observes, with regard to the idea that there should not be a ‘gap’ in Article 82, that: Competition policy would then fail to discriminate between the accumulation of market power by successful product market competition and, on the other hand, the acquisition of market power by merger. It is one thing to win customers over time by serving them well in the product market, quite another to acquire them at a stroke in the stock market … From the point of view of the economic incentives of firms, and on general grounds, that would seem rather odd. The threshold of market power that triggers intervention to maintain competitive incentives by preventing anti-competitive structural changes in markets ought to be lower than that which triggers liability for the breach of competition law prohibitions on firms that have become dominant … Put differently, merger regulation would seem logically to be a closer relative of Article 81 than Article 82. A merger is the ultimate agreement between firms.15

This justification seems closer to form-based thinking than to an economic effects-based approach. One could justify the dichotomy between the obligations that burden a dominant firm and the forbearance regime that applies for the same practice to non-dominant firms by reference to the concept of the special responsibility of dominant firms to preserve the competitive process.16 It 14 Council Regulation (EC) 139/2004 on the control of concentrations between undertakings (the EC Merger Regulation) [2004] OJ L24/1, Art 2; Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between undertakings [2004] OJ C31/3, paras 24–38. 15 John Vickers, ‘How to Reform the EC Merger Test?’, OFT, 8 November 2002, available at http://www.oft.gov.uk/shared_oft/speeches/spe0802.pdf. 16 Case C-322/81 NV Nederlandsche Banden Industrie Michelin v Commission [1983] ECR 3461, para 57; Case T-201/04 Microsoft v Commission [2007] ECR II-3601, para 775; Case T-203/01 Michelin II [2003] ECR II-4071, para 97. See also, Opinion of A-G Kokott, Case C-95/04P British Airways v Commission [2007] ECR I-2331, para 23.

24

Ioannis Lianos

follows from the nature of the obligations imposed by Article 82 of the Treaty that, in specific circumstances, undertakings in a dominant position may be deprived of the right to adopt a course of conduct or take measures which are not in themselves abuses, and which would be unobjectionable if adopted or taken by non-dominant undertakings.17 The case law does not, however, make explicit the theoretical and practical reasons that impose this special responsibility on dominant firms. The likelihood of anticompetitive effects does not seem to be the main reason for this special responsibility. If this were the case, all firms holding market power, not just dominant firms, would bear a special responsibility to preserve the competitive process. The concept of dominant position has recently found an economic alter ego, the concept of substantial market power.18 The emphasis is on the ability of ‘dominant’ undertakings to ignore competitive constraints, rather than on a more structured interpretation of the indicators of market shares and barriers to entry.19 If market power and dominant position are independent concepts, this is not a difference in kind but a difference of degree. It seems difficult in this case to identify a clear dividing line between dominance and market power, or, in other words, between undertakings having a ‘special responsibility’ and undertakings that do not. A different reason may explain this distinction.20 Recent case law has confirmed the continuing importance of the distinction between unilateral and collusive practices. In Van den Bergh Foods, Unilever’s distribution practices were found to infringe both Article 81 and Article 82.21 Article 81 applied to the exclusivity clause that was included in the distribution agreements concerning the use of the freezer cabinets provided by Unilever (then HB Ice Cream) to members of its distribution network. Article 82 applied to the practice of inducing the retailers who did not have a freezer cabinet to enter into freezer-cabinet agreements with HB, subject to a condition of exclusivity. The latter practice could not fall within the scope of Article 81 as it was purely unilateral conduct—an invitation to participate in an agreement, which naturally pre-dates the conclusion of the agreement itself. This interpretation is supported by the explicit reference of the Court to the retailers that did not have their own freezer cabinets or cabinets from other suppliers, ie retailers that had not yet

17

Case T-111/96, ITT Promedia v Commission [1998] ECR II-2937, para 139. Discussion Paper, above n 2, paras 23–24; Guidance, above n 1, para 10. The analogy between dominant position and substantial market power was also employed bv the European Commission in earlier cases, eg Decision 78/670, GB-Inno-BM/Fedetab [1978] OJ L224/29. 19 Guidance, above n 1, para 11. 20 See Ioannis Lianos, ‘Classification of abuses: a “Straight story”?’ forthcoming, UCL CLGE Working Paper, 2009. 21 Case T-65/98 Van den Bergh Foods Ltd v Commission [2003] ECR II-4653, para 118 (Article 81), paras 159–60 (Article 82). 18

Categorical Thinking in Competition Law 25 concluded a distribution agreement with HB.22 Both the CFI and the ECJ rejected Unilever’s argument that the Commission and the CFI had recycled the same arguments for the purposes of applying Articles 81 and 82, and affirmed that these claims involved a ‘separate analysis’.23 What is the purpose of the distinction between unilateral and collusive practices, in terms of logical analysis? One could have argued, prior to the adoption of a legal exception regime, that Articles 81 and 82 establish a different kind of antitrust control: Article 82 institutes an abuse control, while Article 81 prohibits restrictive business practices.24 However, following the entry into force of Regulation 1/2003, there is no difference between the legal implications of a finding of liability under Article 81 or under Article 82. One could also argue that there are different steps in the competition law assessment under Articles 81 and 82. This view, however, ignores the fact that the European courts have progressively developed a similar analytical framework for Articles 81 and 82. This is true even though two analytical steps are only explicitly identified in Article 81—the prohibition principle of Article 81(1) and the exception principle of Article 81(3)—while no such distinction is mentioned in the context of Article 82. The situation is not very different in practice when it comes to the application of Article 82, although there are multiple standards depending on the nature of the specific abuse. The first step of the analysis largely consists in examining the existence of anticompetitive effects, while the second step of the analysis focuses on the analysis of the effects on the market, including possible justifications for these anticompetitive effects.25 The justifications may take different forms: meeting competition defence, efficiency gains or objective justifications.26 The Guidance’s section on objective necessity and efficiencies illustrates the similarities with the ‘efficiency defence’ under Article 81(3) by applying similar cumulative requirements for the 22 Ibid para 159. The infringement of Article 82 ‘takes the form, in this case, of an offer to supply freezer cabinets to the retailers and to maintain the cabinets free of any direct charge to the retailers’. This important element of the Court’s decision is not taken into account by Ekaterina Rousseva, ‘Modernizing by Eradicating: How the Commission’s New Approach to Article 81 EC Dispenses with the Need to Apply Article 82 to Vertical Restraints’ (2005) 42 CML Rev 587, at 636, who concludes that inducement was not a ‘meaningful point of distinction’ between Article 81 and Article 82, and that the reasoning of the Court ‘was a mere recycling of the arguments made under Article 81’. 23 Case C-552/03 (Order of the Court) Unilever Bestfoods Ireland v Commission [2006] ECR I-9091, para 136. 24 Commission (EC), White Paper on the Modernization of the Rules Implementing Articles 85 and 86 of the EC Treaty, COM (99) 101 final, 28 April 1999, para 18. 25 Case C-95/04P British Airways v Commission, above n 16, paras 67 and 69. 26 Ekaterina Rousseva, ‘Abuse of Dominant Position Defenses—Objective Justification and Article 82 EC in the Era of Modernization’ in Giuliano Amato and Claus-Dieter Ehlermann (eds), EC Competition Law—A Critical Assessment (Oxford, Hart Publishing, 2007); Albertina Albors-Llorens, ‘The Role of Objective Justifications and Efficiencies in the Application of Article 82 EC’ (2007) 44 CML Rev 1727.

26

Ioannis Lianos

acceptance of the alleged efficiency gains in Article 82 to the conditions the Treaty requires in the context of Article 81(3).27 Another important difference could be the allocation of the burden of proof between claimants and defendants. Article 2 of Regulation 1/2003 provides that In any national or Community proceedings for the application of Articles 81 and 82 of the Treaty, the burden of proving an infringement of Article 81(1) or of Article 82 of the Treaty shall rest on the party or the authority alleging the infringement. The undertaking or association of undertakings claiming the benefit of Article 81(3) of the Treaty shall bear the burden of proving that the conditions of that paragraph are fulfilled.

The article refers only to Article 81(3) when it mentions the exception to the rule that it is for the claimant to prove the existence of a competition law infringement. The absence of a provision equivalent to Article 81(3) in the context of Article 82 could be interpreted as indicating that it is for the claimant to prove that the conduct produces anticompetitive effects and that it cannot be justified by efficiency gains. This does not, however, take into account Recital 5 of Regulation 1/2003, which shifts the burden of proof to the defendant not only in the context of Article 81(3), but also, more generally, in all circumstances where justifications are advanced: It should be for the undertaking or association of undertakings invoking the benefit of a defence against a finding of an infringement to demonstrate to the required legal standard that the conditions for applying such defence are satisfied.

It also does not take into account decision-theoretic concerns, such as the fact that it is most cost-effective, in terms of enforcement, to ask the dominant undertakings to substantiate possible efficiency gains/objective justifications, as: a) they possess this information already; b) they have the incentive to provide it in order to escape liability; and c) it is less costly for them to produce (in terms of search costs and economies of scale if there are multiple claimants and a number of alleged infringements). 27 The Commission Guidance employs a slightly different formulation of the third requirement by not explicitly requiring that ‘a fair share of the resulting benefit’ be passed on to consumers. However, this requirement was interpreted restrictively by the Guidelines on Article 81(3), below n 32, para 85 (the net effect of the agreement must at least be neutral from the point of view of those consumers directly or likely to be affected by the agreement). The requirement in the Commission Guidance, para 30 that ‘the likely efficiencies brought about by the conduct outweigh any likely negative effects on competition and consumer welfare in the affected markets’ (emphasis added) has the same meaning.

Categorical Thinking in Competition Law 27 Some national courts distinguish between the legal burden of proof and the evidentiary burden of proof to make the point that, although there is no equivalent to Article 81(3) in the context of the prohibition of an abuse of dominant position, it is still for the defendant to advance justifications of conduct producing anticompetitive effects.28 The CFI has also recognised in Microsoft that It must be borne in mind, as a preliminary point, that although the burden of proof of the existence of the circumstances that constitute an infringement of Article 82 EC is borne by the Commission, it is for the dominant undertaking concerned, and not for the Commission, before the end of the administrative procedure, to raise any plea of objective justification and to support it with arguments and evidence. It then falls to the Commission, where it proposes to make a finding of an abuse of a dominant position, to show that the arguments and evidence relied on by the undertaking cannot prevail and, accordingly, that the justification put forward cannot be accepted.29

The dominant undertaking bears the evidentiary burden of proof with regard to the existence of objective justifications, while the legal burden of proof of the existence of an abuse falls on the Commission or the claimant. That is, the claimant or the Commission bears the risks in finding the existence of an abuse, although it is for the defendant to produce evidence of efficiencies and objective justifications. It follows that if the defendant produces, ‘with a sufficient degree of probability’, ‘verifiable’ evidence of efficiencies which conforms with the cumulative requirements set out in the Guidance,30 if there is still a doubt following a balance of probabilities test, this doubt should benefit the dominant undertaking. This could establish a slight difference between Articles 81 and 82, as the legal burden of proof for an Article 81(3) defence, ie the risk of a doubt on the existence of efficiency gains, is borne by the defendant. If there is no major difference between the two provisions as to the structure of the competition law assessment, there might be a difference with regard to the standard of proof of anticompetitive effects and/or possible justifications. The first step of the competition assessment under Article 81 establishes the existence of an anticompetitive object or effect. The competition authority or the claimant bears the legal burden of proof. The standard of proof required in the first step of the analysis is relatively low. First, it is possible to prove the existence of an anticompetitive object by referring

28 On the distinction between the legal and the evidential burden of proof, see British Horseracing Board v Office of Fair Trading [2005] CAT 29, para 132. In the context of Article 82, see Renato Nazzini, ‘The Wood Began to Move: An Essay on Consumer Welfare, Evidence and Burden of Proof in Article 82 EC Cases’ (2006) 31 EL Rev 518. 29 Case T-201/04 Microsoft v Commission, above n 16, para 1144. 30 Guidance, above n 1, para 30.

28

Ioannis Lianos

to the nature of the conduct and by characterising it as falling within the ‘object box’.31 Secondly, the existence of potential negative effects on consumers, in the form of higher prices, reduced innovation and variety, and lower output, may also be sufficient to identify a practice as being restrictive of competition.32 However, even in this case, the competition assessment is not fully effects-based. The enforcement of Article 81(1) largely emphasises structural elements, such as high market shares or barriers to entry.33 The burden then shifts to the defendant, except in circumstances where the specific commercial practice is covered by a block exemption regulation. The defendant may claim efficiency gains that would outweigh the anticompetitive effects under Article 81(3). Nevertheless, there are strict requirements for substantiating efficiency gains.34 In conclusion, an essential characteristic of the framework of Article 81 is the marked asymmetry between the standard of proof that is required from the claimant and the standard of proof required by the defendant, to the benefit of the former. The situation is not very different when it comes to the application of Article 82. In practice it would be very difficult for a dominant undertaking to prove the existence of objective justifications, the control and the conditions for such a defence being at least as restrictive as the conditions of Article 81(3).35 In contrast, the standard of proof for consumer harm is particularly low, as there is no need to prove the existence of an actual or a direct consumer detriment.36 The Court noted in Microsoft that consumer choice would be affected if rival products of equal or better quality could not compete on equal terms.37 Nevertheless, in other parts of the decision, the Court interprets this condition as requiring only the preservation of the market access of competitors, without requiring that the claimant produce evidence that competitors that are excluded from the market are producing, or would be likely to produce, better quality products than those of the dominant firm. Consumer choice seems to have been equated in this case to the preservation of competitive rivalry on the marketplace.38 The standard of proof for harm to competition and possible defences in Article 82 is therefore also asymmetrical, although there might be some difference of degree in the asymmetry: it is generally more difficult for efficiency gains 31 This highly evocative (in terms of categorical thinking) term has been introduced by Richard Whish, Competition Law, 6th edn (Oxford, OUP, 2009), p 118. 32 Communication from the Commission, Notice, Guidelines on the application of Article 81(3) of the Treaty [2004] OJ C101/97, para 24 (hereafter ‘Guidelines on Article 81(3)’). 33 For an interesting analysis, see M Lankhorst, ‘Improving Accuracy in Effects-based analysis: An Incentive-oriented approach’, Amsterdam Center for Law and Economics Working Paper No 2007-1, SSRN available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=956330. 34 Guidelines on Article 81(3), above n 32, para 51. 35 Case C-95/04P British Airways, above n 16, para 86. 36 Ibid, para 106. 37 Case T-201/04 Microsoft, above n 16, para 652. 38 Ibid, para 664.

Categorical Thinking in Competition Law 29 and other objective justifications to outweigh the anticompetitive effects in the context of Article 82 than in Article 81(3).39 There are different reasons that may explain this asymmetry. Beckner and Salop note, in a different context, that decision makers may be risk-neutral or risk-averse. Risk-neutral decision makers weigh ‘potential harms equally with potential benefits’; while risk-averse decision makers ‘would even reject conduct with a higher expected value in light of the significant downside risk’, in other words they will give additional weight to potential harms.40 In this context, risk could be understood as uncertainty over the effects of the decision in the present case and in future cases.41 The fact that decision makers give more weight to anticompetitive effects than to efficiency gains may indicate that they are risk-averse to possible errors of false negatives. (As a matter of comparison, the recent report of the US Department of Justice (DoJ) on Single-Firm Conduct adopts the exact opposite standard: ‘… conduct should be unlawful under Section 2 if its anticompetitive effects are shown to be substantially disproportionate to any associated procompetitive effects’.42 The asymmetry between the weight attached to anticompetitive effects and that attached to pro-competitive effects indicates that the US DoJ is risk-averse to false positives following the enforcement of Section 2 of the Sherman Act. A risk-neutral decision maker would have given the same weight to both effects.) The asymmetry is also closely related to the availability and cost of evidence. According to the Commission’s Staff Working Paper on Damages: Competition cases are characterised by a very asymmetric distribution of the available information and the necessary evidence: it is often very difficult for claimants to produce the required evidence, since many of the relevant facts are in the possession of the defendant or of third persons and are often not known to claimants in sufficient detail.43

A lower standard of proof for anticompetitive effects than for efficiency gains may therefore seek to establish some degree of formal equality in the starting positions of the claimant and the defendant, by compensating for the initial information disadvantage of the claimant. The principle applies to both provisions: Articles 81 and 82. 39 See Giorgio Monti, EC Competition Law, 1st edn (Cambridge, CUP, 2007), p 203, observing that ‘no firm has yet managed to defend itself successfully’ under Article 82. 40 Beckner III and Salop, above n 7, fn 27. 41 Although the concepts of risk and uncertainty are not, in theory, similar, risk being considered as measurable uncertainty while uncertainty cannot be measured. See Frank H Knight, Risk, Uncertainty and Profit (Boston, Houghton Mifflin Company, 1921), at 25. In this context the two terms will be used interchangeably. 42 US DoJ, Competition and Monopoly: Single-Firm Conduct under Section 2 of the Sherman Act, 2008, at ix and 45–46, available at http://www.usdoj.gov/atr/public/reports/236681.pdf. 43 Commission, Staff Working Paper on Damages Actions for Breach of the EC Antitrust Rules, SEC (2008) 404, 2.4.2008, available at http://ec.europa.eu/comm/competition/antitrust/ actionsdamages/documents.html.

30

Ioannis Lianos

In conclusion, the distinction between unilateral and collusive practices seems to be based on a formalistic assumption: prosecuting unilateral behaviour of firms with market power that does not reach the level of dominance is more likely to lead to substantial enforcement errors. This assumption does not constitute a full effects-based approach, in the sense that in certain circumstances unilateral behaviour by non-dominant firms, with market power, may harm consumers. This example shows that, even in a full effects-based approach, it would be necessary to maintain some degree of formal reasoning, which implies the introduction of specific antitrust categories.

B. A form-based approach through the back door: per se prohibitions in Article 82? The existence of per se prohibitions under Article 82 has been a contentious issue. The case law of the European courts has been ambiguous. Certain forms of practice, such as loyalty rebates, predatory prices and exclusive dealing obligations imposed by dominant firms, were often analysed as being anticompetitive and therefore as infringing Article 82 by their nature and effect.44 It is true that the case law does not always require the examination of the existence of actual anticompetitive effects, as these follow from the qualification of the conduct as being a loyalty rebate or predatory pricing.45 This classification or characterisation process that precedes the assessment of the anticompetitive effects of the conduct plays an important role. It is also difficult in practice to advance objective justifications or efficiency gains for this type of conduct, although it remains theoretically possible to justify predatory pricing under a meeting competition defence.46 This raises the issue of the existence of a per se prohibition for certain categories of unilateral practices. Advocate-General Colomer cited, in his Opinion in Sot Lélos Kai Sia v GlaxoSmithKline, two reasons that explain why there should not be a per se prohibition in Article 82. First, Article 82 does not have a provision equivalent to Article 81(3). The absence of an exception provision indicates that the analysis of abusive behaviour always requires a ‘dialectical debate’

44 Joined Cases C-468/06 to 478/06, Sot Lelos kai Sia v GlaxoSmithKline [2008] ECR 00. (hereafter ‘GSK’), Opinion of A-G Colomer, para 50; Case T-203/01 Michelin II, above n 16, para 241: ‘for the purposes of applying Article 82 EC, establishing the anti-competitive object and the anti-competitive effect are one and the same thing … If it is shown that the object pursued by the conduct of an undertaking in a dominant position is to limit competition, that conduct will also be liable to have such an effect.’ 45 Michelin II, n 44 above, para 239 (loyalty rebates); Case T-340/03 France Télécom SA v Commission [2007] 4 CMLR 21, paras 195 and 197 (predatory pricing). 46 ‘France Télécom, above n 45, paras 182–87.

Categorical Thinking in Competition Law 31 between the dominant undertaking and the competent authorities, whether national or EC, as well as interested parties.47 According to AdvocateGeneral Colomer, if certain types of practice were to give rise to an irrebuttable presumption of abuse, dominant firms would be denied the right of defence. Article 82(a)–(d) do not contain examples of per se abuses. The examples of abuses that are cited constitute presumptions iuris tantum in order to shift the burden of proof from the claimant to the defendant.48 Further, the institution of per se abuses would be contrary to the need to decide each case on the basis of its economic and legal context. A per se approach would commit the sin of excessive formalism.49 In a revealing but nevertheless ambiguous paragraph, Advocate-General Colomer also noted that although there are two different types of abuses— those that affect consumers (exploitative abuses) and those that affect actual or potential competitors (exclusionary abuses)—there is no hierarchy between these two categories. Dominant firms should be able to defend themselves on the basis of economic results obtained.50 According to the Advocate-General, a balancing test will determine the existence of positive and negative effects to competition for the consumers and other economic operators that are situated ‘in the same relevant market’. One could interpret this paragraph as indicating that the efficiency gains should be passed on to consumers, as it is more sensible to consider that the positive effects are examined only from the perspective of the consumers.51 Consequently, Advocate-General Colomer suggested that the Court declare without ambiguity that Article 82 does not apply before examining the existence of positive effects/defences, even if it is clear that the facts of the case do not leave any doubt as to the anticompetitive intent of the parties.52 This is a particularly strong statement, which denies the existence of a per se prohibition rule in Article 82. There are three types of justification of abuse that dominant firms could advance: a) reasons related to the market in which they operate; b) legitimate defence of their commercial interests; and c) proof of a positive balance to competition.53

47

GSK, above n 44, Opinion of A-G Colomer, para 69. Ibid, para 70. 49 Ibid, para 72. 50 Ibid, para 74. 51 See (in the context of Article 81 EC), Case T-168/01, GlaxoSmithKline Unlimited v Commission [2006] ECR II-2969, para 118 (the protection of the welfare of final consumers constitutes the objective of Article 81(1)); Cases C-501, 513, 515 & 519/06P. Appeal brought on 11 December 2006 by GlaxoSmithKline Services Unlimited (GSK) (judgment pending). 52 GSK, above n 44, Opinion of A-G Colomer, para 76. 53 Ibid, para 79. 48

32

Ioannis Lianos

The first one has already been contemplated in the Discussion Paper, but the Advocate-General widens its scope in order to cover all situations of markets (not only products) presenting specific characteristics.54 The second one is the traditional justification of the right of dominant firms to protect their legitimate commercial interests, and the third one is the positive economic trade-off of the practice (efficiency gains passed on to consumers). Of particular interest is the discussion of the role of the analysis of the anticompetitive intent of the dominant firm, which is, according to the Advocate-General, an aggravating circumstance for the finding of an abuse.55 In particular, he advances the view that there are two possible aspects of intent: first, subjective intent, which may in certain circumstances indicate the existence of an abuse; secondly, objective intent when a dominant firm adopts conduct that tends to restrict competition. In the latter situation the qualification of intent does not depend on subjective elements but on the nature of the conduct of the dominant firm: there is a strong presumption that categories of conduct that grossly ignore the objectives of the Treaty constitute a serious infringement of Article 82 and could be presumed abusive.56 The finding of anticompetitive intent should not, however, lead to a per se prohibition under Article 82.57 Courts and competition authorities should systematically examine the positive and negative effects of the practice, even if there is a clear intent to restrict competition. Advocate-General Colomer distinguishes between different categories of conduct, some of them being so grossly anticompetitive that they may justify a different allocation of the burden of proof (for example, if there is a restriction of parallel imports, the burden of proof may shift to the defendant). Other practices will require from the claimant a more complete substantiation of anticompetitive effects before moving to the next step of the analysis, the examination of possible defences.58 Categorical thinking seems, therefore, compatible with an effects-based approach, as it may introduce a different allocation of the burden of proof for certain categories of practice. The position of Advocate-General Colomer, with regard to the nonexistence of per se prohibitions under Article 82, leaves place for the introduction of an object/effect distinction in the context of Article 82, with anticompetitive object being a different concept than per se prohibition, as it would be possible to justify conduct having an anticompetitive 54 Ibid, paras 80–98; Discussion Paper, above n 2, para 80, which extends the justifications for public health and safety to ‘imperfect markets’, meaning markets regulated by the State (economic regulation). 55 GSK, above n 44, paras 47–54. 56 Ibid, para 54. 57 Ibid, para 61. 58 Ibid, para 70.

Categorical Thinking in Competition Law 33 object with efficiency gains.59 One could therefore interpret the reference by the European Commission in its Guidance to circumstances in which it would not be necessary to carry out a detailed assessment before concluding that the conduct in question is likely to result in consumer harm, as not excluding the theoretical possibility of an efficiency defence.60 It is true that the Commission considers that in this case the anticompetitive effect will be inferred, as it appears that the conduct can only raise obstacles to competition and creates no efficiencies. One could argue, however, that this paragraph does not affect the theoretical possibility for dominant undertakings to invoke efficiencies, in particular as it remains possible for conduct leading to anticompetitive foreclosure to be justified by the objective necessity defence and efficiencies, before finding the existence of an abuse of a dominant position. The Commission does not explicitly exclude these practices from Section D of the report on objective necessity and efficiencies that follow, although, in practice, it seems unlikely that the alleged efficiencies will be accepted, for failure to comply at least with the first two cumulative conditions of the efficiency defence. It is a pity that the decision of the ECJ in GSK avoided any discussion of the general framework of Article 82. The Court preferred to start its analysis by referring to the familiar case law on refusals to deal, which opens the door for the defendants to argue objective justifications and escape the finding of an abuse of a dominant position.61 The Court’s decision seems nevertheless implicitly to recognise that certain types of conduct, such as a restriction of parallel trade, may create a presumption of negative effects on consumers, and therefore shift the burden of proof to the defendant, without it being necessary for the claimant to bring additional evidence as to the causal link between the specific conduct and consumer harm.62 According to the Court: In the light of the abovementioned Treaty objective (avoid national divisions in the Internal Market trade) as well as that of ensuring that competition in the internal market is not distorted, there can be no escape from the prohibition laid

59 This seems to be contrary to the interpretation of the anticompetitive object concept (in the context of Article 81(1)) by A-G Kokkott in T-Mobile Netherlands BV and Others, above n 1, para 43, who seems to consider that anticompetitive object and per se prohibition refer to the same thing. This is plainly wrong as a matter of law, as, in a legal exception regime, conduct which is anticompetitive by object may still be justified under Article 81(3) and therefore not prohibited per se. In other words, there could be no prohibition under Article 81 before the antitrust decision maker has examined the possible application of Article 81(3), if the parties argue justifications under Article 81(3). 60 Guidance, above n 1, para 22. The Commission provides some non-exhaustive examples, such as conduct through which the dominant undertaking prevents its customers from testing the conduct of competitors, or pays a distributor or a customer to delay the introduction of a competitor’s product. 61 GSK, above n 44, paras 34, 69 and 76. 62 Ibid, paras 56–57.

34

Ioannis Lianos

down in Article 82 EC for the practices of an undertaking in a dominant position which are aimed at avoiding all parallel exports from a Member State to other Member States, practices which, by partitioning the national markets, neutralise the benefits of effective competition in terms of the supply and the prices that those exports would obtain for final consumers in the other Member States.63 (emphasis added)

Despite the language employed, however, the Court immediately qualifies this statement by accepting that this does not amount to an absolute presumption of consumer harm or a per se prohibition rule. The presumption of anticompetitive effects may still be rebutted by the defendant in limited circumstances: a company must be ‘in a position to take steps that are reasonable and in proportion to the need to protect its own commercial interests’.64 This is a more limited array of objective justifications than those contemplated by Advocate-General Colomer in his Opinion. The full array of objective justifications may not apply in this case (only reasonable and proportionate protection of commercial interests).65 The Court introduces a degree of categorical analysis in Article 82, as it implies that some restrictions of competition are so severe that only a limited range of objective justifications may enter into consideration. Furthermore, the Court immediately qualifies the possibility of dominant firms justifying restrictions on parallel trade by indicating that it is limited to circumstances where: a) State intervention is one of the factors liable to create the opportunities for parallel trade in the first place;66 and b) a different interpretation of Article 82, rejecting any possibility of justification, would have left a dominant firm only the choice ‘not to place its medicines on the market at all in a Member State where the prices of those products are set at a relatively low level’.67 The above passage indicates that a possible interpretation of the Court’s position would be that dominant firms may justify restrictions on parallel 63

Ibid, para 66. Ibid, para 69. 65 The content of this concept is unclear. A restrictive definition of the concept will cover only a meeting competition defence, thus excluding the consideration of efficiency gains. See Ekaterina Rousseva, ‘The Concept of “Objective Justification” of an Abuse of a Dominant Position: Can it Help to Modernise the Analysis under Article 82?’ (2006) 2(2) Competition Law Review 27, at 33–34, who observes the following: ‘… the question is what a legitimate commercial interest of a dominant undertaking is: does this interest mean only a right of a dominant undertaking to survive on the market, ie, to prevent its inefficient operation, or does it also mean a right to carry out a profit oriented policy? How does the prerogative of a dominant firm to protect its interest fit with the essential goal of competition to serve consumers’ interests? These are questions on which views diverge.’ 66 GSK, above n 44, para 67. 67 Ibid, para 68. 64

Categorical Thinking in Competition Law 35 trade falling under Article 82 only in a Distillers type situation68: following the prohibition of the differential pricing system by the Commission, Distillers withdrew Johnnie Walker Red label from the UK market.69 This introduces a degree of categorical analysis in Article 82, as it implies that some restrictions of competition are so severe that only a limited range of objective justifications may enter into consideration, and only in specific circumstances. What the previous analysis has shown is that classification is inherent in the legal process and constitutes an essential feature of the analytical framework of Article 82. The calls for an ‘effects-based’ economic approach do not question the need to classify or create categories of abuses. Do they imply, however, the adoption of a unified analytical framework for Article 82 that could apply to all commercial practices? At an abstract level, this may be the case: it is possible clearly to identify a unified analytical framework that could apply to all forms of unilateral practices. As is well detailed in the EAGCP Report, ‘[a]n economics-based approach will naturally lend itself to a rule of reason approach to competition policy’.70 At a more practical level, though, an effects-based approach will not preclude classification. First, the economic approach and terminology are also conducive to categorisation. For example, the EACGP Report distinguishes between three different categories of abuse, in terms of an effects-based approach, and suggests different standards for each of these categories: it seems useful to distinguish three broad typologies of exclusion that differ in respect to the market position of the firms involved and in respect to the specific features that characterize the exclusionary effects: Exclusion within the same market, where an incumbent forces the exit or prevents the entry of a competitor, exclusion in an adjacent market where the dominant firm excludes producers active in markets different but related to its main market, and exclusion in a vertically related market, where exclusion takes places in different stages of the production process.71

The EAGCP essentially suggests a predatory standard for the first type of practice, an anticompetitive foreclosure standard for the second type of practice, and a cautious standard of antitrust intervention for the third type of practice. The intervention of competition authorities on the basis of 68 European Commission, ‘The Distillers Co Ltd/ (Conditions of Sales and Price Terms)’ [1978] OJ L50/16. See the criticisms of A-G Warner in Case 30/78, Distillers Co v Commission [1980] ECR 2229; Valentine Korah, ‘Goodbye Red Label: Condemnation of dual pricing by Distillers’ (1979) EL Rev 1. 69 Similar issues of differential pricing under Article 81 are pending at the ECJ: Pending Case C-501/06P, appeal of Case T-168/01 GlaxoSmithKline Services v Commission [2006] ECR II-2969. 70 EAGCP Report, above n 3, p 3. 71 Ibid, p 17.

36

Ioannis Lianos

foreclosure may have an adverse impact in the long run on the dominant firm’s incentives to invest or innovate. The Report, however, does not fully address the underlying principles of this new classification, and makes no effort to link it to the objectives of Article 82. One could also cite the example of the vertical/horizontal dichotomy in the context of Article 81, as an illustration of categorical thinking in an effects-based approach. Horizontal restraints are considered more harmful for consumers than vertical restraints, because of the self-restricting effect of the relationship between supplier and retailer.72 The effects-based approach does not preclude classification. Categorisation takes place according to a new criterion: that of likely anticompetitive effects, perceived in this example as being more likely to occur for horizontal agreements than for vertical ones. Categories are designed differently and are subject to revision, according to empirical findings and the evolution of economic theory. This constitutes probably the most important difference between a form-based and an effects-based approach. Secondly, it is possible to make a normative argument in favour of classification in an effects-based approach. One could argue that the existence of categories will preserve the principle of legal certainty from the flexibility and consequential vagueness of a fully-fledged effectsbased approach. Conceptual uncertainty is indeed the most important shortcoming of open-ended regulatory standards.73 The principle of legal certainty requires the non-equivocal formulation of norms. According to the ECJ, which elevated this principle to a general principle of EC law, ‘Community legislation must be certain and its application foreseeable by those subject to it’.74 Foreseeability is particularly important in the context of a legal exception regime where firms are required to proceed on a self-assessment of their commercial practices; but foreseeability does not mean predictability. The principle of legal certainty is compatible with the flexibility of an economic effects-based approach. Legal certainty could be preserved in these circumstances by different means. Jürgen Habermas has notoriously given a procedural content to the principle of legal certainty, which could apply in this case.75 The principle should not be conceived as certainty about results but as a claim for an equitable procedure which will make possible the discursive elucidation of the legal and factual issues that arise. 72 Communication from the Commission on the application of the Community competition rules to vertical restraints—Follow-up to the Green Paper on vertical restraints, COM(98) 544 final, [1998] OJ C365/3. 73 Colin Diver, ‘The Optimal Precision of Administrative Rules’ (1984) 93 Yale Law Journal 65. 74 Case C-325/85 Commission v Ireland [1987] ECR 5041, para 18. 75 Jurgen Habermas, Between Facts and Norms: Contribution to a Discourse Theory of Law and Democracy (Cambridge, MIT, 1996).

Categorical Thinking in Competition Law 37 An effects-based economic approach does not put the principle of legal certainty into question, if the economic analytical framework is set out in such way as to enable the dominant undertakings to assess clearly the decision-making process. Competition authorities may also publish detailed guidelines that could provide information on the different steps of the analysis and the criteria of the trade-off. Thirdly, classification may limit enforcement costs, through an optimal allocation of the burden of proof and an effective design of the standard of proof. It would be a waste of resources to conduct an in-depth analysis of the anticompetitive effects of a practice which, with very few exceptions, produces important anticompetitive effects. This would be particularly true if it were possible for the dominant firm to advance justifications, albeit limited, for its conduct at a later stage of the competition assessment. III. SHORTCOMINGS OF THE CURRENT CLASSIFICATION APPROACH

The European courts have gradually developed specific competition law standards for different unilateral practices: predatory pricing, duty to deal, duty to supply an existing customer, duty to license, tying, excessive prices, above-cost pricing, margin squeeze, selective price cutting and exclusive dealing, to mention but a few.76 Different standards apply for each category of practice, some more lenient, others more restrictive, with the result that claimants and defendants may behave strategically and may attempt to blur the distinctions between the different forms of abuse. The case law on bundling may illustrate the difficulties of classification, as this conduct may be characterised as a rebate or as a tying practice. The more lenient standards for refusals to deal, in comparison with other types of abuses, could also raise an issue of conceptual coherence, as well as highlight the risk of strategic litigation behaviour from the defendants or the claimants.

A. The difficulties of the classification exercise: the case of bundling Bundling may take different forms. The Discussion Paper distinguished three: contractual, technical, and financial (what the Discussion Paper called ‘commercial tying’).77 The competition authorities and courts have developed specific antitrust standards for each category.

76 For a description, see Ariel Ezrachi, EC Competition Law (Oxford, Hart Publishing, 2008), 119–26. 77 Discussion Paper, above n 2, para 182.

38

Ioannis Lianos

For the prohibition of tying under Article 82, the presence of the following elements is usually required: a) b) c) d)

the company concerned is dominant in the tying market; the tying and tied goods are two distinct products; the tying practice is likely to have a market distorting foreclosure effect; the tying practice is not justified objectively or by efficiencies.78

The situation is more complicated with regard to mixed bundling. The case law on bundled discounts is a good example of the difficulties of classification of abuse in the context of Article 82. There is an important divergence in the case law used for this form of bundling. The European courts either apply an anticompetitive foreclosure test to these practices, similar to that employed in recent exclusive dealing cases,79 or they acknowledge, as the CFI has done recently in Microsoft, that bundled discounts may, in some circumstances, have a effect equivalent to tying.80 The Discussion Paper acknowledged that bundled discounts (also referred to as ‘commercial tying’) may have effects on competition similar to tying, and that the distinction between mixed bundling and pure bundling is not ‘necessarily clear-cut’, as mixed bundling may come close to pure bundling when the prices charged for the individual offerings are much higher than the standalone price of each.81 But in other parts of the Discussion Paper, the DG Competition staff remarked that there is a difference between these two practices in the sense that in mixed bundling none of the products is ‘tied in the traditional sense’.82 It was observed in the Discussion Paper that both practices may have similar foreclosure effects: mixed bundling constitutes an indirect measure to achieve the same result as contractual tying, ‘by inducing customers to purchase the tied product through granting bonuses, rebates, discounts or any other commercial advantage’.83 It seems that, for the Commission’s staff, coercion and inducement may produce the same effects, and should therefore be analysed under the same standards. The Commission Guidance on its enforcement priorities includes mixed bundling in the category of tying. However, following the previous proposals of the Discussion Paper,84 the Commission advances a different 78 Case T-201/04 Microsoft, above n 16, paras 842, 859–62, 867 and 869. Commission’s Guidance, above n 1, para 49. 79 Case T-65/98 Van den Bergh Foods, above n 21, paras 157–60; Case C-552/03 Unilever, above n 23. 80 Microsoft, above n 16, para 908. Although the Court does not indicate if it will apply the same standard as for tying, the language used indicates that the CFI embraces the tying analogy. 81 Discussion Paper, above n 2, fn 112. 82 Ibid, fn 113. 83 Ibid, para 182. 84 Ibid, para 189.

Categorical Thinking in Competition Law 39 test for assessing the anticompetitive effects of multi-product rebates (mixed bundling) than other forms of tying. This test is inspired by the sharp dichotomy introduced by the Commission Guidance between price and non-price exclusionary conduct. Only the exclusion of competitors as efficient as the dominant firm will trigger antitrust intervention in the case of price-based exclusionary conduct.85 Certainly, 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 a particular pricebased conduct leads to anticompetitive foreclosure.86

However, this possibility is given less weight than was previously the case in the Discussion Paper.87 Consequently, the Commission suggests cost-price tests that would determine, as the first step of the assessment of the price exclusionary conduct, whether even a hypothetical competitor as efficient as the dominant undertaking would likely be foreclosed by the conduct in question.88 In the case of mixed bundling, the Commission introduces a safe harbour, in terms of enforcement of Article 82. This would cover bundled rebates, when the incremental price that customers pay for each of the dominant undertaking’s products in the bundle remains above the long-run average incremental cost (LRAIC) to the dominant firm of including the product in the bundle. The Commission will not normally intervene, ‘since an equally efficient competitor with only one product should in principle be able to compete profitably against the bundle’.89 If the incremental price that customers pay is below the LRAIC, the Commission will proceed to the assessment of the anticompetitive effect of the practice (under the anticompetitive foreclosure standard).90 However, the Commission adopts the predatory standard for bundled rebates—if the dominant undertakings’ competitors are selling identical bundles, or could do so in a timely way without being deterred by possible additional costs.91

85

Guidance, above n 1, para 26. Ibid, para 23. Note that this possibility is not taken into account at the first step of the competition assessment, the cost-based test (which operates basically as a filter for the consideration of the efficiency of the excluded competitor, but only during the second step of the analysis, the existence of an anticompetitive foreclosure. This indicates that for these cases, the two steps will merge into one, and consequently the fact that the excluded competitor is less efficient than the dominant firm will influence the analysis of the existence of an anticompetitive foreclosure, leading therefore in practice to more stingent conditions for the finding of consumer harm. 87 Discussion Paper, above n 2, para 67, which included it as a possible exception to the operation of the cost-based test. It would have not therefore affected the analysis under the second step of anticompetitive foreclosure. 88 Guidance, above n 1, para 24. 89 Ibid, para 59. 90 Ibid, para 59. 91 Ibid para 60 and fn 38. 86

40

Ioannis Lianos

A difficulty with the test suggested is that it does not make clear how mixed bundling is different from technical and contractual tying. It seems that the equally efficient competitor test does not apply to tying, while it would apply to mixed bundling. This raises the issue of the existence of an effective and practical tool to distinguish between the two practices. It would make no sense to have a cost-based test for mixed bundling if the same practice could also be analysed under the anticompetitive foreclosure approach of tying. This implies that a process of characterisation of the specific facts of the case as constituting tying or mixed bundling should precede the step of examining the anticompetitive effects of the practice under either the anticompetitive foreclosure or the predation-like test. There are difficulties, however, with this operation. For example, there is a slight conceptual line that may separate the characterisation of the facts of the Windows Media Player (WMP) Microsoft case as being a bundled discount rather than a tying case. The WMP was offered for free, which may formally correspond to a bundled discount, a practice that entails the offering by the supplier to the distributor of a discount (100 per cent discount in this case) for accepting a bundle of different products or services. The fact that the courts analysed the facts of the case as tying should not conceal the importance of developing a coherent conceptual framework for all types of bundling. One could also argue that there is a thin conceptual line between mixed bundling and single-product rebates, or that their competition law assessment should be similar. Nick Economides observes that single product rebates and bundled rebates may be conceptualised as referring to a similar leveraging concern.92 In a single-product case, demand is divided between an uncontested part that is always purchased from the dominant firm (for example, because of reputation, fear of punishment for the executives of the firm if they do not buy from the dominant firm, limitations in the production capacity of the competitor) and a contested part of the demand where the customer may buy from any firm. The main difference is that in a multi-product case, sales in market A are leveraged to obtain higher sales in market B, while in the single-product case, the uncontested sales in market A are leveraged to obtain the contested sales also in market A. In conclusion, when deciding to adopt different antitrust tests for each practice, one should take into account the effects of a possible recharacterisation of a practice as falling within the scope of another antitrust category. Otherwise the effectiveness of antitrust enforcement could be jeopardised.

92 Nick Economides, ‘Loyalty/Requirement Rebates and the AMC: What is the Appropriate Liability Standard?’, forthcoming Antitrust Bulletin (2009).

Categorical Thinking in Competition Law 41 B. The risks of strategic behaviour and the conceptual incoherence of the current classification of abuses Classifying a practice as a specific type of abuse may also affect the outcome of the case. The case law on tying and loyalty rebates seems to conflict with the more flexible position adopted by the ECJ in Oscar Bronner on refusals to supply.93 The Court took a restrictive view of when there will be an obligation on a dominant undertaking to grant access to its facilities by imposing a number of conditions: a)

refusal ‘must be likely to eliminate all competition’ on the part of the competitor requesting access; b) access should be indispensable, not merely make it harder for the requesting undertaking to compete; and c) refusal should not be capable of objective justification. With respect to the indispensability condition (in b) above), the ECJ held that access would have been indispensable only if it was not economically viable to create a home-delivery system for a newspaper with a circulation comparable to that of the dominant firm.94 One could argue that the conditions in Bronner set the outer boundaries of the special responsibility of a dominant firm, and consequently the corresponding duty, under Article 82, to abstain from any action that would be likely to exclude rivals from the market. The excluded rival would be granted access only if it was impossible for an undertaking with an output comparable to that of the dominant firm to develop such facility, which indicates that the Court applied a ‘not yet as efficient as’ test. Drawing on this case law, Jean-Yves Art and Gregory McCurdy, Microsoft’s legal advisers, argued that the Microsoft bundling case could be re-interpreted as a refusal to supply case, and that it should be examined in conformity with the limiting principles of Bronner.95 They claimed that the main issue involved in the bundling part of the Microsoft case is that by offering to Original Equipment Manufacturers (OEMs) a bundled version of Windows and WMP, Microsoft had effectively denied to its competitors in the media player market access to the appropriate ‘distribution services’ of the OEMs (the OEMs would operate in this case as intermediaries between the market players on the tied product market and the final consumers). Access to the OEMs was, however, necessary to

93

Case C-7/97, Oscar Bronner GmbH & Co KG v Mediaprint [1998] ECR I-7791, para 41. Ibid, paras 45–46. Jean-Yves Art and Gregory McCurdy, ‘The European Commission’s media player remedy in its Microsoft decision: compulsory code removal despite the absence of tying or foreclosure’ (2004) 25(11) European Competition Law Review 694, 703–07. 94 95

42

Ioannis Lianos

maintain competition in the media player market. Their argument was that the application of the Bronner principles in the factual situation of Microsoft would not have led to an infringement of Article 82, as distribution of the player in Windows was not indispensable and Internet downloading was a feasible and only ‘less advantageous’ distribution channel. If we follow Art and McCurdy’s arguments, Microsoft’s refusal to supply access to the ‘distribution services’ of the OEMs to its rivals could not be considered as the legitimate exercise of a property right, for the simple reason that the OEMs were not owned by Microsoft. Another example constitutes the emergence of the margin squeeze abuse in EC competition law, an illustration of antitrust creativity in setting new categories of abuse.96 In Industrie des poudres sphériques, the CFI recognised that an abuse by margin squeeze could take place if there was a imbalance between an upstream and a downstream price, without any need to demonstrate either that the wholesale price was excessive in itself or that the retail price was predatory in itself.97 The approach followed in EC competition law may be contrasted with the position of some US courts and commentators that margin squeeze should not constitute an independent doctrine of antitrust liability.98 The existence of Article 82(b) and the fact that ‘unfair’ prices could be found abusive under EC competition law may be an explanation for this divergence of approach. The case law of the European courts has not, however, been clear until now on the relationship between the doctrine of margin squeeze and that of refusal to deal. This question may not arise in situations where the dominant undertaking has a regulatory duty to deal, but may be an important issue in other cases. It is possible to argue that the claimant in a margin squeeze case should provide evidence that access to the upstream input was indispensable for rivals and, more generally, for effective competition in the downstream level.99 After

96 See European Commission Decision 88/518/EEC, Napier Brown/British Sugar [1988] OJ L284/41. 97 Case T-5/97 Industrie des poudres sphériques SA v Commission [2000] ECR II-3755, para 179; see also Case T-271/03 Deutsche Telekom AG v Commission [2008] ECR 477, para 167; Case COMP/38.784-Wanadoo España v Telefónica [2007], available at http://ec.europa. eu/comm/competition/antitrust/cases/decisions/38784/dec_en.pdf, para 283. 98 See, Gregory Sidak, ‘Abolishing the Price Squeeze as a Theory of Antitrust Liability’ (2008) 4 Journal of Competition Law and Economics 279–309; Dennis W Carlton, ‘Should “Price Squeeze” be a Recognized Form of Anticompetitive Conduct?’ (2008) 4 Journal of Competition Law and Economics 271–78; Covad Communications Company v Bell Atlantic Corp, 398 F 3d 666 (DC Cir 2005) (US). See, however, Linkline Communications, Inc v SBC California, Inc, 503 F 3d 876 (9th Cir 2007) (US), accepting antitrust liability for margin squeeze. 99 See, Robert O’ Donoghue, ‘Regulating the Regulated: Deutsche Telekom v European Commission’, Global Competition Policy Magazine, May 2008, at 19, referring to Case T-271/03, Deutsche Telekom, above n 97, para 237, where the CFI observes that ‘[h]aving regard to the fact that the applicant’s wholesale services are thus indispensable to enabling a competitor to enter into competition with the applicant on the downstream market in retail

Categorical Thinking in Competition Law 43 all, a firm that has the ability to refuse to deal should also be able to decide to deal on unattractive terms. Other authors advance the view, nevertheless, that margin squeeze case law is less concerned with the ‘indispensable’ nature of the relevant upstream input than with the existence of alternative upstream supplies at competitors’ disposal that could allow them to compete.100

The recent decisions of the Commission are ambiguous. In Telefónica, the Commission held that in the light of the specific factual, economic and legal context of the case, in particular the fact that wholesale access at regional level is mandated since March 1999 and wholesale access at national level is mandated since April 2002 and the fact that the former monopoly’s ex ante incentives to invest in its infrastructure are not at stake in the present case, the legal test applied by the European Court of Justice in Oscar Bronner is not applicable in the present case.101

What would have been the case, however, if access was not mandated, or if the dominant firm did not benefit from the investments of a former legal monopoly and therefore its incentives to invest would have been jeopardised by antitrust liability? Would the Oscar Bronner conditions have applied in this case? It is interesting to note that the Commission Guidance on its enforcement priorities in applying Article 82 establishes a parallel between the margin squeeze doctrine and that of refusals to deal: both are treated under the same chapter and as essentially practices that are substitutable for dominant firms.102 The Guidance does not provide a detailed explanation, but it seems that the Bronner conditions, in particular the fact that the claimant should prove that the refusal to deal or margin squeeze relates to a product or service that is ‘objectively necessary’ for the non-dominant firm to be able to compete effectively on a downstream market, also apply to margin squeeze cases.103

access services, a margin squeeze between the applicant’s wholesale and retail charges will in principle hinder the growth of competition in the downstream markets’ (emphasis added). 100 Simon Genevaz, ‘Margin Squeeze after Deutsche Telekom’, Global Competition Policy Magazine, May 2008. 101 Wanadoo España v Telefónica, above n 95, para 309. 102 Guidance, above n 1, part D. 103 The decision of the majority of the US Supreme Court in Pacific Bell Telephone Company v Linkline Communications, Inc, 555 US (2009), is of particular interest here, as the Supreme Court held that no price-squeeze claim could be brought under §2 of the Sherman Act in the absence of an antitrust duty to deal. The Supreme Court considered that a margin squeeze claim is essentially a composite antitrust liability claim, comprising a refusal to deal claim and a predatory pricing claim. It follows that it is only if the conditions of a refusal to deal claim are present that a plaintiff could bring a margin squeeze case.

44

Ioannis Lianos IV. CATEGORISATION AND THE EMERGING EFFECTS-BASED APPROACH TO ARTICLE 82

A. The exploitative versus exclusionary abuses dichotomy: still relevant? The Commission Guidance on its enforcement priorities is not advancing a full effects-based approach in the enforcement of Article 82. The Commission observes that its enforcement priorities will focus on those ‘types of conduct that are most harmful to consumers’, and it excludes from the scope of this Guidance ‘conduct which is directly exploitative of consumers’, thus maintaining the traditional distinction between exploitative and exclusionary abuses under Article 82. One could adopt two different interpretations of the relationship between these two forms of abuse. There is the view that each abuse forms an autonomous concept. If competitors are excluded by the dominant firm, this is enough to infer, in specific circumstances, consumer harm. It is not necessary to prove that exclusionary abuses will lead to exploitative effects. Modern claims of anticompetitive effect are, however, always supplemented with an ‘explanatory link’ between an exclusionary and an exploitative element: the assumption is that the exclusionary abuse will most likely lead to exploitative effects.104 This does not, though, go as far as requiring empirical evidence of anticompetitive effects. It may instead take the form of a theory of consumer harm, which aims to establish a relationship of causality between the specific practice and potential consumer detriment. One could think in terms of a probability statement, that is, an evaluation of the ‘inferential soundness’ of this relationship.105 The explicit introduction of a consumer harm element for exclusionary abuses further weakens the importance of the distinction between exclusionary and exploitative abuses, the only difference being that for exploitative abuses there is already evidence of an actual consumer detriment, which could, for example, be ‘excessive’ pricing (as compared to a competitive level). The Commission Guidance is nevertheless more effects-based than the case law of the European courts, or even the Discussion Paper. The Commission does not proceed to an exhaustive discussion of the standards/ tests that apply to all different categories of single firm conduct, as these were developed by the case law, but advocates a general approach to exclusionary conduct, before providing more details on the assessment of certain practices, as an illustration of the principles that will apply. There has also been some effort made to group together certain practices that 104 Peter Roth, Christopher Bellamy, Vivien Rose, Graham D Child, Bellamy & Child: European Community Law of Competition (Oxford, OUP, 2008), at 948. 105 Jonathan Cohen, The Probable and the Provable (Oxford, Clarendon Press, 1977), at 27.

Categorical Thinking in Competition Law 45 produce similar effects. For example, under the rubric of ‘exclusive dealing’ the Commission regroups exclusive purchasing practices and conditional rebates; under that of ‘tying and bundling’, all different forms of bundling; refusals to supply and margin squeeze are also examined together. This is reminiscent of the new categorisation of vertical restraints performed in the context of the publication of the vertical restraints guidelines: the Commission grouped all vertical restraints into four categories designed in accordance with their negative effects on competition.106 That was one of the elements that marked the transition from a form-based approach to an effects-based one in the context of Article 81.

B. The price/non-price dichotomy: emergence and criticism The most important new classification introduced by the Commission Guidance is the distinction between price- and non price-related exclusionary abuses. These two categories permeate all others, in the sense that the legal standards are profoundly different depending on the characterisation of a restriction as being related to price or non-price exclusionary conduct. If the efficiency of the undertaking is a major concern for price restraints, it remains a secondary one for non-price restraints, for which even the exclusion of a less efficient competitor could be a matter of concern.107 The adoption of cost-based tests as filters for antitrust enforcement in the area of price-based exclusionary conduct is a direct consequence of this classification.108 The price–cost test excludes from the scope of Article 82 conduct that is exclusionary, and may harm consumers (by raising prices or affecting quality and more broadly consumer choice) for the simple reason that the excluded company is a less efficient firm. The 2005 Discussion Paper also suggested an additional classification according to the horizontal and vertical foreclosure effect of the conduct.109 However, this has not been followed in the Commission Guidance. The distinction between price and non-price abuses introduces a conceptual anomaly in the framework of Article 82. The Discussion Paper recognised that price and non-price abuses may lead to ‘similar foreclosure effects’ and that they are substitutable, from the point of view of dominant

106 Commission Notice, Guidelines on vertical restraints [2000] OJ C291/1, para 104: ‘Agreements which are different in form may have the same substantive impact on competition. To analyse these possible negative effects, it is appropriate to divide vertical restraints into four groups: a single branding group, a limited distribution group, a resale price maintenance group and a market partitioning group. The vertical restraints within each group have largely similar negative effects on competition.’ 107 Guidance, above n 1, para 22, interpretation a contrario. 108 Ibid, para 24. 109 Discussion Paper, above n 2, para 73.

46

Ioannis Lianos

firms, although it also advanced a different theoretical framework for assessing pricing behaviour from non price-based exclusionary conduct.110 The Commission Guidance provides a rather short explanation of this fundamental dichotomy—‘vigorous price competition is generally beneficial to consumers’—thus implying that if a cost-based filter did not apply, the costs of false positives could be prohibitive.111 The Commission does not adopt a cautious approach for non price-related exclusionary conduct, which either indicates that there is a higher probability of false positives for price-based conduct than for non price-based conduct, or that the cost of false positives in the context of price-related conduct is higher than the cost of false positives in the context of non price-based conduct. No empirical evidence is advanced, however, to support either of these two claims. The CFI echoed the dichotomy between price and non-price exclusionary conduct in Deutsche Telekom: following the Commission’s position, it chose to emphasise the likelihood that the pricing practices of the dominant undertaking would have the effect of removing from the market an economic operator that was just as efficient as the dominant firm, in order to determine whether the pricing practices were abusive.112 The CFI cited, inter alia, the case law of the ECJ in Akzo113 as an authority for the proposition that the abusive nature of a dominant undertaking’s pricing practices is determined in principle on the basis of its own charge and costs, rather than on the basis of the situation of actual or potential competitors.114

As is well explained in the judgment: any other approach could be contrary to the general principle of legal certainty. If the lawfulness of the pricing practices of a dominant undertaking depended on the particular situation of competing undertakings, particularly their cost structure— information which is generally not known to the dominant undertaking—the latter would not be in a position to assess the lawfulness of its own activities.115

However, in other parts of the judgment the CFI was less clear on the choice of the benchmark for assessing the efficiency of the competing undertakings. It simply indicated that the legality of the dominant firm’s practice is not only determined on the basis of its own costs but that the situation of the rival companies could also serve as a relevant benchmark.116 The 110

Ibid, para 63. Guidance, above n 1, para 22. 112 Deutsche Telekom, above n 97, paras 187 and 199. See also Case T-5/97 Industrie des poudres sphériques SA v Commission [2000] ECR II-3755, para 180. 113 Case C-62/86 Akzo Chemie BV v Commission [1991] ECR I-3359. 114 Deutsche Telekom, above n 97, para 188. 115 Ibid, para 192. 116 Ibid, para 198, where the CFI emphasises the need to provide equal opportunities to the various economic operators, in particular if the latter have different costs and revenue structures. See Simon Gevenaz, ‘Margin Squeeze after Deutsche Telekom’, Global Competition Policy Magazine, May 2008, at 21–23. The CFI seems to employ a ‘not (yet) as efficient as’ test. 111

Categorical Thinking in Competition Law 47 Commission Guidance is more explicit as to the benchmark: this is the LRAIC of the downstream division of the integrated dominant firm. Only when it is not possible clearly to allocate the dominant firm’s costs to downstream and upstream operations will the LRAIC of a non-integrated competitor downstream be used.117 The underlying principle of this case law is that price-related abuses should be dealt with more leniently than non price-related abuses. However, neither the Commission, nor the CFI provides any explanation for such a distinction. Some authors have advanced the view that enforcing Article 82 as regards price-related behaviour leads to a high risk of false positives, while prohibiting exclusionary contracts does not have the same effect: ‘unlike low prices, exclusionary contracts do not always benefit consumers in the short term, regardless of their long-term effect on competition’.118 This assumption may be questioned. Competition based on quality (Q) or variety (V) increasing investment is equally important. Richard Markovits defines QV investment competition as the process through which rival sellers compete away their potential supernormal profits by introducing additional QV investments until even the most profitable project in the relevant area of product space generates just a normal rate of return.119

The importance of this type of competition has been recognised by the European Court of Justice in Metro.120 It may have also justified the more lenient antitrust regime for vertical restraints, as these could be conceived as contract enforcement mechanisms that ensure the quality of distribution services provided to consumers, even if this also has the effect of reducing intra-brand price competition.121 There is no specific reason advanced to explain why price competition is more important as an antitrust concern than QV investment competition. It has been argued by some economists that a cost-based standard should also apply to technological and contractual tying.122 There is a risk (which for some could represent a desirable objective) that the efficiency-based standard for exclusionary pricing practices will expand to non-pricing 117

Guidance, above n 1, para 79. M Laurence Popofsky and Heller Ehrman, ‘Drawing a Line Between Bundling and Contractual Exclusion under the Sherman Act’, Global Competition Policy Magazine, June 2008, at 5–6. 119 Richard Markovits, Truth or Economics. Is economic efficiency a sound basis upon which to make public policy or legal decisions? (New Haven, Yale University Press, 2008), p 90. 120 Case 26/76 Metro SB-Großmärkte GmbH & Co KG v Commission [1977] ECR 1875, para 21. 121 Benjamin Klein and Kevin Murphy, ‘Vertical restraints as Contract Enforcement Mechanisms’ (1988) 31(8) Journal of Law and Economics 265. 122 Jean Tirole, ‘The Analysis of Tying Cases: A Primer’ (2005) 1 Competition Policy International 1. 118

48

Ioannis Lianos

exclusionary practices, thus replacing the consumer interest-orientated anticompetitive foreclosure standard. This appears incompatible with a real effects-based approach. There are instances where a bundled discount may produce the same degree of exclusivity as contractual tying.123 A rule that would adopt a more lenient standard for price-related bundling practices as opposed to non price-related ones might therefore produce false negatives. What is clear is that the distinction between price and non-price exclusionary practices has been introduced in order to create a safe harbour for dominant undertakings. However, the link between the classification of price- and non price-related conduct and the interpretation of the concept of consumer harm (which is the reference point to measure effects) is missing. The Commission defines the concept of consumer harm broadly, as covering not only restraints that affect competition through lower prices, but also those affecting the possibility for consumers to benefit from better quality products and wider choice of new or improved goods and services.124 The underlying objectives of Article 82, in particular its emphasis on preserving consumer choice, may explain the relatively strict antitrust standards that apply to technical tying in comparison to contractual tying.125 It is clear that the Commission does not establish a hierarchy between these different aspects of consumer harm. The distinction between price- and non price-based exclusionary conduct seems incompatible with the objectives of EC competition law and should be re-examined. The introduction of the price/non-price abuses dichotomy in EC competition law is also at odds with recent developments in US antitrust law. In Pacific Bell v Linkline, the US Supreme Court made clear that ‘there is no reason to distinguish between price and non-price components of a transaction’, and consequently found that its reasoning in Trinko (involving an insufficient assistance claim by a competitor) applied ‘with equal force to price squeezing claims’.126 What counts is the effect of the specific conduct on consumers, not the price or non-price label attached to it by the claimant or the defendant.

123 See Nick Economides and Ioannis Lianos, ‘The Elusive Antitrust Standard on Bundling in Europe and in the United States at the Aftermath of the Microsoft Cases’, forthcoming, (2009) 76(1) Antitrust Law Journal, at 2.3.2.2. It is well accepted in EC competition law that de facto exclusivity (which could take the form of financial inducement through rebates) is considered as harmful as de jure exclusivity: Van den Bergh Foods, above n 21, para 118 (Article 81), paras 159–60 (Article 82); confirmed by Unilever Bestfoods, above n 23. 124 Guidance, above n 1, para 5. 125 Ibid, para 52. The main justifications provided for this classification are that technical tying is costly to reverse and that it also reduces the opportunities for resale of individual components. 126 Pacific Bell Telephone Company v Linkline Communications, Inc, above n 103, at 10.

Categorical Thinking in Competition Law 49 V. CONCLUSION

Classification and the definition of categories is an essential ingredient of legal thinking. The reform of Article 82 towards an effects-based approach cannot change that. It may, however, imply a new effort to use classification to render antitrust categories compatible with economic thinking as well as with the evolving objectives of EC competition law. The adoption by the Commission Guidance of the distinction between price and non-price exclusionary abuses shows precisely the inherent need for classification, even in an effects-based approach. The aim of this study was to show that thinking across the lines of the existing categories of abuse does not provide an adequate conceptual framework for Article 82. The Balkanisation of Article 82, with different tests applying to different categories of conduct that may be interchangeable and produce the same effects on consumers, can hardly ensure legal certainty and the effective protection of competition in the internal market. A more careful analysis of recent EC competition law cases on Article 82 also indicates that an effort of categorisation that would ignore the underlying objective of the case law, the preservation of consumer sovereignty, is subject to failure. The integration of this aim requires a new effort of classification along broader categories than the current conceptual framework for abuses in Article 82. The suitability of the classification of price- versus non price-based conduct, introduced by the recent Commission Guidance, will have to be assessed and compared with other alternatives.127 The same is also true for the relationship between Article 82 and Article 81, in particular if a more economics-orientated definition of antitrust agreement/ collusion emerges in the area of vertical relations.128

127 See Ioannis Lianos, ‘Classification of abuses in Article 82 EC: a straight story?’ UCL Working Paper, forthcoming 2009. 128 Ioannis Lianos, ‘Collusion in vertical relations under Article 81 EC’ (2008) 45 CML Rev 1027–77.

3 The Commission’s Guidance on Article 82 EC and the Effects Based Approach—Legal and Practical Challenges ARIEL EZRACHI*

I. INTRODUCTION

T

HIS CHAPTER EXPLORES the development of the so-called ‘effects-based approach’ to Article 82 EC. In doing so, it considers the difficulties in introducing an effects-based approach and its viability and enforceability. The chapter is divided into four main parts. First, it describes the rise of the ‘effects-based approach’ and its manifestation in the recent Commission Guidance on Article 82 (‘the Guidance’).1 It identifies some of the inherent difficulties associated with the departure from a formalistic approach and the introduction of effects-based analysis. The chapter then considers the Commission’s power to advance an effects-based approach and the legal status of its guidelines. It highlights the fact that the Guidance, to the extent that it departs from traditional case law, exists outside the precedent system. It is argued that this ‘external’ expansion hampers the effectiveness of the Guidance as a vehicle to advance timely reform of Article 82. Furthermore, it may trigger inconsistencies between enforcement by the Commission and case law from the courts. The risk of inconsistency in analysis and enforcement may be reduced once (and if) the European courts include variants of the effects-based

* Director, The University of Oxford Centre for Competition Law and Policy. Slaughter and May Lecturer in Competition Law, Oxford University. Fellow, Pembroke College, Oxford. 1 DG Competition, Communication from the Commission, Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings (Brussels, 3 December 2008), http://ec.europa.eu/comm/competition/ antitrust/art82/guidance.pdf.

52

Ariel Ezrachi

approach in the precedent system. This possibility and the extent to which the European Court of Justice (ECJ) and the Court of First Instance (CFI) can shift their analysis and adopt the effects-based approach, are considered in the third part of the chapter. Absent timely backing from the European courts, the risk of inconsistency becomes more apparent. This is particularly the case in the application of competition law at national level by national courts. The chapter considers the possible impact such inconsistency may have at the national level. It also highlights the capacity constraints which might undermine an effects-based approach even if it were to find its way into the European precedent system. II. THE RISE OF THE EFFECTS-BASED APPROACH

In July 2005 the Commission published its Consultation Paper, Report on an Economic Approach to Article 82 (‘the Consultation Paper’),2 in which the Economic Advisory Group for Competition Policy (EAGCP) questioned the merits of the traditional formalistic approach to Article 82. The Paper advocated an effects-based approach which focused on competitive harm and refrained from treating different categories of behaviour as abusive per se. The proposal stemmed from the understanding that the same type of action may result in different effects on the market, some abusive and others not. Similarly, it was recognised that different types of action may lead to the same anticompetitive effect. The publication of the Consultation Paper coincided with the publication of a Discussion Paper on the Application of Article 82 of the Treaty to Exclusionary Abuses (‘the Discussion Paper’).3 Although this Paper might have been seen as a continuation of the EAGCP Consultation Paper, it was more restrictive in nature. Unlike the EAGCP Consultation Paper, the Discussion Paper generally preserved the analysis of dominance as it existed in the Community case law. One could nevertheless identify hints of an economic approach woven into the Discussion Paper. The most notable example was the Commission’s proposed treatment of efficiencies and the creation of an efficiency-like defence.4 As clearly indicated by the 2 EAGCP, Report on an Economic Approach to Article 82 EC (July 2005), http://europa. eu.int/comm/competition/publications/studies/eagcp_july_21_05.pdf. 3 DG Competition, Discussion Paper on the Application of Article 82 of the Treaty to Exclusionary Abuses (December 2005), http://europa.eu.int/comm/competition/antitrust/ others/discpaper2005.pdf. 4 Commission Guidance, above n 1, paras 84–92; see also Discussion Paper, above n 3, para 8: ‘… if the conduct of a dominant company generates efficiencies and provided that all the other conditions of Article 81(3) are satisfied …, such conduct should not be classified as an abuse under Article 82 of the EC Treaty.’ This statement stems from the CFI decision in Case T-193/02 Laurent Piau v Commission [2005] ECR II-209, para 119, yet it seems to widen the Court’s holding and suggest that an Article 81(3) defence may be available.

The Commission’s Guidance on Article 82 EC 53 Commission, the Discussion Paper did not create any legitimate interest and could not be relied upon to provide guidance on Commission enforcement policy.5 The Papers generated a wide-ranging debate on the scope and nature of analysis in Article 82 cases.6 Commenting on the review of Article 82, Commissioner Neelie Kroes stated that the Commission’s intention was not to propose a radical shift in enforcement policy but rather to ‘develop and explain theories of harm on the basis of a sound economic assessment’.7 In December 2008, the Commission published Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings. The Guidance lays down the Commission’s modern approach to Article 82 cases and delivers several of the innovative elements which were aired in the preceding Papers. In principle, the introduction of effects-based variants in the Commission’s analysis is to be welcomed. It reduces the risk of over-regulation and of chilling competition which may result from formalistic analysis. It aims to distinguish between competition on the merits and anticompetitive unilateral action. Yet the practicality of the effects-based analysis is challenging. The move from a formalistic approach to an economic-based approach creates uncertainty as to the relevant benchmarks used to establish abuse and their limiting principles.8 Uncertainty can be identified at a number of different levels. Part of the uncertainty is unavoidably linked to the shift from one analytical framework to another. Another part of it stems from the open-ended and novel nature of some of the new concepts outlined in the Guidance. Take, for instance, the efficiency defence introduced in the Guidance. The Commission has for some time suggested that this should be broadly

As such it stands at odds with the CFI ruling in Cases T-191/98, T-212/98 and T-214/98, Atlantic Container Line (TACA), [2003] ECR II-3275, para 939, and Case T-395/94 Atlantic Container Line [2002] ECR II-875, para 330, as well as the Commission’s Guidelines on Article 81(3), Guidelines on the Application of Article 81(3) of the Treaty (Text with EEA relevance) [2004] OJ C101/97, para 11-06. 5

Discussion Paper, above n 3, para 7. See, eg, Niels, ‘The Article 82 Discussion Paper: A Comment on the Economic Principles’, April 2006, available at www.oxera.com/cmsDocuments/Agenda_April%2006/The%20Articl e%2082%20discussion%20paper.pdf; Padilla, ‘The Law and Economics of Article 82’ (2007) 44(4) CML Rev 1185–86; Papandropoulos, ‘Implementing an effects-based approach under Article 82’ (2008) Concurrences 1–5. See also the responses to the Discussion Paper on the Commission website: http://ec.europa.eu/competition/antitrust/art82/index.html. 7 Neelie Kroes, Competition Commissioner, ‘Preliminary Thoughts on Policy Review of Article 82’, Speech at the Fordham Corporate Law Institute, 23 September 2005. 8 The notion of ‘an equally efficient competitor’ is heavily used in the Guidance to underpin the effects-based approach. See, eg, Guidance, above n 1, paras 42 ff and 59 ff. 6

54

Ariel Ezrachi

structured in line with Article 81(3) EC.9 This would align the efficiency analysis under Articles 81 EC and 82 EC, and prevent inconsistency. In principle, one would welcome the introduction of such a ‘fine tuning mechanism’.10 However, its scope and limitations are as yet unclear, as it has no grounding in the European courts’ case law or Commission decisions. Another example may be found in the treatment of conditional rebates.11 There the Commission stipulates that the loyalty-inducing nature of the rebate can be assessed by reference to the ‘non contestable’ and ‘contestable’ portions of demand. These variables are affected, among other things, by the capacity constraints of existing competitors and the entry prospects of potential competitors. The proposed analysis may provide a useful measure, yet its practical application is challenging. One can question, for example, the ability of a dominant firm to assess these variables accurately and to obtain information from its existing and potential competitors. Similar difficulties in measurement may be attributed to the Commission proposal to examine conditional rebates by estimating the price a competitor would have to offer a customer in order to compensate the customer for the loss of the conditional rebate if it switched part of its demand away from the dominant firm.12 Here again, lack of access to such information and the difficulty of calculating the relevant measurement make the benchmark a sub-optimal tool for guiding the conduct of dominant undertakings.13 More generally, undertakings may find it difficult to establish whether their competitor is ‘as efficient’, and subsequently whether their activity may amount to ‘anticompetitive foreclosure’. The concept of ‘as efficient’ has been used (arguably, to some extent inconsistently) by the Commission throughout the Guidance to set the benchmark for intervention.14 Confusion surrounding such a pervasive concept in the Guidance is damaging to the principle of legal certainty. These ‘inherent uncertainties’ as to the nature and limitations of some of the concepts in the Guidance and their practical applicability affect the boundaries of what constitutes dominance and abuse. This in turn raises questions as to the nature and scope of the Commission’s intervention in future cases. Admittedly, the above uncertainty is not triggered in clearcut cases where the abusive behaviour is evident, as these do not call for

9 Neelie Kroes, Competition Commissioner, ‘Exclusionary abuses of dominance—the European Commission’s enforcement priorities’, Speech at Fordham University Symposium, 25 September 2008; Kroes, speech of 23 September 2005, above n 7. 10 Commission Guidance, above n 3, para 30. 11 Ibid, paras 37–46. 12 Ibid, para 40. 13 See comment on the relevant range in John Temple Lang, ‘A question of priorities—The European Commission New Guidance on Article 82 is Flawed’, Competition Law Insight, 10 February 2009, p 3. 14 Ibid, on the unsatisfactory nature of this benchmark.

The Commission’s Guidance on Article 82 EC 55 application of the effects-based approach.15 However, in the majority of cases where the undertaking is in doubt as to the legality of its actions or the possible challenge by other undertakings or competition authorities, the uncertainty may be detrimental for businesses. This raises concerns about the chilling of competition; precisely the issue which the Guidance was intended to resolve. Adding to the inherent uncertainty is what may be referred to as ‘external uncertainty’. In what follows, the discussion moves beyond the confined economic-legal analysis of the Guidance. It explores the wider range of variables that affect the success and viability of an effects-based approach. It highlights how the internal ambiguity is amplified by ‘external uncertainty’ which stems from the relationship between the new Guidance and the Community judicature, and from the application of the Guidance at national level by national competition agencies and national courts. This ‘external uncertainty’ plays a paramount role in determining the future success of the Guidance and its effects on welfare. III. THE COMMISSION’S LIMITED MANDATE

It is generally accepted that in Commission guidelines, the Commission determines, generally and abstractly, the method it will use in its assessment.16 Although the Commission’s guidelines may not be regarded as rules of law which the administration is always bound to observe, they nevertheless form rules of practice from which the administration may not depart in an individual case without giving reasons that are compatible with the principle of equal treatment.17 In the context of Article 82, it is interesting to explore the Commission’s mandate to promote an effects-based approach, in as much as this differs from established case law. In the Discussion Paper, the Commission clarified that its interpretation of Article 82 is without prejudice to the interpretation that may be given by the ECJ or the CFI.18 Similarly, the Guidance clarifies that it is not intended to constitute a statement of the law and is without prejudice to the interpretation of Article 82 by the European Court of Justice and the Court of First Instance.19

15 See, eg, Case COMP/E-1/38.113 Prokent/Tomra [2008] OJ C219/12, where the dominant company imposed de facto exclusivity clauses on its customers. 16 Case C-167/04 P, JCB Service, v Commission of the European Communities [2006] ECR I-8935, para 209; Joined cases 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, paragraph 213. 17 Dansk Rørindustri (n 16) para 209; JCB Service (n 16) para 207. 18 Discussion Paper (n 3) para 6. 19 Commission Guidance, above n 1, para 3.

56

Ariel Ezrachi

The practical impact of this was highlighted in British Airways v Commission where Advocate-General Kokott noted that even if the Commission’s administrative practice was to change, it ‘would still have to act within the framework prescribed for it by Article 82 as interpreted by the Court of Justice’.20 In Treuhand AG v Commission, the CFI provided an illustrative account of the limited leeway available to the Commission in its interpretation of the Treaty provisions.21 The case concerned the interpretation of Article 81 EC, but is just as relevant to the subject matter of this chapter. The Court noted that since the interpretation of Article 81(1) EC ultimately falls to be determined by the Community judicature, the Commission does not have any leeway enabling it, where relevant, to forego bringing an action against a consultancy firm which satisfies the criteria for shared liability. On the contrary, by virtue of its duty under Article 85(1) EC, the Commission is required to ensure the application of the principles laid down in Article 81 EC … Accordingly, since … the Commission’s decision-making practice prior to the contested decision could appear to conflict with the above interpretation of Article 81(1) EC, that practice was not capable of giving rise to legitimate expectations on the part of the undertakings concerned.22

This principle effectively curtails the ability of the Commission to advance a meaningful and prompt reform of Article 82. In other words, the Commission’s ‘effects-based agenda’ and the prioritising exercise at the heart of the Guidance, can only exist within the boundaries of Article 82 as set by the European courts. Subsequently, the Commission’s ‘creative leeway’ in advancing an effects-based approach is limited and subrogated to the European courts’ interpretation of Article 82. Cynics might point out that the European courts’ interpretation of Article 82 has been guided to a large extent by the Commission’s historic approach, which was formalistic in nature. Subsequently, the Commission’s creative leeway is restricted by its own historic analysis which was approved by the European courts. Having been embedded in the European case law, this structured analysis cannot easily be changed. For now the fact remains that the Commission’s new thinking is subrogated to the courts’ interpretation of Article 82. Arguably, this limitation led the Commission to a creative approach in positioning the new Guidance on Article 82. The document does not merely describe the case law and the Commission’s practice, but aims to widen the analytical framework and introduce novel thinking to the analysis of Article 82. It is not by chance that the document was not entitled ‘Guidelines’ but rather ‘Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty 20 21 22

Case C-95/04 British Airways v Commission [2007] ECR I-2331, para 28. Case T-99/04, AC-Treuhand AG v Commission [2008] ECR 00. Ibid, para 163.

The Commission’s Guidance on Article 82 EC 57 to Abusive Exclusionary Conduct by Dominant Undertaking’. The title echoes the hybrid nature of the document and the potential tension between some of the new concepts and the case law of the European courts. In this respect the nature of the document and its innovative stance undermine the certainty which other guidelines have achieved in the past. One can, for example, question whether an efficiency defence, as stipulated in the Guidance, is the sole preserve of Commission analysis or could be applied elsewhere. Does this new concept coexist with the European courts’ case law? And if so, would and should national courts and competition agencies factor efficiency arguments into their decision making? The answers to these questions bear directly on the dominant undertaking’s behaviour. Uncertainty has a cost, as the most restrictive approach dictates the commercial unilateral behaviour of the dominant undertaking. IV. THE EUROPEAN COURTS’ REACTION TO THE ‘EFFECTS-BASED APPROACH’

Evidently, a positive reaction from the European courts to the Guidance and the effects-based variants would widen the ‘analytical framework’ and provide the Commission with much-needed impetus to push forward the effects-based approach to Article 82. Admittedly, not every case which comes before the CFI and ECJ provides an adequate opportunity to advance an effects-based approach. Adding to this difficulty is the slow turnover of cases, which undermines the courts’ ability to provide a timely reaction to the introduction of new effectsbased concepts and interpretation. Whereas the Commission’s proactive stance enables it to introduce and develop novel concepts independently of an investigation, this stance is not matched by the European courts’ reactive position which comes into play on appeal.23 Subsequently, at this early stage, it is hard to reach firm conclusions as to the European courts’ reaction to the effects-based approach. Nevertheless, with these caveats in mind, a brief look at recent judgments of the European courts reveals little support for effects-based analysis. Take, for example, the CFI decision in British Airways v Commission, where no acknowledgement of an effects-based approach can be detected.24 In this case, the examination of rebates was based on earlier case law and ignored a real economic assessment of possible objective justification. The Court declined to take an active, positive stance questioning earlier case 23 Note also the reactive position of the European courts in Article 234 references. Whilst this is generally less relevant to the new approach taken by the Commission, it is possible for relevant cases to come before the courts using this procedure. See, eg, Joined Cases C-468/06 to 478/06 Sot Lelos kai Sia v GlaxoSmithKline [2008] ECR 00. 24 British Airways, above n 20.

58

Ariel Ezrachi

law or the scope of anticompetitive rebates and their foreclosing effect. The CFI’s comment in paragraph 293 of its judgment is noteworthy: In the first place, for the purposes of establishing an infringement of Article 82, 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 restrict competition, or, in other words, that the conduct is capable of having, or likely to have, such an effect.25

Similarly, in Microsoft Corp v Commission, the CFI took a rather formalistic approach to the question of abuse.26 The judgment focused narrowly on the facts of the case and adopted, at least in terms of form, the Court’s traditional consideration of exceptional circumstances when analysing a refusal to license intellectual property rights. Interestingly, the CFI refrained from commenting directly on the Commission’s view stated in the decision27 and the Discussion Paper, according to which a refusal to license should not impair the customer’s ability to benefit from innovation brought about by the dominant undertaking’s competitors.28 In France Télécom SA v Commission,29 the CFI dealt with alleged predation by Wanadoo Interactive. The Court rejected arguments concerning the need to demonstrate the actual effects of the practice, and held that there is no need to prove the effects of eliminating competition. Accordingly, it held that for the purposes of applying that article, showing an anti-competitive object and an anti-competitive effect may, in some cases, be one and the same thing. If it is shown that the object pursued by the conduct of an undertaking in a dominant position is to restrict competition, that conduct will also be liable to have such an effect.30 ... Furthermore, it should be added that, where an undertaking in a dominant position actually implements a practice whose object is to oust a competitor, the fact

25 Ibid, para 293. See also comments by the CFI in Case T-203/01 Manufacture Française des Pneumatiques Michelin v Commission (‘Michelin II’) [2003] ECR II-4071, [2004] 4 CMLR 18, para 239: ‘For the purposes of establishing an infringement of Article 82, it is sufficient to show that the abusive conduct of the undertaking in a dominant position tends to restrict competition or, in other words, that the conduct is capable of having that effect’. 26 Case T-201/04 Microsoft Corp v Commission [2007] ECR II-3601. 27 Microsoft Case COMP/C-3/37.792, European Commission. The Commission argued that ‘it follows from paragraph 49 of IMS Health, … that a “new product” is a product which does not limit itself essentially to duplicating the products already offered on the market by the owner of the copyright. It is sufficient, therefore, that the product concerned contains substantial elements that result from the licensee’s own efforts’. On appeal, the CFI did not comment directly on this statement (CFI judgment, para 631). 28 Ibid, para 240. 29 Case T-340/03 France Télécom SA v Commission [2007] ECR II-107. 30 Ibid, para 195.

The Commission’s Guidance on Article 82 EC 59 that the result hoped for is not achieved is not sufficient to prevent that being an abuse of a dominant position within the meaning of Article 82.31

A less formalistic approach may be found in the Opinion of AdvocateGeneral Dámaso Ruiz-Jarabo Colomer in the GlaxoSmithKline case.32 The Advocate-General seems to suggest that certain conducts cannot be deemed abusive even when the circumstances of the case make the undertaking’s intention clear.33 He noted that a formalistic approach to the question of abuse may fail to recognise activities which are of benefit to consumers. Accordingly, the real effect of the activities can be recognised only while comparing their positive and negative consequences.34 The Opinion clearly moves away from a formalistic per se analysis into an effects-based one, in which the positive effects of the conduct are being examined. As such, it provides valuable support to the effects-based approach. Unfortunately, the ECJ favoured a more formalistic style of reasoning and followed the traditional route of analysis, focusing on the abuse and opening the door to objective justification.35 In doing so, the Court did not comment on either the effects approach or the suggested framework laid out by the AdvocateGeneral. Again, it is important to acknowledge that only a short period of time has elapsed since the beginning of the discussion on the effects-based approach. As mentioned above, not every case provides the European courts with a relevant opportunity to advance an effects-based approach. Further, even if the courts were inspired by the effects-based approach, they might nevertheless hesitate to adopt novel approaches prior to their maturation over time. Nevertheless, the ECJ’s analysis in the GSK judgment seems to be a missed opportunity to engage with the debate on the effects approach. It is particularly disappointing since the Advocate-General laid the foundations of an effects-based analysis in his Opinion. Consequently, the signal sent by the Court seems to be a strong one, as the Court actively chose to move away from an effects-based analysis. This signal makes it harder to reconcile the effects-based variants in the Guidance with the analysis of the ECJ. Unless and until the European courts provide a clear and positive signal regarding an effects-based approach, a gap is likely to emerge between the Commission’s proposed analysis and the European courts’ case law. This 31 Ibid, para 196, citing Joined Cases T-24/93 to T-26/93 and T-28/93 Compagnie Maritime Belge Transports and Others v Commission [1996] ECR II-1201, para 149; and Case T-228/97 Irish Sugar v Commission [1999] ECR II-2969, para 191. 32 Case C-468/06–478/06 Sot Lélos kai Sia EE v GlaxoSmithKline AEVE Farmakeftikon Proïonton (hereafter GSK), A-G Opinion dated 1 April 2008. 33 Ibid, A-G Opinion, para 123. 34 Ibid, A-G Opinion, paras 73–75. 35 GSK, above n 32, paras 68–70 and 78–122.

60

Ariel Ezrachi

raises questions as to the legal standing of the Commission Guidance. Do the effects-based variants exist within the case law, and can they coexist with the Community judicature? Or, alternatively, do they depart from it? The way one answers these questions directly affects the boundaries of analysis and the viability of the Guidance’s new effect variants. V. SEVERAL STRINGS OF JURISPRUDENCE?

The perceived gap between the effects-based variants in the Guidance and case law may trigger inconsistency at two distinct levels. First, it may lead to the development of two strings of jurisprudence and analysis: one adopted by the Commission and coloured by the effects-based approach, the other being the traditional analysis manifested in the European courts’ judgments. This inconsistency may lead to legal and commercial uncertainty, since the two strings reflect different degrees of intervention. Possible appeals on Commission decisions may provide an opportunity for the courts to accept or reject the novel approaches in the Guidance. This will however, require the CFI and ECJ to take a proactive stance and move beyond the facts of the case, to discuss the principles at stake. Until then, ‘external uncertainty’ will exist as to whether novel variants in the Guidance exist within, or outside, the case law and as to the limiting principles that govern them. Secondly, additional strings of jurisprudence may develop in cases brought before national courts. By its nature, the Guidance functions outside the precedent system. As such, it provides guidance but has no binding effect on national courts.36 The Commission has described the relationship between Community case law, Commission decisions and Commission guidelines in its Notice on the co-operation between the Commission and the courts of the EU Member States: The application of Articles 81 and 82 EC by national courts often depends on complex economic and legal assessments. When applying EC competition rules, national courts are bound by the case law of the Community courts as well as by Commission regulations applying Article 81(3) EC to certain categories of agreements, decisions or concerted practices. Furthermore, the application of Articles 81 and 82 EC by the Commission in a specific case binds the national courts when they apply EC competition rules in the same case in parallel with or subsequent to the Commission. Finally, and without prejudice to the ultimate interpretation of the EC Treaty by the Court of Justice, national courts may find guidance in Commission regulations and decisions which present elements of analogy with

36 Case 66/86 Ahmed Saeed Flugreisen and Silver Line Reisebüro GmbH v Zentrale zur Bekämpfung unlauteren Wettbewerbs eV [1989] ECR 803, para 27; Case C-234/89 Stergios Delimitis v Henninger Bräu AG [1991] ECR I-935, para 50.

The Commission’s Guidance on Article 82 EC 61 the case they are dealing with, as well as in Commission notices and guidelines relating to the application of Articles 81 and 82 EC and in the annual report on competition policy.37

This position is also reflected in national legislation, for example in the provisions of the UK Competition Act 1998 (‘the Act’). Section 60(2) of the Act stipulates that the national court should avoid inconsistencies with decisions of the European courts. Section 60(3) establishes that the national court ‘must have regard to any relevant decision or statement of the Commission’.38 The lack of binding effect does not resolve the possible inconsistency which may stem from the Guidance. Although not binding on national courts, the Guidance may be taken into account by the courts when formulating their decisions. In the context of Article 82, some national courts may voluntarily adopt ‘effects-based variants’ from the Guidance. This will be possible when the national court concludes that these variants can coexist with decisions of the European courts. Other courts, however, may favour the more formalistic approach manifested in the case law of the ECJ and CFI. A consequence of the above might be the creation of an unlevel playing field, as different national courts, with different approaches and expertise, pick and choose different elements to adopt. VI. CAPACITY CONSTRAINTS

The discussion of the possible impact of the Guidance in the context of private enforcement draws attention to another variable which may impact and affect the nature of analysis, ie the capacity of the national court to engage in effects-based analysis. The national courts of the enlarged European Union differ substantially in their experience of analysing competition cases and the related economic investigations. It is generally accepted that the majority of national courts have limited capacity when it comes to the analysis and deliberation of complex economic concepts and the assessment of ‘consumer harm’. A rule-based approach, albeit formalistic in nature, is often easier to apply and predict. By contrast, an effects-based approach requires the court to consider the true impact of the undertaking’s action and take an active step away from previous ‘formalistic’ case law. This is by no means an easy exercise. It involves grappling with complex economic concepts and large 37 Commission Notice, 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, para 8. 38 Note more generally Council Regulation (EC) 139/2004 on the control of concentrations between undertakings (the EC Merger Regulation) [2004] OJ L24/1, art 16; Inntrepreneur Pub Company (CPC) and others v Crehan, [2006] UKHL 38.

62

Ariel Ezrachi

quantities of empirical evidence. In the context of private enforcement, such economic-based analysis may push some domestic courts into unfamiliar territory. Subsequently, capacity constraints are likely to lead some courts to refrain from adopting an ‘effects-based approach’ as far as such approach is manifested in the Commission Guidance. Other courts, with greater proficiency and resources, may adopt some effects-based variants from the Commission’s guidelines. Subsequently, the different levels of capacity across and between Member States are likely to contribute further to uneven application of an effects-based approach. Capacity constraints may also impact on the ‘quality’ of analyses when courts decide to follow the Commission’s footsteps. Again, as courts differ in their economic and competition law proficiency, so might their ‘effectorientated’ analysis of Article 82. The use of Article 234 references and the ability to appeal the court’s decision can only partially cure the risk of inadequate decisions, in particular because of the time required to resolve an Article 234 reference. This dissonance may lead to divergent interpretations of, and rulings on, consumer harm. One area which illustrates the capacity of national courts to foster an economic-based approach is excessive pricing. This area of law is complex, requiring the courts first to establish the competitive price and then to assess the fairness of the excessive price. Illustrative in this respect are comments made by the English Court of Appeal in At the Races Limited v The British Horse Racing Limits and others.39 The case in question concerned claims that the British Horseracing Board Limited had abused its dominant position, among other things, by excessively and unfairly pricing pre-race information which was in its sole possession. In the judgment, the Court of Appeal noted that the claim of excessive pricing presented the Court with a range of factual and legal challenges of a kind which even specialist lawyers and economists regard as very difficult. Mummery LJ stated that difficulties in assessing the excessiveness and unfairness of the price suggested that these problems might be solved ‘more satisfactorily by arbitration or by a specialist body equipped with appropriate expertise and flexible powers’.40 Arguably, these comments do not apply solely in the context of excessive pricing but reveal the difficulty that courts across Europe may have when prompted to engage in greater economic-orientated analysis of Article 82 cases. Capacity concerns may result in low tolerance to novel effects-based analysis, especially when such an approach has not filtered through the precedent system. In the mid-term, uneven application of effect variants and uneven proficiency in competition law and economics, are likely to contribute to the 39 40

At the Races Limited v The British Horse Racing Limits and others [2007] EWCA Civ 38. Ibid, paras 3–7.

The Commission’s Guidance on Article 82 EC 63 creation of an unlevel playing field. These differences at the national level may give rise to legal uncertainty and trigger forum shopping, as claimants seek to bring their claims in a forum which they perceive as better suited to their purposes.41 The possible inconsistent application at the national level draws attention to the interface between the Guidance and the proposed changes to private enforcement in Europe.42 Interesting in this respect is the proposal in the White Paper in Section 2.3, where the Commission states that: The Commission sees no reason why a final decision on Article 81 or 82 taken by an NCA [National Competition Authority] in the European Competition Network (ECN), and a final judgment by a review court upholding the NCA decision or itself finding an infringement, should not be accepted in every Member State as irrebuttable proof of the infringement in subsequent civil antitrust damages cases.

The interface between two developing areas, the private enforcement of competition law and new concepts in relation to Article 82, may result in friction. In Article 82 cases, the White Paper’s proposal may lead to interesting circumstances where courts with different approaches toward the application of Article 82 have to apply, in follow-on claims, decisions which are based on analysis which they have specifically rejected as applicable in their own country. This difficulty is not unique to Article 82, yet the introduction of the ‘effects-based variants’ makes the potential tension more noticeable. These uncertainties as to the scope of Article 82 and its possible application in national courts may also impact on out-of-court settlements. Settlements represent a significant part of private enforcement but, due to their nature, are often not in the public eye. However, the willingness to settle is derived, at least in part, from the cost and certainty as to the outcome of proceedings. In other words, the alternative cost, namely the price and risk of court litigation, impacts on the economic and legal integrity of these settlements. Shortcomings in one or more of the variables mentioned in this chapter would increase this alternative cost, allowing manipulations and opportunism, rather than calculated risks based on a consistent and principled system of law, to set the tone of out-of-court settlements.

41 Such decision may be based on cost, risk, disclosure provisions, and other procedural and substantive provisions. It may also be affected by the type of abuse, the position of the dominant firm and the injured party, transfer of wealth between jurisdictions and the risk of populist decisions. 42 EC Commission, Damages Actions for Breach of the EC Antitrust Rules (White Paper), COM (2008) 165, 2 April 2008.

64

Ariel Ezrachi VII. CONCLUSION

This chapter draws attention to the legal landscape which surrounds the Guidance and the effects-based variants. While it does not question the merit of an effects-based approach, and the benefits such an approach may carry, it notes how the institutional limitations on the Commission’s powers may trigger uncertainty at substantive levels. Subsequently, it is argued that even if the Commission is successful in advancing a clear effects-based agenda in its Guidance—something which some have argued has not been the case43—such agenda is hampered by jurisdictional limitations. As one moves beyond the confined economiccompetition law review of the Guidance, the tension between new variants and existing case law is revealed. The tension stems from the alleged inconsistency between the European courts’ existing jurisprudence and the ‘effect variants’ in the Guidance. This, in turn, leads to uncertainty as to the boundaries of legality and enforcement. The Commission’s solution was to limit the level of ‘effects-based variants’ and to term the document ‘Guidance’ on enforcement priorities, rather than typical ‘guidelines’. The title reflects the hybrid nature of the document, which must balance the difficult task of advancing new thinking while operating within the legal limitations imposed on the Commission. In an attempt to remove any doubt, the Commission states clearly that the Guidance is subrogated to the Community judicature. This avoids the question as to whether new effects-based variants coexist with existing case law or are external to it. Arguably, by definition, their inclusion in the Guidance reflects the Commission’s stance that the former is true. Yet one would find it difficult to find jurisprudential support for some of the concepts advanced by the Commission, most notably the efficiency defence. By avoiding this key question, the Commission was able to advance its modern thinking in the shape of the Guidance. Yet, similar to the principles of conservation of energy in physics, the question did not disappear—it merely shifted form. It is now for other stakeholders—namely the undertakings, the national courts and the national competition agencies—to judge for themselves and adjust their actions accordingly. This, it is argued, triggers uncertainty as to the principles which will be applied in Article 82 cases, and increases the likelihood of the development of several strings of jurisprudence. In other words, the boundaries of Article 82 lose their clarity, not only because of the introduction of effectsbased variants, but also due to these being advanced outside the precedent system. As the activities at stake often concern ongoing unilateral practices,

43

See, eg, John Temple Lang, above n 13.

The Commission’s Guidance on Article 82 EC 65 the ‘uncertainty multiplier’ might be high and result in the chilling of competition of dominant or ‘potentially’ dominant undertakings. Positive, informal endorsement from the European courts, although encouraging, cannot remedy the system friction and override the formalistic nature of recent judgments. Admittedly, due to their reactive stance, the courts are limited in their ability to endorse changes in analysis, unless the right case finds its way to the court room. Still, the courts appear to have remained wedded to their formalistic approach despite cases which presented opportunities for endorsement and clarification. In the context of the private enforcement of Article 82, this chapter further highlights the possible capacity limitation of some national courts, a fact which may also hinder an effects-based analysis and create inconsistency. This highlights the wider implications for competition law in the EU and the way that reform in one area will impact on others.

4 The Evolution of the Notion of Consumer Interest in Light of the Modernisation of Article 82 EC ORIT DAYAGI-EPSTEIN*

I. INTRODUCTION

I

N DECEMBER 2008, the European Commission (‘the Commission’) published its long-awaited Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings (‘the Commission Guidance’).1 The Commission Guidance followed the Discussion Paper on the Application of Article 82 of the Treaty to Exclusionary Abuses (‘the Discussion Paper’),2 which was published in December 2005, and a Report by the Economic Advisory Group on Competition Policy (EAGCP), entitled An Economic Approach to Article 82 EC, which was published in July 2005 (‘the EAGCP Report’).3 These publications introduce an economic/effects-based approach to Article 82 EC which emphasises the effects of exclusionary conducts on consumer welfare. This chapter sets out the evolution of the notion of consumer interest under Article 82 EC, from a formalistic approach which emphasised the structure of the competitive process and individuals’ economic freedom to an effectsbased approach that emphasises consumer welfare. In this framework the

* Research Affiliate, The University of Oxford Centre for Competition Law and Policy; PhD (King’s College London). 1 DG Competition, Communication from the Commission, Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings (Brussels, 3 December 2008), http://ec.europa.eu/comm/competition/ antitrust/art82/guidance.pdf. 2 DG Competition, Discussion Paper on the Application of Article 82 of the Treaty to Exclusionary Abuses (December 2005), http://europa.eu.int/comm/competition/antitrust/ others/discpaper2005.pdf. 3 EAGCP, Report on an Economic Approach to Article 82 EC (July 2005), http://europa. eu.int/comm/competition/publications/studies/eagcp_july_21_05.pdf, pp 2, 7.

68 Orit Dayagi-Epstein discrepancies between the Discussion Paper and the Commission Guidance in respect of consumer welfare are discussed. The chapter concludes by arguing that despite the modernisation process which has been expounded by the publications, the evolution of consumer interest has yet to be supported by the Community courts. II. PROTECTION OF CONSUMER INTEREST UNDER ARTICLE 82 PRIOR TO THE PROCESS OF MODERNISATION

Article 82 EC prohibits the abuse of a dominant position by one or more undertakings within the Common Market (EU) and provides examples of such an abuse. Arguably, the reason for prohibiting a dominant firm from employing strategies that are not based on competitive merits is the desire ‘not to harm consumers’ by changes in the supply side which can seriously jeopardise consumers’ economic freedom of action in the market.4 However, only Article 82(b) EC, which prohibits the limitation of ‘production markets, or technical development to the prejudice of consumers’5 (emphasis added), mentions consumers explicitly, although without setting out what ‘prejudice to consumers’ means.6 This is perhaps not surprising, given the fact that Community courts have interpreted the term ‘consumer’ in the context of Article 82 as encompassing within it any purchaser regardless of whether it is an intermediate customer or a final consumer.7 Furthermore, Article 82 does not specify a particular objective or standard of harm for its application, unlike Article 81 EC which prohibits practices

4 G Howells and S Weatherill, Consumer Protection Law, 2nd edn (Aldershot, Ashgate, 2005), 550; J Stuyck, ‘The Promotion of the Consumer Interest under Art 86 EEC, Comments’ in M Goyens (ed), EC Competition Policy and the Consumer Interest (LouvainLa-Neuve, Cabay, 1985), 319; Case 6/72 Europemballage Corp. and Continental Can Co Inc v Commission [1973] ECR 215; E M Fox ‘What is Harm to Competition? Exclusionary Practices and Anticompetitive Effect’ (2002) 70 Antitrust Law Journal 371, 392. 5 Article 82(b) relates to exclusionary behaviours which are pursued by a dominant firm and are likely to have a foreclosure effect on the market. R O’Donoghue and A J Padilla, The Law and Economics of Article 82 EC (Oxford, Hart Publishing, 2006), 196–202. 6 Nevertheless, it should be noted that despite the fact that the other sub-clauses of Article 82 EC do not mention the protection of consumer interest explicitly, the protection of consumer interest has been taken into account in practice by the Commission and Community courts in the implementation of the other clauses of Article 82. For example, exploitative abuses under Article 82(a) EC were interpreted as incorporating an element of loss of consumer welfare as a result of excessive pricing; O’Donoghue and Padilla, above n 5, 224–25. 7 See, eg, Case C-53/92P Hilti v Commission [1994] ECR I-667 and Case T-30/89 Eurofix Banco v Hilti [1991] ECR II-1439; Case 22/78 Hugin Kassaregister AB v Commission [1979] ECR I-1869; H Vedder, ‘Competition Law and Consumer Protection: How Competition Law Can be Used to Protect Consumers Even Better—Or Not?’ (2006) European Business Law Review 83, 87; Discussion Paper, above n 2, para 55. For a similar interpretation in the context of Article 81 EC, see Commission Notice (EC) Commission Guidelines on the Application of Article 81(3) of the Treaty [2004] OJ C101/97, para 84 (‘Article 81(3) Guidelines’).

The Evolution of the Notion of Consumer Interest 69 that have ‘as their object or effect the prevention, restriction or distortion of competition’.8 Whereas other areas of EC competition policy (the EC merger control and the control of restrictive agreements under Article 81 EC) have shown greater appreciation of economic analysis over the years,9 the policy under Article 82 has remained unchanged over the last 40 years. One of the main reasons for this stagnation derives from the heavy influence of the Ordoliberal School on Community courts.10 As a result of the Ordoliberal influence on Community courts, Article 82 has been interpreted as a means of protecting the structure of the competitive process as ‘an institution’ which enhances individuals’ (whether final consumers, intermediate customers or competitors) economic freedom and improves economic outcomes.11 Under this formalistic structure-based approach, certain conducts will be considered as abusive and hence detrimental to consumers either because of the form of the conduct (object), or due to their detrimental (or likely detrimental) effects on the structure of the competitive process regardless of their direct effects on consumers.12

8 P Akman, ‘“Consumer Welfare” and Article 82 EC: Practice and Rhetoric’, CCP Working Paper 08-25, July 2008, http://www.uea.ac.uk/polopoly_fs/1.104683!ccp08-25.pdf, accessed 5 January 2009, 2 (Akman (2008a)). 9 The consumer welfare standard has been adopted in the Commission’s Article 81(3) Guidelines and the Horizontal Merger Guidelines. See Article 81(3) Guidelines, above n 7, para 96, point 3.4.3 and para 102; Guidelines on the Assessment of Horizontal Mergers under the Council Regulation on the Control of Concentrations between Undertakings [2004] OJ C31/5, paras 79–80. 10 Ordoliberalism is distinguished by its search for a third way between market economy and centralised economy. See C Ahlborn and D E Evans, ‘The Microsoft Judgment and its Implications for Competition Policy towards Dominant Firms in Europe’, http://ssrn.com/ abstract=1115867, accessed 12 July 2008; B J Rodger, ‘Competition Policy, Liberalism and Globalization: A European Perspective’ (2000) 6 Columbia Journal of European Law 289, 293–98; D J Gerber, ‘Constitutionalizing the Economy: German Neo-liberalism, Competition Law and the “New Europe”’ (1994) 42 American Journal of Comparative Law 25, 31–32; D J Gerber, Law and Competition in Twentieth Century Europe (Oxford, Clarendon Press, 1998), ch VII; K J Cseres, Competition Law and Consumer Protection (The Hague, Kluwer Law International, 2005), 83; W Sauter, Competition Law and Industrial Policy in the EU (Oxford, Clarendon Press, 1997), 28. 11 D J Gerber, ‘The Future of Article 82: Dissecting the Conflict’ in Ehlermann and Marquis (eds), European Competition Law Annual 2007: A Reformed Approach to Article 82 EC, forthcoming 2008, http://www.iue.it/RSCAS/Research/Competition/2007(pdf)/200709COMPed-Gerber.pdf, 7, cited in Akman (2008a), above n 8, 5. 12 P Marsden and P Whelan, ‘“Consumer Detriment” and Its Application in EC and UK Competition Law’ (2006) 27(10) European Competition Law Review 569, 584. Case 85/76 Hoffman-La Roche AG v Commission [1979] ECR I-461, para 91; Case 322/81 Nederlansche Banden-Industrie Michelin NV v Commission [1983] ECR 3461 (‘Michelin I’), para 70; Case 6/72 Continental Can, above n 4, para 26. Case T-203/01 Manufacture Française des Pneumatiques Michelin v EC Commission [2003] ECR II-4071 (‘Michelin II’), para 239. Note, in Michelin II and Irish Sugar, in order to establish an infringement of Article 82, anticompetitive object and anticompetitive effect are one and the same thing: Michelin II, para 241; Case T-228/97 Irish Sugar plc v Commission [1999] ECR II-2969, para 170, upheld on appeal in Case C-497/99 Irish Sugar plc v EC Commission [2001] ECR I-5333; Akman (2008a), above n 8, 7.

70 Orit Dayagi-Epstein An example of the formalistic approach which is concerned only with the form of the conduct and not its effect, is the case of Akzo Chemie BV v Commission.13 There, the European Court of Justice (ECJ) found a predatory pricing practice to be unlawful, since the form of the conduct could exclude efficient competitors from the market, without considering the likelihood of this strategy succeeding in practice.14 This approach was reaffirmed by the ECJ in Tetra Pak International SA v Commission (‘Tetra Pak II’), where the Court found the form of the conduct, namely, the tying of the sales of machinery for packaging with the sales of cartons, to be an abuse, without considering the effect of this conduct.15 Likewise, in the case of Hilti, in order to establish an abuse it was sufficient to prove that consumers’ choice of buying the tied product was impaired, without examining the actual effects of the tying and how it would affect consumer interest.16 In other cases, the Community courts concentrated on the detrimental (or likely detrimental) effects on the structure of the competitive process, regardless of their direct effects on consumers. For example, in the early case of Continental Can v Commission, the ECJ held that17 the provision is not only aimed at practices which may cause damage to consumers directly, but also those which are detrimental to them through their impact on an effective competition structure.

In a similar vein, in Michelin II, the Court of First Instance (CFI) held that in order to establish an infringement, it is enough to show that the conduct of a dominant undertaking restricts competition, or is capable of having that effect, without having to prove the actual effect.18 This was also the approach in British Airways,19 where the CFI held that it was not necessary to demonstrate actual effect on consumers: Article 82 EC does not require it to be demonstrated that the conduct in question had any actual or direct effect on consumers. Competition law concentrates upon

13

Case C-62/86 Akzo Chemie BV v Commission [1991] ECR I-3359. See also Case T-228/97 Irish Sugar, above n 12, para 191; Case T-203/01 Michelin II, above n 12, para 245; Case T-219/99 British Airways v Commission [2003] ECR II-5917 (‘British Airways’), para 297. 15 Case C-333/94 Tetra Pak International SA v Commission [1996] ECR I-5951 (‘Tetra Pak II’); Cseres, above n 10, 103–04; E M Fox, ‘We protect competition you protect competitors’ (2003) 26(2) World Competition 149. 16 Case C-53/92P Hilti, above n 7. 17 Case 6/72 Continental Can, above n 4, para 26. 18 Case T-203/01 Michelin II, above n 12, paras 239 and 241; Ahlborn and Evans, above n 10, 16–17; O’Donoghue and Padilla, above n 5, 218. 19 Case T-219/99 British Airways, above n 14, paras 293 and 295. 14

The Evolution of the Notion of Consumer Interest 71 protecting the market structure from artificial distortions because by doing so the interests of the consumer in the medium to long term are best protected.20

Hence, as a general rule, the protection of consumer interest was considered a by-product of the protection of the competitive process and the prohibitions of certain conducts. This view was supported by Advocate-General Kokott’s Opinion in British Airways: Article 82 EC, like the other competition rules of the Treaty, is not designed only or primarily to protect the immediate interests of individual competitors or consumers, but to protect the structure of the market and thus competition as such (as an institution), which has already been weakened by the presence of the dominant undertaking on the market. In this way, consumers are also indirectly protected. Because where competition as such is damaged, disadvantages for consumers are also to be feared.21 (emphasis added)

In Advocate-General Kokott’s view, intervention may be justified when there is harm to the competitive structure, even if there is no direct or indirect consumer harm.22 In other cases, it was held that it is necessary to consider the effect on the market before finding an abuse of a dominant position. For example, in BPB Industries plc and British Gypsum Ltd v Commission,23 the CFI held that promotional payments made by a dominant supplier of plasterboard to a customer in return for an exclusive purchasing commitment could not be prohibited as a matter of principle and the effect on the market needed to be assessed. Having examined the conduct in question, the CFI found the exclusivity agreement to be abusive as it constituted an ‘unacceptable obstacle to entry’ to the market.24 Likewise, in Van den Bergh Foods Ltd v Commission,25 the CFI found freezer-exclusivity agreements between a dominant manufacturer of ice-cream and retailers which prevented the sale of competing brands of ice-creams (for which there was consumer demand) to be abusive as they gave rise to foreclosure effects that prevented other ice-cream manufacturers from gaining access to the market.26

20 Ibid, para 264; see also Case 85/76 Hoffman-La Roche, above n 12, para 91, where the ECJ held that Article 82 can be applied to prohibit conduct affecting the structure of the market. 21 A-G Kokott, Opinion, Case C-95/04P British Airways v Commission [2007] ECR I-2331, para 68. 22 O’Donoghue and Padilla, above n 5, 221. 23 Case T-65/89, BPB Industries plc and British Gypsum Ltd v Commission (‘British Gypsum’) [1993] ECR II-385, paras 65–66. 24 Ibid, para 68. 25 Case T-65/98 Van den Bergh Foods Ltd v Commission [2003] ECR II-4653. 26 Ibid, para 160.

72 Orit Dayagi-Epstein It should be emphasised that despite the fact that in the cases of British Gypsum and Van den Bergh the courts examined the effect of the abusive conducts, their analysis was limited to the effects of the conduct under scrutiny on the market and did not examine its effects on the interests of consumers. The influence of the Ordoliberal School on the interpretation of Article 82, in addition to the protection of the structure of the competitive process, is also evident in the emphasis on the protection of individuals’ economic freedom to have equal opportunities to participate in the market (by means of free choice and freedom to compete on the merits).27 This has led to the protection of competitors’ freedom to compete as part of the preservation of the competitive process, without examining the effects of particular conduct on the interests of final consumers.28 For example, in the case of Institute Chemioterapico Italiano SpA Commercial Solvents Corp v Commission,29 the ECJ was willing to protect the interests of a competitor without considering the effect of the conduct on the welfare of the final consumers. In that case, Commercial Solvents, a dominant firm, refused to supply Zoja, a manufacturer in the downstream market and a competitor of one of Commercial Solvents’ subsidiaries. This was found to be an abuse of a dominant position because it restricted competition in the market for the final product. The ECJ was concerned with the ‘unfair’ exclusion of a competitor from the market and the protection of the legitimate expectations of a competitor, according to which the ‘rules of the game’ cannot be changed in the middle of the game without objective justification.30 The judgment was criticised for not considering the fact that Commercial Solvents’ intention was to replace Zoja with one of its own subsidiaries (ICI). By limiting the analysis to the relationship between Commercial Solvents and Zoja, the ECJ ignored the fact that 27 W Möschel, ‘The Proper Scope of Government Views from an Ordoliberal Perspective: the Example of Competition Policy’ (2001) 157 Journal of Institutional and Theoretical Economics 3; F Böhm, ‘Rule of Law in Market Economy’ in A Peacock and H Willgerodt (eds), Germany’s Social Market Economy: Origins and Evolution (London, Macmillan Press Ltd, 1989), 57–58; W Möschel, ‘Competition Policy from an Ordo Point of View’, ibid, 146, 149, 151–52; D Hildebrand, The Role of Economic Analysis in the EC Competition Rules (The Hague, Kluwer Law International, 2002), 153–62; A H Shand, Free Market Morality: The Political Economy of the Austrian School (London, Routledge, 1989), 74–75; J Wiseman, ‘Social Policy and Social Market Economy’ in A Peacock and H Wilgerdot (eds), German New-Liberals and the Social Market Economy (London, Macmillian, 1989), 172–74; H O Lemel, ‘Evolution of the Social Market Economy’, ibid, 21; Sauter, above n 10, 28–29; Gerber (1998), above n 10, 117. 28 T Eilmansberger, ‘How to Distinguish Good from Bad Competition Under Article 82: In Search of Clearer More Coherent Standards for Anticompetitive Abuses’ (2005) 42 CML Rev 129, 138. 29 Joined Cases C-6/73 and C-7/73 Institute Chemioterapico Italiano SpA and Commercial Solvents Corp v Commission [1974] ECR 223. 30 E J Hughes, ‘The Left Side of Antitrust: What Fairness Means and Why it Matters’ (1994) 77 Marquette Law Review 265, 299.

The Evolution of the Notion of Consumer Interest 73 ICI might have been more efficient than Zoja because of the possible cost savings when a parent company supplies products to its subsidiary. These savings in turn could be passed on to the final consumers.31 Hence, in this case, restriction of intermediate customers’ freedom to compete did not necessarily restrict final consumers’ free choice, and could even have been beneficial to the interests of final consumers. However, as a general rule, the influence of the notion of economic freedom has led to the preference of competitors’ freedom to compete at the expense of the interests of final consumers.32 Nevertheless, there have been some cases where the Community courts have considered the effects of certain conduct on the interests of final consumers rather than only on the interests of intermediate customers.33 For example, in the case of ITP, RTE and BBC v Commission,34 the CFI explicitly referred to the impairment of an intermediate customer’s choice by limiting the appearance of a new product as a means of limiting final consumers’ choice. The CFI upheld the Commission’s finding, according to which ITP, RTE and BBC had abused their dominant position in the TV listings market by preventing a third party from purchasing TV listings from the broadcasting channels with a view to selling an integrated TV guide. The CFI decision was upheld by the ECJ, which ruled that a refusal to grant a licence that would block the appearance of a new product (an integrated TV guide), for which there was a potential consumer demand, constituted an abuse of a dominant position.35 As seen above, prior to the process of modernisation, the notion of consumer interest was envisioned as equivalent to the notion of economic freedom. This interpretation has led, as a general rule, to the prohibition of certain conducts due to the form or the detrimental effects of conducts on competitors and intermediate customers, or on the structure of the competitive process, without examining their actual or likely effects on final consumers. 31 Stuyck, above n 4, 323; L Lovdahl Gormsen, ‘Article 82: Where are we coming from and where are we going to?’ (2005) 2(2) Competition Law Review 5, 13–14. 32 See, eg, Hildebrand, above n 27, 107, 110–11; Möschel (1989), above n 27, 146; E M Fox, ‘Monopolization and Dominance in the United States and the European Community: Efficiency, Opportunity and Fairness’ (1981) 61 Notre Dame Law Review 981. Note also that Case 85/76 Hoffman-La Roche, above n 12, defended intermediate customers and did not consider whether the final consumer was harmed by the exclusionary conduct of restriction of output which might itself have led to higher prices. 33 Case C-418/01 IMS GmbH Co OHG v NDC Health GmbH Co KG [2004] ECR I-5039, paras 49, 51–52; Case COMP/C-3/37.792, Microsoft, Commission Decision of 24 March 2004, para 782. 34 Magill TV Guide/ITP, BBC and RTE [1989] OJ L78/43, [1989] 4 CMLR 757, upheld on appeal to the CFI in Cases T-69, T-70, T-76/89 RTE v Commission [1991] ECR II-485. 35 Cases C-241 and 242/91 P RTE v Commission [1995] ECR I-743 para 56; R Nazzini, ‘The wood began to move: an essay on consumer welfare, evidence and burden of proof in Article 82 cases’ (2006) 31(4) European Law Review 518, 533.

74 Orit Dayagi-Epstein III. THE MODERNISATION OF ARTICLE 82

A. Background The Commission’s enforcement policy on unilateral conduct was criticised for being formalistic, not grounded in sound economics36 and for its tendency to protect competitors rather than competition.37 This critique led DG Competition to initiate the process of the modernisation of Article 82 EC.38 The Discussion Paper and the Commission Guidance reflect a move towards an effects-based approach which examines the legitimacy of certain business practices according to their impact on consumer welfare.39 The process of the modernisation of Article 82 did not emerge in a vacuum but was rather part of a wider move towards a more economic-effects based approach, which has since been implemented by the Commission in respect to the regulation of mergers and Article 81.40 For example, in the merger case of AOL/Time Warner, the Commission departed from its formalistic approach and held that a foreclosure cannot automatically be assumed without closer examination of its effects.41 The Commission’s approach in AOL/Time Warner can be considered a development from its Article 82 decisions in Tetra Pak II and Hilti.42

36 C Ahlborn, ‘Competition Policy in the New Economy: Is European Competition Law up to the Challenge’ (2001) 22(5) European Competition Law Review 156; M Messina, ‘Article 82 and the New Economy: Need for Modernisation?’ (2005) 2(2) Competition Law Review 73, 74–75. 37 J A Padilla, ‘Another look at DG Comp discussion paper on the application of Article 82 to exclusionary abuses’, PowerPoint presentation, LECG Europe, Milan, 24 February 2006, http://www.lecg.com/Files/Publication/6cd6bfd5-7d71-4043-8e3a-774c0081641b/ Presentation/PublicationAttachment/147373d3-4b31-4c9d-bac0-7f183b3abf47/padilla_ milan.pdf, accessed 5 January 2009. 38 L Lovdahl Gormsen, ‘The European Court of First Instance confirms in Microsoft that it is not ready to embrace a more economics-based approach to Article 82 EC’, Centre for Competition Policy’s Newsletter, 14 May 2008, http://www.uea.ac.uk/polopoly_fs/1.104308! ccpnewsletter14.pdf, accessed 5 January 2009, 7–8; P Akman, ‘The EC Discussion Paper on the Application of Article 82’, http://ec.europa.eu/comm/competition/antitrust/art82/004.pdf, accessed 5 January 2009, 1. 39 S Bishop and P Marsden, ‘Editorial: The Article 82 Discussion Paper: A Missed Opportunity’ (April 2006) 2(1) European Competition Journal 1, 2, http://www.biicl.org/files/1103_ marsden_and_bishop_on_article_82_discussion_paper.pdf, accessed 5 January 2009. 40 The consumer welfare standard has been adopted in the Commission’s Article 81(3) Guidelines and the Horizontal Merger Guidelines. See Article 81(3) Guidelines, above n 7, paras 21, 24, 84, 96, point 3.4.3 and para 102; Horizontal Mergers Guidelines, above n 9, paras 79–86. Also P Lowe, ‘Consumer Welfare and Efficiency—New Guiding Principles of Competition Policy?’, Speech given on 27 March 2007 at the 13th International Conference on Competition and 14th European Day, http://ec.europa.eu/comm/competition/speeches/text/ sp2007_02_en.pdf, 4–5, accessed 5 January 2009 (Lowe (2007a)); O’Donoghue and Padilla, above n 5, 193. 41 AOL/Time Warner (Case IV/M1845), Commission Decision 2001/718/EC, [2001] OJ L268/28. 42 Case C-333/94 Tetra Pak II, above n 15; Case C-53/92 Hilti, above n 7.

The Evolution of the Notion of Consumer Interest 75 B. The consumer welfare standard Consumer welfare is considered by many to be the ultimate goal of competition law and a benchmark in its enforcement.43 Under the effects-based approach, exclusionary conduct will be considered a violation of Article 82 if as a whole it produces anticompetitive effects on consumer welfare. This is the case, for example, when reduced competition has adverse effects on price which are not counterbalanced by creating sufficient improvement in performances.44 Consumer welfare is measured by ‘the difference between the price a consumer is willing to pay and the price he actually pays for a good or service’.45 According to the accepted view, consumer welfare is concerned with the deadweight loss to society, namely, the loss to consumers who would have purchased the product (at a price higher than the marginal cost) but have switched to a less desired substitute as a result of monopolistic pricing (allocative efficiency, also known as the ‘triangle’), and with the loss to consumers who have continued to purchase the product at the monopolistic price (wealth transfer from consumers to producers, also known as producer welfare or the ‘rectangle’).46 The notion of consumer welfare was introduced in paragraph 4 of the Discussion Paper, which states that with regard to exclusionary abuses the objective of Article 82 is the protection of competition in the market as a means of enhancing consumer welfare and ensuring an efficient allocation of resources.47 (emphasis added)

43 See, eg, in the context of Article 82 EC, Lowe (2007a), above n 40; N Kroes, ‘European Competition Policy: Delivering Better Markets and Better Choices’, SPEECH/05/512, European Consumer and Competition Day, London, 15 September 2005, http://europa.eu/rapid/press ReleasesAction.do?reference=SPEECH/05/512&format=HTML&aged=0&language= EN&guiLanguage=en, accessed 5 January 2009. 44 O’Donoghue and Padilla, above n 5, 191; Commission Guidance, above n 1, para 11: ‘… the expression “increase prices” … is used as a shorthand for the various ways in which the parameters of competition—such as prices, output, innovation, the variety or quality or goods and services—can be influenced for the profit of the dominant undertaking and to the detriment of consumers’. 45 H Hovenkamp, ‘Legislation, Well being and Public Choice’ (1990) 57 University of Chicago Law Review 63, 72; R H Lande, ‘Wealth Transfer as the Original and Primary Concern of Antirust: The Efficiency Interpretation Challenged’ (1982) 34 Hastings Law Journal 65, 74–76; J F Brodley, ‘The Economic Goals of Antitrust: Efficiency, Consumer Welfare and Technological Progress’ (1987) 62 New York University Law Review 1020, 1021, 1033. 46 Lande, above n 45, 74–76; S Bishop and M Walker, The Economics of EC Competition Law, 2nd edn (London, Sweet & Maxwell, 2002), 24; D W Barnes, ‘Nonefficiency Goals in the Antitrust Law of Mergers’ (1989) 30 William and Mary Law Review 787, 854–55. 47 Discussion Paper, above n 2, para 4. See also ibid, pp 9, 17–18, para 54; Lowe (2007a), above n 40, 2; L Lovdahl Gormsen, ‘The Conflict between Economic Freedom and Consumer Welfare in the Modernisation of Article 82 EC’ (2007) 3(2) European Competition Journal 329, 338.

76 Orit Dayagi-Epstein It may therefore be argued that in the Discussion Paper, allocative efficiency and consumer welfare are considered as separate notions, and therefore that the notion of consumer welfare is narrower than the accepted view and refers only to the wealth transfer from consumers to producers and not to the deadweight loss.48 Such a distinction between consumer welfare and allocative efficiency is absent from the Commission Guidance, which states in paragraph 19 that The aim of the Commission’s enforcement activity … is to ensure that dominant undertakings do not impair effective competition by foreclosing their rivals in an anticompetitive way and thus having an adverse impact on consumer welfare.49

It therefore seems that the Commission Guidance does not follow the Discussion Paper’s understanding of consumer welfare, but rather implicitly adopts the accepted definition of consumer welfare, which does not distinguish between consumer welfare and allocative efficiency. Following the Discussion Paper’s interpretation of consumer welfare, the Commission might have been expected to pursue excessively high prices (unfair pricing) cases. These types of cases involve detailed examination of the prices under scrutiny, which may turn the Commission into a price regulator. This conflicts with previous Commission statements that it does not wish to pursue such cases.50 In the Commission Guidance, the Commission went on to state that while it may decide to intervene in alleged excessive prices cases51 where the protection of consumers and the good functioning of the internal market cannot otherwise be adequately ensured, at this stage it limits itself to certain types of exclusionary conduct.52 According to the Discussion Paper, the existence of exclusionary conduct requires actual or likely anticompetitive effects in the market and possible direct or indirect consumer harm. This implies that harm to the competitive process (market-distorting foreclosure effect) and harm to consumers are not considered to be identical but rather cumulative.53 The second requirement, namely ‘harm to consumers’, which is a new development compared with the case law, derives from the presumption that unless there is or is 48 This interpretation is in line with Article 81(3) EC Guidelines, above n 7, paras 21, 46; Akman (2008a), above n 8, 12. 49 Commission Guidance, above n 1, para 19. 50 European Commission XXIVth Annual Report on Competition Policy (1994), para 207, cited in Akman (2008a), above n 8, at 12; P Lowe, ‘Concluding Remarks’, European University Institute, Robert Schuman Centre for Advanced Studies 2007 EU Competition Law and Policy Workshop/Proceedings (Brussels, 9 June 2007), to be published in CD Ehlermann and M Marquis (eds), European Competition Law Annual 2007: A Reformed Approach to Article 82 EC (Oxford, Hart Publishing, in preparation), http://www.iue.it/RSCAS/Research/ Competition/2007(pdf)/200709-COMPed-Lowe.pdf, accessed 5 January 2009, 5. 51 O’Donoghue and Padilla, above n 5, 195. 52 Commission Guidance, above n 1, para 7. 53 Discussion Paper, above n 2, paras 54, 55, 56, 58, 88; Akman, above n 38, 8; O’Donoghue and Padilla, above n 5, 200.

The Evolution of the Notion of Consumer Interest 77 likely to be consumer harm, the harm to the competitive process is no longer relevant.54 The Commission Guidance seems to adopt a similar approach regarding exclusionary conduct, and states: The aim of the Commission’s enforcement activity in relation to exclusionary conduct is to ensure that dominant undertakings do not impair effective competition by foreclosing their rivals in an anticompetitive way and thus having an adverse impact on consumer welfare.55

Interestingly, despite the importance afforded to ‘harm to consumers’, there is no definition of this concept or of consumer welfare either in the Discussion Paper or in the Commission Guidance, although both set out several examples of adverse and beneficial effects on consumer welfare, such as prices, quality, innovation and consumer choice.56 The absence of these definitions is in keeping with their absence (together with the definition of ‘consumer benefit’) from the EC Treaty and from other soft law provisions.57

C. The standard of harm under Article 82 Arguably, the consumer welfare standard can become a benchmark in the enforcement of Article 82 EC only once the Commission’s statements are translated into clear enforcement criteria regarding the standard of harm that should be applied. Such criteria would include the manner in which consumer harm should be assessed, how conduct pursued by a dominant undertaking leads to foreclosure of competitors, and the appropriate separation between consumer harm, competitor harm and harm to competition. This is particularly important in order to avoid the tendency to protect competitors rather than competition, and thereby consumers.58 The absence of clear criteria is evident in the Discussion Paper, where there seems to be an ambiguity regarding the standard of harm that should be applied. For example, paragraph 54 of the Discussion Paper adopts the position that the protection of the interests of consumers is considered a by-product of the protection of the competitive process: The concern is to prevent exclusionary conduct of the dominant firm which is likely to limit the remaining competitive constraints on the dominant company …

54 55 56 57 58

O’Donoghue and Padilla, above n 5, 221. Commission Guidance, above n 1, para 19. Ibid, paras 5, 11, 19, 86; Discussion Paper, above n 2, para 4. Marsden and Whelan, above n 12, 584. Bishop and Marsden, above n 39, 3, 6.

78 Orit Dayagi-Epstein so as to avoid that consumers are harmed. This means that it is competition and not competitors as such, that is to be protected.59 (emphasis added).

However, despite the explicit statement above, a detailed reading of the Discussion Paper reveals a somewhat less clear-cut position on this matter. The definition of ‘foreclosure’ in paragraph 58 of the Discussion Paper concentrates on the effects on competitors rather than on harm to competition.60 This may obscure the manner in which harm to competition should be differentiated from harm to competitors, and implies the continuing use of the formalistic approach which implicitly assumes that harm to the competitive process necessarily harms consumers, without providing a clear alternative to the assessment of the anticompetitive effects.61 Nevertheless, the assumption that the protection of the competitive process via the protection of competitors benefits consumers is not always sound. For instance, under the ‘meeting competition defence’,62 a dominant firm will be allowed to react proportionately in response to a lowering of prices by competitors but it will not be allowed to beat its competitors (say, by further reduction of prices). This protects competitors at the expense of consumer welfare, since consumers will no longer be able to benefit from a further reduction of prices.63 Arguably, despite the central role which is associated with the consumer welfare standard in the Discussion Paper,64 the test proposed therein hardly seeks proof of any actual or possible harm to consumers, and does not address how the assessment of the effects on consumers is to be made.65 Indeed, instead of concentrating on likely effects on the structure of the market and competitors, the Discussion Paper should have concentrated on actual market evidence.66 The latter may include empirical evidence regarding customers’ behaviour (eg, in the case of loyalty rebates) and competitors’ market shares which may indicate whether the conduct under

59 Discussion Paper, above n 2, para 54; N Kroes, European Commissioner for Competition Policy, ‘The Commission’s Review of Exclusionary Abuses of Dominant Position’, Speech before the Korean Competition Forum organised on the occasion of the Fourth Annual Bilateral Meeting on 26–27 June 2006 in Seoul, 5, http://ec.europa.eu/competition/speeches/ text/sp2006_012_en.pdf, accessed 5 January 2009; Akman (2008a), above n 8, 5, n 18; N Kroes, European Commissioner for Competition Policy, ‘Preliminary Thoughts on Policy Review of Article 82’, Speech at the Fordham Corporate Law Institute New York, 23 September 2005, http://europa.eu/rapid/pressReleasesAction.do?reference=SPEECH/05/537, accessed 5 January 2009, 3. 60 Para 58 of the Discussion Paper, above n 2, states: ‘By foreclosure is meant that actual or potential competitors are completely or partially denied profitable access to a market.’ 61 Discussion Paper, above n 2, paras 55, 56, 58; Padilla, above n 37; Akman, above n 38, 2. 62 Discussion Paper, above n 2, para 81. 63 Ibid, para 82; Bishop and Marsden, above n 39, 6. 64 Discussion Paper, above n 2, paras 4, 54, 55, 88. 65 Akman (2008a), above n 8, 11. 66 O’Donoghue and Padilla, above n 5, 222–24. For the importance of sound economic analysis which is based on facts, see EAGCP Report, above n 3, 13.

The Evolution of the Notion of Consumer Interest 79 scrutiny is exclusionary. For example, an increase in a competitor’s market share may indicate that the conduct under scrutiny does not foreclose competition.67 The Commission Guidance took on board some of the above critique and spells out the main theories of harm to competition and the corresponding identification of key facts. In this vein, the Commission Guidance sets out the considerations which will be taken into account when determining whether to intervene. Among these considerations the Commission Guidance lists the degree of dominance and the position of competitors, network effects and entry barriers, the position of customers and input suppliers, the extent of the allegedly abusive conduct, possible evidence of actual foreclosure and evidence of an explicit exclusionary conduct.68 The Commission Guidance also addresses the standard of harm that should be applied under Article 82, and reveals greater consumer orientation and rejection of competitor protection. For instance, paragraph 6 of the Commission Guidance states that what really matters is to protect an effective competitive process and not simply protecting competitors. This may well mean that competitors who deliver less to consumers … will leave the market.69

Indeed, the component of consumer harm receives a central role in the Commission Guidance, as can be learned from the general approach towards exclusionary conduct (the definition of anticompetitive foreclosure relates to ‘foreclosure leading to consumer harm’)70 and from the fact that consumer harm constitutes a vital component of specific forms of abuse (such as in the case of refusal to supply and margin squeeze).71 In addition, the Commission Guidance elaborates on how consumer harm can be identified, and states that ‘likely consumer harm can rely on qualitative and, where possible and appropriate, quantitative evidence’.72 The latter suggests the intended use of economic evidence to substantiate the findings of likely consumer harm. Such an elaborate consumer harm-based approach may imply the introduction of a higher burden of proof regarding likely consumer harm which will be required in order to support findings of an abuse. Despite the above, it should be noted that the Commission Guidance at times upholds the formalistic approach. For example, it states that 67 Bishop and Marsden, above n 39, 4. However, see Case T-219/99 British Airways, above n 14, para 298, where the fact that British Airways’ competitors gained market shares during the period of the alleged abuse did not affect the CFI’s conclusion regarding the exclusionary conduct. 68 Commission Guidance, above n 1, para 20. 69 Ibid, para 6. 70 Ibid, para 19. 71 Ibid, paras 85–87. 72 Ibid, para 19.

80 Orit Dayagi-Epstein in cases that where it appears that a conduct can only raise obstacles to competition and that it creates no efficiencies, it may not be necessary for the Commission to carry out a detailed assessment before concluding that the conduct in question is likely to result in consumer harm. According to the Commission Guidance, such conducts may occur when the dominant undertaking prevents its customers from testing competitors’ products, or provides financial incentives to its customers on condition that they do not test such products, or pays a distributor or a customer to delay the introduction of a competitor’s product.73 Importantly, the Commission Guidance is more consistent in its interpretation of the consumer welfare standard. An example of the inconsistency of the Discussion Paper in relation to the standard of harm is the case of predation. The fact that predation is condemned as such without the need of proving the possibilities of recoupment,74 may indicate the adoption of the total welfare standard which considers pricing below marginal costs to be inefficient. However, as discussed above, according to the Discussion Paper, the notion of consumer welfare comprises only wealth transfer. Accordingly, predation will not be condemned as long as there is no recoupment, since predation enhances consumer welfare in the short-run by enabling consumers to benefit from lower prices.75 However, the Commission Guidance presents a consistent view in respect to the standard of harm in the case of predation, and states that consumers are likely to be harmed where the dominant undertaking incurs a ‘sacrifice’ (making an avoidable loss or forgoing profits). Pricing below average avoidable costs (AAC), or pricing above AAC which leads to a loss that could be avoided (pricing below average total costs), will be presumed to constitute such a sacrifice. Where the dominant undertaking is ‘likely to be in a position to benefit from the sacrifice’ (ie to recoup) in the form of increased prices, delay or prevention of a fall in prices, consumers are likely to be harmed and the Commission is likely to intervene.76 The requirement of recoupment in the Commission Guidance provides further evidence for the preference for the consumer welfare standard in the analysis of exclusionary conducts. As pointed out in the Commission Guidance, the notion of ‘consumer’ under the consumer welfare standard encompasses direct and indirect users of the product affected by the conduct.77 In other words, the notion of ‘consumer’ encompasses competitors, intermediate customers and final consumers. This broad definition may obscure the manner in which harm

73 74 75 76 77

Ibid, para 21. Discussion Paper, above n 2, para 122; Akman (2008a), above n 8, 12. Akman (2008a), above n 8, 12. Commission Guidance, above n 1, para 69. Ibid, p 8, n 15.

The Evolution of the Notion of Consumer Interest 81 to consumers can be identified. Conscious of the fact that harm or benefit to intermediate customers will not necessarily affect the interests of the final consumers, the EAGCP Report pointed out the need to explain the trade-offs between different groups of consumers. However, the Discussion Paper does not refer to this point and seems to assume that the impact on the interests of intermediate customers will eventually coincide with those of the final consumers.78 Unfortunately, this is not always the case. For example, when intermediate customers benefit from a certain degree of market power, they may not pass on the efficiency gains in the form of lower prices to the final consumers.79 Unlike the Discussion Paper, the Commission Guidance seems to distinguish between anticompetitive foreclosure effects at the intermediate level and at the level of the final consumer.80 It also differentiates between situations in which the intermediate users (customers) ‘are actual or potential competitors of the dominant undertaking,’ where ‘the assessment focuses on the effects of the conduct on users further downstream,’ and cases in which the intermediate customers are not competitors of the dominant undertaking.81 Arguably, in the latter case, the assessment of the effects on consumers further downstream is not mandatory.82 The recognition of the need to assess the effects of the conduct under scrutiny on the interests of the final consumers when the intermediate customers are competitors of the dominant undertaking, is another indication of the importance awarded in the Commission Guidance to the consumer welfare standard. Had this approach been applied in the case of Commercial Solvents,83 where a competitor’s freedom to compete was preferred over the welfare of the final consumers, the Court might have reached a different verdict. However, it may be argued that the Commission Guidance should have gone even further by accepting that when intermediate customers are not competitors of the dominant undertaking, their interests do not necessarily coincide with those of the final consumers.84 For example, this may be the situation in the case of non-linear pricing such as rebates, discounts and tying practices, where the price paid varies according to the number 78

EAGCP Report, above n 3, 9–10; Discussion Paper, above n 2, para 55. K J Cseres, ‘The Controversies of the Consumer Welfare Standard’ (2007) 3(2) Competition Law Review 121, 132. For a discussion of other instances in which final consumer welfare and intermediate customer welfare clash, see P Akman, ‘“Consumer” versus “Customer”: the Devil in the Detail’, CCP Working Paper 08-34, http://www.uea.ac.uk/polopoly_fs/1.104698! ccp08-34.pdf, accessed 5 January 2009, 14–25 (Akman (2008b)). 80 Commission Guidance, above n 1, para 19. 81 Ibid, para 19, p 8, n 15. 82 Akman (2008b), 4, 10. 83 Joined Cases C-6/73 and C-7/73 Commercial Solvents, above n 29. 84 An example of a case in which the interests of intermediate customers and final consumers coincide is Magill TV Guide, above n 34. 79

82 Orit Dayagi-Epstein of units purchased, and the (average) price per unit changes according to the number of units purchased.85 This derives from the fact that non-linear pricing may increase the intermediate customer’s opportunity costs of purchasing from another seller. This situation was apparent in the case of British Airways, in which the CFI and the ECJ held that increased commission bonuses granted by British Airways to travel agents selling its tickets, on the basis of a comparison of sales between the current month and the corresponding month in the previous year, constituted an abuse. The fact that the bonuses were awarded in respect of all the tickets that were sold was found to restrict the agents’ freedom to sell their services to other airlines, and thereby hindered the access of competitors.86 Hence, although the Courts based their judgments in this case on the foreclosure of competitors (other airlines) and did not consider the detriment to the final consumers, the case demonstrates that the interests of the intermediate customers do not necessarily coincide with the interests of the final consumers, as the travel agents (intermediate customers) benefited from the bonuses while the final consumers’ welfare was impaired. IV. THE IMPLEMENTATION OF THE EFFECTS-BASED APPROACH BY COMMUNITY COURTS

The Discussion Paper and the Commission Guidance are not considered authoritative sources,87 therefore any prospective change in the enforcement policy of Article 82 towards a more economic effects-based approach will be dependent upon the support of the Community courts.88 However, as is evident from some recent judgments,89 it seems that the Community courts are not yet ready to embrace fully an economic-based approach which concentrates on the consumer welfare standard as the guiding principle in the context of Article 82. For example, in British Airways,90 although the ECJ mentioned the need to establish some form of competitive impact and an appreciation of 85 Akman (2008b), above n 79, 6–7; M Motta, Competition Policy Theory and Practice (Cambridge, CUP, 2004), 307; D W Carlton and J M Perloff, Modern Industrial Organization, 3rd edn (Reading MA, Addison Wesley Longman, 1999), 274, 297, cited in Akman (2008b), 6, n 13. 86 Case T-219/99 British Airways, above n 14, paras 234, 270; Akman (2008b), above n 79, 27. 87 According to Article 249 EC, only regulations, directives and decisions are legally binding measures in the EU. See Lovdahl Gormsen, above n 47, 330. 88 A-G Kokott’s Opinion in Case C-95/04P British Airways, above n 21, para 28; D S Evans and C Grave, ‘The Changing Role of Economics in Competition Policy Decisions by the European Commission during the Monti Years’ (2005) 1(1) Competition Policy International 133. 89 Case C-95/04P British Airways, above n 21; Case T-201/04 Microsoft Corp v Commission [2007] ECR II-3601; Case T-340/03 France Télécom SA (formerly Wanadoo Interactive SA) v EC Commission [2007] 4 CMLR 21. 90 Case C-95/04P British Airways, above n 21, para 107.

The Evolution of the Notion of Consumer Interest 83 economic benefit, it nevertheless followed the rationale in the early case of Continental Can, according to which practices that harm the effective competitive structure may be considered to be indirectly detrimental to consumers in both the medium and long term,91 and restated that there is no requirement of proving direct harm to consumers.92 In France Télécom SA (formerly Wanadoo Interactive SA) v Commission, which dealt with predatory pricing in the market for ADSL-based Internet access services, the CFI adopted a formalistic approach and held that lack of consumer harm (as the predatory pricing scheme benefited consumers in the short run) did not prevent a finding of abuse.93 In addition, the Court held that it was not necessary to show a realistic chance of recoupment in order to establish an abuse.94 The CFI’s preference for the formalistic approach stands in contrast to the effects-based approach which is advocated in the context of predation in the EAGCP Report95 and the Commission Guidance.96 In September 2007, the CFI upheld the Commission’s decision in the case of Microsoft.97 The Commission found that Microsoft had breached Article 82 EC by refusing to supply interoperability information to competing providers of work group server operating systems (thereby hindering the ability of its competitors to develop and distribute interoperable work group server operating system products),98 and by tying its Windows Media Player (WMP) with its Windows PC operating system. The Commission’s decision on the tying component embraced a rule of reason approach and held that tying of WMP to Windows could not automatically be assumed to foreclose competition without examining the effect of the tying on consumer welfare.99 In its decision the Commission deviated from the judgments in Tetra Pak II and Hilti,100 where it was sufficient to prove that consumer choice was impaired as a result of the tying, without examining the effect of the tying on consumer welfare. Although on the merits the CFI upheld the Commission’s finding of abuse, it chose to follow the formalistic approach which was adopted in the cases of Hoffman-La Roche AG v Commission, British Airways and Michelin II,

91

Ibid, paras 106, 264. Ibid, para 30; Case T-219/99 British Airways, above n 14 paras 107, 293. 93 Case T-340/03 France Télécom, above n 90, paras 266, 195–97. 94 Ibid, paras 227, 224–28. 95 EAGCP Report, above n 3, 50–53. 96 Commission Guidance, above n 1, para 69. 97 Microsoft, above n 33; Case T-201/04 Microsoft, above n 90. 98 Microsoft, above n 33, recitals 546, 794–95. 99 Ibid, recitals 839, 841, 891, 955; Case T-201/04 Microsoft, above n 90, paras 971, 1034, 1040, 1058. 100 Case C-333/94 Tetra Pak II, above n 15; Case C-53/92 Hilti, above n 7. 92

84 Orit Dayagi-Epstein and reaffirmed that no direct effects on consumers were required for the finding of an abuse:101 It is settled case law that Article 82 EC covers not only practices which may prejudice consumers directly but also those which indirectly prejudice them by impairing an effective competitive structure … In this case, Microsoft impaired the effective competitive structure on the work group server operating systems market by acquiring a significant market share on that market.

In respect to the tying segment, the CFI held that it was sufficient to demonstrate the advantage over competitors (that did not have the ability to bundle) in order to satisfy the ‘foreclosure of competition’ element of the tying. The effect of such an advantage on competition, and thereby on consumer welfare, will then be presumed and need not be established separately.102 Likewise, in the refusal to supply segment, the CFI followed the ECJ’s cumulative test which was applied in the cases of Magill TV Guide103 and IMS Health,104 and focused on the fact that competitors were placed at a competitive disadvantage, rejecting the need to establish any actual or likely harm to consumer welfare.105 The post-Discussion Paper’s approach of the Community courts in respect of Article 82 EC stands in contrast to the shift towards a more economic-based approach in the context of Article 81 EC and merger control. A striking example of this is the two GlaxoSmithKline cases, one at the CFI concerning a breach of Article 81 EC and the other at the ECJ concerning a breach of Article 82 EC. Despite the similarity of the facts in both cases, a breach was found in only one of them. The CFI case (GlaxsoSmithKline Services v Commission)106 considered a dual pricing mechanism clause in agreements between a producer and its wholesalers that restricted parallel imports between Spain and other Member States. The CFI analysed the question of whether the agreements had the effect of adversely affecting the welfare of the final consumers.107 101 Case 85/76 Hoffman-La Roche, above n 12; Case T-203/01 Michelin II, above n 12; Case C-95/04P British Airways, above n 21; Microsoft, above n 33, para 839; Case T-201/04 Microsoft, above n 90, paras 1089–90; Ahlborn and Evans, above n 10, 13. 102 Case T-201/04 Microsoft, above n 90, paras 1055, 1058; Ahlborn and Evans, above n 10, 14, 18–19. 103 Magill TV Guide, above n 34. 104 Case C-418/01 IMS Health, above n 33; Case T-201/04 Microsoft, above n 90; Ahlborn and Evans, above n 10, 6, n 35. 105 Case T-201/04 Microsoft, above n 90, 19–20. 106 Case T-168/01 GlaxsoSmithKline Services v Commission [2006] ECR II-2629. 107 Ibid, paras 171, 118–19, 244, 273. In this respect it may be argued that CFI’s statement stands in contrast to para 84 of the Article 81(3) Guidelines, above n 7, according to which, ‘[c]onsumers within the meaning of Article 81(3) are the customers of the parties to the agreement and subsequent purchasers’. See also Joined Cases T-213/01 and 214/01 Österreichische Postsparkasse AG and Bank für Arbeit und Wirtscaft AG v EC Commission [2006] ECR II-1601, para 115, where in the context of Article 81 EC the CFI held that the ultimate goal of the competition rules is to increase the well-being of consumers.

The Evolution of the Notion of Consumer Interest 85 In September 2008, the ECJ handed down its judgment (GlaxoSmithKline AEVE Farmakeftikon Proionton)108 in a dispute between GlaxoSmithKline and a number of independent Greek wholesalers of prescription medicines regarding refusal to supply and the supply of restricted quantities of medicines, effectively restricting the wholesalers’ ability to export to Member States in which medicine was sold for higher prices than in Greece. Relying on the cases of United Brands v Commission109 and Commercial Solvents,110 the ECJ held that a refusal to supply medicines to wholesalers in order to prevent parallel exports constitutes an abuse unless the practice is considered reasonable and proportionate in order to protect the company’s legitimate commercial interests. The ECJ concentrated on the form of the conduct rather than on the effects of the conduct on the final consumers. V. CONCLUSIONS

The chapter has examined the evolution of the notion of consumer interest in the context of Article 82 EC. According to the traditional formalistic approach, alleged abusive conducts were assessed, as a general rule, according to their likely effects on the competitive structure of the market or the form of the conduct, regardless of their actual effects on the final consumers.111 In this framework, consumer interest was interpreted as equivalent to the notion of economic freedom. This view of the consumer interest, together with the lack of differentiation between competitor, intermediate customers and final consumers’ freedom, led, at times, to the preference of competitors’ interests over the interests of final consumers. The new effects-based approach to Article 82 EC, which was introduced by the Commission as part of the Discussion Paper and the Commission Guidance, emphasises the consumer welfare standard and deviates from the formalistic approach through the requirement to provide evidence regarding actual or likely anticompetitive effects in the market and possible direct or indirect consumer harm when establishing an abuse. Despite the emphasis on the consumer welfare standard in the Discussion Paper, a detailed reading of it reveals some inconsistencies regarding the standard of harm that should be applied. In addition, the Discussion Paper rarely seeks proof of any actual or possible harm to consumers, and does not explain how the assessment of the effects on consumers is to be made. 108 Joined Cases C-468/06 to C-478/06 GlaxoSmithKline AEVE Farmakeftikon Proionton formerly Glaxowellcome AEVE, judgment dated 16 September 2008. 109 Case C-27/76 United Brands v Commission [1978] ECR 207. 110 Joined Cases C-6/73 and C-7/73 Commercial Solvents, above n 29. 111 Case C-62/86 Akzo, above n 13; Michelin II, above n 12; Case 85/76 Hoffman-La Roche, above n 12.

86 Orit Dayagi-Epstein The Commission Guidance attempts to address some of these flaws, and appears to present greater consumer orientation and rejection of competitors’ protection. Within this framework, the Commission Guidance has differentiated between harm to intermediate customers that are competitors of the dominant undertaking and harm to final consumers. However, arguably, the Commission Guidance should have gone a step further and not assumed that when intermediate customers are not competitors of the dominant undertaking, their interests coincide with those of the final consumers, as this may not always be the case. Despite the developments above, the Community courts chose not to take part in the modernisation process of Article 82 EC and continue to implement a formalistic approach. This is perhaps not surprising, considering the incoherence of the Discussion Paper regarding the standard of harm which should be applied in the case of unilateral conduct. It is quite possible that the attempt to address some of the flaws of the Discussion Paper in the Commission Guidance will encourage the Community courts to reconsider their position on this matter, and may lead to greater consistency between the courts and the Commission.112 It will indeed be interesting to see if the Community courts, national competition authorities and national courts will take the opportunity to use the Commission Guidance, which is not binding upon them, and implement it in a manner which will place Article 82 EC within the modernisation process already witnessed in the areas of restrictive agreements and merger control.

112

Lowe, above n 50, 2.

5 The Epithet That Dares Not Speak its Name: The Essential Facilities Concept in Article 82 EC and IPRs after the Microsoft Case STEVEN ANDERMAN*

I. INTRODUCTION

T

HE CASE OF Microsoft v Commission1 in the Court of First Instance (CFI) has raised in bold relief the question of what is left of the ‘exceptional circumstances’ test. This test determines when the refusal to license an Intellectual Property Right (IPR) by its owner can be regarded as infringing Article 82 EC. The competition rules stress that in ‘normal’ cases the owner of an IPR is entitled to refuse to license because exclusivity is the very substance of the right. However, in ‘exceptional’ circumstances, an IPR owner may be subject to a competition law remedy of compulsory licensing because the refusal infringes Article 82. Prior to Microsoft, the test of ‘exceptional circumstances’ tended to be defined by the criteria established by the European Court of Justice (ECJ) in its Magill2 and IMS Health judgments.3 After Microsoft, it appears that the test of exceptional circumstances may have been widened to include another parameter, which extends the test under Article 82(b) to certain forms of conduct limiting the technical development of secondary markets. However wide the exceptional circumstances test of abuse may now be, it is important to reiterate that this test is itself subject to a further, unusual precondition: the special nature of the dominance that must be established as a precondition to a finding of a refusal to supply/license abuse. Described *

Professor of Law, University of Essex. Case T-201/04 Microsoft v Commission [2007] ECR II-3601. 2 Joined Cases C-241/91P and C-242/91P RTE and ITV v Commission [1995] ECR I-743. 3 Case C-418/01 IMS Health GmbH v NDC Health GmbH & Co KG [2004] ECRI-5039. 1

88

Steven Anderman

as ‘essential facilities’ dominance or ‘super-dominance’, this precondition is exceptional both in terms of the nature of its definition of dominance and in terms of the effect of this type of dominance upon the concept of abuse. II. THE NATURE OF ESSENTIAL FACILITIES DOMINANCE

As we know all too well, for most forms of abuse, the concept of dominance in Article 82 often seems to be a separate finding before the concept of abuse is considered. Once an undertaking is found to be dominant, the investigation moves on to determine whether the undertaking’s conduct amounts to an abuse. In most cases, the degree of dominance seems to be relatively unimportant as long as it is established that the undertaking’s market power exceeds the legal definitional threshold of dominance. The actual degree of dominance in the cases has varied from market shares of 39 per cent to 100 per cent.4 In certain refusal to supply cases, however, including those involving compulsory licensing of IPRs, there is a definite link between the nature of the degree of dominance and the concept of abuse. In such cases, the nature of the dominance required is that the product protected by an IPR must be an ‘essential facility’, or rather, as the ECJ has chosen to define it, both a de facto monopoly, with no available substitutes, and an indispensable input into a second market. If the product is a monopoly but is not an indispensable input into a second market, its owner has no obligation to supply or license unless there is an abuse of the United Brands type.5 The reasons for the linkage with a particular finding of dominance is that once a product reaches the stage that its market power or degree of dominance amounts to an essential facility, a refusal to supply or license can be abusive because it will result in the exclusion of access or elimination of competition in the markets dependent upon the input. Yet at the same time, the imposition of a stringent definition of dominance in such cases ensures that regulation is kept to the margin of business conduct. The tight definition of dominance creates a wide scope for lawful refusals to supply or license. Anyone reading Advocate-General Jacob’s Opinion, or the judgment of the ECJ, in Oscar Bronner6 can see evidence of a desire to limit the scope of essential facilities reasoning to truly egregious cases of misuse of market power. In the context of Article 82 cases involving IPRs, essential facilities dominance has become something of an ‘epithet that dares not speak its 4 See, eg, Case T-219/99 British Airways v Commission [2003] ECR II-5917 (BA found dominant with a 40–46% market share); and Case COMP/38.233 Wanadoo Interactive (dominant with a market share of 39%). 5 Case 27/96 United Brands v Commission [1978] ECR 207. 6 Case C-7/97 Oscar Bronner GmbH & Co KG v Mediaprint [1998] ECR I-7791.

The Epithet That Dares Not Speak its Name 89 name’.7 Instead, ‘de facto monopoly’ and ‘indispensable’ are the phrases used by the courts. Yet whatever the label, the stricter definition of dominance has been recognised as essential in cases involving an IPR, as well as in other refusal to supply cases under Article 82. Certainly, it serves to remind us that IPRs are not being subjected to the bad old days of competition law when IPRs were viewed as monopolies because the legal monopoly was mistakenly equated with market power. Now, a refusal to license by an IPR owner can be challenged only if the IPR protects a product which meets the test of essential facility dominance. III. THE REGULATORY CONCEPT OF ESSENTIAL FACILITIES IN THE EU

The concept of ‘essential facilities’ now prevalent in EU competition law had its origins in the USA in 1912, in the case of Terminal Railroad Association.8 It was viewed as a variation of refusal to supply under Section 2 of the Sherman Act which prohibits the acquisition or maintenance of monopoly power. However, the US essential facilities doctrine, now severely restricted by the US Supreme Court,9 is a single market concept. In contrast, the essential facilities doctrine in the EU has evolved as a two market concept: a variation of the abuse of refusal to supply by a dominant firm into a second, dependent market. The foundations for this exclusionary abuse were established in Commercial Solvents,10 where a refusal to supply by a firm enjoying an essential facility type of dominance was found to infringe Article 82 because it eliminated competition in a secondary dependent market. Elimination of competition was held to be a fortiorari a distortion of competition. The abuse applies where: a) there are two well-defined markets; b) the owner of an asset in the first market enjoys a monopoly which is indispensable to competitors in a secondary dependent market (or to another level of economic activity). Its essence is the concern that the refusal will be used as a means to leverage a monopolist’s power in one market to exclude competitors and competition in a second market. It often presupposes narrow market definitions by the regulatory authority. 7 See Areeda, ‘Essential Facilities: An Epithet in Need of Limiting Principles’ (1990) 58 Antitrust Law Journal 841. 8 United States v Terminal Railroad Assn of St Louis, 224 US 383 (1912). 9 See Verizon Communications v Trinko 124 S Ct 872 (US). 10 Cases 6&7/73 ICI and Commercial Solvents v Commission [1974] ECR 223.

90

Steven Anderman

Once the constituent elements of essential facility dominance are established, the doctrine applies quite widely in EU law and regulation. It applies in telecoms and other regulated sectors, underpinning sector-specific regulation of privatised monopolies.11 It has resulted in a number of cases compelling access to ports, railways and other physical facilities.12 And it applies to the interface between IPRs and the competition rules. There are controversial features of the doctrine in the EU as well as in the USA. It raises issues about the appropriateness of competition law to apply compulsory access rules to successful private firms, who have built up a thriving business based on their acquisition of IPRs, their own entrepreneurial acumen and their carefully weighted risk-taking, as opposed to the inheritors of national monopolies. Its presence in competition law raises issues about the pricing of compulsory access, with the methodology of the FRAND formula remaining rather obscure to courts and even competition agency officials.13 Nevertheless, the essential facilities doctrine reflects a need to further the purposes of EC competition law so as to maintain effective competition on markets and to prevent the possession of overwhelming market power in one market to be used to leverage the dominant firm into dominance, or even monopoly, in a dependent market. In the context of IPRs, it is significant that the definition of the nature of dominance does not focus on the prior history of the property or its owner. So whether the owner’s asset is a privatised nationalised monopoly or a product developed from the start, investment in research and development and IP protection is irrelevant to a finding of its market power. As long as the product is both found to enjoy a monopoly in an economic sense, ie that there is no substitute available, and to be an indispensable input into a secondary dependent market, it meets the definition. The evidence for indispensability consists of proof of a lack of available substitutes and a finding that the competitor cannot itself invest and replicate the facility of the incumbent firm. In the Bronner case, Advocate-General Jacobs was concerned to ensure that the definition of indispensability was sufficiently tight to prevent opportunistic conduct by competitors. The ECJ agreed, and offered a test of indispensability that meant that only if it were impossible to replicate the ‘essential’ product by an undertaking with the resources

11 See, eg, Notice on the Application to Access Agreements in the Telecommunication Sector [1998] OJ C265/2. 12 See Sea Containers Ltd/Stena Sealink [1994] OJ L15/8; Port of Rodby [1994] OJ L55/52. See also Temple Lang, ‘Defining Legitimate Competition: Companies’ Duties to Supply Competitors and Access to Essential Facilities’ (1994) 18 Fordham International Law Journal 437; Ridyard, ‘Essential Facilities and the Obligation to Supply Competitors’ [1996] European Competition Law Review 438. 13 The FRAND formula (fair, reasonable and non-discriminatory terms) refers to the obligation that is often required by Standards Setting Organisations (SSOs) for members which participate in the standard-setting process.

The Epithet That Dares Not Speak its Name 91 of the incumbent would the product be viewed as indispensable.14 This suggested either a physical impossibility (ports), or a legal barrier (IPRs or regulation). In Magill,15 the Court of Justice found that the TV companies enjoyed a de facto monopoly over the listings, which was an indispensable input into the TV guide market. The existence of a copyright in the programme listings operated as a barrier to the competitor, Magill, obtaining access to the listings, and required a remedy of compulsory licensing rather than a simple order to supply. In Microsoft,16 a further factor proving indispensability was the finding of a network effects barrier to entry. This factor meant that any competitor attempting to replicate the Windows product would fail unless it could win over the tens of thousands of applications makers to interoperate with it. One significant aspect of the Microsoft case was that the monopoly enjoyed by Microsoft was an economic one, underpinned not only by the owner’s IP in the Windows software but also by the applications barrier to entry that prevented substitution by competitors. IV. THE SPECIAL RESPONSIBILITIES OF AN OWNER OF AN ESSENTIAL FACILITY

Since Michelin I,17 the special responsibilities of a dominant undertaking have been well documented in the jurisprudence of Article 82. What is sometimes less clear are the special responsibilities of an owner of an essential facility. The owners of essential facilities not involving IPRs seem to have a duty to supply new entrants as well as existing customers, simply by virtue of the essential facility dominance of their products. In other words, the nature of the dominance defines the abuse. Only in the case of IPRs is there a further test of exceptional circumstances before a refusal to supply/license can be considered to infringe Article 82. What is at stake, in the concept of abuse, is the need to monitor undertakings enjoying ownership of an essential facility. Essential facility owners may compete in secondary markets but, if so, they must compete on the merits, and that includes supplying competitors who wish access to that secondary market. It would be an abuse for an owner of an essential facility to leverage its market power in the first market by refusing to supply without objective justification, or to supply on discriminatory terms and perhaps even to supply on any terms other than FRAND terms. On the other hand, the incumbent owner of the essential facility can reduce or eliminate

14 15 16 17

Case C-7/97 Oscar Bronner (n 6). Above n 2. Above n 1. Case 32/81 Michelin v Commission [1983] ECR 3461.

92

Steven Anderman

competition on the second market by competing on price (at non-predatory levels), on quality and on the basis of brand loyalty. V. THE EXCEPTIONAL CIRCUMSTANCES TEST AND IPRS

Owners of essential facilities where IPRs are not involved may have special responsibilities that extend to offering access to competitors whether new or old. As mentioned above, the existence of the special form of dominance dictates the special responsibility under Article 82. In the case of IPRs, however, there must be a further finding of exceptional circumstances in addition to the finding of essential facility dominance. This additional constituent element to the abuse reflects the accommodation with IPRs incorporated in the Article 82 case law. Article 82 recognises both the legitimacy of the ‘existence’ of IPRs, ie the criteria chosen by Member States for the grant of such rights, and certain features of their ‘exercise.’ In Volvo v Veng,18 the ECJ indicated that the ‘normal exercise’ of exclusivity associated with IPRs would be respected under Article 82 in the primary market of manufacturing and selling new Volvo spare parts. The Court’s reasoning was that the right of an IP owner to prevent third parties from manufacturing and selling or importing, without consent, products incorporating the IP right, constitutes the very subject matter of his exclusive right, and any other view would deprive him of ‘the substance on his exclusive right’.19 However, in a secondary dependent market, the dominant IPR owner has special responsibilities. For example, he may not arbitrarily refuse to supply spare parts to independent repairers in the maintenance market. In Magill,20 the ECJ went on to maintain that where a new entrant in a secondary market asked for access to an IP-protected essential input, it was necessary to show that there must be exceptional circumstances present before Article 82 could be used to override the free exercise of an IPR. The exceptional circumstances test, as expounded in Magill and reaffirmed in IMS Health,21 consisted of three cumulative conditions over and above a finding of essential facilities dominance. The first condition was that the undertaking asking for a compulsory licence must offer a new product for which there was potential or demonstrable consumer demand. Secondly, that the refusal to license would lead to an elimination of all competition on the secondary market. Thirdly, that the dominant undertaking had no objective justification for refusing to license.

18 19 20 21

Case 238/87 Volvo v Veng (UK) Ltd [1988] ECR 6211. Ibid, para 8. C-241/91P RTE and ITPV v Commission, above n 2. Case C-418/01 IMS Health, above n 3.

The Epithet That Dares Not Speak its Name 93 In the Microsoft case, Microsoft argued that the test of exceptional circumstances should be a strict test based on the ‘new product’ rule as set out in the IMS case.22 This argument was not entirely convincing, since the ECJ in IMS said that the facts of Magill were ‘sufficient’ to meet the exceptional circumstances test, implying that other categories of exceptional circumstances could exist.23 Microsoft also attempted to establish that competitors would not be providing a new product but replicating MS Windows server operating systems, ‘mimicking their functionality’, and that therefore consumers would not be harmed by its refusal to share interface information. Had Microsoft succeeded in proving this assertion, the case might have been otherwise decided. However, the products offered by Sun, particularly its Solaris server, were different and had added functionality. The Commission maintained that Microsoft’s competitors could use the interface disclosures to develop the advanced features of their own products.24 It argued, therefore, that the refusal to continue to supply interface information stopped ‘follow on’ innovation. The facts as found supported this point, since the Solaris workgroup server offered more functions, eg reliability/availability of the network and security, than the Microsoft workgroup servers.25 On appeal, the Commission refined its new product argument by arguing that the definition of ‘new product’ is a product that does not duplicate existing products but adds substantial elements arising from competitor’s own research efforts. The Commission took a far more effective line on the crucial issue of adapting the exceptional circumstances categories to the facts of the case by reiterating that IMS Health did not establish an exhaustive list of exceptional circumstances.26 It made the strategic point that the framework for analysing the test of refusal to supply was provided by Article 82(b), which declares it ‘an abuse to limit technical development’ to the prejudice of consumers. It argued that the refusal to provide the indispensable interface information amounted to a blockage of interoperability in the specific sector of information technology and a blow to ‘follow on’ innovation in the secondary market. It also explained that this factor differentiated the facts of the case from the IMS Health and Magill type of case.27 Lastly, it also drew attention to the fact that Microsoft was disrupting an existing pattern of a history of supplying interface information to the sector, and the discontinuation of supply had occurred after Microsoft 22 At one point, Microsoft also argued that none of the criteria of the Bronner case was satisfied, but this was on the basis that Bronner should apply if there were no IPRs at stake. See Microsoft, above n 1, para 301. 23 IMS Health, above n 3, para 107. 24 Microsoft, above n 1, para 695. 25 See eg, ibid, para 652. 26 Ibid, para 303. 27 Ibid, para 305.

94

Steven Anderman

had made use of interoperability in the sector to build up its market for Windows operating systems.28 This also reminds us that on this point too, the facts of Microsoft can be significantly distinguished from a refusal to supply a new entrant to the sector (as was the case in the facts of Magill and IMS Health).29 The CFI accepted almost all of the Commission’s arguments, and indeed developed them in certain respects. The Court emphasised that the test of ‘new product’ must be considered under Article 82, which makes it ‘an abuse to limit … technical developments to the prejudice of consumers’. It pointed out that the new product rule in Magill was meant to protect consumers from the harm of preventing the appearance of a totally new product, one not offered by the dominant firm and one for which there was a potential consumer demand. It also stressed that in IMS Health, the Court of Justice had emphasised the damage to consumer interests that can arise where the refusal to grant a licence prevents the development of a secondary market to the detriment of consumers. It then concluded that the Magill and IMS Health new product rule cannot be the only parameter which determines whether a refusal to license an IPR is capable of causing prejudice to consumers within the meaning of Article 82(b) EC. That provision refers not only to limiting production or markets, but also to a limitation of technical development.30 The CFI went on to find that the Commission had evidence for its finding that the refusal to supply would prevent the competitors from ‘developing the advanced features of their own products available in the web of interoperability relationships that underpin the Windows domain architecture’, and evidence for its finding that ‘the refusal caused prejudice to consumers’.31 Hence, the CFI concluded that the Commission’s findings were not manifestly incorrect.32 VI. THE EXCEPTIONAL CIRCUMSTANCES TEST AFTER MICROSOFT

It is clear that after the CFI judgment, the exceptional circumstances test has wider parameters than the ‘new product’ rule in Magill. This was implied in IMS Health, but is not confirmed. However, the question remains: how wide are the parameters?

28

Ibid, para 307. See Anderman, ‘Does the Microsoft Case Offer a New Paradigm for the “Exceptional Circumstances” Test and Compulsory Licenses under EC Competition Law?’ (2004) 1(2) Competition Law Review 13–14. 30 Microsoft, above n 1, para 647. 31 Ibid, paras 651–56. 32 Ibid, para 665. 29

The Epithet That Dares Not Speak its Name 95 The crucial feature of the exceptional circumstances test after the Microsoft judgment is to discern how the new product test has been modified. In Microsoft there were two factors that seemed to be influential in the Court’s judgment. First, this was a case concerning the need to maintain interoperability in the information technology sector to ensure the continuation of demonstrable plural sources of innovation, sources which were also demonstrably satisfying consumer needs that were not satisfied by Microsoft. Sun Microsystems had demonstrated that it was a real source of innovation, and promised to continue to be so. This case did not concern a new entrant seeking access. Sun Microsystems and other work server makers had enjoyed a long period of interoperability with Windows based on its willing supply of interface information, albeit to a predecessor of Sun Microsystems. Microsoft’s decision to discontinue this practice after receiving the letter from Sun, whatever its obvious irritation at the scope of the request, could objectively be viewed as an abuse comparable to the definition of abuse in a line of cases since Commercial Solvents. Moreover, the refusal in such circumstances could legitimately be construed as special circumstances modifying the new product test, ‘owing to its limitation of technical development in the sector to the prejudice of consumers’.33 After Microsoft, the category of exceptional circumstances, when applied to new entrants, will continue to require that the new entrant offers a new product. The case law on the definition of ‘new product’ continues to be opaque: all we know is that the new entrant may not offer a replica or clone of the product already offered by the owner of an IP-protected industrial standard. What is unclear is the extent to which, and the way in which, the new product must be different from the old. In the Microsoft case, a close look at the facts of work group servers indicated a recognisable difference in the functionalities of the Solaris servers from those of Microsoft. But of course the Microsoft case was not a case of a new entrant similar to Magill and IMS Health. It was a case of a dominant firm disrupting an existing relationship of interoperability on grounds which could be found to be anti-competitive. Following Microsoft, we can assume that where an incumbent has already been dealing with undertakings operating in a dependent secondary market, and the competitor in the secondary market: a) already has advanced features in its own products; and b) requires access to the interface information in order viably to continue to develop such advanced features, Article 82(b) will apply. Innovators will get such protection because a refusal in such circumstances will limit the technical development of the market.

33

Ibid.

96

Steven Anderman

Yet will it stop there? In the Microsoft case, the Commission did not seem to consider the distinction between new entrant and established relationship particularly significant.34 It merely drew attention to the fact that this was the case in Microsoft. The CFI thought it more significant, but only as an additional factor. Yet the CFI judgment in Microsoft drew attention to the fact that the language and purpose of Article 82(b) now guides us in defining the outer limits of the competition rules. And in respect of monopoly undertakings with an indispensable product which have historically provided access to competitors in secondary markets, the test of exceptional circumstances may be even wider, and possibly moving closer to cases of essential facilities which are not IP-protected. Commercial Solvents continues to provide possible authority for the interpretation of Article 82(b), and that provision refers to limiting markets, as well as technical development, to the prejudice of consumers. Where a monopolist with an indispensable input disrupts a relationship of previous supply for reasons that could be viewed as anticompetitive, it may be sufficiently egregious behaviour to bypass the established exceptional circumstances test for IP owners.35 VII. CONCLUSIONS: THE EFFECT OF A FINDING OF ESSENTIAL FACILITIES DOMINANCE ON THE TEST OF ABUSE

The finding of essential facilities dominance had a very strong effect on the test of abuse. The fact that the test of exceptional circumstances set out in ECJ judgments presupposes that the IP-protected product in the first market must be both a monopoly and an indispensable input into the second market, influences the issue of the standard of proof of the effect of the abusive conduct. A constituent element of the abuse must be that there is strong evidence that the practice of the IP owner will give rise to the risk of ‘eliminating effective competition’ in the market. In Microsoft, the Court accepted the Commission’s argument that it was enough to prove that the refusal to supply gave rise to ‘a risk of elimination of competition’ in the second market in order to establish an abuse under Article 82. The Court pointed out that the test did not require the Commission to establish that competition had been eliminated, or that its elimination was imminent.36 Microsoft argued that in IP cases, following the case law, there must be a stricter test: it must be shown that the conduct was ‘likely to eliminate’ all competition, or that the conduct of the dominant undertaking had a ‘high 34

Ibid, para 307. Or, as John Kallaugher has suggested, it might be viewed as yet another exceptional circumstance. See Kallaugher, ‘Microsoft and More—Developments Under Article 81 and 82 EC in 2007’ (2008) 29(7) European Competition Law Review 418. 36 Microsoft, above n 1, para 457. 35

The Epithet That Dares Not Speak its Name 97 probability’ of elimination of all competition. The CFI, however, agreed with the Commission, saying that ‘Microsoft’s complaint is purely one of terminology and wholly irrelevant’.37 The CFI said the two tests reflect the same idea, namely that Article 82 EC does not only apply in markets where competition has been eliminated or elimination is imminent.38 This would be counter to the objective of the provision, to maintain undistorted competition in the market and safeguard the competition that still exists. The Court also noted that ‘network effects’ would make it harder for competitors, once ejected, to come back into the market. This was a reference to the applications barrier to entry that had been found to underpin the monopoly of Windows. The CFI emphasised that under Article 82, what matters to establish an infringement is that the refusal ‘is liable to or is likely to eliminate all effective competition on the market’.39 The fact that some competitors ‘retain a marginal presence in some niches of the market’ does not constitute the existence of such competition.40 Of course, technically there is a considerable difference between the two standards of proof: ‘liable to’ and ‘likely to’ eliminate competition. The CFI’s judgment, however, appeared to wish to emphasise that the test of Article 82 is not concerned with establishing a particular standard of proof of effects. Instead, it wanted to stress that it was enough to show conduct that could possibly produce such effects, and this is in line with previous ECJ case law on Article 82. For example, in Commercial Solvents v Commission41 the ECJ referred to the test as ‘risks eliminating competition’. In Telemarketing,42 the Court used the standard of ‘possibility of eliminating all competition’. Yet this part of the decision must be read in the light of the fact that the existence of essential facilities dominance offered strong evidence that the likely effect of the abuse would have been to eliminate competition. The Commission had found that access to MS Windows OS was ‘necessary to maintain viability in the market for work servers’,43 and that that such information was essential to viability on the market.44 Indeed, in so far as the interface information was indispensable to a continued presence on the work group server market, it was in fact not only likely that competitors would be eliminated from the market, but also that once ejected, competitors would find it impossible to return to the market because of the existence of the ‘network effects’ barrier to entry. 37 38 39 40 41 42 43 44

Ibid, para 561. Ibid. Ibid, para 563. Ibid. Case 6/73 Commercial Solvents v Commission, above n 10. Case 311/84 CBEM Telemarketing v CLT [1985] ECR 3261. Microsoft, above n 1, paras 392–93. Ibid, paras 421–22.

98

Steven Anderman

To economists as well as lawyers, one implication of this judgment is that the CFI is not particularly impressed by the concept of reforming Article 82 to add a stronger test of effects as a constituent element of this particular abuse. The test of abuse under Article 82 is conduct to which objectively abusive intent could be attributed and which creates a plausible risk of the harm of elimination of effective competition. Once essential facilities dominance is established, it will not be difficult for the Commission to prove the probable exclusionary effects of a refusal to license.

6 The Microsoft Case—at the Heart of the IP/Antitrust Intersection DAN EKLÖF*

I. INTRODUCTION

A

LTHOUGH THE INTELLECTUAL property (IP) laws, on the one hand, establish far-reaching exclusive rights and institute legal monopolies while antitrust law, on the other hand, curtails the undue exercise of market power, there is no fundamental incongruity between the two regimes. They share the same goal: to enhance consumer welfare and innovation. Due to its deep-seated conformity with free market principles, IP law only rarely gives rise to entry barriers and market power. Legal monopoly does not automatically translate into economic monopoly. Even if IP occasionally does generate significant market power, the mere possession and exercise of such influence is not in itself a violation of Article 82 EC. Indeed, the motto ‘a successful competitor, having been urged to compete, must not be turned upon when he wins’ certainly applies as much in IP affected markets as in any other market. The general, abstract harmony between IP and antitrust does not, however, preclude the occasional existence of glaring dissonances when descending from the abstract level and entering the lower arena of imperfect real-world markets. Tension does exist on the fringe. A unilateral refusal to license can cause substantial detrimental effects, and may warrant antitrust intervention by means of compulsory licensing. This chapter will explore, through the lens of the Microsoft proceedings, how IP jurisprudence, with its nuances and underlying purposes, is, or at least ought to be, fully incorporated into antitrust analysis.1 Since antitrust *

Associate Professor, Stockholm University; Associate, Sandart & Partners law firm. This chapter focuses on the CFI’s final judgment in Case T-201/04 Microsoft v Commission [2007] ECR II-3601, in which Microsoft contested the Commission Decision COMP C-3/37.792 Microsoft [2007] OJ L32/23. A number of decisions have been taken in the subsequent enforcement proceedings. Microsoft has brought an appeal against one of these decisions, see Case T-167/08, nyd. For a collection of public documents, see http://ec.europa.eu/ competition/antitrust/cases/microsoft. 1

100 Dan Eklöf is not applied in the abstract or in a vacuum, it seems wholly appropriate that all IP rights should not be approached in the same way under Article 82, but that antitrust should instead differentiate between individual IP rights under scrutiny and be attuned to the underlying purposes of the IP regulation at hand. After an introductory discussion of the general relationship between Article 82 and IP (sections I–IV), the chapter analyses and reflects on the legal standards applied in Microsoft (sections V–VII). The following sections explore the factual setting and legislative backdrop most prominently featured in the case, ie the software environment and its legal foundation in the EC Computer Programs Directive, and the impact this had in the proceedings (sections VIII–X). Thereafter, it is argued that Microsoft follows a pattern of challenges to questionable IP rights under Article 82 (section XI). This is followed by analysis and discussion of the internal copyright provision targeting interoperability between different makes of software (decompilation), leading to the conclusion that the presence of such a provision supports rather than negates Article 82 intervention (sections XII–XIII). The chapter concludes with reflections in support of antitrust-based compulsory licensing, albeit limited and fine-tuned, since the ultimate beneficiary of such a policy is the IP system itself, which has to be assumed to work in the public interest (section XIV). II. THE QUESTION OF INTELLECTUAL PROPERTY IMMUNITY

Although sometimes advocated, the notion of absolute immunity for IP holders from any kind of Article 82 intervention has not been endorsed by the Commission or the Community courts. Absolute immunity has been suggested on formalistic, legalistic grounds, including Article 295, and certain understandings of the relationship and hierarchy between Community and national law, the Community’s and Member States’ international IP commitments, such as the TRIPs agreement, etc. These arguments have been rejected. Absolute immunity for IP rights has also been argued on substantive policy grounds. One such argument relies on the misty notion that the design of the IP systems is perfect in terms of fine-tuning: the system consistently attains the optimal economic outcome. The exercise of exclusivity, ie refusals to license an IP right, consequently always has a net beneficial effect on the competitive process. Hence, rationality compels the legal order to adopt a presumption to this effect in order to avoid wasteful litigation about a known fact. Absolute immunity is warranted.2 The formalistic, legalistic argument mentioned above in practice merges with this substantive argument 2 See, eg, the discussion in the US case Data General Corp v Grumman Sys Support Corp, 36 F 3d 1147 (1st Cir 1994), where absolute immunity for copyright was rejected.

The Microsoft Case

101

for immunity—it would be completely rational to apply absolute immunity for IP rights as a matter of law. An alternative argument advocates absolute immunity under the more reasonable assumption that forced sharing of IP rights generally does more harm than good. The instances where intervention might be justified are so few, and of such small magnitude, that it would be wise to operate with absolute immunity. The vast majority of refusals—or rather, the preservation of the IP holder’s exclusivity—protect underlying incentives and ultimately promote consumer welfare. That the IP system is almost perfectly calibrated seems to be the assumption under this line of reasoning. Although this second, substantive argument for immunity has some merit, it has not been accepted by the Community judicature. For good reasons, one might argue. Like all artefacts, the IP laws are not flawless. They result from a mixture of sophisticated deliberations in combination with traditions, historical accidents and political compromises in a complicated law-making process with an international, Community and national dimension. Further down the chain, the precise content and scope of protection is developed and refined in complex judicial decision-making in individual IP cases before the courts and IP agencies. The latter institutions are hardly equipped always to get it right in terms of achieving the legal structure for an optimal economic outcome. They probably do not even regard this as their first priority, but focus on developing a clear, consistent, systematically intact, predictable and administrable IP system rather than paying sole attention to the efficiency parameter. Furthermore, IP rights operate in a constantly changing society in terms of technology, business models, consumer preferences and other societal patterns. Against this backdrop, it would be optimistic to assume that the entire IP regime in all respects operates in the public interest all the time. The fact that the IP regime as a whole is beneficial and pro-competitive does not necessarily mean that each and every constituent part of it—the exercise of every individual IP right—is necessarily so. On the contrary, it seems quite unavoidable that individual rights sporadically pop up as ‘bugs’ in an otherwise well-functioning system. One such example of a ‘bug’ might be the Microsoft Corporation’s Windows program emerging as a de facto standard for PC operating systems. This market includes features such as network effects and consumer inertia, both operating in favour of the incumbent standard. Another example might occur in some aftermarkets. Both the ECJ in Volvo v Veng3 and the Commission in its Discussion Paper on Article 824 have identified problems. Consumers, often unsophisticated and poorly informed when making 3

Case 238/87 Volvo v Veng [1988] ECR 6211. DG Competition, Discussion Paper on the Application of Article 82 of the Treaty to Exclusionary Abuses (December 2005), paras 252 and 264. DG COMP has a record of dealing with IPs and aftermarkets but has not yet issued any compulsory licence under Article 82. 4

102 Dan Eklöf purchase decisions, find themselves locked in to a secondary market after having bought an expensive product in the primary market, eg a car. Not having counted on the aftermarket costs—including service, upgrading and repairs—they are stuck. High switching costs prevent them from purchasing a substitutable product on the primary market. This leaves the consumers vulnerable to exploitation by the vendors, who are repeat players possessing much more information than consumers ex ante and ex post the initial purchase. Competitors’ market entry on the aftermarket can be blocked by the presence of IP rights. III. SCEPTICISM OF ABSOLUTE ANTITRUST IMMUNITY

Although conveying no immunity on IP holders, the Community courts— and, in all fairness, the Commission—have displayed great reluctance to afford Article 82 considerable latitude for compulsory licensing. According to established case law, forced sharing is triggered only in ‘exceptional circumstances’. A significant antitrust restraint applies. The notion of absolute immunity—if indeed it ever was founded in EC law—has been replaced with a principle of caution and restraint, although licensing might well be imposed. Absolute immunity is unsound for several reasons. First, a refusal to license can bring about serious negative effects which could not be remedied by means other than forced sharing. Secondly, although dominance problems due to IP possession are normally selfcorrecting in the long term, this is not necessarily the case, as some IPs generate high entry barriers. The dominance can also be upheld by artificial means; in addition, medium- and even short-term effects may be important antitrust concerns. Thirdly, immunity could result in the systematic, artificial inclusion of IP rights in products and business models, which does not necessarily conform to efficiency considerations. Fourthly, in high-tech sectors—with a myriad of IPs, quasi IPs and unprotected items close to protection intersecting—it is highly difficult to ascertain exactly which IP rights are involved and precisely how an Article 82 order affects them, as evidenced by the Microsoft proceedings. The presence of IP rights was, understandably, though regrettably, presumed in a sweeping manner instead of established in any detail.5 An absolute immunity might shift the litigation’s focus towards formalistic definitions and distinctions. This would replace the traditional antitrust concern with the peculiarities and workings of the market under scrutiny—and ultimately the effect on consumer welfare.6 5

Microsoft v Commission, above n 1, paras 283 ff. See Melamed and Stoeppelwerth, ‘The CSU Case: Facts, Formalism and the Intersection of Antitrust and Intellectual Property Law’ (2001–2002) 10 Geo Mason L Rev 407; Boyle, Lister and Everett, ‘Antitrust Law at the Federal Circuit: Red Light or Green Light at the IP/Antitrust Intersection?’ (2002) 70 Antitrust Law Journal 739. 6

The Microsoft Case

103

IV. REASONS FOR RESTRAINT RATHER THAN ABSOLUTE IMMUNITY

A restrictive approach rather than unqualified immunity may be advocated instead, and reflects Community jurisprudence. To reverse this policy completely and deny the IP holder any privileged position under competition law would go too far. Exclusivity effectively conveys upon the proprietor almost total control over the IP right, which forms the foundation for his implementation of business strategies and models, and the basis for his bargaining position on the market. To transform this exclusivity to merely a right to receive remuneration would seriously disturb this basic structure, which ultimately serves the public interest and, it has to be assumed, in the vast majority of cases supports consumer welfare. A considerable antitrust restraint seems motivated for the following reasons. First, protecting the fundamental incentives on which the IP systems are largely based. Secondly, legal-systematic considerations: exclusive rights to an IP have been introduced by legislation and the integrity of the legislative process should be upheld; the legislator has the institutional capacity as well as constitutional authority to carry out an appropriate balancing act. Thirdly, the ‘procedural economic’ reason: antitrust litigation is costly and time-consuming, and could—if frequently invoked against rightholders—create uncertainty, dampen investment and restrict innovation. Admittedly, the same three arguments could be made in favour of absolute immunity. Such an approach would, however, downplay the detrimental effects on consumer welfare caused by some licence refusals, overstate the general need for the rightholders’ continuous and absolute exclusivity regardless of the circumstances, and overrate the legislative capability to devise IP systems which adequately meets all the imperfections of realworld markets. So if absolute immunity is not an option, how should a restrictive approach be defined and applied? The ECJ’s attempt to solve this obvious problem has been to employ the inexact phrase ‘exceptional circumstances’. V. ‘EXCEPTIONAL CIRCUMSTANCES’—THE LEGAL STANDARD FOR COMPULSORY LICENSING THROUGH THE LENS OF MICROSOFT

Roughly speaking, two competing interpretations of ‘exceptional circumstances’, as employed by the ECJ, can be identified. Both positions were indeed claimed to be correct during the Microsoft proceedings.7

7

Microsoft, above n 1, paras 291–307.

104 Dan Eklöf According to the Microsoft Corporation, a Magill/IMS Health8 legal paradigm exclusively applies. Only in such a scenario can ‘exceptional circumstances’ exist. The criteria set forth under this standard should be construed narrowly in favour of the antitrust defendant. Only if all the conditions in the Magill/IMS Health list are met might a forced sharing be acceptable, the four cumulative criteria being: a) the refusal relates to a product or service indispensable to the exercise of a particular activity on a neighbouring market; b) the refusal is of such a kind as to exclude any effective competition on that neighbouring market; c) the refusal prevents the appearance of a new product for which there is potential consumer demand; and d) the absence of an objective justification for the refusal. The Commission took the opposite view. It employed an ‘entirety of the circumstances’ standard, and did not recognise in ECJ case law the existence of a mechanical application of an exhaustive checklist such as the one proposed by Microsoft. Rather, standard Article 82 jurisprudence applied to licence refusals, although, in addition, the presence of ‘exceptional circumstances’ had to be established. It should be added that the Commission successfully maintained that even if a Magill/IMS Health list did indeed apply, the Commission had proven its case by this legal standard too (‘heads, I win—tails, you lose’).9 The CFI occupied the middle ground. It did not take an absolute position as to the applicable legal standard. The Court certainly did apply the four criteria of the Magill/IMS Health list, but without an outright rejection of the Commission’s legal analysis in this respect.10 It was not ruled out that another scenario might prove ‘exceptional’ and trigger Article 82 liability.11 But since the Magill/IMS Health list was in any event satisfied, it was unnecessary for the Court to address the Commission’s ‘entirety of the circumstances’ approach. That issue was left for another day. More importantly, the CFI construed the Magill/IMS Health criteria in an elastic manner, certainly not axiomatically in favour of the antitrust defendant. This might prove to be the most enduring and vital legacy of the proceedings as a matter of law: the Magill/IMS Health list is a flexible yet 8 Joined Cases C-241/91P and C-242/91P RTE and ITV v Commission [1995] ECR I-743; and Case C-418/01 IMS Health GmbH v NDC Health GmbH & Co KG [2004] ECR I-5039. 9 See, eg, Microsoft, above n 1, para 107. 10 Ibid, para 336. 11 See eg ibid, para 330 (‘sufficient’); para 332 (‘in particular’); para 333 (‘may infringe Article 82’); para 691 (‘exceptional circumstances such as those hitherto envisaged in the case-law’).

The Microsoft Case

105

sharp legal tool. In this regard the outcome represents a significant victory for the Commission, which claimed that it had proven its case even if the Magill/IMS Health standard did apply. What legal standard, then, applies after Microsoft? Although probably not exclusively encompassing all possible factual and legal settings, the Magill/IMS Health list is likely to form the safest basis for successful Article 82 actions in the future. Forced sharing might follow under other circumstances, but the Community judiciary (and national courts) will probably hesitate to deviate from the Magill/IMS Health paradigm. On the other hand, the interpretation of this list has proven to be flexible, at least until the ECJ rules otherwise. To friends of compulsory licensing, the need for an ‘entirety of the circumstances’ standard as advocated by the Commission has thereby been reduced. In its Article 82 Guidance of December 2008, the Commission relies on Microsoft to defuse the ‘new product’ criterion.12 The Commission’s formulations suggest that no bright line separates this criterion from the ‘exclusion of effective competition’ criterion. The message is that these criteria overlap, in effect even merge, and that the focus is the ultimate effect on consumers. The crucial welfare issue is not short-term product improvement, but rather what preserved rivalry or credible threat of entry means for the medium- and long-term rate of innovation in a sector. The Guidance uses the notion of ‘innovation’ in a rather abstract and ultimately undefined manner. It is not necessarily confined to what can be predicted or anticipated, but instead encompasses both incremental, step-by-step innovation and the potential for more radical, unforeseeable changes in terms of products, business models and organisation. The Commission’s reliance on rivalry coupled with an undefined innovation concept arguably reflects a conservative approach. This is probably due to the general theoretical and empirical uncertainty about the sources of innovation, ranging from the traditional discourse on innovation as being either supply-pushed or demand-led to modern process-based theories. The overall impression of the Article 82 Guidance is that the Commission tries to play down the IP/antitrust issue as a special case, instead ‘normalising’ it and bringing it in line with standard Article 82 jurisprudence (as modified or at least clarified by the Guidance, integrating innovation and protection of incentives as natural and important components). Yet it cannot be inferred from Microsoft or the Article 82 Guidance that compulsory licensing will be ordered lightly. Throughout the CFI judgment, repeated references are made to the need for a very strong and enduring 12 DG Competition, Communication from the Commission, Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings (Brussels, 3 December 2008), paras 19 ff, 27 ff, and 74 ff.

106 Dan Eklöf market position. Few IP holders possess a position comparable to that of the Microsoft Corporation. Although dominance is an absolute term in the sense that one is either dominant or not, it is also a relative term. Firms which only just cross the dominance threshold are unlikely to be caught, provided the Magill/IMS Health criteria, particularly the objective justification/protection of incentives defence, are construed and applied sensibly. Despite the fact that neither the Commission nor the CFI clearly articulated that Microsoft was a highly exceptional case of limited general relevance, this conclusion seems reasonable. VI. NO SUBSTANTIAL CHANGE IN GENERAL POLICY LIKELY—THE ‘MISSING’ OBJECTIVE JUSTIFICATION

The Microsoft judgment does not signal the beginning of an antitrust era signified by frequent and massive mandatory licensing. It was a peculiar case, with a very strong company operating in an unusual factual and legal setting, including questionable—or at least weak—IP rights. Where substantial and risky investments lie beneath the challenged IP rights, an objective justification defence should operate in favour of the defendant, in order to ensure that Article 82 pays proper deference to the basic investment protection structure inherent in IP law. Microsoft does not appear to have produced any hard evidence that the material subjected to forced sharing was the result of large and risky investments, given the magnitude of the markets concerned.13 The CFI never dissected the investment protection issue in-depth but seems rather to have avoided it. Perhaps the Court did not want to find itself caught up in the very difficult quantification and trade-off between, on the one hand, preserving ex ante incentives by denying a licence and, on the other, promoting ex post efficiency by allowing a licence. If Microsoft had indeed produced hard evidence of substantial research and development (R&D) and investments in relation to the material covered by the mandatory licence order, this trade-off might not have lent itself to judicial circumvention. Without here moving into the discussion of the ‘new balancing test’ in this part of the judgment, the Microsoft case seems underdeveloped in terms of the objective justification defence. The CFI’s treatment is rather brief, abstract, and in parts somewhat legalistic, rather than being an application 13 See, eg, Case T-201/04 Microsoft Corp v Commission [2004] ECR II-4463, Interim Decision by the President of the CFI, 22 December 2004, para 257: ‘… the vague allegation that Microsoft’s communications protocols have “cost tens of millions of [United States] dollars”, even if it were well founded, is not supported by any evidence’; cf Microsoft’s position in the final judgment, Microsoft, above n 1, paras 274 and 666; cf observations by the Commission, Microsoft, ibid, paras 310 and 683.

The Microsoft Case

107

of the law to hard facts. This limits the value of Microsoft as a precedent, and implies that the objective justification defence needs further elaboration and clarification. A reasonable prediction, albeit speculative, is that the CFI’s flexible application of the first three criteria in the Magill/IMS Health list, will work in favour of the claimant. However, the objective justification criterion is likely to be developed to emphasise the need to secure incentives. This will serve the interests of IP holders, provided that they can establish that investments deserving protection have been made.14 VII. THE ‘EXCEPTIONAL CIRCUMSTANCES’ OR ‘CHARACTERISTICS’ ACCORDING TO THE COMMISSION

Which, then, were the ‘exceptional circumstances’ identified by the Commission (see section V above), supporting an Article 82 intervention, besides the application of standard Article 82 jurisprudence? Three circumstances qualifying as ‘exceptional’, or at least being ‘characteristic’, were invoked by the Commission (as the CFI interpreted it): a)

The information which Microsoft refused to disclose to its competitors related to interoperability in the software industry, a matter to which the Community legislature attaches particular importance. b) Microsoft used its extraordinary power on the PC operating systems market to eliminate competition on the adjacent work group server operating systems market. c) Disruption of previous levels of supply. The second and third factors will not be discussed in depth here. Suffice it to say that ‘leverage’ appears too vague to operate as a stand-alone basis for inferences on Article 82 liability. Although making a good common-sense argument in the Microsoft proceedings, ‘leverage’ as a separate antitrust concept is questionable due to its imprecision. It would be perfectly logical to bring ‘leverage’ into the IP/antitrust equation if indeed the IP regimes operated with a defined, consistent and clear concept of ‘leverage’, confining the exercise of the exclusive right to particular economic activities or markets. But the IP regimes do not. No specified

14 See the Article 82 Guidance, above n 12, paras 20 ff, 74, 77, 81, 83, 85–86 and 88–89; see also the Discussion Paper on Article 82, above n 4, paras 213 and 235 ff (note also paras 230 and 241 ff which appear tailor-made for the Microsoft proceedings). In the context of Article 81, see eg Case 262/81 Coditel II [1982] ECR 3381, para 19; and Commission’s Guidelines on the application of Article 81 of the EC Treaty to technology transfer agreements [2004] OJ C101/2, paras 8 and 147 (protection of investments and incentives forming an integral part of the assessment under Article 81).

108 Dan Eklöf limits exist as to which economic activities and markets are covered by an individual right, at least as far as copyright and patent law are concerned. The rights can—somewhat simplified—be relied on in any market and economic activity, regardless of size, shape or form. Earlier antitrust jurisprudence under Article 81 elaborated with a ‘scope of the grant’15 concept when assessing licensing agreements under Article 81, somewhat akin to a ‘leverage’ theory. The ‘scope of the grant’ framework and equivalent approaches seem to have been abandoned. A revival under Article 82 would be ill-advised. As regards ‘disruption of previous levels of supply’, it is arguably correct that an antitrust defendant’s sudden change of policy works for the claimant’s case. The treatment of old, occasional and new customers in terms of refusal to deal has been differentiated in previous case law. But the law is far from clear in this respect.16 The application of Article 82 on fast-moving, high-tech markets might differ considerably from its application on more mature, stable and nondynamic markets, such as the ones for bananas and oil. While a change of business policy or deviations from commonplace business practices might indicate abusive conduct, this certainly does not hold true as a general proposition on dynamic markets, where past behaviour is often a poor indicator of future developments. The first ‘exceptional circumstance’ or ‘characteristic’ invoked by the Commission, stressing the Community policy of interoperability in the software industry, is a strong one. It emphasises that the Microsoft case is situated at the heart of the IP/antitrust intersection, and it will be discussed in the following section. VIII. FORCED SHARING AND EC SECONDARY LEGISLATION IN THE IP FIELD—INTEROPERABILITY AND THE SOFTWARE DIRECTIVE

The EC legislator has not been prolific, in the secondary law context, in articulating a position on how Article 82 should be applied to licence refusals. The issue has been left to the courts—ultimately to the ECJ. However, two more explicit statements are found in the Database Directive17 from 1996 and the earlier Software Directive from 1991.18 15 See Anderman, EC Competition Law and Intellectual Property Rights (Oxford, Clarendon Press, 1998), 53–54. 16 Note the rather sweeping statements in this context by the Commission in the Article 82 Guidance, above n 12, paras 83 and 89. 17 European Parliament and Council Directive 96/9/EC of 11 March 1996 on the legal protection of databases, [1996] OJ L77/20. 18 Council Directive 91/250/EEC of 14 May 1991 on the legal protection of computer programs, [1991] OJ L122/42, ‘the Software Directive’.

The Microsoft Case

109

The Database Directive addresses the issue in relation to the new so-called sui generis right introduced by the Directive: Whereas, in the interests of competition between suppliers of information products and services, protection by the sui generis right must not be afforded in such a way as to facilitate abuses of a dominant position, in particular as regards the creation and distribution of new products and services which have an intellectual, documentary, technical, economic or commercial added value; whereas, therefore, the provisions of this Directive are without prejudice to the application of Community or national competition rules.19

The notion that an individual IP right might be used to reduce output and innovation, although the reason for establishing the sui generis right of course was the very opposite, was clearly before the law-maker. It might be added that the original concern that the sui generis right might hamper product development and innovation does not seem to have materialised on a large scale, although a number of successful forced sharing proceedings have taken place before the Member States’ courts. The clear message from the legislator about a consumer welfare concern is still there. A second example was incorporated into the earlier Software Directive. In 1991 it was correctly assumed that software would be the engine of the next technological revolution. A rapidly introduced, harmonised legal regime across the Community was a priority.20 At the same time the legislator was keen to ensure that rivalry was not completely eliminated by granting this relatively new phenomenon the full status of copyright protection. The backdrop was the IBM proceedings from the mid-1980s, active until 1994.21 This antitrust experience, albeit limited, with the growing computer market had shown that de facto standardisation could indeed occur. Article 82, an ‘external’ market correction mechanism, was—together with ‘internal’ copyright measures—thought of as a proper tool to address possible future competition problems. Consequently, the preamble to the Software Directive contains the following passage: Whereas the provisions of this Directive are without prejudice to the application of the competition rules under Articles [81] and [82] of the Treaty if a dominant supplier refuses to make information available which is necessary for interoperability as defined in this Directive.22

There was a serious concern that a dominant player’s conduct could work as an obstacle to the very technological progress and increased rate of 19

Database Directive, above n 17, Recital 47, preamble. See Software Directive, above n 18, introductory statements in the preamble on investment protection, the importance of software technology and harmonisation. 21 See Lomholt, ‘The 1984 IBM Undertaking—Commission’s Monitoring and Practical Effects’, Commission’s Competition Policy Newsletter, No 3, 1998, p 6. 22 The preamble does not have an official numbering system; this passage is referred to as ‘recital 26’ in Microsoft, para 1337. 20

110 Dan Eklöf innovation that the Directive was intended to secure. It seems reasonable to assume that the legislator’s vision was a software market with largely open interoperability standards. Performance-based competition resembling more normal competitive conditions between manufacturers was preferred, without the hurdles of compatibility obstacles raised by the very nature of software technology. Admittedly, directives—and in particular their preamble statements—do not and cannot dictate, confine or expand the application of Article 82, a superior rule in terms of hierarchy. Notwithstanding this, it seems perfectly reasonable that a competition law assessment incorporates considerations and policies expressed in secondary legislation. At least, this seems reasonable unless subsequent technological developments, accumulated experience and largely undisputed economic theory suggest that the original concern is outdated, was never justified in the first place or has been overtaken for other reasons. However, this does not appear to be the case with interoperability in the software industry, as indicated by the Microsoft case. Admittedly, the risk of engaging in circular reasoning in this context is obvious. Nonetheless, it is here submitted that when the Community legislature granted a generous IP protection to a relatively new phenomenon such as software in the early 1990s, this was done in an ad hoc fashion with limited experience of the technology and how software markets would evolve. But the issue of interoperability was observed and dealt with. There seems to be no obvious reason in the Microsoft proceedings to question the Software Directive’s presumptively valid concern about the lack of interoperability as a potential problem that can be redressed by Article 82. The ‘onus of proof’ to establish that the concerns raised in the Directive are no longer relevant in an antirust context cannot rest with the claimant. IX. THE INDIRECT RELEVANCE OF THE SOFTWARE DIRECTIVE

Competition law is not an autonomous system. It does not apply in a vacuum but instead in a setting of public regulation, rights and entitlements defined by other laws. It should carefully approach and be attuned to the particular, even in many cases peculiar, characteristics of the IP regime at hand. An antitrust analysis accordingly has to include IP considerations as well as antitrust considerations in order to avoid either false clearance or false condemnation.23 The direct relevance of the Software Directive’s concerns was sidestepped by the CFI. Although the Software Directive plays a prominent role in the Commission’s decision, this is based on Article 82. Technically, therefore, 23 Cf the general statements on refusals to deal by the US Supreme Court in Trinko, 540 US 398 (2004).

The Microsoft Case

111

the concept of interoperability in the Software Directive was not decisive in the current proceedings. Rather, the CFI emphasised that the interoperability required to maintain a competitive market structure was the result of an Article 82 standard, autonomously applied, and not defined by the Software Directive.24 Although the CFI is undoubtedly correct, it also seems correct of the Commission to invoke the legislator’s clear signal about a key concern as a characteristic feature supporting its case. And it certainly cannot be ruled out that the CFI was indeed deeply affected by the legal context provided by the Software Directive, although it chose to circumvent a direct discussion about the Directive’s relevance. This does not imply that the lack of any competition policy statements in the relevant IP legislation rules out compulsory licensing. Each IP/antitrust case has to be judged on its own merits. Examples are the hypothetical scenarios dealt with by the Advocates-General in Laserdisken25 and Metronome.26 The litigation concerned laser discs, a predecessor of the DVD, and the refusal to grant rental rights. Under copyright law, the rightholder has the exclusive right to authorise rental licences for films. The EC legislator instituted such a right in the Rental and Lending Directive,27 partly in order to secure the development of a smooth rental market for videocassettes and DVDs. This new technology had brought about a revolution in the way the people watch films. Previously, consumers were confined to cinemas, television and film clubs. The technology and the legal protection brought about a wholly new product. A user-friendly rental market with affordable prices emerged, clearly an increase in output and consumer welfare. According to the Advocate-General in Laserdisken, it was ‘patently clear’ that if film companies were to exercise their exclusive rental rights in a purely negative fashion, solely in order to block the development of a videodisc rental market—with the result that consumers would have to purchase, at a high price, products that they would prefer to rent—that conduct could be curtailed by competition law.28 In the same vein, the Advocate-General in Metronome stated, with reference to Magill, that Article 82 might be used if the sole purpose of a refusal to grant rental licences was to eliminate those already engaged in the rental

24

Microsoft, above n 1, paras 227–29. Case C-61/97 Laserdisken [1998] ECR I-5171. 26 Case C-200/96 Metronome [1998] ECR I-1953. 27 Council Directive 92/100/EEC of 19 November 1992 on rental right and lending right and on certain rights related to copyright in the field of intellectual property, [1992] OJ L346/61 (subsequently replaced by European Parliament and Council Directive 2006/115/EC of 12 December 2006 on rental right and lending right and on certain rights related to copyright in the field of intellectual property [2006] OJ L376/28). 28 A-G Opinion in Laserdisken, above n 25, para 17. 25

112 Dan Eklöf business from the market so that, subsequently, the rental market could be occupied by the producers themselves.29 These elaborations were brief and hypothetical. The ECJ never addressed them. Being Opinions of Advocates-General, they carry no authority; but besides being of general interest to the IP/antitrust discourse, they represent examples of intervention being discussed—even advocated—although the new and carefully drafted Rental and Lending Directive was completely silent on the application of Articles 81 and 82. Rather, the Opinions were based on a conclusion that the film owners’ (hypothetical) conduct conflicted with the reason IP rights in general, and the rental right in particular, were introduced in the first place, namely to create win–win markets by protecting copyright holders and simultaneously promote consumer welfare. No inferences should be drawn from the fact that secondary legislation in the IP field does not discuss or even mention competition law: Articles 81 and 82 nonetheless apply. On the other hand, if the relevant IP legislation directly addresses competition concerns, or such problems clearly follow from a systematic analysis of the IP regulation at hand, representing a policy emanating from the competent community institutions, such considerations should clearly form an integral part of an IP/antitrust appraisal. X. PECULIAR FEATURES OF SOFTWARE—IMPLICATIONS FOR THE MICROSOFT PROCEEDINGS

From a pure IP perspective, it is hardly surprising that Article 82 is used to curtail market power stemming from the copyright protection of software. This essentially technological product—consisting of features normally denied protection under copyright law—does not sit comfortably within copyright jurisprudence. Rather, computer programs were pressed into the copyright sphere under considerable friction from a legal-systematic point of view. It was due to pragmatic concerns, rather than traditional IP jurisprudence, that the copyright regime was chosen as the smoothest vehicle to protect investments and secure the emerging markets. Software is the only kind of copyrightable subject-matter which cannot easily be scrutinised and analysed by third parties, be they consumers or competitors. Compare a movie or a book; they can be scrutinised in detail. Nothing is hidden. Competitors are perfectly free to analyse a competitor’s successful product and incorporate into their own products those elements of the original which are not subject to protection. This includes ideas, principles, concepts and facts. The universal copyright 29

A-G Opinion in Metronome, above n 26, para 33.

The Microsoft Case

113

doctrine of idea/expression dichotomy applies. Such imitation occurs all the time—unprotected features from competitors’ successful works operate as an essential inspiration and direct input in many copyright industries. The same cannot easily take place within the software industry. The reason is the ‘secretive’ nature of software. Provided that the proprietor chooses not to publish it, the source code—ie the protected subject-matter under copyright law—remains non-public. The exact knowledge of the code design and the inner workings of the program rest with the owner alone. This so called proprietary software can be contrasted with open source software, where the entire source code is published and normally free of charge to use. An example is the well-known Linux operating system. Another key feature distinguishing software from other types of copyrighted works is the notion of interoperability. As far as traditional works, such as books and films, are concerned, no technological interfaces exist between the work of authorship and the user: one reads a book, one watches a film. The user relies directly on his human senses in the perception of the protected work. Of course watching a film requires intermediary technical devices, such as a DVD player and a TV set. But that is not the point being made here: between the protected work as such and the user, no interfaces exist. In any event, such intermediary devices are normally standardised under open technical protocols. The copyrighted works function commercially and technically on a stand-alone basis. A DVD can be replaced with another DVD, etc. The contrary applies in a software environment. Interaction and interoperability between works are crucial elements. A computer program is virtually meaningless and worthless as a stand-alone product. Technological interaction is necessary. It is made to interoperate. The program has constantly to communicate seamlessly with other software and hardware as a constituent part of a larger software system, otherwise the program serves no technical or commercial purpose. The program simply does not work. Hence, interoperability in a given environment is a key condition for the successful launching of a product. The interoperability concern holds particularly true if one is—as the EC legislator and Commission appear to be—committed to the view of a vibrant inter-brand competition as a fundamental driving force behind innovation in software markets. Simplified (or rather, grossly simplified), the components of a computer program laying down the technical rules for interoperability in its intended environment are called ‘interface specifications’ or ‘communication protocols’. These are documented as an integral part of software coding. This unique feature of software—the tricky picture surrounding technical compatibility—is fertile soil for competition problems. As such, the interface specifications are normally denied copyright protection. A third party wishing to enter a market needs compatibility—ie access to the interface specifications—with the reigning de facto standard (if indeed there is one).

114 Dan Eklöf However, the interface specifications, just as the source code, are not public knowledge. Instead they are of a secret, hidden nature, unless the software proprietor provides or publishes them. Practically speaking, in today’s software market, if a company wants to enter the word-processing or spreadsheet market, it needs compatibility with Windows if the entrant wants to sell its products in a Windows environment. This environment is commercially very attractive, since currently well over 90 per cent of the world’s PCs have Windows installed.30 Being frozen out from such a market is, therefore, a serious detriment. When introducing the harmonised copyright regime for software in 1991, the Community legislator predicted a Microsoft scenario. A powerful de facto standard did indeed emerge, coupled with the dominant player’s reluctance to share interface information, at least with independent manufacturers in some markets. In the case of PC operating systems, Microsoft’s Windows acquired this position. And, as envisaged, Article 82 was triggered and applied. The company was forced to release the information, thus opening up for inter-brand competition by enabling independent software producers to attain compatibility with the mighty Windows program. In casu, manufacturers of work group servers operating in a Windows environment were the beneficiaries of the Commission’s intervention, improving their opportunity to challenge Microsoft on the work group server market. But as a precedent, Microsoft probably covers a wider range of sectors where technical compatibility with Windows is an essential condition for viable market entry. XI. MICROSOFT IN THE LIGHT OF PREVIOUS CASE LAW—ONE SOLE CONSISTENCY

The sole consistency in the leading refusal cases, starting with Volvo v Veng and followed by Magill, IMS Health and Microsoft, is the IP rights affected. They have not been near the core of intended protection under any definition of traditional IP jurisprudence. Rather, their location has been at the periphery: near, or even possibly across, the outer boundary of the protected zone.31 This might be a pure coincidence, or it might reflect that the intrinsic qualities of the case-specific IP rights are a factor taken into

30

Commission’s decision in Microsoft, above n 1, para 435. In Microsoft, the number and status of the alleged IP rights are difficult to ascertain, since the record of the proceedings does not indicate this with any precision, see paras 267 ff. In the subsequent enforcement proceedings the status and nature of the IP rights (and quasi-IP rights, such as trade secrets) are more thoroughly identified and discussed. Yet it is clear that we are nowhere near the traditional core of IP protection. 31

The Microsoft Case

115

account while not explicitly addressed.32 It appears systematically adequate to incorporate such an assessment, although it might occasionally prove a difficult task in practice. The IP laws themselves do not protect all content in a stereotypical, inflexible fashion. While some of the parameters of protection remain fixed, such as the duration in time, others, such as the scope of protection against third party imitation, in effect often vary. Intellectual property rights represent a continuum, a sliding scale, where some objects enjoy a wide scope of protection against infringers whereas other rights are considerably thinner. Not all individual IP rights are afforded the same protection in all the various parameters which together constitute IP protection. For instance, the basic conditions for qualifying as IP, such as originality in copyright and novelty/inventive step in patent law, can be viewed as fixed thresholds. In isolation, something is either protected or it is not. Yet these criteria have a relative dimension. Simplified, some works of copyright are highly original and some patentable inventions are very innovative. Consequently, they enjoy a wider scope of protection against infringers than rather trivial creations having just passed the threshold. Typically, or at least as a reasonable working antitrust hypothesis and assumption, there is a correlation between investments made and degree of originality and innovation embodied in an IP right. If called upon to decide a case in the IP/antitrust intersection, the decisionmaker should arguably incorporate considerations from both sets of law that are intersecting, in order to avoid enforcement errors and aim for a balanced, optimal outcome. Antitrust law should be receptive to the individual IP right(s) under scrutiny and carefully avoid a one-dimensional, static view of IP law. Failure to do so would mean that antitrust disregards the fact that all the millions of IP rights granted are not given equal weight within the IP system itself. However difficult such an assessment might be from a practical perspective, it hardly seems more problematic than many of the meticulous enquiries into the factual and legal context forming part of Article 82 litigation. Arguably, antitrust should, to the extent this is not already done intuitively, take into account the nature of the IP right at issue, ie if it involves serious creative, resourceful efforts and/or substantial investments. The more characteristics of this kind that are displayed and contained in the right/rights under assessment, the harder it should be to invoke Article 82. Conversely, works at or near the outer boundary of protection should, in principle, be more easily subjected to intervention. Whether such a doctrine will emerge as an expressed, integral part of IP/antitrust analysis remains to be seen. 32 See the Opinions of the Advocates-General in Magill, paras 83 and 126–27, and in Case C-7/97 Oscar Bronner GmbH & Co KG v Mediaprint [1998] ECR I-7791, para 63 (referring to Magill).

116 Dan Eklöf XII. INTERNAL COPYRIGHT MEASURES— THE DECOMPILATION PROVISION

What seems less controversial to include in antitrust analysis is the presence of an explicit IP exception to the exclusivity, specifically targeting a competition concern. Article 6 of the Software Directive33 is such a rare example. It provides for decompilation (reverse engineering) of software. It is a specific limitation of the rightholder’s exclusive right to reproduction. Unauthorised third parties are permitted to engage in reverse engineering, in order to obtain the interface information for interoperability purposes. It is impossible to reverse engineer software without making a large number of copies, given the digital environment in which software operates. Hence, we see the need to limit the software owner’s exclusive reproduction right. An unusual step has also been taken by stipulating that contractual provisions in licence agreements negating the licensee’s right to decompile are null and void as a matter of EC law.34 In addition, unauthorised third parties are specifically allowed to circumvent technological protection measures (TPMs), based on copyright, patents and trade secrets. A TPM is a technology designed to protect unauthorised copying and dissemination, including anti-copying devices and scrambling systems. Normally, breaking through these ‘electronic locks’—a process known as circumvention—is illegal. However, if the purpose of circumvention is to achieve interface information for interoperability purposes, such conduct is permissible.35 Moving from the copyright-based Software Directive to patent law, this field is as yet only partially harmonised in the Community. Proposals do, however, identify and take precautionary steps to secure continuing opportunities to decompile. Two important patent proposals contain explicit references to the Software Directive’s Article 6. A specific limitation of the exclusivity stipulates that a patent right cannot be enforced so as to block decompilation.36 The patent proposals have proven controversial and have not been enacted. However, this is not due to the provisions relating to decompilation, which are uncontroversial. Although not yet part of 33 See above n 18. Article 6 does not require the target program’s rightholder to hold a position of market power, emphasising the provision’s purpose of achieving open communication standards in general. Nonetheless, the primary targets of the provision are software brands enjoying a strong market position. 34 Software Directive, above n 18, Art 9. 35 Ibid, Art 7. See also Rec 50 in the preamble of European Parliament and Council Directive 2001/29/EC of 22 May 2001 on the harmonisation of certain aspects of copyright and related rights in the information society (‘Infosoc’), [2001] OJ L167/70. 36 See Council’s proposal for a council regulation on the Community patent—Revised text, 8539/03, Brussels 2003, Art 9; see also Commission’s proposal for a directive on the patentability of computer-implemented inventions, COM (2002) 92 Final, Rec 18 and Art 6.

The Microsoft Case

117

Community law, the proposals signal the persistent unease among the EC legislative bodies about the lack of interoperability reducing inter-brand competition, following the intrinsic nature of software technology and its legal protection.

XIII. THE DECOMPILATION PROVISION AND MICROSOFT

When a specific exception to the exclusive right is introduced on the clear policy ground of promoting inter-brand competition, it seems odd not to bring this into the IP/antitrust equation. It could of course be argued that since the Software Directive contains an ‘internal’ IP mechanism to correct a problem, there should be less room for ‘external’ intervention by competition law. However, that would be a false logic, both generally and in the specific context of the Software Directive. The Directive expressly identifies Article 82 intervention as a safety valve that might be activated if dominant companies withheld interface information. That an IP regulation specifically targets a problem by means of an explicit exception does not imply that all problems are thereby solved and recourse to antitrust consequently is ruled out. No such logic exists; rather, it is the other way around. Competition law should more easily be put into play under such a scenario. The decompilation provision seeks to promote open communication standards. It was (and is) realised, however, that this provision does not necessarily do the job adequately, given the complexity and imperfections of real-world markets. Decompilation is a sometimes slow, unreliable and time-consuming process, which does not necessarily provide all the adequate interface information necessary for a timely, commercially viable entry into the marketplace for all who wish to enter. Hence, the internal IP mechanism might need to be supplemented by antitrust intervention, construed and applied in tune with the systematic considerations underlying the IP regulation at hand, including specified limitations and exceptions to exclusivity. The two sets of law, IP and antitrust, should truly interact. Arguably, this is exactly what happened in the Microsoft case.

XIV. CONCLUDING REMARKS

Antitrust intervention on an ad hoc basis in IP law seems justified on rare occasions, for the reason that IP regulation cannot be assumed always to be perfectly designed to achieve its own purpose of maximising dynamic efficiency while adequately balancing the static notions of efficiency. While no individual IP right escapes the ambit of Article 82 per se, weak or trivial rights in terms of investment and intrinsic qualities are a more

118 Dan Eklöf appropriate antitrust target than subject-matter located in the core of protection. This does not imply a general ‘second-guessing’ of the legislator’s decision to protect an object by means of IP. Rather, it represents a realistic approach to the IP system, recognising that although the IP regime normally serves the interest of consumer welfare, the system is not perfectly calibrated to meet all the imperfections of real-world markets. Furthermore, such an approach aligns the two fields of law, recognising that all IP rights are not afforded the same status and protection within the IP system itself. A systematic analysis of the relevant IP systems and IP rights is required to make sure that the antitrust decision-maker is fully informed as to the factual and legal context in Article 82 litigation. The most obvious example is when the IP regime involved clearly articulates a competition concern. In the Microsoft proceedings, the most prominent legal context was the Software Directive. It contains a far-reaching provision aiming for the release of interface information by means of decompilation, which was not sufficient on the facts to deal with the interoperability concern. The Directive even makes a direct reference to Article 82, addressing the dissemination of interface information. Most legal instruments in the IP field contain few or no explicit competition-related exceptions or references, but the various IP regulations can always be subjected to in-depth analysis through a competition lens, yielding valuable insights. A systematic analysis of what can be protected, how it is protected, exceptions and limitations, etc allows the antitrust decision-maker to achieve a balanced outcome, incorporating sophisticated IP jurisprudence. In the vast majority of cases, this will obviously lead to a non-intervention decision. It seems likely that IP rights of a more genuine character—and being the result of substantial, identifiable investments— than the ones thus far attacked under Article 82 are less susceptible to interference and will generally withstand scrutiny. Ex post Article 82 intervention, transforming the exclusivity into a reasonable remuneration, does not necessarily reduce ex ante incentives or threaten the IP system, provided that the investments and commercial risks involved are limited. This seems especially true if such an approach is explicitly incorporated, processed and accounted for in the antitrust analysis, thereby openly communicating such considerations to the market actors. Paradoxically, the system that has most to gain from an appropriate antitrust intervention policy is probably the IP system itself. Ultimately, this system relies on the general confidence that it works in the public interest. The current IP regime is not sacrosanct; we are not dealing with natural rights here. The system has, in parts, been seriously questioned by the public and in prominent political circles. So far, this has most loudly taken place in the fields of copyright file sharing, software patents and biotechnology.

The Microsoft Case

119

Intellectual property has moved from the esoteric arena into the public eye. It is likely to continue to be the subject of an expanded debate. A more general perception that IP does not work in the public interest might be fuelled by the absence of safety valves such as Article 82, removing excesses resulting from a disproportion between the financial and intellectual input underlying a right and the resulting market power.

7 A Reformed Approach to Article 82 and the Special Responsibility not to Distort Competition KATHRYN MCMAHON*

I. INTRODUCTION

T

HE PUBLICATION OF the Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings1 (hereafter ‘the Commission Guidance’) represents the culmination of sustained discussion2 and reflection by the European Commission on the scope and place of Article 82 within EC competition law and policy.3 This reform process has been variously and frequently described as a desire to move towards a more economic4 interpretation of the provision and the assessment of conduct by its effects in the market5 and impact on consumer welfare.6 But * Associate Professor, School of Law, University of Warwick. My thanks to Sara-Louise Khabazian for very helpful research assistance. 1 DG Competition, Communication from the Commission, Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings (Brussels, 3 December 2008), http://ec.europa.eu/comm/competition/ antitrust/art82/guidance.pdf. 2 DG Competition, Discussion Paper on the Application of Article 82 of the Treaty to Exclusionary Abuses (December 2005), http://europa.eu.int/comm/competition/antitrust/ others/discpaper2005.pdf. 3 The Guidance is confined to the discussion of exclusionary abuses, not those directly exploitative of consumers: Commission Guidance, above n 1, para 7. 4 See Philip Lowe, Speech delivered at the Thirtieth Annual Conference on International Antitrust Law and Policy, Fordham Antitrust Conference, 23 October 2003, at 2; EAGCP, Report on an Economic Approach to Article 82 EC (July 2005), http://europa.eu.int/ comm/competition/publications/studies/eagcp_july_21_05.pdf; D J Gerber, ‘Two Forms of Modernization in European Competition Law’ (2008) 31 Fordham International Law Journal 1235, at 1247–48. 5 See the Discussion Paper, above n 2, para 4; Neelie Kroes, Competition Commissioner, ‘Preliminary Thoughts on Policy Review of Article 82’, Speech at the Fordham Corporate Law Institute, 23 September 2005, at 2. 6 Commission Guidance, above n 1, para 5.

122 Kathryn McMahon any assessment of reform must be measured against the mischief it claims to replace. Is the new Commission Guidance merely an evolution, or does it mark a genuine break with previous enforcement strategies and interpretation by the Community courts? The Commission is correct to point out that its document ‘is not intended to constitute a statement of law and is without prejudice to the interpretation of Article 82 by the European Court of Justice or the Court of First Instance’.7 However, it does cite Community court decisions in support of its approach in circumstances where it may actually be difficult to reconcile a closer reading of those decisions with the Commission’s stated position. This chapter does not purport to address these many unresolved issues but instead focuses on one aspect of the debate that the Commission has placed at the forefront of its approach, namely, that a dominant undertaking has ‘a special responsibility not to allow its conduct to impair genuine undistorted competition on the common market’.8 What does this concept tell us about the goals of Article 82, and what significance does it have within the reformed approach to the abuse of dominance? II. THE CONCEPT OF SPECIAL RESPONSIBILITY

The possession of market power clearly allows a firm to withhold output and increase price, but this conduct, in itself, has never been an antitrust abuse.9 A dominant firm’s conduct will always have a greater effect on the market than if the same conduct were pursued by a non-dominant firm. The possession of market power therefore provides greater opportunities for the firm to engage in conduct that is more damaging to the competitive process and more likely to result in reduced output and higher prices.10 Conduct such as predatory pricing is generally a rational strategy only for a firm with market power, because only a large firm is likely to be in a position to expand output and have the ability to recoup its investment.11 The 7

Ibid, para 3. Ibid, paras 1 and 9, citing Case 322/81 Nederlandsche Banden Industrie Michelin v Commission [1983] ECR 3461 (‘Michelin I’), para 57; Case T-83/91 Tetra Pak v Commission [1993] ECR II-755 (‘Tetra Pak II’), para 114; Case T-111/96 ITT Promedia v Commission [1998] ECR II-2937, para 139; Case T-228/97 Irish Sugar v Commission [1999] ECR II-2969, para 112; and Case T-203/01 Michelin v Commission [2003] ECR II-4071 (‘Michelin II’), para 97. 9 See, eg, Berkey Photo v Eastman Kodak Co 603 F 2d 263 (2nd Cir, 1979); cert denied 444 US 1093 (1980). 10 Learned Hand CJ, in United States v Aluminum Co of America 148 F 2d 416 (1945), at 430 (citing US v Swift & Co 286 US 106, at 116 (Cardozo J)), stated: ‘… size carries with it an opportunity for abuse that is not to be ignored when the opportunity is proved to have been utilized in the past’. 11 See generally Brooke Group v Brown & Williamson Tobacco Corp 509 US 209 (1993); Weyerhaeuser Co v Ross-Simmons Hardwood Lumber Co 127 S Ct 1069 (2007); G Hay 8

A Reformed Approach to Article 82 and the Special Responsibility 123 potential to do harm explains why the legislature examines the behaviour of dominant firms in more detail. As Justice Scalia noted in dissent in the US Supreme Court decision in Eastman Kodak v Image Technical Services: Where a defendant maintains substantial market power, his activities are examined through a special lens: Behavior that might otherwise not be of concern to the antitrust laws—or that might even be viewed as procompetitive—can take on exclusionary connotations when practiced by a monopolist.12

The ‘special responsibility’ of a dominant undertaking not to impair undistorted competition was first set out by the ECJ in Michelin I.13 This special responsibility does not prevent a dominant firm from taking reasonable steps to protect its own commercial interests, but it requires it to act in a manner proportionate to its economic strength. The ECJ in United Brands stated that Even if the possibility of a counter-attack is acceptable that attack must still be proportionate to the threat taking into account the economic strength of the undertakings confronting each other.14

How useful is the concept of ‘special responsibility’ for identifying abuse under Article 82 such as exclusionary foreclosure by undertakings with market power? The term is oft-repeated in the courts and appears twice in the Commission Guidance.15 However, there is no further analysis of how this concept relates to the objectives of Article 82, and how it may manifest itself in rules for the identification of exclusionary conduct. The seminal statement of abuse under Article 82 is that of the ECJ in Hoffmann-La Roche v Commission: An objective concept relating to the behaviour of an undertaking in a dominant position which is such as to influence the structure of a market where, as a result of the very presence of the undertaking in question the degree of competition is weakened and which, through recourse to methods different from those which condition normal competition in products or services on the basis of the transaction of commercial operators, has the effect of hindering the maintenance of the degree of competition still existing in the market or the growth of that competition.16

Similarly, the ECJ asks in United Brands ‘whether the dominant undertaking has made use of the opportunities arising out of its dominant position’.17

and K McMahon, ‘Predatory Pricing in an Oligopoly Context’ (1995) 3 Competition and Consumer Law Journal 144. 12

Eastman Kodak v Image Technical Services 504 US 451 (1992), at 488. Above n 8, paras 57 and 10. 14 Case C-27/76 United Brands v Commission [1978] ECR 207, para 190; see also BBI/ Boosey & Hawkes: Interim Measures [1987] OJ L286/36, [1988] 4 CMLR 67. 15 Commission Guidance, above n 1, paras 1 and 9. 16 Case 85/76 Hoffmann- La Roche v Commission [1979] ECR 461, para 91. 17 United Brands, above n 14, para 249. 13

124 Kathryn McMahon The emphasis on ‘recourse to methods different from those which condition normal competition’ would seem to imply some causality between dominance and the abuse, and yet this is not an element of liability under Article 82. The ECJ in Hoffmann-La Roche stated that the interpretation suggested by the applicant that an abuse implies that the use of the economic power bestowed by a dominant position is the means whereby the abuse has been brought about cannot be accepted.18

In Tetra Pak II, however, the ECJ declared that the ‘application of Article [82] presupposes a link between the dominant position and the alleged abusive conduct’.19 But this has not been taken to mean the need to establish that but for the dominance the exclusionary effect or leverage in the secondary market could not be achieved. Instead the Court merely requires that the abuse could occur in a secondary (but related) market and not the market where dominance was established.20 In the final analysis, whether a dominant firm is held to a higher a priori duty, requiring it to act in proportion to its strength, or not, is still not determinative of whether abuse has occurred in the particular case. For example, the Commission, in its Microsoft decision, asked whether Microsoft ‘provides to its competitors in the work group server operating system market the interoperability information that it has a special responsibility to provide’.21 This statement merely collapses under its own circularity and does not specify a justiciable standard for the scope of the duty. The idea of special responsibility is potentially broad in its application because the EU has consistently applied a fairly low threshold to find a dominant position. For example, an undertaking with a 40 per cent market share can be in a dominant position.22 The Commission also adopts the broad categories of barriers to entry, as suggested by Bain23 and others, which go beyond the narrow categories (mainly legal barriers) preferred by the Chicago School. The Commission’s approach therefore includes, inter alia, the minimum efficient scale of the firm, economies of scale and scope, network effects, access to supply or distribution and the dominant undertaking’s own conduct.24

18

Hoffmann-La Roche, above n 16, para 9. Case C-333/94 Tetra Pak International SA v Commission [1996] ECR I-5951 (Tetra Pak II), para 27. 20 See R Whish, Competition Law, 6th edn (Oxford, OUP, 2009), at 201. For a different approach to this issue under Australian law, see Queensland Wire v BHP (1989) 167 CLR 177, 202; Melway Publishing Pty Ltd v Robert Hicks Pty Ltd (2001) 205 CLR 1, para 23. 21 Microsoft, Case COMP/C-3/37.792, European Commission, para 33 (emphasis added). 22 Commission Guidance, above n 1, para 14; Case 62/86 AKZO v Commission [1991] ECR I-3359. 23 See generally, J S Bain, Barriers to New Competition: their character and consequences in manufacturing industries (Cambridge, Massachusetts, Harvard University Press, 1956). 24 Commission Guidance, above n 1, para 17. 19

A Reformed Approach to Article 82 and the Special Responsibility 125 If the court asks the correct competition law questions about whether the conduct has an exclusionary effect on the market, any special responsibility will be irrelevant to the determination of liability. On the other hand, if the court does not formulate the questions in this way and merely seeks to restrain the conduct of a monopolist, it may be punishing an undertaking for merely being in, but not necessarily for abusing, that dominant position.25 The concept of special responsibility has also been linked to the equally problematic concept of ‘super-dominance’. In its Microsoft decision the Commission found that Microsoft’s market share of over 90 per cent placed it in a ‘quasi-monopoly’ position, an ‘overwhelmingly dominant position’.26 The Commission, citing Compagnie Maritime Belge v Commission,27 referred to super-dominance,28 which it defined as the ‘particularly onerous special obligation’ attaching to undertakings which enjoy a ‘dominance verging on monopoly’.29 Whish notes that ‘if a dominant undertaking has a “special” responsibility, a super-dominant has one that is even greater’.30 The scope of the special responsibility is considered in the light of the special circumstances of each case.31 This has been found by the Commission to be related to the degree of dominance held by the firm.32 But it is very difficult to identify exactly what duties attach to the special responsibility, and to determine how these duties manifest themselves in rules for the identification of predatory behaviour. This problem is only exacerbated by use of the term ‘super-dominance’. As Appeldoorn notes, attaching greater duties to ‘super-dominance’ is confusing and sets the legal boundaries around such a firm based ‘on its size alone, and not on its behaviour’.33

25 Learned Hand CJ, in United States v Aluminum Co of America, above n 10, at 430, stated, ‘the Act does not mean to condemn the resultant of those very forces which it is its prime object to foster … The successful competitor, having been urged to compete, must not be turned upon when he wins’. 26 Microsoft, Commission Decision, above n 21, para 435. 27 Ibid, n 560, citing the Opinion of A-G Fennelly in Joined Cases C-395/96P & C-396/ 96P Compagnie Maritime Belge and others v Commission [2000] ECR I-1365 (‘Cewal’), para 136. 28 Microsoft, Commission Decision, above n 21, n 560. On the concept of ‘super-dominance’, see generally Whish, above n 20, at 185–86. 29 Microsoft, Commission Decision, above n 21, n 560. See also Napp Pharmaceutical Holdings Ltd v Director General of Fair Trading [2002] Comp Appeal Reports 13, available at: http://www.catribunal.org.uk/files/JdgNapp150102.pdf, para 214. 30 Whish, above n 20, at 185. 31 Tetra Pak II, above n 8, para 24. 32 See Deutsche Post AG—Interception of cross-border mail [2001] OJ L331/40, [2002] 4 CMLR 598, para 103. 33 J Appeldoorn, ‘He who spareth his rod, hateth his son? Microsoft, super-dominance and Article 82 EC’ (2005) 26 European Competition Law Review 653, at 656 and 657.

126 Kathryn McMahon The Commission in its Article 82 Guidance also suggests, citing Cewal that the degree of dominance will be a relevant factor in establishing foreclosure effects: Experience suggests that the higher the market share and the longer the period of time over which it is held, the more likely it is that it constitutes an important preliminary indication of the existence of a dominant position and, in certain circumstances, of possible serious effects of abusive conduct, justifying an intervention by the Commission under Article 82.34

It further states that ‘[i]n general, the stronger the dominant position, the higher the likelihood that conduct protecting that position leads to anticompetitive foreclosure’.35 Does this mean that ‘super-dominance’ sets up a rebuttable presumption of foreclosure? If it does not set up a presumption then what is the substance of the more onerous duty?36 III. THE DIVERGENCE BETWEEN THE OBJECTIVES OF EU AND US COMPETITION POLICY

The EU approach, where undertakings in a dominant position are subject to a special responsibility, contrasts with the US approach, where the courts have repeatedly held that a monopolist has no special duties under Section 2 of the Sherman Act with respect to other market participants. The notion that a monopolist should have to co-operate with its rivals, exercise any special restraint or be held to a standard of behaviour which differs from that of other competitors was dismissed by Judge Posner in Olympia Equipment: Forty years ago it was thought that even a firm with a lawful monopoly … could not be allowed to defend its monopoly against would-be competitors by tactics otherwise legitimate; it had to exercise special restraint … Later, as the emphasis of antitrust policy shifted from the protection of competition as a process of rivalry to the protection of competition as a means of promoting economic efficiency, it became recognized that the lawful monopolist should be free to compete like everyone else; otherwise the antitrust laws would be holding an umbrella over inefficient competitors. A monopolist, no less than any other competitor, is permitted and indeed encouraged to compete aggressively on the merits …37

34 Commission Guidance, above n 1, para 15, citing Cewal n 27 above, para 119 and Case T-228/97 Irish Sugar plc v Commission [1999] ECR II-2969, para 186. 35 Commission Guidance, above n 1, para 20. 36 See Discussion Paper, above n 2, para 91. 37 Olympia Equipment Leasing v Western Union Telegraph 797 F 2d 370 (7th Cir 1986), at 375; cf Berkey Photo Co v Eastman Kodak Co 603 F 2d 263 (2nd Cir 1979); certiorari denied 444 US 1093 (1980), at 281.

A Reformed Approach to Article 82 and the Special Responsibility 127 More recently, in Verizon Communications Inc v Trinko, the Supreme Court concluded that Verizon’s failure to provide sufficient assistance in the supply of services to its rivals was not a recognised antitrust claim under the Court’s refusal to deal precedents, as ‘there is no duty to aid competitors’.38 In particular: Compelling such firms to share the source of their advantage is in some tension with the underlying purpose of antitrust law, since it may lessen the incentive for the monopolist, the rival, or both to invest in those economically beneficial facilities.39

What accounts for the difference in treatment of the special responsibility concept under EU and US Law? In Olympia Equipment, Judge Posner saw the rejection of the concept as directly linked to the evolution of antitrust policy in the US, and more particularly to the shift from the protection of competition as a process of rivalry to the promotion of economic efficiency. Posner’s account can be questioned, as the Chicago School has been accused of rewriting the history and intent of the US Sherman Act to more closely resemble their own efficiency model.40 Nevertheless, the special duty concept has been regarded as inconsistent with the Chicagoan focus on ‘consumer welfare as efficiency’ and ideas concerning the ‘self-correcting’ power of markets.41 In the EU the view has also been expressed that the special responsibility concept is the last bastion of an obsolete approach which should now be jettisoned in favour of the modernisation of Article 82, namely, a tighter focus on ‘consumer welfare’ and ‘a more economic approach’.42 Notwithstanding attempts at trans-Atlantic convergence, major differences are still perceived between the two competition models. Certainly, in the wake of the GE/ Honeywell43 and Microsoft44 decisions, the US strongly criticised the EU

38 Verizon Communications Inc v Law Offices of Curtis V Trinko 540 US 398 (2004), 124 S Ct 872, at 411. 39 Ibid, at 407–08. 40 See generally R Bork, The Antitrust Paradox: A Policy at War with Itself (New York, Basic Books, 1978; rev edn 1993), at 50–66, 90–91; R J R Peritz, Competition Policy in America 1888–1992 (Oxford, OUP, 1996), ch 5; R L Lande, ‘Wealth Transfers as the Original And Primary Concern of Antitrust: The Efficiency Interpretation Challenged’ (1982) 34 Hastings Law Journal 65. 41 See Bork, above n 40, at 81–92. Section 2 of the Sherman Act has also recently been subject to review: see US DoJ, Competition and Monopoly: Single-Firm Conduct under Section 2 of the Sherman Act, available at http://www.usdoj.gov/atr/public/reports/236681.pdf. 42 R Allendesalazar, ‘Can We Finally Say Farewell to the “Special Responsibility” of Dominant Companies?’ in C-D Ehlermann and M Marquis (eds), European Competition Law Annual 2007: A Reformed Approach to Article 82 EC (Oxford, Hart Publishing, 2008), at 319. 43 Case No COMP/M.2220; Case T-210/01 General Electric v Commission [2005] ECR II-5575. 44 Case T-201/04 Microsoft v Commission [2007] ECR II-3601.

128 Kathryn McMahon for a perceived failure to protect consumer welfare and an unwarranted intervention in the global market strategies of US firms.45 In their amicus brief in the Linkline appeal from the Ninth Circuit46 to the US Supreme Court,47 a group of eminent US antitrust professors, led by Robert Bork and J Gregory Sidak, set out what they perceived to be the differences between the EU and US approaches to antitrust: The alternative to consumer-welfare maximization is the view that antitrust law is simply one more tool of industrial policy, and thus its application may permissibly compromise consumer welfare to advance the welfare of competitors. Other nations evidently consider this normative proposition to be appropriate, if recent developments in the European Union are a valid indication. More than ever before, the United States and Europe appear to be at a fork in the road over whether the law of monopolization exists to protect consumers or to ensure that a specified number of firms will profitably populate a market.48

Does the amicus brief correctly identify the differences between the US and EU law? It is not the intention of this chapter to argue for the reassertion of the status or significance of the special responsibility concept, which the author contends has little substantive content. The unanswered question remains whether its continued invocation by the European courts and the Commission is somehow representative of a distinct approach to EC competition law? These differences may be traced to a greater concern under EU law with the exclusionary effects arising from the distortion of market structure, as encapsulated in some idea of a special responsibility, as opposed to more Chicagoan ideas concerning self-correcting markets and consumer welfare as efficiency.49 Amato suggests that the EU concept of special responsibility of dominant firms has its origins in fairness50 and that it imposes a seemingly

45 See, eg, Thomas O Barnett, Statement on European Microsoft Decision (17 September 2007), http://www.usdoj.gov/atr/public/press_releases/2007/226070.htm; E M Fox, ‘We Protect Competition, You Protect Competitors’ (2003) 26 World Competition 149. 46 Linkline Communications, Inc v Pacific Bell Telephone Co, AT&T 503 F 3d 876 (9th Cir 2007). 47 Pacific Bell Telephone Co v Linkline Communications, Inc (No. 07-512) 503 F. 3d 876 US (2009), decided 25 February 2009. 48 Brief of Amici Curiae Professors and Scholars in Law and Economics in support of the petitioners, Pacific Bell Telephone Company v Linkline Communications (No 07-512), at 5. Other academic signatories included Baumol, Demsetz, Elzinga, Epstein, Fisher, Hahn, Hausman and Jorde. 49 See E M Fox, ‘Monopolization, Abuse of Dominance, and the Indeterminancy of Economics: The US/EU Divide’ (2006) Utah Law Review 725. 50 See G Amato, Antitrust and the Bounds of Power: The Dilemma of Liberal Democracy in the History of the Market (Oxford, Hart Publishing, 1997), at 65–66; D J Gerber, Law and Competition in Twentieth Century Europe, Protecting Prometheus (Oxford, OUP, 1998), at 367–68; see discussion in D Howarth and K McMahon, ‘Windows has performed an illegal operation: The Court of First Instance’s Judgment in Microsoft v Commission’ [2008] European Competition Law Review 117, at 130–31; K McMahon,

A Reformed Approach to Article 82 and the Special Responsibility 129 public rather than a private law burden.51 Gerber refers to the development of the ‘abuse’ concept under Article 82 in several refusal to supply decisions as consistent with the protection of small and medium-sized firms and concerns about the ability of large firms to extract unfair prices and terms from smaller enterprises.52 He refers specifically to notions of economic dependency developed in German competition law where a supplier did not have ‘sufficient and reasonable possibilities’ to shift to another purchaser.53 This included the concept of ‘relative dominance’, where economic power was a problem even where there was no dominance of a market. Gerber notes that if a firm was ‘dependent’ on another firm, the firm with this ‘relative power’ was in a position to harm or destroy the dependent, and thus it should be prevented from abusing this power …54

Amato explains the difference between the EU and US standards of assessment in the following terms: ... in the USA leading to illegality only when both the effect of exclusion and of reduction of consumer welfare appear together, while in Europe they are oriented to the special responsibility of firms in dominant positions to protect existing small competitors and hence in any case take account of possible damage to such competitors.55

Those in the US would no doubt regard these EU goals as a return to an earlier interpretation of the Sherman Act which was excessively concerned with deconcentration and oligopoly theory,56 the foreclosing of competition which may ‘deprive rivals of a fair opportunity to compete’,57 and a purpose ‘to perpetuate and preserve, for its own sake and in spite of possible cost, an organization of industry in small units’.58

‘Interoperability: “Indispensability” and “Special Responsibility” in High Technology Markets’ (2007) 9 Tulane Journal of Technology & Intellectual Property 123, at 161–66. 51

Amato, above n 50, at 66. Gerber, above n 50, at 367–368, citing United Brands, above 14, and Hoffman-La Roche, above n 16; Amato, above n 50, at 70. 53 The prohibition against discriminatory or obstructive conduct in the German Act against Restraints of Competition is extended by s 20 to certain undertakings without a dominant position. It includes undertakings on which small and medium-sized enterprises are economically dependent (s 20(2)). The finding of a dominant position in cases such as Case 22/78 Hugin v Commission [1979] ECR 1869, in the aftermarket for proprietary spare parts in circumstances where there are no available substitutes, is also based on some notion of ‘dependency’. 54 Gerber, above n 50, at 316 and n 140. Gerber states the concept was developed primarily by the economist Helmut Arndt in Markt und Macht, 2nd edn (Tübingen, JCB Mohr 1973) and Wirtschaftliche Macht, 3rd edn (Munich, Beck, 1980). 55 Amato, above n 50, at 70–71. 56 See, eg, US DoJ Merger Guidelines (1968); R J R Peritz, above n 40, ch 4. 57 Brown Shoe Co v United States 370 US 294 (1962), at 324. 58 Judge Learned Hand in United States v Aluminium Co of America, above n 10, at 429. 52

130 Kathryn McMahon IV. THE GOALS OF ARTICLE 82

The EC focus on preserving rivalry, preventing foreclosure and ensuring ‘competition on the merits’59 is embedded in a unique institutional and political history which prioritises market integration and, in Article 3(1)(g) of the EC Treaty, sets out a system ensuring that competition in the internal market is not distorted.60 In particular, the theoretical foundations of EC competition law are strongly linked to the ideas of ordoliberalism61 and its concern with individual economic freedom as a value in itself. Under this view, allocative efficiency is ‘but an indirect and derived goal’.62 Competition law has an essential role in the maintenance of the ideal of an ‘economic constitution’, whereby the structure of the market is given a legal form. The State has an important role in safeguarding individual economic freedom against the exercise of private power.63 This is linked to ensuring that markets are not foreclosed, forming an obstacle to an integrated market.64 The ideas of equality of opportunity and freedom of contract65 have arguably had an impact, for example, on the notion of a ‘consumer’ in EC competition law. This concept has been interpreted within Article 81(3) as meaning not merely the final consumer but all direct or indirect users of the products, wholesalers and retailers.66 In other words, EC competition law analysis includes a consideration of foreclosure and preserving freedom to compete at any level in the chain of distribution. This approach does not merely focus on the impact on the ultimate consumer as understood in the US antitrust concept of consumer welfare and the Chicago School utilitarian ‘wealth maximisation’.67 The Commission Guidance, while insisting that its approach will examine conduct ‘most harmful to consumers’, is also clearly

59 See generally OCED, ‘Competition on the Merits’ DAF/COMP(2005)27, available at http://www.oecd.org/dataoecd/7/13/35911017.pdf. 60 See generally H Schweitzer, ‘The History, Interpretation and Underlying Principles of Section 2 Sherman Act and Article 82 EC’ in Ehlermann and Marquis (eds), above n 42, at 119. 61 See generally, Gerber, above n 50, ch VII; cf P Akman, ‘Searching for the Long-Lost Soul of Article 82 EC’ (2007) CCP Working Paper 07-5. 62 See W Möschel, ‘The Proper Scope of Government Viewed from an Ordoliberal Perspective: The Example of Competition Policy’ (2001) 62 Journal of Institutional and Theoretical Economics 3, at 4; L L Gormsen, ‘The Conflict between Economic Freedom and Consumer Welfare in the Modernisation of Article 82 EC’ (2007) 3 European Competition Journal 329. 63 Gormsen, above n 62. 64 See Schweitzer, above n 60, at 123. 65 See Möschel, above n 62, at 10. 66 See Commission Notice, Guidelines on the application of Article 81(3) of the Treaty (2004/C101/08), para 84. Against see Case T-168/01 GlaxoSmithKline v Commission [2006] ECR II-2969, para 118. See also Peritz, above n 40, at 256–58. 67 Möschel argues that such a utilitarian calculus, implicit in the Chicago model, is incompatible with the ordoliberal system of values which treats individuals as ends in themselves and not as the means of another’s welfare: Möschel, above n 62, at 7.

A Reformed Approach to Article 82 and the Special Responsibility 131 distinguishing the concept from that understood in the US. The Commission states that it will address anticompetitive foreclosure both at the ‘intermediate level and/or at the level of final consumers’,68 defining consumers as all direct or indirect users of the products affected by the conduct, including intermediate producers that use the products as an input, as well as distributors and final consumers both of the immediate product and of products provided by intermediate producers. Where intermediate users are actual or potential competitors of the dominant undertaking, the assessment focuses on the effects of the conduct on users further downstream.69

The proper identification and correct evaluation of goals is essential for any system of competition law. As Bork has argued, ‘Only when the issue of goals has been settled is it possible to frame a coherent body of substantive rules’.70 What can we discern about the goals of Article 82 as defined by the Commission? In its Guidance the Commission states that it will focus on those types of conduct that are most harmful to consumers ... The emphasis of the Commission’s enforcement activity in relation to exclusionary conduct is on safeguarding the competitive process in the internal market and ensuring that undertakings which hold a dominant position do not exclude their rivals by other means than competing on the merits of the products or services they provide. In doing so the Commission is mindful that what really matters is to protect an effective competitive process and not simply protecting competitors. This may well mean that competitors who deliver less to consumers in terms of price, choice, quality and innovation will leave the market.71

That this approach can be distinguished from the Chicago School’s ‘consumer welfare as economic efficiency’ approach is further exemplified by the Commission’s concern for ‘residual competitors’ even if they are not as efficient as the dominant firm. Commissioner Kroes has stated: ‘[w]e need to take into account not only short term harm, but also medium and long term harm arising from the exclusion of competitors ... should residual competitors be foreclosed’.72 It is considered important to ensure that these residual competitors ‘be able to improve in efficiency and thereby increase competition to the dominant company’.73 The focus is on competition in the long run, a belief that the competitive process, if undistorted, will eventually deliver consumer benefits. The protection of a less efficient residual competitor is clearly a competition policy which protects rivalry without further investigation into the

68

Commission Guidance, above n 1, para 19. Ibid, n 15. 70 Bork, above n 40, at 50. 71 Commission Guidance, above n 1, paras 5 and 6. See also Discussion Paper, above n 2, paras 4, 54. 72 Kroes, above n 5. 73 Ibid. 69

132 Kathryn McMahon effect on ‘consumer welfare’ in the US sense. In many ways it is a reassertion of different and special duties being placed on dominant firms due to their ‘very presence’74 in the market. As Kroes pointed out: A famous phrase in the Hoffman-La Roche judgment is ‘through recourse to methods different from those which condition normal competition’. This is, of course, connected to the well-known issue of what constitutes ‘competition on the merits’.75

The difficulty with this approach is that the Commission does not always translate the objectives of the protection of rivalry and less efficient residual competitors into consistent and transparent rules or standards for application to specific cases. For example, the Commission Guidance states that price-based exclusionary abuse will be found if the conduct is capable of hampering competition from a competitor who is ‘as efficient’ as the dominant undertaking.76 At the same time it goes on to set out circumstances in which a ‘less efficient’ competitor should also be protected: 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 a particular price-based conduct leads to anticompetitive foreclosure. The Commission will take a dynamic view of this constraint, given that in the absence of an abusive practice such a competitor may benefit from demand-related advantages, such as network and learning effects, which will tend to enhance its efficiency.77

At other times the Commission will use a ‘profit sacrifice’ test,78 which condemns conduct which is rational (that is, profit-maximising) merely on the premise that competitors are to be driven from the market or damaged.79 For example, in the determination of predatory pricing the Commission states that a failure to cover Average Avoidable Cost (AAC) indicates that the dominant undertaking is ‘sacrificing profits’ in the short term and that an as-efficient competitor cannot serve the targeted customers without incurring a loss.80 But later, when dealing with ‘conditional rebates’, it states that foreclosure can occur ‘without necessarily entailing a sacrifice for the dominant undertaking’.81

74

See Hoffmann-La Roche v Commission, above n 16, para 91. Kroes, above n 5. Commission Guidance, above n 1, para 22. 77 Ibid, para 23. 78 Ibid, paras 63–65. 79 See generally A Gavil, ‘Symposium: Integrating new economic learning with antitrust doctrine—Exclusionary distribution strategies by dominant firms: striking a better balance’ (2005) 72 Antitrust Law Journal 3, at 5; G J Werden, ‘Identifying Exclusionary Conduct Under Section 2: The “No Economic Sense” Test’ (2005) 73 Antitrust Law Journal 413; E M Fox, ‘Is There Life in Aspen after Trinko? The Silent Revolution of Section 2 of the Sherman Act’ (2005) 73 Antitrust Law Journal 3. 80 Commission Guidance, above n 1, paras 25, 62–65. 81 Ibid, para 36. See discussion of rebates below. 75 76

A Reformed Approach to Article 82 and the Special Responsibility 133 Potentially conflicting standards are also apparent in the Commission’s assessment of abusive conduct amounting to a refusal to supply, where the Commission requires in addition to the establishment of ‘indispensability’ that ‘the refusal is likely to lead to the elimination of effective competition on the downstream market; and the refusal is likely to lead to consumer harm’.82 In other words, establishing foreclosure/reduction in rivalry is insufficient to determine liability83; there is an additional, but separate, requirement of the likelihood of ‘consumer harm’, which is defined as follows: In examining the likely impact of a refusal to supply on consumer welfare, the Commission will examine whether for the consumers, the likely negative consequences of the refusal to supply in the relevant market outweigh over time the negative consequences of imposing an obligation to supply.84

Consumer harm or detriment to consumer welfare is therefore not implicit in the establishment of foreclosure. Are we seeing a different standard from that of ‘protection of rivalry’ applied here? It can be argued that the concept of ‘consumer welfare’ applied in this case is closer to that of the Chicago School, where the focus is on the final consumer and not, as the Commission previously found, on both final and intermediate ones. The Commission further defines ‘consumer harm’ as occurring where competitors are, as a result of the refusal, prevented from bringing innovative goods or services to the market,85 or where they intend to produce new or improved goods or services for which there is a potential consumer demand.86 This brings the test for a general refusal to supply into line with that generally reserved for a refusal to license an intellectual property right.87 These are clearly more complicated considerations than a mere quantitative measurement of foreclosure, reduction in rivalry or inequality of opportunity. In highlighting the apparently conflicting and variable standards used by the Commission it is not intended to make an argument for one unifying rule which is applicable to all abusive conduct. The Commission Guidance is already a vast improvement on the exposition of principles and standards with respect to the formalistic categories of conduct set out in their Article 82 Discussion Paper.88 What remains to be done is to make a greater effort to explain, on a case-by-case basis, in what way these rules are designed to achieve the goals of Article 82. 82

Ibid, para 80. Ibid, para 84. Ibid, para 85. 85 Ibid, para 86. 86 Ibid, para 87. 87 Case C-418/01 IMS Health GmbH & Co OHG v NDC Health GmbH & Co KG [2004] ECR I-5039, para 49; Joined Cases C-241/91P and C-242/91 RTE & ITP v Commission [1995] ECR I-743; Case T-201/04 Microsoft v Commission [2007] ECR II-3601. 88 Discussion Paper, above n 2. 83 84

134 Kathryn McMahon V. THE IMPORTANCE PLACED ON THESE CONCEPTS IN RECENT CASE LAW

What can we discern about the purposes of Article 82 and the emphasis placed on the concepts of special responsibility and undistorted competition in the recent case law of the Community courts? These principles have been applied in the decisions in Wanadoo, British Airways and Microsoft, to hold dominant firms to an apparently higher standard than those in a non-dominant position. In Wanadoo the Commission,89 and the Court of First Instance (CFI) affirming on appeal,90 applied the special responsibility concept in a predatory pricing case to set out a higher standard for dominant firms in meeting competition. A subsidiary of France Télécom (now Wanadoo) was fined €10.3 million by the Commission for pricing its broadband Internet connection fee for residential customers below cost for a period of time and then pricing at cost together with an overall predatory plan to destroy competition.91 On appeal to the CFI, Wanadoo argued that even if it was found to be pricing below cost, it was merely ‘meeting competition’ in aligning its prices in good faith to its competitors’ prices.92 The CFI stated, citing United Brands, that the competition provisions cannot be used to deprive a dominant undertaking of the right to protect its own commercial interests if they are attacked and such an undertaking must be allowed the right to take such reasonable steps as it deems appropriate to protect those interests, such behaviour cannot be countenanced if its actual purpose is to strengthen this dominant position and abuse it.93

The CFI went on to note, however, the ‘specific obligations’ imposed on undertakings in a dominant position,94 which may in specific circumstances deprive such undertakings of the right to adopt a course of conduct or take measures which are not in themselves abuses and which would even be unobjectionable if adopted or taken by a non-dominant undertaking …95

It is difficult to know what conduct the CFI is condemning. The conduct would not be considered predatory pricing under the AKZO standard,96

89

COMP/38.233, [2005] 5 CMLR 120. Case T-340/03 France Télécom SA v Commission [2007] ECR II-107. An appeal to the ECJ was dismissed: Case C-202/07 France Télécom SA v Commission, decided 2 April 2009. 91 Applying the test for predatory pricing in Case 62/86 AKZO v Commission [1991] ECR I-3359, paras 71 and 72. 92 France Télécom, above n 90, para 171. See also para 181. 93 Ibid, para 185. 94 Citing Case T-111/96 ITT Promedia v Commission [1998] ECR II-2937, para 139. 95 France Télécom, above n 90, para 186. 96 Case C-62/86 AKZO v Commission, above n 91, para 72. 90

A Reformed Approach to Article 82 and the Special Responsibility 135 and the CFI did not object to the aligning of prices as long as they do not fall below cost.97 The CFI went on to state that while the alignment of prices is not in itself abusive or objectionable, it might become so where it is aimed not only at protecting its interests but also at strengthening and abusing its dominant position.98

This is arguably a modification of the subjective element of the AKZO standard. Where a dominant undertaking prices above Average Variable Cost (AVC) but below Average Total Cost (ATC), it may no longer need be shown to have a subjective ‘predatory plan’ to be found guilty of an abuse. This is replaced by an arguably lesser, more objective standard of ‘strengthening and abusing its dominant position’. Could this new element also be used to render above ATC pricing abusive? When is a dominant firm deprived of the right to adopt measures which are ‘not in themselves abuses’? In the absence of clearly-defined criteria or further explanation, this seems to impose a form of per se liability on a firm in a dominant position. The lack of clarity of what constitutes abusive pricing for the dominant firm is particularly problematic in the context of high-technology markets, such as the one at issue in Wanadoo. The market for broadband is characterised by high fixed costs and low marginal costs, making the measurement and allocation of costs for different services particularly problematic for the purposes of determining predatory pricing. These are emerging and developing markets where we encounter learning effects or economies of scale, so that efficient output levels (which are pro-competitive) may be achieved only after an initial period of low or below cost prices, which can include promotional pricing. Costs are expected to be recouped in the long run when an optimal level of connections is achieved, taking into account network effects.99 The below AVC AKZO standard appears to be less useful here. Elzinga and Mills point out that temporary price wars may be triggered by entry where buyers incur set-up or switching costs when dealing with new suppliers. In this model, the entrant must discount its price below AAC to induce buyers to absorb set-up costs. The incumbent drops its price simultaneously to retain some of its buyers. Once the entrant has locked in its buyers, prices rise, albeit not to pre-entry levels.100

97

France Télécom, above n 90, para 183. Ibid, para 187. 99 K G Elzinga and D E Mills, ‘Predatory Pricing and Strategic Theory’ (2001) 89 Georgetown Law Journal 2475; see also P Bolton, J F Brodley and M H Riordan, ‘Predatory Pricing: Strategic Theory and Legal Policy’ (2000) 88 Georgetown Law Journal 2239. 100 Elzinga and Mills, above n 99, at 2485. AAC is average avoidable cost meaning those costs that could have been avoided had the firm not produced the predatory increment of output. 98

136 Kathryn McMahon Whether this behaviour is abusive will largely depend on the availability of substitutes and potential competition. The switching costs of substituting a new provider may be relatively low in broadband markets: while the customer may have entered into a long-term service agreement with the incumbent, the modem is usually provided and the connection fee may be minimal. It was precisely these arguments based on learning effects and economies of scale which Wanadoo used to justify its below cost pricing. But the CFI rejected them101 and instead focused on Wanadoo’s recognition of ‘the period of profitability only coming later’102 to establish liability for an overall ‘predatory plan’. The CFI found documents evidencing a detailed plan of pre-emption—‘our pre-emption of the ADSL market is imperative’103—which would not be achieved if prices were too high. The CFI stated: An undertaking which charges predatory prices may enjoy economies of scale and learning effects on account of increased production precisely because of such pricing. The economies of scale and learning effects cannot therefore exempt that undertaking from liability under Article 82.104

But such conduct is equally consistent with competition in network markets, and the CFI does not appear to acknowledge those circumstances where such pricing may be justifiable. The Commission Guidance does recognise that while pricing below AAC would be a clear indication of profit sacrifice,105 there are circumstances in which undertakings should not be penalised for incurring ex post losses where the ex ante decision to engage in the conduct was taken in good faith, ie if they can provide conclusive evidence that they could reasonably expect that the activity would be profitable.106

This seems to be confined to a genuine mistake about expected profits, perhaps related to the erroneous marketing of a new product, rather than including cases of an overall rational strategy. The Commission does concede, however, unlike the CFI in Wanadoo, that it will ‘consider claims by dominant undertakings that the low pricing enables it to achieve economies of scale or efficiencies related to expanding the market.’107

101 102 103 104 105 106 107

France Télécom, above n 90, para 217. Ibid, para 215. Ibid, para 199. Ibid, para 217. Commission Guidance, above n 1, para 62. Ibid, n 43. Ibid, para 73.

A Reformed Approach to Article 82 and the Special Responsibility 137 The CFI in its Microsoft108 decision again too readily found a breach of Article 82 by conduct which distorted structure, rather than undertaking a more focused examination of the effect on competitors and consumer welfare. Microsoft was found to have refused to supply interoperability information,109 contrary to Article 82, to Sun Microsystems (and others) who required this information to create work group server operating systems that fully interoperated with Microsoft’s Windows client PC operating system. The CFI seemed to equate consumer detriment with disadvantage to competitors. Microsoft’s ‘artificial advantage’ by way of interoperability placed those competitors ‘at a disadvantage by comparison with Microsoft’.110 The CFI stated that ‘Microsoft impaired the effective competitive structure on the work group server operating systems market by acquiring a significant market share on that market’.111 The CFI also placed the obligation to supply the information within the concept of special responsibility: Microsoft ‘had not taken sufficiently into account its special responsibility not to hinder effective and undistorted competition in the common market’.112 But the acquisition of market share, in itself, is not a competition law issue. The decline in market share of the competitors on the work group server market could be attributed to a number of sources other than the refusal to supply interoperability information, including the superior efficiency of Microsoft products, consumer preference for a homogeneous market solution, or inertia from lock-in due to switching costs in a network environment. While structure can clearly affect outcomes, a focus on this issue alone without a closer examination of the impact of conduct on consumer welfare (and efficiency) is problematic in network environments. It does not grant sufficient deference to the manner in which competition actually occurs in such high-technology markets.113 Gormsen argues that the CFI judgment in Microsoft can only be understood in the context of the ordoliberal concern ‘for the economic freedom of rivals and the structure of competition more than consumer welfare’,114

108

Microsoft v Commission, above n 44. See generally Howarth and McMahon, above

n 50. 109 ‘Interoperability’ generally arises in high technology markets where software producers may be dependent on interface information concerning the operation of other software (eg platform operating systems) in order to produce products (applications) which are compatible. 110 Microsoft v Commission, above n 44, para 653. 111 Ibid, para 664. 112 Ibid, para 775. See also para 229. 113 See generally M Motta, Competition Policy, Theory and Practice (Cambridge, CUP, 2004) at 483–85; C Ahlborn, D Evans and A Padilla, ‘Competition Policy in the New Economy: Is Competition Law Up to the Challenge’ [2001] European Competition Law Review 156. 114 Gormsen, above n 62, at 342.

138 Kathryn McMahon because otherwise it is ‘hard to find a coherent theory on consumer harm’.115 But if this reading is correct, is a sole concern for economic freedom, in the absence of an explicit finding with respect to consumer harm, enough to impose liability? Similar questions concerning the scope of the duty applied to dominant undertakings arose in British Airways. British Airways, the dominant buyer in the air travel agency services, was fined €6.8 million by the Commission for offering travel agents loyalty-inducing rebates for meeting or exceeding previous year sales targets.116 The decision was affirmed on appeal by the CFI117 and the ECJ.118 British Airways applied a ‘rolled-back’ scheme, whereby once the target number of sales was reached, commission was applied not just on the tickets sold thereafter, but also retrospectively on all the tickets sold in the relevant period.119 It was found that the marginal sales needed to reach the target had a disproportionate and exclusionary effect on the market for UK air travel agent services. The rebates were also found to be discriminatory contrary to Article 82(c). It is not the intention of this chapter to assess the EC law concerning discrimination, rebates and loyalties.120 Instead the focus is on assessing the apparent willingness of the courts to assert foreclosure/abuse as a simple outcome of market structure. The CFI acknowledged that Article 82 does not require that the conduct result in an anticompetitive effect but rather that the abusive conduct ‘tends to restrict competition, or ... that the conduct is capable of having, or likely to have, such an effect.’121 But the proof of this tendency or likelihood seems to be established by the mere finding of dominance, and exclusionary effect is too readily equated with a mere loss in equality of opportunity and freedom to compete. The CFI stated that since Article 82 EC is aimed at penalising even an objective detriment to the structure of competition itself ... BA’s argument that there is no proof of damage caused to consumers by its reward schemes cannot be accepted …122

115

Ibid. Virgin/British Airways [2000] OJ L 30/1. 117 Case T-219/99 British Airways v Commission [2003] ECR II-5917 (CFI). 118 Case C-95/04P British Airways v Commission [2007] ECR II-4071 (ECJ). 119 Ibid, para 73. 120 See, eg, G Faella, ‘The Antitrust Assessment of Loyalty Discounts and Rebates’ (2008) 4 Journal of Competition Law and Economics 373; OECD, Loyalty and Fidelity Discounts and Rebates (2002), available at http://www.oecd.org/dataoecd/18/27/2493106.pdf. 121 British Airways, above n 117, para 293. 122 Ibid, para 311, citing Case 6/72 Europemballage and Continental Can v Commission [1973] ECR 215. The CFI went on to find that as travel agents carried out 85% of all air ticket sales in the UK, BA’s abusive conduct ‘cannot fail to have had the effect of excluding competing airlines’: ibid, para 295. 116

A Reformed Approach to Article 82 and the Special Responsibility 139 Similarly, the ECJ asked whether the discount ‘tends to remove or restrict the buyer’s freedom to choose his sources of supply’,123 and whether the effect is to make it ‘more difficult or impossible for its co-contractors to choose between various sources of supply or commercial partners’.124 The ECJ stated that the pressure exerted on resellers is ‘further strengthened where that undertaking holds a very much larger market share than its competitors.’125 But the exclusionary effect seems to be assessed as a mere by-product of BA’s market share: by its ‘significantly higher market share, the undertaking in a dominant position generally constitutes an unavoidable business partner in the market’.126 The ECJ appears to be introducing the notion of ‘indispensability’, as understood as part of the test for a refusal to supply, but without further examination of the actual competition requirement under that test that ‘there is no actual or potential substitute’.127 According to the ECJ, Article 82 is aimed not only at practices which may cause prejudice to consumers directly, but also ‘at those which are detrimental to them through their impact on an effective competition structure, such as is mentioned in Article 3(1)(g) EC’.128 Because of this, the CFI was entitled not to examine whether BA’s conduct had caused prejudice to consumers under Article 82(b) but could instead inquire whether the schemes ‘had a restrictive effect on competition’.129 The sale of an identical number of BA tickets being remunerated at different levels according to whether or not the travel agents had attained sales targets was also found discriminatory contrary to Article 82(c).130 But this conclusion too seemed to be determined by the simple assessment that the different levels of commission distorted competition between them and the mere finding that the bonus schemes ‘could lead to exponential changes in the revenue of travel agents’.131 The determination of undistorted competition is merely equated with ‘hindering the competitive position of some of the business partners of that undertaking in relation to the others’.132 No further proof was required of

123

British Airways, above n 118, para 67. Ibid, para 68. 125 Ibid, para 75. 126 Ibid. 127 See Joined Cases C-241/91P and C-242/91 RTE & ITP v Commission, above n 87, paras 52 and 53; Case 7/97 Oscar Bronner v Mediaprint [1998] ECR I-7791, paras 44–45; Case T-201/04 Microsoft v Commission [2007] ECR II-3601, para 421; see also Commission Guidance, above n 1, para 82, n 57. 128 British Airways, above n 118, para 106, citing Continental Can, above n 123. 129 Ibid, para 170. 130 Ibid, para 139. 131 Ibid, para 147. 132 Ibid, para 144, citing Case 40/73 Suiker Unie v Commission [1975] ECR 1663, paras 523 and 524. 124

140 Kathryn McMahon an ‘actual quantifiable deterioration in the competitive position of the business partners’.133 The ECJ agreed that the CFI could come without any detailed intermediate stage, to the conclusion that the possibilities for those agents to compete with each other had been affected by the discriminatory conditions for remuneration implemented by BA …134

The likely anticompetitive effect of the conduct appears to have been established by BA’s significant market share, co-extensive with a finding of dominance. This is in fact a conclusion from structure, ie the idea that a firm by its very presence is foreclosing the market. The Court’s conclusion may well be correct that the conduct resulted in foreclosure but the market share should be the starting point for that analysis, not its conclusion. It is too often used as a shortcut to liability, implicit in the ECJ’s view that there was no need for a ‘detailed intermediate stage’.135 This mode of analysis is employed to the detriment of more formal, and possibly more accurate, economic reasoning. This is not to say that conduct which distorts rivalry is not a legitimate competition concern (which has an important foundation in the Treaty). The emphasis placed by the ECJ on protecting the competitive process in British Airways is similar to the issues raised in the Opinion by Advocate-General Kokott in that case. She referred to competition as an ‘institution’,136 and highlighted the need to protect the competitive process, an approach under which consumers’ interests are only indirectly protected: Article 82 EC, like the other competition rules in the Treaty, is not designed only or primarily to protect the immediate interests of individual competitors or consumers, but to protect the structure of the market and thus competition as such (as an institution) …137

Similarly, Gerber describes how the objective of ‘competition on the merits’ is to allow the market to develop without being subject to distortions caused by a dominant firm’s ‘non-competitive’ conduct ... to create a legal framework within which markets can develop on the basis of competitive conduct. It thus protects the conditions of operation on the market from distortions in order to allow the market to develop competitively.138

However, the difficulty in practice is that often the cases have not formally assessed these competitive concerns. They too readily define the protection 133

Ibid, para 145. Ibid, para 148. 135 Ibid. 136 British Airways, above n 118, A-G Kokott, Opinion, para 69. 137 Ibid, paras 44, 68 and 86. 138 D J Gerber, ‘The Future of Article 82: Dissecting the Conflict’ in Ehlermann and Marquis (eds), above n 42, 37, at 42. 134

A Reformed Approach to Article 82 and the Special Responsibility 141 of the ‘competitive process’ and exclusionary conduct merely as inequality of opportunity. This approach is evident in the decision in Motosykletistiki Omospondia Ellados NPID,139 where, on a reference under Article 234, the ECJ found that an undertaking which was entrusted with administrative power to authorise motorcycling events, but which also, at other times, competed in the organising of such events, could give rise to an issue under Articles 82 and 86. The ECJ stated: ‘A system of undistorted competition, such as that provided for by the Treaty, can be guaranteed only if equality of opportunity is secured as between economic operators’.140 The Court seemed to invoke Article 82 for any disruption to a ‘level playing field’ or ‘competitive neutrality’.141 This arguably casts a very broad net of liability and is not a useful standard for distinguishing competitive from abusive behaviour. This broad net cast by a finding of foreclosure also makes it more difficult to assess whether there are any countervailing (and narrowly construed) efficiencies as an ‘objective justification’, and then whether these sufficiently outweigh the exclusionary effect.142 The Commission Guidance sets out a high threshold for any justification on the grounds of efficiencies. The dominant undertaking must establish a number of cumulative conditions, including that the conduct does not eliminate effective competition, by removing all or most existing sources of actual or potential competition. Rivalry between firms is an essential driver of economic efficiency, including dynamic efficiencies in the form of innovation. In its absence the dominant firm will lack adequate incentives to continue to create and pass on efficiency gains. Where there is no residual competition and no foreseeable threat of entry, the protection of rivalry and the competitive process outweighs possible efficiency gains.143

The Commission Guidance does, however, seem to apply a more economic approach to the abusive effect of rebates than that applied by the Court in British Airways. It examines whether the retroactive rebates are capable of hindering the expansion or entry of ‘as efficient’ competitors.144 The assessment is more clearly focused on the relationship between price and costs, and account is taken of the switching costs to rivals once the rebate

139 Case C-49/07 Motosykletistiki Omospondia Ellados NPID (MOTOE) v Greece [2008] 5 CMLR 11. 140 Ibid, para 51. See also Case C-202/88 France v Commission [1991] ECR I-1223, para 51. 141 On the drafting of the EC Treaty and the rejection of a broad non-discrimination provision, see Schweitzer, above n 60, at 131–32. 142 BA had argued that the high fixed costs in air transport and the importance of aircraft occupancy rates justified the rebates, but this was not accepted by the Court: British Airways, above n 118, paras 86–87. 143 Commission Guidance, above n 1, para 29. 144 Ibid, para 40.

142 Kathryn McMahon has been taken into account. If the effective price is above the Long Run Average Incremental Cost (LRAIC) of the dominant undertaking, this would normally allow an equally efficient competitor to compete profitability notwithstanding the rebate.145 If the price is below AAC the scheme is capable of foreclosing an ‘as efficient’ competitor, although the Commission does not go as far as to state that the requisite test is that of predatory pricing.146 VI. CONCLUSION

This discussion has highlighted the variable standards applied by both the courts and the Commission in the assessment of abuse under Article 82. The Commission Guidance, while professing to shift towards a more economic approach and a focus on consumer welfare, has not always established standards which are consistent with those terms as understood in US antitrust or in traditional economic theories of competition law. Rather, in the EC we observe the pursuit of other goals, such as the protection of the competitive process and rivalry. It is these goals which are probably more consistent with Article 3(g) of the Treaty, market integration and the theoretical foundations of EC competition law. These goals are valid and relevant objectives for competition law and policy. But there is a failure to articulate properly how these goals translate into transparent, predictable and justiciable rules for the identification of abusive conduct. Instead, it appears that the Commission sets out apparently conflicting and variable standards, sometime protecting ‘as efficient’ competitors and at other times protecting ‘less efficient’ competitors. The Commission will also in some instances define abuse as harm to the ‘final consumer’, while at other times it will more clearly focus on rivalry and define consumers to include both the ‘intermediate’ and the ‘final consumer’. As Ahlborn and Padilla point out with respect to the Commission’s Article 82 Discussion Paper: Are dominant firms still bound by the special responsibility principle? ... only the application of this principle can explain the wide notion of foreclosure adopted in the text—one that requires dominant players to consider the impact of their actions on the rate of growth of their rivals.147

145

Ibid, para 42. Ibid, para 43. See also Virgin’s similar complaint in the US courts, which was dismissed on the grounds that the rebates must meet the predatory pricing rules, including the possibility of recoupment: Virgin Atlantic Airways Ltd v British Airways plc 257 F 3d 256 (2d Cir 2001). 147 C Ahlborn and A J Padilla, ‘From Fairness to Welfare: Implications for the Assessment of Unilateral Conduct under EC Competition Law’ in Ehlermann and Marquis (eds), above n 42, at 88. 146

A Reformed Approach to Article 82 and the Special Responsibility 143 The courts too have readily equated mere damage to equality of opportunity to abuse. This truncated analysis for the finding of abuse is very much linked to a determination based on the restraint of the dominant undertaking, its ‘mere presence’ and ‘special responsibility’, rather than a full examination of the links between its conduct and damage to the competitive process (and ultimately the consumer). This is not an argument for jettisoning this approach in favour of the adoption of a US model of consumer welfare, narrowly construed as wealth maximisation, and the deference to ‘self-correcting’ markets. Nor does this analysis argue for a shift to a purely ‘effects-based’ approach as measured by merely elevated price or restricted output. As Balto and Nagata point out: a price/output effects tests can yield a false positive. Even if there is evidence of elevated price or restricted output, that does not necessarily mean that the conduct in question was exclusionary. Monopoly power resides in both lawfully attained monopolies and those attained or maintained through exclusionary means. Any exercise of market power by a monopolist can be expected to result in elevated price or restricted output. Thus, an observation of elevated price or restricted output is insufficient to label the conduct exclusionary.148

Rather, there is need for a clearer, more economic exposition of how and when consumer welfare, in the wider sense, is damaged by the distortion of competition. This articulation of the goals and standards of abuse under Article 82 may also go some way to explaining the differences between the US and EU models. In particular it might serve to reject the view that the EU approach merely ‘protects competitors’, or is ‘simply one more tool of industrial policy’ which compromises ‘consumer welfare to advance the welfare of competitors’.149 In any event, the more interventionist approach of the EU is arguably more relevant in these times of global financial crisis which have led some to reject simple economic models and the belief in ‘self-correcting’ markets. This latter approach, characterised by the Chicago School, has resulted in a succession of US Supreme Court decisions150 which have rendered Section 2 of the Sherman Act almost completely irrelevant as an instrument for the control of abuse of monopoly.

148 D Balto and E Nagata, ‘Proof of Competitive Effects in Monopolization Cases: A Response to Professor Muris’ (2000) 68 Antitrust Law Journal 309, at 313. 149 Brief of Amici Curiae in Linkline, above n 48. 150 See Brooke Group v Brown & Williamson Tobacco Corp 509 US 209 (1993); Verizon Communications Inc v Law Offices of Curtis V Trinko, above n 38; Weyerhaeuser Co v RossSimmons Hardwood Lumber Co 127 S Ct 1069 (2007); Pacific Bell Telephone Co v Linkline Communications, Inc, above n 47.

144 Kathryn McMahon It is certainly the recognition that markets are more complex than the simple ideological assumptions which underlines much of the Chicago School approach that has led to the development of the so-called postChicago School.151 This approach uses the insights from modern industrial organisation and game theory to demonstrate how the strategic practices of dominant firms152 could be profitable and therefore predatory.153 In particular they demonstrate how abuse provisions can be applied to deal with complicated issues such as asymmetric information,154 network effects and technological dependency in high-technology industries.155 This approach is characterised by a richer factual analysis of individual cases and the application of more complex rules based on repeated interaction, rather than reliance on more theoretical models and per se tests.156 For some, however, this increased complexity and discretion raises the problem of false positives157 and wrongful convictions, which can be detrimental to consumer welfare and the expectations of rational economic actors. The more interventionist EU model of competition law has also been more readily adopted over the US model in post-liberalised and emerging economies.158 While in some Eastern European countries this can be attributed to a desire eventually to join the European Union, others have chosen the EU model as one they consider more conducive to their more complex and highly concentrated economies as they dismantle State monopolies and/ or require closer regulation of previously privatised utilities and networks. Lastly, a deeper understanding of the theoretical foundations and purposes of EC competition law may be increasingly important in the context of the Lisbon Reform Treaty which no longer prioritises ‘distortion of competition’ as a singular goal but rather makes it one among many, including

151 See J B Baker, ‘A Preface to post-Chicago Antitrust’ in A Cucinotta, R Pardolesi and R Van Den Bergh (eds), Post-Chicago Developments in Antitrust Law (Cheltenham, Edward Elgar, 2002), ch 3; E M Fox, ‘Post-Chicago, post-Seattle and the dilemma of globalization’, ibid, ch 4. 152 Against, see E Elhauge ‘Why Above-Cost Price Cuts to Drive Out Entrants are Not Predatory—And Implications for Defining Costs and Market Power’ (2003) 112 Yale Law Journal 681. 153 See generally Bolton et al, above n 99. 154 See eg Eastman Kodak Co v Image Technical Services Inc 504 US 451 (1992). 155 See eg United States v Microsoft Corp 87 F Supp 2d 30 (DDC); 253 F 3d 34 (DC Cir 2001). 156 California Dental Association v FTC 128 F 3d 720 (9th Cir 1997), rev’d 119 S Ct 1604 (1999). The Commission Guidance indicates that it is receptive to these ideas in the context of predation, above n 1, para 67. 157 See generally R A Cass and K N Hylton, ‘Preserving Competition: Economic Analysis, Legal Standards and Microsoft’ (1999) 8 George Mason Law Review 1; F S McChesney, ‘Talking ‘Bout My Antitrust Generation: Competition for and in the Field of Competition Law’ (2003) 52 Emory Law Journal 1401; F Easterbrook, ‘Predatory Strategies and Counterstrategies’ (1981) 48 University of Chicago Law Review 263. 158 See generally E M Fox, ‘Antitrust and regulatory federalism: races up, down, and sideways’ (2000) 75 New York University Law Review 1781.

A Reformed Approach to Article 82 and the Special Responsibility 145 non-economic and social purposes. While Chicago School antitrust scholars would assert that non-economic goals are unworkable and cannot be translated into coherent, administrable legal doctrine,159 it may be the EU competition model, rather than a ‘consumer welfare as efficiency’ model, which will prove to be more attuned to accommodating these new complexities and multiple goals.

159

See R A Posner, Antitrust Law, 2nd edn (Chicago, University of Chicago, 2001), at 35.

8 Rhetoric or Reform: Does the Law of Tying and Bundling Reflect the Economic Theory? PRANVERA KËLLEZI*

I. INTRODUCTION

I

N RECENT YEARS, both tying and bundling have been a concern for competition authorities. In the Microsoft proceedings, the European Commission considered whether the tying of Windows Media Player (WMP) to Windows constituted an abuse of a dominant position under Article 82 EC.1 Their finding of an abuse was upheld by the Court of First Instance (CFI) on appeal.2 The case stressed the need for further clarification of the criteria defining harmful tying and bundling. The Commission is now considering action against Microsoft in the web browser market because of its policy of tying Internet Explorer to Windows.3 Tying has been seen as harmful for a long time, and has even been considered illegal in the United States.4 In EC law, tying is specifically covered by Article 82(d) of the Treaty, which lists as an abuse ‘making the conclusion of contracts subject to acceptance by other parties of supplementary obligations which, by their nature or according to commercial usage, have no connection with the subject of such contracts’. Tying may take the form of contractual arrangements, a refusal to supply the tying product,

* PhD in law (University of Geneva), LLM in EC law (College of Europe, Bruges). Legal adviser for competition law at European Broadcasting Union (Geneva, Switzerland). 1 Case COMP/C-3/37.792 Microsoft 2007/53/EC, [2007] OJ L32/23, Commission Decision (summary). 2 Case T-201/04 Microsoft v Commission [2007] ECR II-3601. 3 The Commission sent a statement of objections to Microsoft on the tying of Internet Explorer to Windows, MEMO/09/15 of 17 January 2009. 4 Standard Oil Co (Cal) v United States, 337 US 293, 305–306 (1949): the Supreme Court held that ‘tying agreements serve hardly any purpose beyond suppression of competition’. Note that in United States, contrary to European law, the per se illegality is a non-rebuttable presumption.

148 Pranvera Këllezi or pricing practices that have tying effects, such as rebates and discounts on a package of goods (bundling).5 Article 82(d) EC, read in conjunction with Article 82(b) of the Treaty, prohibits the foreclosure of competitors ‘to the prejudice of consumers’.6 The fundamental concern is the ability of dominant firms to leverage market power from the market in which they were dominant to the tied one. As a consequence, other producers of the tied product could be seriously hampered, or foreclosed.7 Foreclosures resulting from tying have traditionally been treated more harshly than other types of abuse, because it was thought that there is no reason to tie products other than to reduce competition.8 However, recent court decisions have more extensively analysed the competitive effects of tying and reflected a progressive abandonment of the per se rule.9 In this regard, economics has played a significant role, especially in promoting the view that there are economic justifications for tie-ins.10 In the US Microsoft case, the Circuit Court of Columbia declared that ‘the rule of reason, rather than per se analysis, should govern the legality of tying arrangements involving platform software products’.11 Similarly in Europe, the CFI confirmed the effects-based approach applied by the Commission in the Microsoft case.12 Dominant undertakings commit an abuse if they engage in conduct that restricts competition. The trend toward an effects-based approach aims to analyse in depth the anticompetitive effects of conduct. Economic thinking is crucial to this analysis. The CFI made it clear, with respect to Article 82(d), that while it is true that neither that provision nor, more generally, Article 82 EC as a whole contains any reference to the anti-competitive effect of bundling, the fact remains that, in principle, conduct will be regarded as abusive only if it is capable of restricting competition.13

5

Richard Whish, Competition Law, 6th edn (Oxford, OUP, 2008), 679. See John Temple Lang, ‘Anticompetitive Non-pricing Abuses under European and National Antitrust Law’ in Barry E Hawk (ed), Annual Proceedings of the Fordham Corporate Law Institute 2003 (New York, Juris Publishing, 2004), 241–42. 7 Phillip Areeda, Antitrust Law: An Analysis of Antitrust Principles and their Application (1991), Vol 9, Part III, Chapter 17, 1700d. 8 Ibid, 1700j. 9 Herbert Hovenkamp, Federal Antitrust Policy: The Law of Competition and Its Practice, 2nd edn (St Paul, Minn, West Publishing Co, 1999), 393. 10 See Robert Pitofsky, ‘Policy Objectives of Competition Law and Enforcement’ (2003) EU Competition Law and Policy Workshop Proceedings, European University Institute of Florence, 6. 11 United States v Microsoft Corp, 2001 US App LEXIS 14324, HN36. Nevertheless, the ruling was criticised: see eg Herbert Hovenkamp, ‘IP Ties and Microsoft’s Rule of Reason’ (2002) 47(2) Antitrust Bulletin 369, at 370. 12 Microsoft v Commission, above n 2, para 859. 13 Ibid, para 867. 6

Rhetoric or Reform? 149 Therefore, the CFI does not consider the effects-based approach as a new legal theory. Crucially, the Commission is entitled not only to analyse the actual foreclosure effects, but also to assess the likelihood that tying would ‘lead to a lessening of competition so that the maintenance of an effective competition structure would not be ensured in the foreseeable future’.14 This is a welcome trend, since it is the effects-based approach, and not the form-based approach, that creates the appropriate legal framework for the introduction of the new economic approach. The European Commission Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings15 (hereafter ‘the Commission Guidance’) and the Discussion Paper on the Application of Article 82 EC to Exclusionary Abuses (‘the Discussion Paper’16) produced by the Commission intend to contribute to the process of introducing an economic-based approach into European competition law enforcement. Economic criteria improve the methodology for the assessment of exclusionary behaviour, and give guidance on the conditions under which an undertaking’s conduct might cause harm to competition and consumers, thereby promoting clarity and predictability.17 When deciding its enforcement priorities, the Commission will consider a number of factors; the Guidance precisely sets out these factors, together with a general analytical framework for considering anticompetitive conduct. The economic-based approach indisputably contributes to distinguishing harmful tying and bundling from benign tying and bundling, and focuses on the anticompetitive effects while taking into account efficiencies. Nevertheless, the business community and competition law enforcers are also afraid that it might reduce legal certainty. The interest of undertakings in a form-based approach was apparent in the Microsoft case.18 Indeed, the relationship between law and economics is not straightforward. Competition law is governed by the law of evidence: competition authorities should prove anticompetitive foreclosure, using direct or circumstantial 14

Ibid, para 1089. DG Competition, Communication from the Commission, Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings (Brussels, 3 December 2008), http://ec.europa.eu/comm/competition/ antitrust/art82/guidance.pdf. 16 DG Competition, Discussion Paper on the Application of Article 82 of the Treaty to Exclusionary Abuses (December 2005), http://europa.eu.int/comm/competition/antitrust/ others/discpaper2005.pdf. 17 Philip Lowe, ‘The European Commission formulates its enforcement priorities as regards exclusionary conduct by dominant undertakings’ February 2009 (1) Global Competition Policy, 3. 18 Microsoft v Commission, above n 2, paras 866, 1032 and 1035. The CFI in Microsoft fully endorsed the effects-based approach, by rejecting the claim made by Microsoft that Article 82 does not include a separate condition on the foreclosure of competitors, and would therefore be adopting a new speculative theory. 15

150 Pranvera Këllezi evidence. While economic theory greatly contributes to building a general framework and to improving understanding of companies’ business strategies, their incentives, and the likely impact of their behaviour on the market, the evidential value of economic models is questionable.19 Therefore, the issue is not only whether the law is consistent with economic theory, but also whether the latter can be reconciled with the requirements of the law. This chapter analyses the recent developments in European competition law with regard to tying and bundling, and assesses them in the light of economic theory. I consider the role of economic theory, whether the case law or the new Commission Guidance is consistent with economic thinking, and whether it is flexible enough to allow for economic learning to be taken into account. I begin with a discussion of different economic theories of tying and bundling. There then follows an analysis of the recent developments in European Community competition law, and the chapter concludes with some remarks on the consistency of the Commission activity with economic theory. II. ECONOMIC ANALYSIS OF TYING AND BUNDLING

Before presenting a brief economic analysis of tying and bundling, some remarks are required on the use of economic theory in competition law. The use of formal economic models as an indispensable part of the antitrust process requires a good understanding of what a model can offer. Economic models offer a formalisation of a foreseeable market situation and its resulting effects on competition. However, they cannot consider all the parameters that have an influence in real markets; these models are thus necessarily simplified. A genuine economic approach requires the implementation of theories in real-world situations, and the construction of a framework able to direct authorities or judges in their work.20 With these limitations in mind, it is now possible to consider four specific economic theories that relate to tying and bundling. The underlying idea of the ‘leverage theory’ is that tying and bundling can be used by a firm with a dominant position in one market to extend or leverage its power in a second market (offensive leverage). A variant of the theory has been 19 See eg ‘Suggested best practice for submissions of technical economic analysis from parties to the Competition Commission’, available at http://www.competition-commission. org.uk/rep_pub/corporate_documents/corporate_policies/best_practice.pdf. The CC requires the submissions to present all the assumptions and the reasons for making them. Indeed, it reminds that ‘all economic models make assumptions for convenience and tractability’ (point 28), and often these assumptions are of questionable evidential value. 20 See Jay Pil Choi, ‘Antitrust Analysis of Tying Arrangements’ in Choi (ed), Recent Developments in Antitrust. Theory and Evidence (Cambridge, Mass; London, MIT Press, 2007), at 8, who expresses the general consensus that the appropriate antitrust policy depends on the specifics of the case.

Rhetoric or Reform? 151 developed: the foreclosure on the tied market may effectively protect the market power on the tying market (defensive leverage). Economic theory has also studied the use of tying and bundling as an entry deterrent device, and finally their likely impact on innovation. These theories will now be analysed to consider the impact which they have had on the development of EC competition law.

A. Offensive leveraging Leverage theory in its classical form has come under attack by scholars of the Chicago School. According to this School, a monopolist would not tie a good supplied competitively to his own product, since the only monopoly rent he could extract would be on his own product.21 In short, the profit of the firm would not increase as a result of tying; therefore the firm will have no incentive to do so. Other explanations for bundling, for instance its use as a price discrimination device, have reinforced the view that there are other motivations for tying and bundling much more plausible than the leverage of market power.22 Nevertheless, the fact that, in general, firms have no incentive to tie, is irrelevant when assessing real cases where tying and bundling might have anticompetitive effects. A later generation of scholars studied the market conditions under which tying and bundling may foreclose competition, which is very useful for competition law enforcers. In this section, I examine a number of these studies. The criticisms of the Chicago School were examined by Whinston, who questioned the assumptions made about the nature of the tied market.23 While the Chicago School assumed that the tied market has a competitive, constant, returns-to-scale structure, Whinston builds a model with an oligopolistic structure of the tied good market. He shows that in the presence of economies of scale, tying can lead to the monopolisation of the tied good market through foreclosure. Existing competitors will be excluded as a consequence of the reduction of their residual demand, and thus of their sales. The focus of attention in this model is not pricing but the effect on the structure of the tied good market, which is much closer to the concerns expressed by the European Commission with regard to anticompetitive foreclosure. 21 See eg George Stigler, ‘A Note on Block Booking’ in George Stigler (ed), The Organisation of Industry (Homewood, Richard D Irvin, 1968), 165. 22 See eg Arthur Lewbel, ‘Bundling of Substitutes or Complements’ (1985) 3(1) International Journal of Industrial Organisation 106, concluding that a single product monopoly cannot increase its profits by leveraging (on a competitive market), ‘unless it is coupled with either price discrimination via mixed bundling, or with some joint production cost savings’. 23 Michael Whinston, ‘Tying, Foreclosure, and Exclusion’ (1990) 80 American Economic Review 837, at 838. He also relaxed the fixed-proportion assumption.

152 Pranvera Këllezi The altering of the market structure of the tied good market is linked with profitability, and this depends on whether the monopolist is able to make a pre-commitment to tie, for instance through product design or designing incompatible interfaces between his products and those of his competitors.24 When tying leads to exclusion of competitors, consumers may be worse off because prices may rise, and choice in the tied market necessarily falls.25 The effect on total welfare is not always negative. It is recognised by the author that whereas the model presents a coherent leverage hypothesis, its normative implications are less clear.26 The study of Martin focuses specifically on welfare effects, and shows that in the absence of reductions of the marginal and fixed costs, bundling reduces social welfare compared with the case of no bundling.27 Independently of the considerations related to welfare effects, the work of Whinston explored the conditions under which foreclosure is likely to occur, and highlighted the importance of the market structure and the study of the impact of a company’s behaviour on that structure. In a market characterised by economies of scale, tying and bundling are liable to have a negative impact on competitors, and consequently on market structure and the market’s ability to ensure effective competition. The approach of the Commission in focusing on the effect of the firm’s behaviour on the market structure is consistent with the work of Whinston demonstrating competition policy that takes into account economic thinking.

B. Defensive leveraging The modern concept of defensive leverage was developed by Carlton and Waldman.28 The key insight of this dynamic model is that foreclosure today

24 Ibid, 839. For independent products, tying is not useful absent pre-commitment, but can be a profitable strategy with pre-commitment. In this latter case, the competitors in the tied product market can be excluded by the ‘strategic foreclosure’ effect of tying (ibid, 840). In the presence of heterogeneous preferences among consumers for the tying good, especially when a significant number of consumers in the tied market have low valuation for the tying good, tying does not necessarily result in strategic foreclosure. For complementary products used in fixed proportions, a monopolist may find the presence of other competitors in the tied market more profitable (ibid, 850). 25 Ibid, 845. Although the incentive to lower the price of the tied product is also present when the monopolist in the tying market succeeds in monopolising the tied market, one should expect the price for the tied product (and the bundle) to rise when the gains from this market are substantial. In this case, the consumers are worst off. 26 Ibid, 855. 27 Stephen Martin, ‘Strategic and Welfare Implications of Bundling’ (1998) CIE Working Paper 98-14. 28 Dennis Carlton and Michael Waldman, ‘The Strategic Use of Tying to Preserve and Create Market Power in Evolving Industries’ (2002) 33 RAND Journal of Economics 194–220; see

Rhetoric or Reform? 153 in the tied market for a complementary product, may protect the monopoly in the tying market later on. Since the model is of a dynamic nature, it could realistically be applied in the Microsoft case (bundling of Windows with Internet Explorer). Strategic behaviour to reduce the scale of a rival in the tied market in the first period, can create significant competitive advantages in the second period, especially in the presence of network effects,29 scale economies and installed base.30 The latter assumptions and the possibility of future competition in the tying market distinguish this model from previous ones.31 Reducing the scale of a rival in the tied market, or reducing the products available in that market, has the effect that inefficient production or reduced choice in the tied market make entry into the tying market more difficult. This is because future competitors in the tying market must enter both markets (or contract with the supplier of the tied market), and it will be difficult for them to have an efficient supply source or a sufficient number of complementary products on the tied market.32 When the complementary product may become a substitute in the second period, as was the case with Internet Explorer, monopolisation of the ‘future substitute’ will protect the position of the firm in the tying market. The overall effect is that the strategic behaviour in the secondary (tied) market is aimed to preserve or defend the existing market power in the primary (tying) market. Moreover, the model shows how a firm can transfer its monopoly power to other complementary products.33 Although the Commission does not refer explicitly to the work of Carlton and Waldman, the defensive leverage is the underlying hypothesis of its finding that the control of the media player market would enable Microsoft to extend its power to a range of other related markets.34 The Commission’s analysis also Dennis Carlton, ‘A General Analysis of Exclusionary Conduct and Refusal to Deal—why Aspen and Kodak are Misguided’ (1998) NBER Working Paper Nr 8105. 29 Network effect (or network externalities) means that the benefit to a user of a product depends on how many other customers use the same product and on how many other complementary products are available. Dennis Carlton and Jeffrey Perloff, Modern Industrial Organization, 3rd edn (Reading, Mass; Harlow, Addison-Wesley, 2000), 374. 30 Installed base is a measure of the number of units of a particular type of system (usually a computing platform) actually in use, as opposed to market share which only reflects sales over a particular period. Because installed base includes machines that may have been in use for many years, it is usually a higher figure than market share. Carlton and Waldman, above n 28, at 205; Carlton, above n 28, at 13. 31 See eg Whinston, above n 23, at 852 ff. 32 Carlton, above n 28, at 14. 33 Carlton and Waldman, above n 28, at 212; Carlton, above n 28, at 16. The hypothesis applies in the case where the newly emerging market for, say, product C is associated with the same complementary good (B) as is the primary market (A). After leveraging the market power from A to B, the firm can leverage its market power to C by tying B to C. In this case the firm transfers its market power from A to C, and if C is a future substitute for A (or B), this strategy can protect the market power it has on both A and B markets. 34 Microsoft (Commission Decision), above n 1, para 212. See considerations in section III.A below.

154 Pranvera Këllezi focuses on the same mechanisms, which again shows that the concerns and the approach of the Commission are consistent with economic theory.

C. Tying and bundling as an entry deterrent device Tying and bundling have also been considered as an entry deterrent device.35 Nalebuff focuses specifically on a model in which bundling is used not to leverage a monopoly from one market to another, but rather to protect both markets of a multigood monopolist against entry. A rival who produces only one of these goods finds it harder to enter one of these markets. Unlike previous authors, Nalebuff’s work suggests that bundling is credible even without any commitment device.36 This strategy can protect both markets not only when entry is deterred, but also in situations where entry succeeds or when a one-product competitor already exists in the market. In the former situation, Nalebuff’s model shows that the profits of the incumbent can more than double. In both situations the result is consistent with the findings of Whinston: bundling reduces the profitability of the competitor through foreclosure. The concerns of the Commission on the ability of tying and bundling to raise barriers to entry are therefore justified. Entry deterrence is more effective37 when the same group of consumers is used to buying both products; entry with only one of the products cannot satisfy customers, who will prefer buying the bundle. With no discount by the incumbent, the bundle reduces the entrant’s profit and, as the bundle is sold at a discount relative to the independent pricing, entry becomes even less profitable.38 Similarly, complementarities in consumption make bundling more profitable, and consumers are even less attracted by only one product.39 While the presence of negative correlation makes the price discrimination strategy profitable, it makes the entry-deterrent effect less valuable. Nevertheless, even if this effect is diminished, it does not disappear with substitutes.40 The Commission reflects these lessons in its Guidance, and considers that when the products composing the bundle are complements, bundling raises barriers to entry because customers prefer buying both products

35 See eg Whinston, above n 23; Jay Pil Choi and Christodoulos Stefanadis, ‘Tying, Investment, and Dynamic Leverage Theory’ (2001) 32 RAND Journal of Economics 52–71; Carlton and Waldman, above n 28; Barry Nalebuff, ‘Bundling as an Entry Barrier’ (2004) 119 The Quarterly Journal of Economics 159–87. 36 Nalebuff, above n 35, at 162–63. 37 Ibid, 160. 38 Ibid, 168–69. 39 Ibid, 178. 40 Ibid, 174.

Rhetoric or Reform? 155 together, rather than acquiring them from separate producers.41 Conversely, the Commission considers that tying and bundling are not liable to lead to foreclosure when customers are interested in buying only one of the products.42 The entry-deterrent effect increases if the incumbent adds more goods to the bundle.43 Thus, bundling becomes an effective competitive tool against a competitor with a limited product line. Nalebuff analyses the effects of large bundles in another of his papers.44 He suggests that the price advantage of a firm that sells bundles of complementary products will grow as the bundle grows in scale. When four or more products are aggregated to a bundle, the advantage of an incumbent is difficult to contest for actual or potential rivals. The difficulty for competitors to replicate bundles is also stressed by the Commission Guidance.45 In the case where actual rivals respond by bundling their products (a bundle-versus-bundle competition), the profits will fall for both of them, and the result is worse as the size of the bundle grows.46 The competitive effects of large bundles were analysed in detail by Bakos and Brynjolfsson.47 The authors examine the case of information goods available on the Internet. The Internet has radically reduced the marginal cost of reproducing and distributing information goods to consumers, and this difference makes possible the bundling of a large number of information goods, such as different types of software and Internet content. The authors analyse how a firm can exploit ‘economies of aggregation’, a different type of demand-side economies which may be created when the information goods are ‘aggregated’ and offered in large-scale bundles.48 These economies of aggregation can significantly increase the profit of the firm. The increased profitability is made possible precisely because of the low (almost zero) marginal cost. The authors conclude that the firm that offers a large bundle can price it very low, which has the effect of making the entry of a single-product firm (or a firm with a small bundle) very unattractive, despite a superior cost structure or quality level.49

41

Commission Guidance, above n 15, para 57. Ibid, para 54. 43 Nevertheless, according to Nalebuff, above n 35, the value of the bundle does not seem to grow proportionately. 44 Barry Nalebuff, ‘Competing against Bundles’ (2000), Yale School of Management. 45 Commission Guidance, above n 15, para 53. 46 Nalebuff above n 44, 11–12. 47 Yannis Bakos and Erik Brynjolfsson, ‘Bundling and Competition on the Internet’ (2000) 19(1) Marketing Science 63–82. 48 Ibid, 64. 49 Ibid, 77. The authors emphasise that this entry-deterrent effect is not based on strategic behaviour, for instance by lowering the price in the short run. The incumbent chooses a price level that maximises its profits in the current period. 42

156 Pranvera Këllezi D. Impact on innovation Tying and bundling can affect not only prices and the competitive structure of the market, but also innovation. Among others, Choi has explored the dynamic aspect of bundling.50 Extending Whinston’s model, Choi focuses on the long-term effects of tying on incentives to innovate. In contrast to earlier work, his model does not rely on the exit of rivals. Even in the absence of exit, the model shows that tying can be a profitable strategy in the long term, since it increases the tying firm’s incentive to innovate and decreases the rivals’ incentives. Tying allows a dominant firm to extend its market power from the tying to the tied good market, by expanding its market share in the latter market. In doing so, it can spread the sunk cost of research and development over a larger number of units, which allows the firm to recover its costs and profit from the research. Conversely, since the market shares of the rivals will decrease, they will have little possibility to recover their costs, decreasing their incentives to innovate in the long term.51 In this model, tying does not necessarily lead to foreclosure in the product market, but rather to foreclosure in the research and development market. In dynamic markets, if this long-term effect outweighs the reduction of the profitability from reduced prices, bundling may be profitable for the dominant firm even in the absence of exit by rival firms. The same reasoning applies to mergers: since a merger with bundling can reduce the scale of the existing competitors, bundling reduces their incentive to invest in research and development.52 Choi also analyses the welfare effects of bundling. He concludes that the effects of tying in his model on social welfare were ‘unambiguously negative’: social welfare decreases with bundling.53 The Commission Guidance does not make any statement on the impact of bundling on the competitor’s incentive to innovate. However, in highly innovative markets, the maintenance of a high incentive to innovate for all market participants is crucial not only because innovation is liable to increase consumer welfare in the long run, but also because the entry of innovative new products would challenge the market power of incumbents by limiting their ability to foreclose competition. To summarise, the likelihood of anticompetitive foreclosure increases if an undertaking has substantial market power in the tying market, and the 50 Jay Pil Choi, ‘Tying and Innovation: A Dynamic Analysis of Tying Arrangements’ (2004) 114 The Economic Journal 83. 51 Ibid, 93. 52 Jay Pil Choi, ‘Antitrust Analysis of Mergers with Bundling in Complementary Markets: Implications for Pricing, Innovation, and Compatibility Choice’ (October 2003) Net Institute Working Paper #03-02, http://www.netinst.org/Choi.pdf, 15. 53 Choi, above n 50, 97–98. He emphasises that this strong result is due to various assumptions made in his model.

Rhetoric or Reform? 157 tied good market structure is of an oligopolistic nature, characterised by economies of scale, network effects or problems of installed base. In these situations, tying and bundling are likely to have a negative effect on the structure of the tied good market, increase barriers to entry into the tied and the tying markets, and have a negative impact on innovation. We shall see below that this condition is not sufficiently discussed in the Commission Guidance. Further, the economic theory shows that under these conditions, the more products added to the bundle, the more effective the entry-deterrent effect. A commitment to tie is not always a necessary condition, although it increases the likelihood of a negative impact. We shall now discuss in greater detail how these economic principles are reflected in the Commission Guidance and its decision-making practice. III. RECENT DEVELOPMENTS REGARDING TYING AND BUNDLING IN EC COMPETITION LAW

The reform of EC competition law on tying and bundling is visible in individual decisions and in the Commission Guidance. This section discusses the influence of economic theory on the analysis of the Commission in the Microsoft case and on the new Guidance. In line with the economic approach, the Commission’s recent policy with regard to Article 82 emphasises the protection of consumers.54 Nevertheless, while economic theory relies on consumer welfare and, therefore, harm to final consumers,55 competition law should take into account the restriction of competition at any level of the market, and assess the harm to consumers and any other customer of the dominant undertaking.56 Article 82 covers ‘not only practices which may prejudice consumers directly but also those which indirectly prejudice them by impairing an effective competitive structure’.57 Moreover, the Microsoft case emphasised that consumer prejudice captures not only parameters, such as price and quantity, but also quality, choice and, most importantly, innovation.58 These latter criteria are crucial for the assessment of tying and bundling practices in dynamic industries characterised by rapid innovation; they also constitute the basis of the Commission’s intervention regarding the business strategy of Microsoft. Consequently,

54 Antitrust: consumer welfare at heart of Commission fight against abuses by dominant undertakings, IP/08/1877 of 3 December 2008. 55 See eg EAGCP, Report on an Economic Approach to Article 82 EC (July 2005), http:// europa.eu.int/comm/competition/publications/studies/eagcp_july_21_05.pdf, 8–9. 56 Commission Guidance, above n 15, para 19, n 2. 57 Microsoft v Commission, above n 2, para 664. 58 Ibid, paras 648–49. The Court considers that the prejudice to consumers relates also to the limitation of innovation, and not only to the limitation of markets and production.

158 Pranvera Këllezi prejudice to the final consumer does not constitute an additional condition for the finding of an abuse of a dominant position; the foreclosure of competitors as such is liable to have an adverse impact on consumer welfare. This broad interpretation of consumer prejudice in EC competition law does not mean that it is inconsistent with economic theory. The work of economists on tying and bundling has also been focused on the impact of such practices on the market structure, and on the impact on the ability of competitors to compete effectively. On the other hand, it should also be stressed that economics is a positive science; although it offers concepts like ‘efficiency’ or ‘welfare,’ these are only tools to help build an idea of how the markets work. Any conclusion about how competitors and regulators should act requires the involvement of institutions which express value judgements, and should be the result of a democratic process. The main contribution of economic analysis is to reduce the margin of error and the risk of over- and under-intervention. In this respect, the consumer welfare concept is a useful tool to ensure that enforcing bodies remain focused on the need to protect competition as a process and do not get sidetracked into protecting individual competitors. Two areas of particular importance are highlighted in the Guidance. First, the ‘equally efficient competitor’ test adopted by the Commission in its Guidance shows concern to protect competition, and not competitors, and promotes economic efficiency. Similarly, the possibility of taking into account efficiencies and the benefits for consumers at the justification stage is an important step in ensuring that the law reflects the economic theory, and will be further discussed below.

A. Integrating economic theories in individual cases Microsoft was one of the first Commission investigations to take into account the new economic thinking with regard to tying and bundling. The Commission explicitly mentions four general conditions for abusive tying: a) the tying and tied goods are two separate products; b) the undertaking is dominant in the tying product market; c) the undertaking does not give customers a choice to obtain the tying product without the tied product; and d) tying forecloses competition.59 Its analysis has been confirmed by the CFI, and the Commission adopts the same conditions in the new Guidance, with the exception of the absence of

59

Microsoft (Commission Decision), above n 1, para 794.

Rhetoric or Reform? 159 choice for customers.60 The main developments relate to the confirmation of the effects-based approach, to the distinct analysis of the existence of separate products, and to the assessment of the anticompetitive effects. The requirement that separate products exist has been implicit in the previous cases handled by the Commission, but has proved to be central to the finding of abusive tying in the Microsoft case. Generally, Article 82(d) is applicable to separate products or services, rather than to integrated ones.61 Intuitively, the mere act of tying or bundling two or more products shows that the latter have an existence of their own in the market, otherwise there would be no conduct that could be considered abusive. This condition can be used to identify genuine efficiencies and to identify separate markets and, therefore, market power. Moreover, it contributes to the understanding of the strategy of incumbent undertakings and of the mechanism of anticompetitive foreclosure. Assembling components in a bundle can create real value for consumers. This is reflected in competition law62 by distinguishing between packages that constitute ‘inherently one product’ and those that do not.63 Consequently, the finding of a composed product identifies genuine efficiencies for consumers, and serves as a ‘screening device’ to distinguish between useful ‘assembling’ and other kinds of tying or bundling that might impact the market. A plausible explanation of the wide acceptance of tying and bundling is consumers’ desire not to assemble the products themselves. Put in different terms, there is no need to unbundle, because there is almost no demand for it.64 One of the efficiency arguments put forward by Microsoft was that it offered an integrated product; in its view, there were not two separate markets. In addressing this, the Commission and the Court focused on consumer demand65; if there is no separate demand for distinct components, the assembling of components in one product would result in a distinct product that is of value for consumers.66

60

Commission Guidance, above n 15, para 50. Bellamy and Child, European Law of Competition, 5th edn (London, Sweet & Maxwell, 2001), 747. 62 See Robert Dansby and Cecilia Conrad, ‘Commodity Bundling’ (1984) 74(2) American Economic Review 377. 63 Robert Bork, The Antitrust Paradox. A Policy at War with Itself (New York, The Free Press, 1978), 379. See also Microsoft (Commission Decision), above n 1, para 955. 64 Barry Nalebuff, ‘Bundling, Tying, and Portfolio Power, Part I’. DTI Economics Paper Nr 1 (2003) 31; Carlton and Perloff, above n 29, at 304. Obviously, for many bundled products the manufacturer can integrate the products better than customers can. 65 Commission Guidance, above n 15, para 50. 66 This is the case for ‘composed products’, such as radio stations, car bodies or computers, where the choice left to consumers to assemble or to pick and choose different package options would result in increased complexity and higher costs for both manufacturers and consumers. 61

160 Pranvera Këllezi The Commission’s analysis in Microsoft is consistent with other Commission publications, and shows that the reform of Article 82 is part of wider reform to integrate economic analysis in competition law. According to the Commission Guidelines on vertical restraints, two products are distinct if the demand side (the buyers) considers them to be part of two distinct relevant markets; in other words, they would buy them separately in the absence of tying.67 The customers’ preference for the simultaneous use of two products induces producers to offer them together, considered in other terms ‘commercial usage’ or ‘accepted practice’. In those latter cases, there is no tying or bundling involved.68 Another key area where the influence of economic analysis is obvious in the Microsoft case, is the assessment of anticompetitive effects. In its decision, the Commission held that Microsoft infringed Article 82(d) by tying the Windows Media Player (WMP) to the Windows PC operating system.69 Considering the fact that the users can obtain media players from rivals through the Internet, it held that ‘[t]here are … good reasons not to assume without further analysis that tying WMP constitutes conduct which by its very nature is liable to foreclose competition’.70 First, the Commission considers the market power in the tying market, where Microsoft has a market share of 90 to 95 per cent and which gives Microsoft the ability to control the distribution of both products through PC sales.71 Since downloading and other distribution channels are not as effective, the practice used by Microsoft is likely to have a harmful effect on the structure of the market for media players. The negative effects will also expand onto the markets for complementary software and content. It should be noted that exit of rivals is not necessary to prove foreclosure effects.72 Although Original Equipment Manufacturers (OEMs) like DELL could offer another media player, the technical integration operated by Microsoft creates disincentives for the latter to bundle an additional product,73 and one of the remedies obliges Microsoft to provide OEMs with a version of Windows without WMP. The Commission’s decision demonstrated that it took into account the economic theory on defensive leverage and those on the entry-deterrence effects of bundling. First, the Commission rejected the arguments put forward by Microsoft that media players are not substitutes for the operating

67 Commission Notice, Guidelines on Vertical Restraints [2000] OJ C291/1, para 216. The Commission takes the same view at consideration 191 of the Guidelines on the Application of Article 81 of the EC Treaty to Technology Transfer Agreements [2004] OJ C101/2. 68 Guidelines on Vertical Restraints, above n 67, para 216. 69 Microsoft (Commission Decision), above n 1, paras 792 ff. 70 Ibid, para 841. 71 Ibid, paras 843–44. 72 Ibid, para 946. 73 Ibid, para 851.

Rhetoric or Reform? 161 system and could not threaten Windows; the Commission maintains, for example, that in combination with Java, they could be a ‘general platform substitute’.74 Here we find all the assumptions of the defensive leverage theory.75 Secondly, the Commission adopts the second hypothesis of Carlton and Waldman: the transfer of the monopoly power to complementary products of the tied product.76 According to the Commission, the media player market is a gateway to a range of related markets, such as those for content encoding software, wireless device software and online music delivery.77 Gaining a leading position in the media player market will give Microsoft the possibility to transfer its market power to these complementary products. Lastly, the decision is consistent with the theory on entrydeterrent effects of tying and the literature that considers the effect on innovation: incompatibilities may reduce the prospect of successful entry in both the operating system and the applications market, and reduce innovation incentives for rival technologies.78 The Microsoft case clearly identifies the elements of a harmful tying, and recognisably articulates the theories underlying the concerns of the Commission. It is, therefore, a very welcome development in EC competition law and suggests that not only is the Commission fully committed to the move to a more economic approach, but also that it has begun to develop the tools to implement it.

B. Commission’s Guidance: identifying criteria for harmful tying and bundling The Commission Guidance gives a general overview of the factors that increase the likelihood of a finding of anticompetitive foreclosure. The Guidance takes a different approach from the Discussion Paper, in that it starts with a description of relevant factors for the assessment of any kind of abusive exclusionary conduct.79 This is helpful in that it identifies the common grounds between different types of abuse. However, the Guidance does not explicitly clarify the relationship between these general criteria and the specific factors relevant to each particular type of exclusionary conduct. The clarification of this link is important to explain the extent to which individual market conditions contribute to anticompetitive foreclosure of tying and bundling.

74 75 76 77 78 79

Ibid, paras 971–72. See Carlton and Waldman, above n 28, 198 ff. Ibid, 212 ff. Microsoft (Commission Decision), above n 1, para 975. Ibid, paras 974 and 979 ff. Commission Guidance, above n 15, para 20.

162 Pranvera Këllezi The Guidance makes it clear that its objective is only to present the factors which increase the likelihood of intervention.80 The two general relevant criteria with regard to tying and bundling are, first, the undertakings’ dominant position in one of the product markets81 and, secondly, the existence of economies of scale and/or scope and network effects on the relevant market.82 As discussed above, foreclosure effects are likely to occur when the tied product market is characterised by scale economies, network effects or an installed base, since only in those cases are tying or bundling able to reduce the ability of competitors to achieve scale economies, to reduce their profitability and their ability to grow or innovate. The Discussion Paper section on tying and bundling explicitly mentioned scale economies, learning curve or network effects on the tied market, and the foreclosure mechanism. It is unfortunate that this was not carried through into the Guidance on tying and bundling but was left in the general principles section.83 Another general factor relates to the proportion of the foreclosed sales.84 The Guidance discusses the extent of the allegedly abusive conduct, stating that the higher the total sales in the (tied) market affected by the conduct, and the longer the duration of the conduct, the greater is the likely foreclosure effect. Such a criterion was also present in the Discussion Paper, and was expanded in relation to tying and bundling.85 There the Commission considered that when ‘the dominant company ties a sufficient part of the market, it raises a rebuttable presumption that the tying practice has a market distorting foreclosure effect’.86 The main issue, which also arises in the context of other kinds of exclusionary conduct, is what constitutes a ‘sufficient part of the market’, taking into account that the foreclosure of competitors depends on their own cost structure and economies of scale. Economic theory has not given an answer to these questions, although it is clear that in the presence of substantial 80

Ibid, para 1. The Commission has always stressed the importance of a second condition, namely the existence of market power in the tying product market. See eg Case T-30/89 Hilti v Commission [1991] ECR II-1439; Case C-333/94 Tetra Pak International SA v Commission [1996] ECR I-5951 (‘Tetra Pak II’). These cases involved high market shares in the tying market: Hilti had 70% of market shares in the cartridge market, while Tetra Pak had 90% of the aseptic sector (both carton and packaging machines). See also Case 311/84, Centre Belge d’Etudes de Marché-Télémarketing v CLT [1985] ECR 3261, at 26. The Court held that the fact that CBEMT, a statutory monopoly, subjected the sale of broadcasting time to the condition that the advertisers should use the telephone lines of an advertising agent belonging to the same group as the television station, amounted to an abuse. 82 Commission Guidance, above n 15, para 20. 83 Discussion Paper, above n 16, para 199. Although these factors were mentioned, it is submitted that the Discussion Paper on Article 82 EC did not attached a lot of importance to the conditions in the tied product market. 84 Commission Guidance, above n 15, para 20. 85 Discussion Paper, above n 16, paras 196 ff. 86 Ibid, para 188. 81

Rhetoric or Reform? 163 economies of scope or network effects, tying or bundling artificially reduces the scale and the profitability of competitors. Nevertheless, concentrating on the proportion of the tied sales without making any reference to the required level of economies of scale, or similar market conditions, on a given market is insufficient. This might lead to an automatic finding of exclusionary effect even in highly competitive markets, and therefore to over-intervention. The Guidance sets out a number of specific factors that would contribute to foreclosure effects of tying and bundling. First, the Guidance states that a lasting tying strategy—for instance through technical tying—would increase the risk of anticompetitive foreclosure.87 This criterion relates to coercion, or the absence of choice for customers, which is one of the conditions set out in the Microsoft decision but which is not made explicit in the Guidance. The law on tying and bundling has evolved to cover any direct or indirect form of coercion. In rejecting the argument of Microsoft, the CFI held that for this condition to be fulfilled, the facts that the consumers do not pay a price for the tied product, and that the consumer is not forced to use the tied product or prevented from using the same product when supplied by a competitor of the dominant undertaking, are not relevant criteria.88 Most importantly, the Court confirmed that bundling acts as a deterrent to the pre-installing of other media players, and as a disincentive to use other competitors’ media players, even though the latter may be of better quality.89 The CFI is also right to point out that, as long as WMP remains pre-installed, consumers will have a limited incentive to install and use competing products.90 Therefore, even if technical tying might be easy to reverse, the effect on competition will remain the same if consumers have no incentive to choose another competing product. The emphasis on consumers’ incentives enables economic theory to play its role in full, and to assess the incentive mechanisms and the situations where tying and bundling may harm competition. For instance, economics shows that mixed bundling forecloses competition, even when consumers have a choice to buy the components of the bundle separately.91 What counts is the exclusionary effect on the market. In this respect, the fact that the Guidance does not expressly refer to coercion shows greater flexibility from a legal point of view. The Guidance gives limited direction concerning bundled rebates,92 or rebates coupled with mixed bundling. When undertakings offer large 87

Commission Guidance, above n 15, para 53. Microsoft v Commission, above n 2, paras 967 and 970. 89 Ibid, para 971. 90 Ibid, para 974. 91 The Discussion Paper, above n 16, paras 182 and 189, dealt expressly with mixed bundling, and presented it as a form of economic incentive to buy the bundle. 92 Commission Guidance, above n 15, para 59. 88

164 Pranvera Këllezi rebates on a bundle, there is no incentive for consumers to buy competing products separately, so that even equally efficient competitors would risk being excluded from the market. To assess this probable effect of the rebate, the Guidance refers to the incremental costs as a benchmark,93 but recognises that the assessment is complex in practice. The Commission does not provide a complete analysis of the incentives and the market conditions that would make tying and bundling a harmful practice. The Guidance only gives the example of technical tying, a practice that can produce significant efficiencies.94 Moreover, the Guidance merely indicates how companies can price bundles in order to reduce antitrust risk: if the price of the bundle covers the long-run incremental cost of including such a product in a bundle, the antitrust risks are minimised.95 This approach is, however, limited in some cases; for example, it is not appropriate for information goods, which present an almost zero marginal cost. Another situation which increases the likelihood of intervention occurs when the firm is dominant on more than one product market.96 However, the Guidance describes the situation without presenting the market conditions that may contribute to anticompetitive effects. While intuitively such analysis may be correct, in the absence of economies of scale or scope, the network effect or installed base in the tied market, competitors can compete for each of the markets without offering a large bundle of products. This is particularly the case if the consumers are not used to buying and consuming these products together. With regard to the bundling of substitute products, the Guidance does not exclude possible anticompetitive effects but limits its analysis to substitute input products used in variable proportions.97 Indeed, bundling of substitute products may be less frequent, but the Guidance rightly states that the effect on competition is directly related to increases in price. The Guidance takes into account the main conclusions of the economic analysis and remains at the same time flexible to avoid a straitjacket effect in individual cases. There is no change in the outcomes of tying and bundling cases—to date there have not been any tying cases which have been criticised for over-intervention, nor does the Guidance signal a more lenient approach. Rather, the Commission is reforming its decision-making process by integrating economic analysis in a case-by-case and effects-based approach.

93 94 95 96 97

Ibid, para 60. See Daniel Gaynor, ‘Technological tying’ (2006) FTC Working Paper 284. See Guidance, above n 15, para 59. Ibid, para 53. Ibid, para 56.

Rhetoric or Reform? 165 C. Taking into account efficiencies Contrary to Article 81(3), Article 82 does not provide for an explicit legal basis for considering efficiencies. However, the courts have accepted that the behaviour of defendant undertakings may be objectively justified provided that it is proportional.98 If this is found to be the case then the conduct in question is not considered abusive.99 However, even if objective justifications can be invoked by the undertakings and are part of the analysis—indeed, in tying cases, ‘there are almost always efficiency arguments’100—the courts have been reluctant to take these reasons into account.101 In recent years, the Commission’s policy towards dominant undertakings attempting to justify their apparently abusive conduct on efficiency grounds has evolved. In the first guidelines influenced by economic analysis, Article 81(3) EC was interpreted as precluding any application of this provision to restrictive agreements concluded by dominant undertakings.102 In later guidelines, the Commission clarified that this limitation applies only to agreements that constitute an abuse of a dominant position.103 The application of Article 81(3) to the conduct of dominant undertakings is a step toward the integration of efficiencies in the overall assessment. It also suggests that Article 82 should have the ability to take into account possible efficiencies in a coherent way. The Commission has abandoned the view that the exclusionary conduct of dominant undertakings cannot be justified because it causes substantial harm to the competitive structure of the market which cannot be counterbalanced by any efficiency. These developments are welcomed, first, because tying and bundling have great potential to produce economic efficiencies; and, secondly, because the finding of a dominant position in European competition law does not necessarily require a large market share.104 Consequently, even limited anticompetitive effects can be captured

98

See Whish, above n 5, 206–07. Ibid, p 207. 100 John Temple Lang, above n 6, 324. 101 For instance, the courts rejected as disproportionate the claims of Hilti and Tetra Pak (see above n 81) on the grounds of health and safety obligations of these undertakings. 102 Guidelines on Vertical Restraints, above n 67, paras 135, 153, 221 and 222; Communication for the Commission, Notice, Guidelines on the applicability of Article 81 of the EC Treaty to horizontal cooperation agreements [2001] OJ C3/2, paras 36, 71, 105, 134 and 155. 103 Communication from the Commission, Notice, Guidelines on the application of Article 81(3) of the Treaty [2004] OJ C101/97, para 106; Communication from the Commission, Notice, Guidelines on the application of Article 81 of the EC Treaty to Technology Transfer Agreements [2004] OJ C101/2, para 151. 104 See eg Case IV/D-2/34.780 Virgin/British Airways [2000] OJ L30/1, Commission Decision. Upheld by the CFI, Case T-219/99 British Airways v Commission [2003] ECR II-5917, where the Court held that a market share of 39.7% is large enough to support a finding that a firm holds a dominant position. 99

166 Pranvera Këllezi by Article 82, and the possibility of putting forward justifications based on efficiency reasons mitigates the risk of over-intervention. On the other hand, the Commission takes the necessary precautions to avoid justification when the undertaking has a significant market position. It stresses that exclusionary conduct which maintains, creates or strengthens a market position approaching that of a monopoly can normally not be justified on the grounds that it also creates efficiency gains …105

Bundling may be used to increase efficiency, especially in the form of cost savings or quality improvement. For instance, selling two goods together may lower the transaction costs, and may also be valuable to consumers, in particular when consumers would buy both goods when sold separately.106 Efficiencies usually arise in the form of economies of scale or scope in production, distribution, specialisation, and research and development; synergies, in so far as cost and knowledge, can be shared, and learning by doing spills over from one product to another.107 It is recognised by the Guidance that tying and bundling are common practices that can be intended to provide customers with better products or offerings in a more cost-effective way.108 As discussed above, efficiencies could be taken into account at the stage of the assessment of the existence of separate products. If at this stage efficiencies are not so significant as to render the bundle a ‘composed product’, in that there is also an individual demand for separate products, then the negative effect of bundling should be considered. Another step toward recognition of the importance of efficiencies is the use of the as efficient competitor test for price-based exclusionary conduct. Bundling is regarded as harmful when it is able to exclude or foreclose competitors which are as efficient as the incumbent company.109 Alternatively, when anticompetitive exclusionary effects are identified, they can be taken into account at the stage of justification on economic efficiency grounds if the weighing-up of negative effects against substantial efficiencies shows no net harm to consumers.110 If efficiencies are present, they must outweigh the negative effects of tying or bundling. Although tying and bundling may reduce costs and thereby improve production efficiency, the foreclosure of competitors strengthens

105

Commission Guidance, above n 15, para 30. Ibid, para 62. 107 Patrick Rey, Paul Seabright and Jean Tirole, ‘The Activities of a Monopoly Firm in Adjacent Competitive Markets: Economic Consequences and Implications for Competition Policy’ (2001) Institut d’Economie Industrielle, Université de Toulouse 1, 9. 108 Commission Guidance, above n 15 para 49; Discussion Paper, above n 16, para 178. 109 Commission Guidance, above n 15, para 22. 110 Ibid, paras 29 and 31. 106

Rhetoric or Reform? 167 the market power of the dominant undertaking, which may reduce allocative efficiency and ultimately harm consumers. The improved framework for taking into account economic efficiencies is a significant improvement and a real reform toward integrating economic analysis and reducing intervention errors. IV. CONCLUDING REMARKS

Tying and bundling may have anticompetitive effects by transferring market power from one market to another and raising barriers to entry in one or more markets where the dominant undertaking is operating, thereby protecting its position. An economic analysis offers an explanation of mechanisms that lead to anticompetitive foreclosure, and identifies a number of factors that increase the likelihood of anticompetitive effects. The effects-based approach will improve the methodology for finding anticompetitive effects, and the possibility to consider efficiencies will act as a supplementary screening device to avoid over-intervention. The Microsoft case and the Commission Guidance integrate economic analysis to differing degrees. By definition, the Guidance gives general directions on criteria for harmful tying and bundling. On the other hand, in the Microsoft case, the Commission carried out a detailed analysis on a specific set of facts, guided by economic thinking. This demonstrates what economic analysis can offer to competition law: general guidance which should be adapted to the specific facts of the case. In both instances, the Commission shows that it is carrying out a real reform in terms of approaching tying and bundling cases, and is consolidating the principles governing tying and bundling. The risk of competitors’ foreclosure is linked to their reduced capacity to realise economies of scale or scope, or to create and maintain a sufficient base of consumers for profitability, to continue to compete in a market, and to introduce new products or bring innovation to the market. In Microsoft, such economies of scale and network effects were present and fully analysed by the Commission. In contrast, the Guidance fails to clarify that anticompetitive foreclosure may arise only if the tied market structure is not competitive. However, the general framework for finding anticompetitive exclusionary conduct is much improved, and the flexibility of the CFI judgment in the Microsoft case allows the Commission to analyse market conditions and a company’s conduct on a case-by-case basis and fully to consider economic theory.

9 The Darker Side of the Moon: Assessment of Excessive Pricing and Proposal for a Post-entry Price-cut Benchmark *

ARIEL EZRACHI AND DAVID GILO

**

I. INTRODUCTION

E

XCESSIVE PRICING BY a dominant firm is considered one of the most blatant forms of abuse. It is the dominant firm’s simplest possible form of exploitation. Despite this, in many instances, competition authorities refrain from intervening against excessive pricing. Similarly, and to some extent consequently, such cases rarely find their way to the national courts.1 The recent debate on Article 82 EC has focused predominantly on exclusionary forms of abuse.2 That is, it has focused on practices used by the dominant firm to entrench or reinforce its position in the market. On the other hand, the treatment of exploitative abuse has received little attention from the Commission and academics alike. Moreover, in the context of excessive pricing, recent statements from the Commission reflect a lack of enthusiasm in confronting such cases.3 * Slaughter and May lecturer in Competition Law, Oxford University. Director, The University of Oxford Centre for Competition Law and Policy. ** Associate Professor, Buchmann Faculty of Law, Tel Aviv University. David Gilo wishes to thank the Cegla Center for Interdisciplinary Research of the Law, Buchmann Faculty of Law, for financial assistance. 1 The small number of cases being brought by the competition agencies influences the number of follow-on claims in national courts. 2 See, eg, DG Competition, Communication from the Commission, Guidance on the Commission’s Enforcement Priorities in Applying Article 82 EC Treaty to Abusive Exclusionary Conduct by Dominant Undertakings (‘Guidance’) (Brussels, 3 December 2008); DG Competition, Discussion Paper on the Application of Article 82 of the Treaty to Exclusionary Abuses (December 2005). 3 See nn 12 and 13 below.

170 Ariel Ezrachi and David Gilo This chapter explores the main grounds commonly used to justify nonintervention in cases of alleged excessive pricing. In doing so it questions the weight attributed to some of these grounds. While acknowledging the difficulties at stake, the discussion focuses on one of the grounds for non-intervention—the complexity of evaluating whether a price is excessive. It examines the practical challenges in establishing excessiveness and puts forward a proposal for a post-entry price-cut benchmark which may facilitate the finding of excessive pricing. The benchmark is based on the evaluation of the difference between the price a dominant firm had charged before entry of new firms onto its market, and the price that prevailed after such entry. In discussing the proposal, we highlight the advantages and disadvantages of such a benchmark. We explain why competition agencies have refrained from using this benchmark, and consider how such a benchmark could be used by private litigants in national courts.

II. THE MAIN GROUNDS FOR NON-INTERVENTION

Three justifications are commonly cited to explain competition agencies’ reluctance to pursue actions against dominant firms for excessive pricing. The first is that excessive prices attract entry and are therefore selfcorrecting. The second is that the prohibition of excessive pricing might chill firms’ incentive to innovate or invest ex ante.4 The third is the difficulty in determining whether the price is excessive.5 These three considerations serve as the main rationale behind US scepticism of the merit of prohibiting excessive pricing. In several instances, the United States Supreme Court has held that US antitrust law does not prohibit the charging of high prices per se.6 Accordingly, excessive pricing by a dominant firm is not considered a violation, and firms in the US may demand whatever rates they can obtain in the marketplace.7 Illustrative in

4 See, eg, Motta and De Streel, ‘Exploitative and Exclusionary Excessive Prices in EU Law’ (2006), 15, available online at http://professorgeradin.blogs.com/professor_geradins_weblog/ files/ExcessivePrices18122003.pdf; R O’Donoghue and A J Padilla, The Law and Economics of Article 82 EC (Oxford, Hart Publishing, 2006), 607; M Motta, Competition Policy: Theory and Practice (Cambridge, CUP, 2004), 69; Verizon Communications, Inc v Trinko LLP, 157 L Ed 2d 823, 836 (2004) (US). 5 See, eg, O’Donoghue and Padilla, above n 4, at 627. 6 Berkey Photo, Inc v Eastman Kodak Co, 603 F 2d 263, 274, n12 (2d Cir 1979) (US); Blue Cross & Blue Shield United v Marshfield Clinic, 65 F 3d 1406, 1413 (7th Cir 1995) (US); Verizon Communications Inc v Trinko LLP, 157 L Ed 2d 823, 836 (2004) (US). See also the earlier case of United States v American Can Co, 230 F 859. 7 See, eg, Areeda and Hovenkamp, Antitrust law: an analysis of antitrust principles and their application (New York, Aspen, 2006), para 720a, at 254–55; see also Chicago Professional Sports Ltd Partnership v NBA, 95 F 3d 593, 597 (7th Cir 1996) (US), overturning

The Darker Side of the Moon: Assessment of Excessive Pricing 171 this respect are comments made by the US Supreme Court in the landmark case of United States v Aluminum Co of America. There, the Court held that although non-intervention may expose the public to the evils of monopoly, US antitrust law will not condemn the resultant of those very forces which it is its prime object to foster: finis opus coronat. The successful competitor, having been urged to compete, must not be turned upon when he wins.8

Under European competition law, intervention in cases of excessive prices is not excluded. Article 82 of the EC Treaty targets any abuse by undertakings in a dominant position, and refers to the possibility that such abuse may consist of ‘directly or indirectly imposing unfair purchase or selling prices or other unfair trading conditions’.9 The word ‘unfair’ in Article 82(a) EC was held by the European courts and Commission to encompass excessive pricing.10 European Union jurisprudence generally describes an excessive price as one that bears no reasonable relation to economic value.11 Despite the existence of a provision aimed at monitoring excessive pricing, the three grounds identified above have led to a very limited number of excessive price cases in Europe. The reluctance to engage in price regulation was echoed by the Commission in its yearly reports on competition policy. There the Commission noted that the existence of a dominant position is not in itself against the rules of competition. Consumers can suffer from a dominant company exploiting this position, the most likely way being through prices higher than would be found if the market were subject to effective competition. The Commission in its decision-making practice does not normally control or condemn the high level of prices as such. Rather it examines the behaviour of the dominant company designed to preserve its dominance, usually directly against competitors or new entrants who would

a decision that the NBA’s telecast fees were too high and noting that ‘the antitrust laws do not deputize district judges as one-man regulatory agencies’; E M Fox, ‘Monopolization and Dominance in the United States and the European Community: Efficiency, Opportunity, and Fairness’, Notre Dame Law Review 61, 981–1020, at 985 and 993; Michal Gal, ‘Monopoly pricing as an antitrust offense in the US and the EC: two systems of belief about monopoly’, Antitrust Bulletin, Spring 2004, at 343. 8

United States v Aluminum Co of America (Aloca), 148 F 2d 416, 444 (1945). Article 82 EC. In addition the Article targets abuse which involves the limitation of production to the detriment of consumers. 10 See, eg, the ECJ decisions in Case 40-70 Sirena v Eda [1971] ECR 69; Case 26/75 General Motors v Commission [1975] ECR 1367; Case 27/76 United Brands v Commission of the European Communities [1978] ECR 207, appeal on Commission decision [1976] OJ L95/1; also see Commission decision COMP/C-1/36.915 British Post Office v Deutsche Post AG [2001] OJ L331/40; [2002] 4 CMLR 17. 11 Case 27/76 United Brands v Commission [1978] ECR 207; Case 26/75 General Motors v Commission [1975] ECR 1367. 9

172 Ariel Ezrachi and David Gilo normally bring about effective competition and the price level associated with it.12 (emphasis added)

Commenting on the difficulty in measuring excessive pricing, Philip Lowe, Director-General DG Competition, noted that: We are … aware that it is extremely difficult to measure what constitutes an excessive price. In practice, most of our enforcement focuses therefore as in the US on exclusionary abuses, ie those which seek to harm consumers indirectly by changing the competitive structure or process of the market.13

In what follows, we critically assess the three grounds for non-intervention that have had such a notable effect on the Commission’s enforcement policy. III. REFLECTING ON THE NON-INTERVENTIONIST APPROACH

Despite the challenges of excessive pricing cases, it is important to assess critically the three main grounds for non-intervention. The assessment questions the weight attributed at times to some of these grounds. While we do not categorically recommend intervention, such examination may widen the range of cases in which the assessment of excessive pricing is merited and possible. At the very least, it should highlight the arguments for nonintervention which should receive the most weight in a particular case.

A. The self-correcting nature of excessive pricing The first ground for non-intervention is that excessive prices generate excessive profits and therefore attract entry into the market. The premise is that the regulation of excessive pricing is often redundant because prices would be forced down by a new entrant.14 Moreover, it is argued that 12 XXIVth Commission Report on Competition Policy 1994, para 207. See also Vth Commission Report on Competition Policy 1975, para 76; XXVIIth Commission Report on Competition Policy 1997, para 77. 13 Philip Lowe, D-G DG Competition, ‘How different is EU anti-trust? A route map for advisors—An overview of EU competition law and policy on commercial practices’, Speech, ABA 2003, Fall Meeting. 14 See Areeda and Hovenkamp, Antitrust law: an analysis of antitrust principles and their application, (New York, Aspen, 2006), para 720b: ‘[W]hile permitting the monopolist to charge its profit-maximizing price encourages new competition, forcing it to price at a judicially administered “competitive” level would discourage entry and thus prolong the period of such pricing.’; Berkey Photo Inc v Eastman Kodak Co 603 F 2d 263, 294 (2nd Cir, 1979) (US); Whish, Competition Law, 5th edn (LexisNexis, 2003), at 688–89: ‘[I]f normal market forces have their way, the fact that a monopolist is able to earn large profits should inevitably, in the absence of barriers to entry, attract new entrants to the market. In this case the extraction of monopoly profits will be self-deterring in the long run and can act as an important economic indicator to potential entrants to enter the market. If one accepts this view of the way that markets operate, one should accept with equanimity periods during which a firm

The Darker Side of the Moon: Assessment of Excessive Pricing 173 the mere prospect of new entry would often suffice to prevent excessive pricing, as it would deter the dominant undertaking from pricing excessively in the first place. Accordingly, it is claimed that intervention is merited only where significant barriers to entry exist that hinder entry to the market.15 Elsewhere, we have shown in detail that the premise that excessive prices are self-correcting does not justify non-intervention.16 Our principal argument is based on the claim that high prices, in and of themselves, do not attract new entry. It is the post-entry price, and not the pre-entry price, that potential entrants consider when deciding whether to enter. Since the dominant firm can usually cut prices immediately upon a rival’s entry, its excessive pre-entry prices do not affect the potential entrants’ decision to enter, unless they signal to potential entrants that the dominant firm is relatively inefficient, making entry profitable.17 Accordingly, if a potential entrant has sufficient information regarding the incumbent’s advantages, and particularly the incumbent’s marginal costs,18 and the entrant perceives the incumbent to be comparatively more efficient, it is unlikely to enter, even if the incumbent charges an excessive price. Conversely, if a potential entrant perceives the incumbent to be comparatively less efficient, it is likely to enter, but not because of the excessive price. Such an entrant would have entered regardless of pre-entry price levels. We also explore the case in which pre-entry prices could serve as a signal to uninformed entrants, and we show that such a signal could be equally clear (or sometimes clearer) in a regime in which excessive prices are prohibited.19 Consequently, the alleged self-correcting nature of excessive prices should not be overstated, as by itself it does not justify non-intervention. This shifts

earns monopoly profit: the market will in due course correct itself and intervention by the competition authorities will have the effect of undesirably distorting this process.’; Korah, EC Competition Law and Practice, 6th edn (Oxford, Hart Publishing, 1997), at 113: ‘The cost price approach ignores the function of pricing as a signal encouraging new entrants. If prices and profits are high, new firms may be attracted into the market over, at least, modest entry barriers.’; O’Donoghue and Padilla, above n 4, at 635–36; M Motta and De Streel, above n 4, at 15; Report by the Economic Advisory Group on Competition Policy, An Economic Approach to Article 82, July 2005, at 11: ‘Such a policy intervention [against monopolistic pricing] drastically reduces, and may even forego the chance to protect consumers in the future by competition rather than policy intervention.’; Napp Pharmaceutical Holdings Ltd, Decision of the D-G OFT, CA98/2/2001: ‘The Director considers that a price is excessive and an abuse … where it is clear that high profits will not stimulate successful new entry within a reasonable period.’ 15

See sources cited ibid. See Ariel Ezrachi and David Gilo, ‘Are Excessive Prices Really Self-correcting?’ Journal of Competition Law and Economics (2009) 5(2) 249–268. 17 Ibid. 18 A firm’s marginal cost is its cost of supplying one additional unit. 19 Ezrachi and Gilo, above n 16. 16

174 Ariel Ezrachi and David Gilo the focus from the self-correcting nature of excessive pricing to the other two grounds for non-intervention.

B. Chilling effect on investment The second ground used to justify non-intervention is that the prohibition of excessive pricing might chill firms’ incentive to innovate or invest ex ante.20 Such a concern supports a ‘hands-off’ approach in cases in which the court or competition authority holds that prohibiting high profits would harm ex ante investment incentives.21 This concern is no doubt a valid one. We wish to stress, however, that it arises only in markets in which competition, as envisioned by the competition laws, fails to provide the desired outcome. These are markets in which, had the dominant firm faced healthy competition enabling it to earn only competitive profits, it would not be induced to engage in presumably valuable investment. In such markets, regulation limiting entry of new competitors and expansion of small competitors could be welfare-enhancing, as new entry of viable competitors or expansion of small competitors would erode the dominant firm’s supra-competitive profits and again harm its ex ante investment incentives. Examples for such industries are those involving intellectual property protection. In order to encourage innovation and R&D, intellectual property rights help the inventor insulate itself from competition, enabling the inventor reasonably to expect supra-competitive profits. Note, however, that even in such a market, the argument against intervention loses much of its force once the dominant firm’s investment has been recouped, through a sufficiently long period of dominance and abovecost pricing. A firm whose patent or monopolistic government licence has expired can serve as a good example for such a case. Once a patent or a monopolistic licence expires, not only does viable entry by new firms into the market often become plausible, but the assumption is that the dominant firm has already recouped its ex ante investment.

20 See, eg, Motta and De Streel, above n 4, at 15; O’Donohue and Paddilla, above n 4, at 607; M Motta, above n 4, at 69; Verizon Communications, Inc v Trinko LLP 157 L Ed 2d 823, 836 (2004) (US). 21 See, eg, Motta and De Streel, above n 4, at 16, who argue that when such investment considerations exist, there should be no intervention against excessive pricing. See also David S Evans and A Jorge Padilla, ‘Excessive Prices: Using Economics to Define Administrable Legal Rules’ CEMFI Working Paper No 0416 (September 2004), proposing an interventionist approach only in ‘exceptional circumstances’ where the firm enjoys near monopoly position, the price charged by the firm widely exceeds its average total cost, and there is a risk that the prices would prevent the emergence of new goods and services in adjacent markets.

The Darker Side of the Moon: Assessment of Excessive Pricing 175 C. Difficulty in assessment The third ground for competition agencies’ reluctance to attack excessive pricing is the difficulty in determining what constitutes an excessive price.22 The difficulty in determining the competitive price makes the prohibition of excessive pricing complex, cumbersome and inaccurate. A finding of excessive pricing requires the competition agency to establish what the competitive price might have been, and whether the difference between this price and the price charged by the dominant undertaking is excessive.23 This provides two difficulties: a) different jurisdictions may have different views as to what amounts to an unfair difference between the competitive price and the price actually charged;24 b) the difficulties in assessment could result in inconsistency and unpredictability, which could be connected with the second ground for nonintervention concerning the need to stimulate valuable investment. The larger such investment, the more permissible the price difference between the actual price and the ‘competitive’ price. In what follows we focus on the third ground for non-intervention and explore the difficulties in establishing excessiveness. To ease some of the difficulties in assessment we put forward a proposal for a ‘post-entry pricecut’ benchmark which can in some circumstances alleviate problems of assessment in a welfare-enhancing way. IV. EVALUATING EXCESSIVENESS

Article 82 EC stipulates that an abuse may consist of ‘directly or indirectly imposing unfair purchase or selling prices or other unfair trading conditions’.25 The word ‘unfair’ in Article 82(a) was held by the European courts and the Commission to cover a wide range of pricing practices, including excessive pricing.26 European Union jurisprudence describes an excessive 22 See, eg O’Donohue and Padilla, above n 4, at 627; R Whish, Competition Law, 4th edn (London, Butterworths, 2001), 634–35. 23 The US Supreme Court stressed the practical difficulty in measuring the excessiveness of prices as early as 1897. See United States v Trans-Missouri Freight Assn, 166 US 290 (1897). 24 On the range of approaches to unfairness, see Evans and Padilla, above n 21. See also O’Donoghue and Padilla, above n 4, at 621–38. 25 Similarly, Article 82 targets abuse involving the limitation of production to the detriment of consumers. 26 See, eg, the ECJ decisions in Sirena v Eda, above n 10; General Motors v Commission, above n 10; United Brands, above n 10. Also see Commission Decision COMP/C-1/36.915 British Post Office v Deutsche Post AG, above n 10.

176 Ariel Ezrachi and David Gilo price as one which ‘has no reasonable relation to the economic value of the product supplied’.27 The Court added that this excess could, inter alia, be determined objectively if it were possible for it to be calculated by making a comparison between the selling price of the product in question and its cost of production, which would disclose the amount of the profit margin. The questions therefore to be determined are whether the difference between the costs actually incurred and the price actually charged is excessive, and, if the answer to this question is in the affirmative, whether a price has been imposed which is either unfair in itself or when compared to competing products.28

The UK’s Office of Fair Trading (OFT) also indicated that the measurement of supra-normal profit may involve a cost/revenue comparison, return on capital/weighted average cost comparison, or a comparison of the ‘certainty equivalent accounting rate of return’ (CARR) over a number of years with the risk free rate of interest.29 A crucial problem which results from the test described above concerns the finding of the economic value of the product through the assessment of costs. The determination of cost is often prohibitively difficult.30 This task is even more formidable when dealing with a multi-product undertaking, where allocation of costs to the different products the firm produces becomes difficult, if not impossible.31 Furthermore, cost is not a good approximation of the price that would have prevailed under viable competition. The difficulty arises in at least two respects. First, as is well known, a firm’s fixed costs—which are not affected by its level of production—do not affect the price charged in the marketplace.32 Laddie J, in British Horseracing Board v Victor Chandler International,33 stressed this difficulty. In his

27

United Brands v Commission, above n 10, para 250. Ibid, para 251-2. 29 OFT 414, ‘UK Guidelines on Assessment of Individual Agreements and Conduct’, paras 2.1–2.23. Also note OFT 657, ‘Assessing Profitability in Competition Policy Analysis’. See also Martin Graham and Anthony Steele, ‘The Assessment of Profitability by Competition Authorities’ (1997), OFT Research Paper No 10. 30 Note that the need to evaluate cost within the application of competition law is not unique to excessive pricing and is also conducted in cases of predatory pricing. According to current doctrine in the EU, the low prices of a dominant firm can be condemned as predatory if they are below Average Variable Costs (AVC). When prices are above AVC but below Average Total Costs (ATC), the conduct may be held to be abusive when the undertaking intended to eliminate a competitor. See Case C-62/86 AKZO Chemie BV v Commission [1991] ECR I3359, [1993] 5 CMLR 215; Case C-333/94P Tetra Pak International SA v Commission [1996] ECR I-5951, [1997] 4 CMLR 662; DG Competition, Guidance, above n 2. 31 See O’Donohue and Padilla, above n 4. 32 The reason is that a profit-maximising firm supplies additional units (and, accordingly, reduces its price) as long as the revenue from supplying the additional unit is greater than the cost of supplying this unit. Only the costs per unit and the demand structure, and not fixed costs, affect this mechanism. See, generally, Jean Tirole, The Theory of Industrial Organization (Cambridge, Mass, MIT Press, 1988) at 66. 33 [2005] EWHC 1074 (Ch). 28

The Darker Side of the Moon: Assessment of Excessive Pricing 177 decision, Laddie J questioned whether in competitive markets prices should end up covering the cost of capital. He noted: I do not see that there is any necessary correlation between the cost of production and the cost of capital and the price which can be achieved in the market place. … this [approach] breaks down as soon as one applies it in the real world. What happens if there are only a few customers? Must the cost of production, including all research and development, be recovered from them? If so, does that mean that the price varies depending on the number of customers one has? Does it also mean that the price must go down once all the research and development costs have been recovered? Does it mean that traders cannot increase the price if they engage in successful advertising campaigns which whet the consumer’s appetite?34

Secondly, the price that would have prevailed under viable competition does not even resemble the variable cost—the portion of costs that does depend on the level of production. Even with viable competition, competition is seldom perfect. It typically does not drive prices all the way down to the costs of supplying each unit. The main reasons are product differentiation (consumers do not see competing products as perfect substitutes)35 and capacity constraints (the competing firms have constraints as to the quantity they can supply, given the size of their plant, distribution and supply networks, etc).36 Hence, variable cost, or the cost per unit, cannot serve as a proxy to the price that would have prevailed in a competitive market.37 Given that cost is extremely difficult to assess and serves as an imperfect proxy for the competitive price, Community courts have at times considered the question of excessiveness through comparative analysis of prices charged in competitive markets and prices charged by the dominant undertaking.38 It is often possible to look at prices charged in similar industries in other geographic markets, which do seem to enjoy competitive conditions. For example, in Lucazeau v SACEM, a comparison of prices charged for the same service in different Member States was held to form a reasonable basis for establishing excessiveness of price.39 The accuracy of such a comparative benchmark may increase when comparing prices charged by the same undertaking in different markets, some competitive, some not. Such comparison eliminates possible differences between undertakings as it focuses on 34

Ibid, paras 47–49. See Tirole, above n 33, ch 7. 36 Ibid, ch 5. 37 See generally n 4 above. 38 But see, eg, United Brands, above n 11, overruling a Commission holding of excessive pricing due to lack of a cost study while other benchmarks were available. See also the Competition Appeal Tribunal (CAT) decision in Albion Water Limited v Water Services Regulation Authority [2006] CAT 36, where the CAT required further investigation by the Authority, notably to determine the extent to which the price was above cost, while other benchmarks were available. 39 Case 110/99 Lucazeau v SACEM [1989] ECR 2811. 35

178 Ariel Ezrachi and David Gilo the same dominant undertaking and its pricing strategy in different markets. The dominant firm could justify the price difference only if it faced higher costs in providing the product in the high-priced market. In Bodson,40 for example, the excessiveness of price was assessed by comparisons to pricing in other regions where the market was competitive. In this case the ECJ assessed the fairness of the price charged for funeral services in a particular area, by comparing these charges with other areas in which the undertaking did not hold a dominant position. In British Post Office v Deutsche Post AG, a comparison between price for cross-border and domestic mail was used to determine the excessiveness of price.41 In Ministère Public v JeanLouis Tournier, the ECJ indicated that when a dominant undertaking charges higher fees for its services in one Member State than in others: where a comparison of the fee levels has been made on a consistent basis, that difference must be regarded as indicative of an abuse of a dominant position. In such a case it is for the undertaking in question to justify the difference by reference to objective dissimilarities between the situation in the Member State concerned and the situation prevailing in all the other Member States.42

In the case of Napp Pharmaceutical Holding,43 the OFT established the excessiveness of Napp’s price by using a double benchmark, which included both a cost-price comparison and a price comparison across markets. Napp appealed the OFT’s decision to the CAT.44 On appeal, the CAT assessed in detail the OFT finding of excessive price and considered that comparisons of: (i) Napp’s prices with Napp’s costs, (ii) Napp’s prices with the costs of its next most profitable competitor, (iii) Napp’s prices with those of its other competitors and (iv) Napp’s prices with prices charged by Napp in other markets, were among the approaches that could reasonably be used to establish excessive prices in that case.45

The CAT upheld the OFT decision and stated that ‘during the period of the infringement, Napp’s prices in the community segment were significantly higher than would be expected in a competitive market’.46 In general, the use of the comparative benchmark resolves the difficulties associated with establishing cost structure and economic value. Such benchmarks aim to identify the competitive price and compare it to the

40 Case 30/87 Corinne Bodson v SA Pompes Funèbres des Regions Liberées, [1988] ECR 2479. 41 Case COMP/C-1/36.915 British Post Office v Deutsche Post AG, above n 10. 42 Case 395/87 Ministère Public v Jean-Louis Tournier [1989] ECR 2521, para 38. 43 Napp Pharmaceutical Holdings Ltd, above n 14. 44 Napp Pharmaceutical Holdings Ltd v Director-General of Fair Trading [2002] CAT 1. 45 Ibid, para 392. 46 Ibid, para 403; more generally note the OFT draft guidelines on assessment of conduct, where the OFT provides a list of possible benchmarks which may be used to assess whether prices are excessively high, and these include comparative benchmarks. See OFT 414, paras 2.7–2.15.

The Darker Side of the Moon: Assessment of Excessive Pricing 179 alleged excessive price. Note, however, the limitation of this benchmark, as it compares price between different geographies or different undertakings. Although it offers a valuable proxy of the competitive price, either on its own or in addition to a cost-based analysis, it does not provide definitive evidence of the competitive price that would have been charged by the dominant undertaking in the same market. V. PROPOSAL FOR A POST-ENTRY PRICE-CUT BENCHMARK

As noted, a comparative benchmark provides a valuable tool for approximating the competitive price. The court or competition agency can use it to examine whether the price charged by the dominant undertaking is excessively high in comparison to that proxy. However, one notable difficulty with such a benchmark lies in its reliance on certain assumptions. It must either be based on market realities different from those of the dominant firm’s market, ie where it is comparing prices charged by the dominant undertaking in a different market, or be based on undertakings different from the dominant firm. At times, it may be possible to improve the quality of the comparative benchmark by comparing different pricing strategies by the same undertaking in the same market. In British Leyland v Commission,47 the ECJ upheld a Commission decision in which British Leyland was found to abuse its dominant position by charging excessive fees for its services. British Leyland enjoyed a State monopoly for the issuance of certificates of conformity to vehicles in Great Britain. It systematically increased the price charged for issuing the certificates from 1981 to 1983, irrespective of the costs of its activities. Absent justification for the price increase, the Court found that the fees were fixed at a level which was disproportionate to the economic value of the service provided.48 The decision is interesting, in the sense that the comparison of price was conducted in the same market and with reference to the same undertaking. Yet in a case such as British Leyland, in which prices had increased over time, it was still open to the dominant undertaking to show that prices had increased due to a parallel increase in costs. In this section we put forward a proposal for a comparative benchmark which similarly considers pricing by the same undertaking and in the same market. Our proposed benchmark is applied in cases where the dominant undertaking cuts its price due to new entry into the market. 47

Case 226/84 British Leyland Plc v Commission [1986] ECR 3263, [1987] 1 CMLR 185. In its decision, the Court cited the General Motors judgment, above n 10, in which it was held that an undertaking abuses its dominant position where it has an administrative monopoly and charges for its services fees which are disproportionate to the economic value of the service provided. 48

180 Ariel Ezrachi and David Gilo Suppose that after enjoying several years of dominance, a new entrant (or new entrants) finally enters the dominant firm’s market. Naturally, the new entrant (or entrants) is expected to charge prices lower than those charged by the dominant firm, in order to gain a customer base and market share. The dominant firm, for its part, is expected to react by cutting its own price, in order to try to retain its market share. Absent tacit or explicit collusion between the dominant firm and its new rival(s), such competitive tension typically lowers the price down to the competitive price. This competitive price can be a useful benchmark in showing that the dominant firm charged an excessive price before entry. According to a post-entry price-cut benchmark, when a dominant firm significantly cuts its prices upon new entry into its market, it could be found to have priced excessively prior to entry. This will be the case when the difference between the dominant firm’s pre-entry prices and post-entry prices exceeds a threshold, defined by the competition authority or court. For such a finding to be made, the low post-entry price would have to persist for a considerable period of time. As an illustration, suppose that the dominant firm is a former patent holder whose patent has expired, or a former government-created monopoly whose market is now opened to competition. Suppose further that this dominant firm had priced at a certain level for several years following the expiration of the patent or the monopolistic licence, but then, following new entry into its market, the dominant firm significantly cut its prices in order to defend its market share. We explore whether the price difference between the pre-entry price and the post-entry price can serve as evidence as to the excessiveness of the pre-entry price in the years following the expiration of the patent or monopoly licence. This benchmark is particularly interesting, in that the dominant firm might have relatively few excuses for significant post-entry price cuts other than pricing excessively before entry. It would seems quite a coincidence that the same firm would charge significantly different prices in the same market, where the sudden change in its pricing behaviour occurred only after competitive entry into its market. To be sure, in rare cases, the dominant company might be able to justify such behaviour. The dominant firm could try to show that an exogenous change other than new competition justified the price drop.49 The use of a post-entry price-cut benchmark for showing that pre-entry prices were excessive bears some resemblance to proposals made by scholars in the context of predatory pricing. Edlin asserted that a dominant firm should be blocked from significantly cutting its price for a period of

49 For example, an excessive price difference could be justified by proof of a sudden drop in the dominant firm’s costs that occurred at the same time as entry.

The Darker Side of the Moon: Assessment of Excessive Pricing 181 12 to 18 months following substantial entry into its market.50 Williamson proposed a rule according to which, in the post-entry period, the dominant firm should not be able to increase output above the pre-entry level for 12 to 18 months.51 Baumol suggested that if an entrant is driven out of the market following a price cut by an incumbent firm then the incumbent should not be allowed to raise its price again unless justified by cost or demand changes.52 We, however, do not consider a post-entry price cut by a dominant firm as possible predatory pricing but rather merely as evidence, among other items of evidence, to support the claim that the pre-entry price was excessive.53 A potential problem with using a post-entry price-cut benchmark to help reveal excessive pre-entry pricing is that it could alter the behaviour of the dominant firm and of new entrants into its market. In particular, if the dominant firm knows that a large difference between pre-entry and postentry price could be used as evidence of excessive pricing, it might not only lower its pre-entry prices but also raise its post-entry prices so as to minimise this difference. Moreover, if the new entrant foresees such behaviour, the entrant too could raise the price it charges after entry, compared to the price it would have charged absent the use of this benchmark. However, this disadvantage of the post-entry price-cut benchmark should not be overestimated. First, existing benchmarks which compare the dominant firm’s own pricing in several markets share a similar disadvantage. Consider, for example, the widely-used benchmark which compares the dominant firm’s price in the market in which it is dominant, to its price in another market in which the firm is subject to competition. Knowing that its different prices in both markets could support an excessive pricing claim, the dominant firm may want to raise the price it charges in the competitive market. The firm’s rivals in the competitive market may raise their prices accordingly. Secondly, if the dominant firm is expected to 50 Aaron S Edlin, ‘Do Guaranteed-Low-Price Policies Guarantee High Prices, and Can Antitrust Rise to The Challenge?’ (1997) 111 Harvard Law Review 528, at 563. Against, see Elhauge, ‘Why Above-Cost Price Cuts to Drive out Entrants Are Not Predatory—and the Implications for Defining Costs and Market Power’ (2003) Yale Law Journal, 112, 681–827. 51 Williamson, ‘Predatory Pricing: A Strategic and Welfare Analysis’ (1977) Yale Law Journal, 87, 284. 52 Baumol, ‘Quasi-permanence of Price Reductions: A Policy for Prevention of Predatory Pricing’ (1979) Yale Law Journal, 89, 1–26. 53 Note, the damages to be calculated in a predatory pricing suit are those of the entrant, harmed by post-entry price cuts. Such damages stem from lost sales. In contrast, when the pre-entry price is attacked as excessive, damages equal the difference between the incumbent’s pre- and post-entry prices, multiplied by the number of units sold under the pre-entry price (admittedly, such a measure of damages is understated, in the sense that it does not grasp the harm to consumers who, due to the excessive price, refrained from buying the product). In any case, calculating damages from excessive pre-entry pricing seems easier than calculating the hypothetical loss of sales of a new entrant under a predatory pricing claim. The former measure also better reflects the social loss (rather than the loss to a particular firm) stemming from the dominant firm’s behaviour.

182 Ariel Ezrachi and David Gilo lower its post-entry price more moderately, this in itself could attract entry of firms that otherwise might have hesitated to enter. More entry could, naturally, benefit consumers.54 Thirdly, the post-entry price-cut benchmark is typically only one of the potential pieces of evidence that could support an excessive pricing claim. Hence, the dominant firm knows that even if it keeps its post-entry price relatively high, its pre-entry price may still be successfully attacked using other benchmarks and methods. Reacting softly to entry only in order to weaken somewhat the excessive pricing claim against it thus may not be worth the significant sacrifice of post-entry profits and market share. Of course, the more numerous the new entrants into the dominant firm’s market, the less troubling is the fear of soft post-entry competition. Even if the dominant firm would be reluctant to cut prices substantially after entry, competition among the new entrants would considerably drive prices down. VI. ENFORCEMENT CHOICES

Having outlined the proposed post-entry price-cut benchmark and explained its benefits and limitations, we now turn to explore how it may be used by competition agencies or private litigants. Let us distinguish between public and private enforcement.

A. Public enforcement By its nature the post-entry price cut benchmark comes into play only following successful entry. Consequently, although the benchmark relieves many of the difficulties in assessment, it does so in circumstances in which the European Commission might deem intervention inappropriate and superfluous. Indeed, we are not aware of EU cases which have used this benchmark. The Commission is unlikely to use a post-entry price-cut benchmark in excessive pricing cases, because it is typically expected to intervene in cases which involve ongoing abuses of a dominant position, while the benchmark that we explore involves a former abuse of a dominant position (namely, the excessive prices the dominant firm had charged in the pre-entry period).55 That said, we see no reason why competition 54 New entrants attracted by above-cost pricing of incumbent firms may be less efficient than the incumbent firm, but even the entry of less efficient firms could promote consumer welfare, which is the main focus of competition law. 55 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, OJ L1/1—which constitutes the legal basis for the Commission’s powers to impose remedies for infringements of Article 82—provides in Art 7(1) that the Commission ‘may impose … any behavioural or

The Darker Side of the Moon: Assessment of Excessive Pricing 183 agencies should limit themselves to remedying existing infringements. Their expertise and access to relevant market data place them in a superior position to consider cases of excessive pricing and use the post-entry price-cut benchmark with the aim of deterring infringements of competition law. B. Private enforcement The use of competition law by private parties in national courts may increase social welfare by supplementing public enforcement, enhancing its deterrent effect and by providing a channel for corrective justice through compensation and injunctive relief.56 The public value of the complementary role of private enforcement was echoed in Courage v Crehan, where the ECJ held that the right to sue for damages against violations of competition law strengthens the working of the Community competition rules and discourages agreements or practices, which are able to restrict or distort competition. From that point of view actions for damages before the national courts can make a significant contribution to the maintenance of effective competition in the Community.57

In the context of a dominant firm that allegedly charged excessive prices for a significant period (say, a few years), and then ceased to do so when entry occurred, it seems particularly important to enable private action on account of the pre-entry abuse. As noted, competition agencies may lack incentive to act in such cases. Accordingly, absent private enforcement, the dominant firm would have no reason to refrain from pricing excessively, while it can, before entry into its market occurs. Due to the cost and risks associated with such private actions, and the fact that the victims of abuse are typically dispersed consumers, each of which has a small stake, one would expect such suits to be lead by representative or collective actions. In its White Paper on Damages Actions for Breach of the EC Antitrust Rules, the Commission considered the question of collective redress and suggested two complementary mechanisms58: a) representative actions, which are brought by qualified entities; and b) opt-in collective actions, in which victims decide to combine their individual claims for harm they have suffered into one single action. structural remedies which are proportionate to the infringement committed and necessary to bring the infringement effectively to an end.’ (emphasis added) 56 Case 26/62 NV Algemene Transport- en Expeditie Onderneming van Gend & Loos v Netherlands Inland Revenue Administration [1963] ECR 1: ‘the vigilance of individuals concerned to protect their rights amounts to an effective supervision in addition to the supervision entrusted … to the diligence of the Commission and of the Member States.’ 57 Case C-453/99 Courage Ltd v Bernard Crehan [2001] ECR I-6297, para 27. 58 http://ec.europa.eu/comm/competition/antitrust/actionsdamages/files_white_paper/ whitepaper_en.pdf.

184 Ariel Ezrachi and David Gilo If accepted, these proposals would facilitate private enforcement, and may serve as the main vehicle to advance post-entry excessive pricing claims. A notable hurdle that a private claimant is likely to face using the postentry price-cut benchmark is the lack of a public investigation on which to base its claim. This can materially impact on the likelihood of such claims reaching the national court. A stand-alone claim requires the claimant to prove all the elements of the infringements, and to bear the associated costs and risks. The claimant would need to establish the defendant’s dominant position in a properly defined market, and that the dominant firm had indeed priced excessively. In order to apply the post-entry price-cut benchmark, the claimant (say, of a class action by consumers) would need to acquire information as to the quantity the dominant firm had sold for the allegedly excessive price before entry. It would also have to show that the post-entry price is indeed a good proxy for the industry’s competitive price, rather than a price which is temporarily below the so-called ‘competitive’ price. For example, a new entrant may temporarily charge a particularly low price, below the industry’s true competitive price, in order to promote its product and get a foothold in the market. The national court would have to be persuaded that the difference between the pre-entry price and the post-entry (‘competitive’) price is excessive. This could be a difficult task, particularly when the dominant firm had made significant welfare-enhancing investments (such as in R&D) that could justify a relatively high profit margin until entry occurred. Typically, all of the relevant information regarding ex ante investment, cost structure and plausible explanations justifying the post-entry price cut is in the hands of the dominant firm. Accordingly, it could be sound policy to shift the burden of proof to the dominant firm once the plaintiff establishes that the difference between the pre- and post-entry prices is above a certain threshold.

VII. CONCLUSION

The recent review of the scope and application of Article 82 has predominantly focused on exclusionary abuse, leaving aside the ‘darker side of the moon’, namely exploitative abuse. In this chapter we have attempted to shed some light on the most blatant form of exploitative abuse: excessive pricing. Despite the fact that excessive pricing causes direct harm to consumers, competition agencies have often preferred to limit their intervention in such cases. The most convincing reason for non-intervention is the difficulty of measuring the competitive price and subsequently establishing that the price charged is excessive.

The Darker Side of the Moon: Assessment of Excessive Pricing 185 We have put forward a proposal for a post-entry price-cut benchmark which can be used to detect excessive pricing. This benchmark, we believe, can enrich the analytical toolbox available to competition authorities or private plaintiffs, and assist in identifying cases of excessive pricing, albeit that there are limitations associated with its use. Moreover, it may contribute to deterring dominant companies from pricing excessively in the first place.

10 The Sanction of Voidness Under Article 82 EC and its Relation to the Right to Damages ULF BERNITZ*

I. INTRODUCTION

T

HE SANCTION OF voidness is of fundamental importance in relation to the prohibition on restrictive agreements in Article 81 EC. However, as is well known, Article 82 EC, prohibiting abuse of dominant position, is silent on the private law consequences and does not include a provision on the voidness of agreements covered by that article. In this chapter, the issue of voidness under Article 82 will be revisited in the light of the recent Commission White Paper on Damages under competition law.1 In the author’s view, there is a strong link between the right to damages and the issue of voidness. II. PRIVATE REMEDIES IN CONTRACT AND IN TORT

The White Paper, Damages actions for breach of the EC antitrust rules, published in 2008, focuses on the importance of the right to damages for private enforcement of European competition law. The Commission stresses the importance of effective remedies for private parties and their role in combating illegal restrictions of competition. However, the White Paper is completely silent on the other major part of the private law sanctions

* Professor of European Law, Stockholm University; Director, Oxford/Stockholm Wallenberg Venture in European Law, Associate Senior Fellow, Pembroke College, Oxford. The text is partially based on a Round Table Seminar in Competition Law at Pembroke College, Oxford, 8 May 2008. 1 Commission (EC), Damages actions for breach of the EC antitrust rules (White Paper), COM (2008) 165 final, 2 April 2008; Commission (EC) Damages actions for breach of the EC antitrust rules (Staff Working Paper accompanying the White Paper) SEC (2008) 404, 2 April 2008. The White Paper was preceded by Damages action for breach of the EC antitrust rules (Green Paper), COM (2005) 672 final, 19 December 2005.

188 Ulf Bernitz against anticompetitive agreements and misuse, namely, the sanction of voidness and its legal consequences. Consequently, the relationship between contractual nullity and the right to claim damages is not treated in the White Paper. This can be regarded as a weakness in the Paper: obviously there is a close connection in practice between the private law remedies in contract and in tort. As is well known, Article 81(2) EC states: ‘Any agreements or decisions prohibited pursuant to this Article shall be automatically void’. There is no similar provision in Article 82. However, abuses falling within the Article 82 prohibitions often take the form of contract terms imposed by the party having a dominant position. The Treaty test itself mentions a number of such abuses, including: imposing unfair purchase or selling prices; imposing unfair trading conditions; applying dissimilar conditions to equivalent transactions; and applying tying agreements or conditions. Although the private law effects of prohibited abuses are not treated in Article 82, there is a general consensus that agreements and contract terms contravening Article 82 are also void and unenforceable between the parties. This legal rule is considered to follow primarily from the direct effect of the provision2 and also by analogy with Article 81(2). As the ECJ stated in the Courage case, Article 82, like Article 81(1), produces direct effects in relations between individuals and creates rights for the individuals concerned which the national courts must safeguard.3 However, there is no explicit decision of the European Court of Justice (ECJ) on the voidness of contract terms contravening Article 82. An early case pointing in this direction is BRT v SABAM (BRT II) of 1973, where the Court said: If abusive practices are exposed, it is for the [national] court to decide whether and to what extent they affect the interests of authors or third parties concerned, with a view to deciding the consequences with regard to the validity and effect of the contracts in dispute or certain of their provisions …4

The view that the same principles of invalidity are applicable to Article 82 as to Article 81(1) is also predominant in legal literature, albeit it that the precise limits of this legal principle are unclear and disputed.5 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, Art 1(3): the abuse of a dominant position referred to in Article 82 of the Treaty shall be prohibited, no prior decision to that effect being required. 3 Case C-453/99 Courage v Crehan [2001] ECR I-6297, para 23. 4 Case 127/73 Belgische Radio en Televisie v SV SABAM and NV Fonior [1974] ECR 313, para 14. See also Case 66/86 Ahmed Saeed Flugreisen v Zentrale zur Bekämpfung unlauteren Wettbewerbs [1989] ECR 803, para 45. 5 See eg R Whish, Competition Law, 6th edn (London, LexisNexis UK, 2008), 290 ff; Bellamy and Child, European Community Law of Competition, 6th edn (London, Sweet & Maxwell, 2008), 1431, n 300; Jones and Sufrin, EC Competition Law: Text, Cases and Materials, 3rd edn (Oxford, OUP, 2008); Komninos, EC Private Antitrust Enforcement (Oxford, Hart Publishing, 2008), 158 ff, with rich references to national jurisdictions.

The Sanction of Voidness Under Article 82 EC 189 However, there exists a Swedish Court of Appeal case confirming the state of the law: the Arlanda Terminal 2 case.6 It is to be deplored that this interesting case was neither referred to the ECJ, nor granted leave to proceed to the Supreme Court.7 It is instructive to consider it in more detail. At Arlanda Airport, the major airport in the Stockholm region, the owner of the airport, the State-owned Swedish Board of Civil Aviation, had built a new terminal, Terminal 2, originally designed for the use of the Scandinavian Airlines System (SAS). After some time, SAS found the new terminal unsuitable for its needs and moved out. However, the Aviation Board forced SAS to continue to pay for Terminal 2 in accordance with a very costly tariff system which had been abandoned in relation to other airlines using the Arlanda airport. The SAS sued the Aviation Board in a civil case, invoking European and Swedish competition law on misuse of dominant position. The Court of Appeal clearly spelt out that the competition law provisions are prescriptive and cannot be derogated from by way of agreement between the parties. The Court found that the Aviation Board was applying dissimilar conditions to equivalent transactions to the disadvantage of SAS, without any proven cost justification. It concluded that the pricing conditions constituted an abuse of dominant position contravening Article 82.8 The Court explicitly addressed the question whether invalidity under civil law can follow from breaches of Article 82. Referring to the solidarity principle in Article 10 EC and the direct effect of Article 82, and also quoting the case of BRT v SABAM, the Court found that contract terms comprising abuse of a dominant position should be regarded as having no legal effect. III. EFFECT AND EXTENT OF NULLITY

What does it actually mean that an agreement or contract term is void under EC competition law? Presumably, the same legal principles are to be applied irrespective of whether the invalidity occurs under Article 81 or Article 82. At present, the effect and extent of nullity are decided by the national courts in the Member States. In practice, this result has followed from the

6 Scandinavian Airlines System v Swedish Civil Aviation Administration (unreported), 27 April 2001, Case No T 33/00, Gota Court of Appeal, Sweden. 7 For a more comprehensive presentation of the Arlanda case, see Bernitz, ‘The Arlanda Terminal 2 Case: Substantial Damages for Breach of Article 82’ [2003] Competition Law Journal 195; see also Gustafsson, ‘What are the prospects for enhanced private antitrust litigation? A Swedish perspective’ (2005) 30 European Law Review 490. 8 See on this point Pettersson and Alwall, ‘Discriminatory Pricing: Comments on a Swedish Case’ (2003) 24(7) European Competition Law Review 295.

190 Ulf Bernitz decisions by the ECJ in BRT v SABAM and other cases, referring these issues back to be decided by the national court which had requested the preliminary ruling.9 It follows from the ECJ case law that the nullity in European competition law is of a strong, absolute character. Thus, the ECJ stated in the Manfredi case: Since the invalidity ... is absolute, an agreement which is null and void by virtue of this provision has no effect as between the contracting parties and cannot be invoked against third parties …10

The Court also stated that the ‘principle of invalidity can be relied on by anyone’,11 hence the nullity has effect erga omnes and can be invoked by co-contractors. Thus, in European law the nullity has a specific, Community law meaning which might be more or less different from the concept of nullity as applied in different national jurisdictions.12 It is worth noting the issue of voidness is touched upon in the recently presented Draft Common Frame of Reference prepared by the Study Group on a European Civil Code and the Research Group on EC Private Law (the Acquis Group).13 In this first draft of a possible future European Civil Code, one finds a provision on contracts infringing fundamental principles: A contract is void to the extent that: a)

It infringes a principle recognised as fundamental in the Member States of the European Union; and b) Nullity is required to give effect to that principle.14

Most, and possibly all, Member States of the EU have today a prohibition on misuse of dominant position in their competition law. It seems highly likely that this provision would cover contract terms constituting abuse of a dominant position. It is an established principle that nullity applies only to those individual elements of an agreement which fall within the competition law prohibition. Contractual provisions which are not affected by the prohibition are considered to fall outside its scope and are enforceable, provided those

9

Bellamy and Child, above n 5, 1432, with further references. Joined Cases C-295/04 to C-298/04 Vincenzo Manfredi v Lloyd Adriatico Assicurazioni SpA et al [2006] ECR I-6619, para 57. 11 Ibid. 12 Komninos, above n 5, 150 ff. 13 C von Bar, E Clive and H Schulte-Nölke (eds), Principles, Definitions and Model Rules of European Private Law (Draft Common Frame of Reference 2008) (Munich, Sellier European Law Publishers, 2008). 14 Ibid, Book II, II-7:301. 10

The Sanction of Voidness Under Article 82 EC 191 elements are severable from the agreement as a whole.15 Interestingly, the Draft Common Frame of Reference contains such a general provision: Where only part of a contract or other juridical act is invalid or ineffective, the remaining part continues in effect if it can reasonably be maintained without the invalid or ineffective part.16

However, the application of these principles is to be decided under applicable national law17 and varies to some extent between the Member States depending on the structure of their contract law. It is worth mentioning that the Nordic countries18 have a more flexible approach than, for example, English law. In the Nordic countries the relevant statutory provision is section 36 of the Contract Act, similarly worded in all five countries, which reads: A contract term may be adjusted or held unenforceable if the term is unreasonable with respect to the contract’s contents, circumstances at the formation of the contract, subsequent events or other circumstances. If the term is of such significance for the contract that the remaining part of the contract cannot reasonably be upheld in its original form, the contract may be adjusted in other respects or held unenforceable in its entirety.19

This offers a more flexible solution than the English ‘blue pencil rule’. The Nordic courts are given a broad power to model the effects of unenforceable contract clauses on the rest of the contract, including the possibility to let clauses remain in force after adjustment. Thus, the court can adjust an excessive price by lowering it to a level considered reasonable. To take another example: suppose a restrictive covenant prohibits a party to a contract from competing for seven years within a broadly defined area of activity. If a Nordic court finds this provision to be contrary to competition law, it would not be necessary to rule out the whole restrictive covenant as invalid; the court would have the power to adjust the covenant so it would be reasonable and acceptable under Community law, for example by holding it valid for two or three years and invalid thereafter, and by restricting its scope. It should be noted, however, that in the Arlanda Terminal 2 case, the Swedish Court of Appeal held the contract term on pricing conditions to be invalid in its entirety and upheld the remaining parts of the contract.

15 Case 56/65 Société La Technique Minière (LTM) v Machinenbau Ulm [1966] ECR 235; Whish, above n 5, 289 ff; Komninos, above n 5, 156 ff; Jones and Sufrin, above n 5, 1319 ff. 16 Draft Common Frame of Reference 2008, above n 13, Book II, II-1:109. 17 Komninos, above n 5, 156. 18 The EU countries Denmark, Finland and Sweden, and the EEA countries Iceland and Norway. 19 Translation from the Swedish text version.

192 Ulf Bernitz IV. THE RIGHT TO DAMAGES

The White Paper on damages actions is in a favourable position to build strongly on the ECJ’s recent decisions in the Courage and Manfredi cases. Both cases dealt with Article 81(1) issues, but there seems to be no reason to assume that the Court would have taken a different view had the cases dealt with Article 82 issues. In Courage, the Court proclaimed the private right to damages in competition law, using a very open argument: The full effectiveness of Article [81] of the Treaty and, in particular, the practical effect of the prohibition laid down in Article [81(1)] 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.20

The Court went on to state that the existence of such a right to damages would strengthen the working of the competition rules and discourage unlawful agreements or practices. It is interesting to note how explicitly the ECJ based its decision on legal policy arguments. It did not try to argue that a right to damages in this kind of case would follow from accepted legal principles in the Member States, the obvious reason being that such principles could not yet claim to be established.21 Neither did the Court refer to general principles of Community law. A particular reason why the ECJ did not try to link the right to damages and the effects of nullity in contract law in the Courage case, was probably the fact that under English law, the national law of the case, a party to an illegal agreement is barred from claiming damages from the other party under the principle in pari delicto; the party is considered equally responsible for the wrongdoing.22 As is well known, the Court did not find this, rather particular, common law rule to be applicable in Community law. Under the latter, the Court found a party not to be barred from claiming damages for loss caused by performance of a contract contravening competition law on the sole ground that the claimant himself is a party to that contract. However, the situation would be different if it were established that the claimant party bore significant responsibility for the distortion of competition, as he should not be able to rely on his own unlawful actions to obtain damages.23

20

Courage v Crehan, above n 3, para 26. Basedow (ed), ‘Private Enforcement of EC Competition Law’ (Alphen aan den Rijn, Kluwer Law International, 2007) presents the rather different approaches taken in German, French and Italian law. 22 See eg Whish, above n 5, 291 f. 23 Courage v Crehan, above n 3, paras 17 ff and the Court’s conclusions. 21

The Sanction of Voidness Under Article 82 EC 193 In its 2006 Manfredi judgment, referring to Courage, the ECJ went further and stated that any individual can claim compensation for harm suffered where there is a causal relationship between that harm and a prohibited agreement or practice.24 In Manfredi, the Court was also asked to explain in more detail how damages under competition law should be determined. Primarily, the ECJ was asked to answer the very interesting question whether the claimant would be entitled to punitive damages in order to establish a deterrent effect. Several Member State governments, intervening in the case, declared punitive damages under competition law to be foreign to their legal systems. The Court pointed to the absence of any Community rules governing the matter, and found it to be for each Member State to set the criteria for determining the extent of damages, provided the principles of equivalence and effectiveness are observed. However, the ECJ clarified that Community law requires damages giving full compensation to be awarded. The compensation has to include not only compensation for actual loss, damnum emergens, but also loss of profit, lucrum cessans, plus interest.25 The Court found this rule to follow from the principle of effectiveness and the right of the individual to compensation. In reality, by expressing this so clearly, the Court seems to have established a general principle of Community law on the assessment of damages, capable of being applicable far outside the field of competition law. In addition, the ECJ underlined that Community law does not prevent national courts from taking steps to ensure that the protection of the rights guaranteed by Community law does not entail the unjust enrichment of those who enjoy them. The principle of unjust enrichment is closely linked to the right to restitution or refund, and to the right to obtain damages.26 In the Arlanda Terminal 2 case, discussed above, the Aviation Board was ordered by the Court of Appeal to refund to the SAS the fees paid to the extent that these were found to be discriminatory and in contravention of Article 82, a sum of approximately €55,000,000. Many times, claims for restitution of money which the claimant should not have been forced to pay might be combined with damage claims for actual losses and loss of profit. In competition law disputes between contracting parties, there is no clear distinction between the right to restitution and the right to damages. These different legal concepts might cover the same realities. It is to be hoped that the relationship between damages and the right to restitution under competition law will be clarified more fully during the continued work on European damages rules.

24

Manfredi, above n 10, para 61. Ibid, para 95. 26 On the right to refund, see Case C-242/95 GT-Link v De Danske Statsbaner [1997] ECR I-4449. The rationale of this case should be applicable not only to public undertakings, but also to dominant firms in general. On this point, see Whish, above n 5, 296 f. 25

194 Ulf Bernitz The White Paper builds on Courage and Manfredi. In particular, the Commission states in the White Paper that it proposes balanced measures rooted in European culture and traditions.27 Obviously, deterrence, promoting fewer infringements and greater compliance with EC antitrust rules, is the Commission’s primary policy objective. However, the Commission also stresses the importance of effective compensation to victims of antitrust infringements and law-abiding businesses. In line with Manfredi and existing European legal culture, being different from American, the Commission rejects the introduction of multiple damages.28 According to the White Paper, full compensation must be the guiding principle, and it must be available for all categories of victim, all types of breach and all sectors of the economy.29 In line with Courage, full compensation must be available not only for third parties, but also for parties to a restrictive agreement or contractual requirement who are not responsible for the illegal action. According to the Commission, the standard of liability is neither strict liability nor negligence: the Commission proposes that the infringer should be liable for damages caused, unless he is able to demonstrate that the infringement was the result of a genuinely excusable error. However, an error would be excusable only if a reasonable person applying a high standard of care could not have been aware that the conduct restricted competition. In reality, the White Paper proposal seems to come close to a strict liability rule. The Commission declares that it intends to draw up guidelines for the calculation of damages based on approximate methods of calculation and simplified rules on estimating the loss. Such rules might have a very substantial effect on future court practice in the Member States, in particular when taking into account the fact that most European courts have little experience of competition law damage cases. The Commission’s current approach of applying a more efficiency-based analysis to the application of Article 82, will certainly not make the task easier for national courts.30 V. FINAL REFLECTIONS ON THE RELATIONSHIP BETWEEN CONTRACTUAL REMEDIES AND THE RIGHT TO DAMAGES

The Commission invited comments on the White Paper to be sent by 15 July 2008. Presumably, the Commission will continue its work and present its proposal for a legally binding text. Issues being particularly observed

27

White Paper, above n 1, para 1.2. On the difference between American and European approaches to multiple damages, see C A Jones, Private Enforcement of Antitrust Law in the EU, UK and USA (Oxford, OUP, 1999). 29 White Paper, above n 1, para 1.2. 30 See Ezrachi, ‘The interplay between the economic approach to Article 82 EC and private enforcement’ (2008) Global Competition Litigation Review 148. 28

The Sanction of Voidness Under Article 82 EC 195 during this ongoing legislative process include the position of the consumers and the possibilities for collective redress, problems related to disclosure and access to evidence, rules on the passing-on of overcharges, and the interaction between leniency programmes and actions for damages. Practically nothing is said in the White Paper and the accompanying, more comprehensive Commission Staff Working Paper about the relationship between the right to damages and contractual remedies. The two Papers state only what is the status of the acquis communautaire: no further analysis is to be found. This is a clear weakness in the White Paper. The silence is probably to be attributed to Commission reluctance to be criticised for intervening in the national private law of the Member States more than strictly necessary. However, certain conclusions with regard to the contractual remedies can be drawn from the White Paper on damages actions. The most important one refers to the sanction of voidness under Article 82. As mentioned above, the White Paper makes no distinction between agreements and actions contrary to Article 81 and Article 82; the same rules on the right to damages would be applicable. This position presupposes that the same rules on nullity of contract terms are applicable under Article 82 as under Article 81(2); a different position would not make sense, and would be inconsistent with the law on the right to damages which places infringements of the two articles on an equal footing. Thus, given that Community legislation based on the White Paper will be enacted, we would in reality have a legal rule on the nullity of contract terms, imposed by abuse of dominant position, introduced indirectly. As also mentioned, the White Paper makes no distinction between the right to damages for third parties and damages in contractual relations. The principle of the right to full compensation will also apply in contractual relations, provided the claimant has not been responsible for the illegal action. The White Paper does not discuss the relationship between the right of a party to damages under European competition law and the claims that might follow from the nullity of the contract terms at issue, including the right to claim restitution of payments made in fulfilment of illegal contract terms. In relation to the problem of passing on overcharges, the Commission takes the general position that unjust enrichment should not be accepted.31 However, it remains unclear whether principles of unjust enrichment are meant to be applicable in contractual relations in combination with the right to damages. The unjust enrichment of the wrong-doer might easily amount to a larger sum of money than the damage suffered by the contracting party. These aspects, as well as the relationship between claims for restitution and for damages, deserve more attention before the White Paper results in legislation.

31

White Paper, above, n 1, para 2.6.

196 Ulf Bernitz VI. CONCLUSION

The Commission’s White Paper, Damages actions for breach of the EC antitrust rules, involves more complex issues of private law remedies than are recognised in the White Paper itself, in particular the relationship between damages, unjust enrichment and restitution. Also, as this chapter has demonstrated, there is a close link between a Community law right to damages and a sanction of voidness under Article 82. From the viewpoint of clarity, in addition to the efficiency of private enforcement, the forthcoming legislation should also include a position on the voidness of agreements and decisions which contravene the prohibition on abuse of dominant position in Article 82.

Index abuse of dominance (Art 82) see also specific forms of abuse 2005 Commission review, 1–2, 19–20, 52–3 burden of proof, 26–9, 70–1 categorical thinking see categorical thinking categories of abuse, 37, 44–5 competitive neutrality, 141 damages, 187–96 direct effect, 189 essential facilities see essential facilities EU case law, 1–8 exclusionary abuses, 31, 44–5, 169 exploitative abuses, 31, 44–5, 169 formalism, 20, 70–1 level playing fields, 61, 63, 141 objectives, 71, 130–3 lack of coherence, 142–5 recent EU case law and, 134–42 US comparisons, 126–9, 143–5 ordoliberalism, 3, 69, 72, 130, 137–8 reform need, 2–8 relative dominance, 129 responsibilities, 3 unilateral v collusive practices, 24–30 Acquis Group, 190 Ahlborn, C, 142 allocative efficiency, 75–6, 130 Amato, G, 128–9 anti-competitive agreements (Art 81): burden of proof, 26–9 Commission Guidance, 25–6 nullity, 188 principles, 25 vertical/horizontal dichotomy, 36 Appeldoorn, J, 125 Areeda, Philip, 172n14 Art, Jean-Yves, 41–2 Article 82 see abuse of dominance (Art 82) Bain, JS, 124 Bakos, Yannis, 155 Balto, D, 143 Baumol, William, 181 Beckner, Frederick, 29 Bork, Robert, 128, 131 British Airways: consumer welfare and, 70–1, 82–3, 138–9 effects-based approach and, 57–8

loyalty rebates, 2, 7–8, 138 special responsibility, 138–41 status of guidance, 56 Brynjoffsson, Erik, 155 bundling: bundled discounts, 38–40 classification problems, 37–40 Commission Guidance, 154–5, 158–9, 161–4 efficiencies, 165–7 consumer welfare, 157–8, 159 economic analysis, 150–7 defensive leveraging, 152–4, 161 entry deterrence, 154–5, 161 impact on innovation, 156–7 Microsoft and, 158–61 offensive leveraging, 151–2 welfare, 156 economic approach, 148, 149–50 leverage theory, 148, 150–1 mixed bundling, 38–40, 163–4 recent EU developments, 157–67 burden of proof: EU competition principle, 26–9 formalistic approach, 70–1 capacity constraints, 61–3 Carlton, Dennis, 152–3, 161 categorical thinking: back-door formalism, 30–7 bundling example, 37–40 categories of abuse, 37 conceptual incoherence, 41–3 difficulties of classification, 37–40 effects-based approach and, 21, 44–8 enforcement costs, 37 exploitative v exclusionary dichotomy, 31, 44–5, 169 limiting effects-based approach, 22–30 overview, 21–37 price v non-price dichotomy, 45–8 risk of strategical behaviour, 41–3 shortcomings, 37–43 certainty, 13–14, 36–7, 53–5, 64–5 Chicago School, 124, 127, 128, 130, 131, 133, 143–5, 151 Choi, Jay Pil, 156 classification see categorical thinking

198 Index Commission Guidance: Art 81, 25–6 Commission’s limited mandate, 55–7 conditional rebates, 54, 132 consumer harm, 79–82 concept of consumer, 80–1 definition, 43, 133 predatory pricing, 80 consumer welfare, 76–7, 130–1 critique, 142 degree of dominance, 9–10, 126 effects-based approach, 8–14, 53–5, 149 efficiencies, 33, 53–4, 132, 141, 142 enforcement priorities, 2, 8–14 exclusionary v exploitative abuses, 44–5 foreclosure, 132 landmark, 121–2 margin squeeze, 43 price v non-price dichotomy, 45–8 profit sacrifice test, 132, 136 publication, 53 refusal to supply, 43, 133 special responsibility of dominant firms, 122, 123 status, 55–7, 82 tying and bundling, 154–5, 158–9, 161–4 efficiencies, 165–7 mixed bundling, 38–9 vertical restraints, 160 comparative law, US and EU, 126–9, 143–5 competition law: damages and nullity, 187–96 US v EU objectives, 126–9, 143–5 compulsory licensing: Database Directive, 108–9 exceptional circumstances circumstances or characteristics, 107–8 effect on abuse test, 96–8 entirety of circumstances, 105 innovation criterion, 105 meaning, 87 Microsoft, 87, 93–4, 102–8 objective justification, 106–7 post-Microsoft, 94–6 pre-Microsoft case law, 114–15 test, 92–4 software Database Directive, 108–9 Software Directive, 109–12 conditional rebates, 10, 45, 54, 132 consumer welfare: Article 82 and, 68–73 Chicago School, 130, 131, 133 Commission definition of consumer harm, 48, 133 effects-based approach, 20, 82–5 EU case law and, 82–5, 137–8 EU v US approach, 127–8, 130–2, 143 measurement, 75

modernisation of Art 82, 74–82 allocative efficiency, 75–6 concept of consumer, 80–1 intermediate consumers, 81 non-linear pricing, 81–2 predatory pricing, 80 standard of harm, 77–82 welfare standard, 75–7 sovereignty, 49 tying and bundling, 157–8, 159 contracts: nullity under EC competition law, 189–91 void terms under Art 82, damages, 188–9, 192–4 copyright: software decompilation, 100, 116 software features, 112–14 damages: excessive pricing, 183–4 nullity under EC competition law, 187–9, 192–6 predatory pricing, 181n53 Database Directive, compulsory licensing, 108–9 decompilation, 100, 116–17, 118 Draft Common Frame of Reference, 190, 191 e-commerce, software bundling, 155 EAGCP Report: on formalism, 20 predatory pricing, 83 publication, 20 rule of reason, 35 types of exclusion, 35–6 Economides, Nick, 40 Edlin, Aaron, 180–1 effects-based approach: Article 82 limits, 22–30 capacity constraints, 61–3 categorical thinking and, 21, 44–8 Commission Guidance, 8–14, 53–5, 149 consumer welfare and, 74–82 degree of dominance, 9–10, 126 degree of foreclosure, 10–11 effects on consumers, meaning, 20 EU case law and, 57–60, 134–42, 148–50 European Commission and, 8–14 exclusion of as-efficient behaviour, 11–12 formalism versus, 20 inconsistency risk, 51–2 introduction, 52–5, 67 legal certainty and, 13–14, 36–7, 53–5, 64–5 meaning, 20 national inconsistencies, 60–1 objectives, 53 shortcomings, 21 tying and bundling, 150–7 vertical/horizontal dichotomy and, 36

Index 199 efficiency defence: allocative efficiency, 75–6, 130 Commission Guidance, 33, 53–4, 141, 142 effects-based approach and, 12–14 equal efficiency test, 11–12, 54, 132, 142, 158, 166 predatory pricing and, 5, 12–14 tying and bundling, 159 Commission Guidance, 165–7 Elzinga, KG, 135 enforcement costs, 37 entry deterrence, tying and bundling, 154–5, 161 essential facilities see also compulsory licensing EU regulation, 89–91 intellectual property rights, 92–4 effect on abuse test, 96–8 Microsoft case, 93–4 post-Microsoft, 94–6 nature of dominance, 88–9 special responsibilities of owners, 91–2 terminology, 89 United States, 89 European Commission: on damages, 29 guidance see Commission Guidance review of Article 82, 1–2, 19–20, 52–3 European Union: competition goals, US comparisons, 126–9 fairness principle, 128–9 market integration objective, 130 exceptional circumstances see compulsory licensing excessive pricing: Art 82 and, 171–2 assessing excess, 170, 172, 175–9 chilling effect on investment, 170, 174 EU case law, 177–8, 179 EU definition, 171, 175–6 form of abuse, 169 non-intervention rationale, 170–5 OFT measurement, 176 post-entry price-cut benchmark, 170, 179–82 private enforcement, 183–4 public law enforcement, 182–3 self-correcting nature, 170, 172–4 UK case law, 176–8 United States, 170–1 exclusivity agreements, burden of proof, 71 fairness principle, 128–9 films, rental rights, 111 foreclosure: Commission Guidance, 132 degree, 10–11

meaning, 10 tying, 151, 152–3, 162, 163 France Telecom case: consumer welfare and, 83 effects-based approach and, 58–9 formalism, 15 landmark case, 2 law reform and, 14–17 recoupment test, 4–5, 15–16, 83 special responsibility, 134–6 Gerber, DJ, 129, 140 Germany, concept of economic dependency, 129 GlaxoSmithKline case: anticompetitive intent, 32 consumer welfare and, 84–5 defences, 31–2 effects-based approach and, 59 parallel trade restrictions, 34–5 presumptions of abuse and, 30–1, 33–4 types of abuses, 31 Gormsen, LL, 137 Habermas, Jürgen, 36 Hovenkamp, Herbert, 172n14 information costs, 21 innovation, 105, 156–7 intellectual property: absolute immunity debate, 100–3 competition and, 99–100 compulsory licensing see compulsory licensing continuum of rights, 115 excessive pricing, 174 secondary markets, 102 software features, 112–14 Internet Explorer, 147, 153 Kroes, Neelie, 53, 131, 132 legal certainty, 13–14, 36–7, 53–5, 64–5 level playing fields, 61, 63, 141 Linux, 113 Lisbon Treaty, 144–5 Lowe, Philip, 172 loyalty rebates: Commission test, 10–11 consumer behaviour, 78–9 EU case law, 7–8, 138–41 rebate test, 10n36 McCurdy, Gregory, 41–2 margin squeeze, 42–3 market power concept, 3, 24 Markovits, Richard, 47 Martin, Stephen, 152

200 Index mergers: burden of proof, 74 current economic thinking, 1 recent EU approach, 23 Microsoft cases: consumer welfare and, 83, 127–8 decompilation provisions of Software Directive and, 117, 118 effects-based approach and, 58, 93–4, 148 essential facilities, 91 exceptional circumstances consistency with previous case law, 114–15 objective justification, 106–7 post-Microsoft, 94–6 test, 87, 93–4, 102–8 formalism, 149 landmark, 2 market access, 28–9, 41–2 risk of elimination of competition, 96–7 special responsibility, 124, 137–8 standard of proof, 97 super-dominance, 125 tying, 38, 40, 147, 148, 153 consumer harm criteria, 157–8 economic approach and, 158–61, 163 Nagata, E, 143 Nalebuff, Barry, 154, 155 non-linear pricing, 81–2 nullity: anticompetitive agreements, 188 EC competition law, 189–91 damages, 187–9, 192–6 Nordic countries, 191 Ofcom, 12 Office of Fair Trading, excessive pricing, 176, 178 ordoliberalism, 3, 69, 72, 130, 137–8 Padilla, AJ, 142 parallel trade, 32, 34–5, 84, 85 patents, 116–17, 174 Posner, Richard, 22 post-Chicago School, 144 predatory pricing: consumer harm, Guidance, 80 damages, 181n53 degree of foreclosure, 11 efficiency grounds, 5, 12–14 exclusion of as-efficient behaviour, 11–12 formalistic approach, 70 France Telecom case, 14–17, 134–6 profit sacrifice test, 132, 136 rationality, 122 recoupment test, 4–5, 15–17, 83 targeted discounting, 4–7, 11

prices see excessive pricing; predatory pricing profit sacrifice test, 132, 136 promotional payments, 71 recoupment test, 4–5, 15–17, 83 refusal to deal: burden of proof, 72–3 Commission Guidance, 43, 133 consumer welfare and, 85 essential facilities see essential facilities indispensability, 139 IPRs see compulsory licensing protection of SMEs, 129 relative dominance, 129 reverse engineering, 116 risk, adversity v neutrality, 29 Salop, Steven, 29 Sidak, Gregory, 128 software see also Microsoft cases bugs, 101 compulsory licensing Database Directive, 108–9 Software Directive, 109–12 decompilation, Software Directive, 100, 116–17, 118 e-commerce, 155 peculiar copyright features, 112–14 secretive nature, 113 technological protection measures (TPMs), 116 solidarity principle, 189 special responsibility: dominant undertakings, 122–6 EU case law, 123–6, 134–42, 143 super-dominance, 125–6 US-EU comparisons, 126–9 Study Group on a European Civil Code, 190 substantial market power, 24 super-dominance, 125–6 Sweden, void contracts under Art 82, 189, 191, 193 technological protection measures (TPMs), 116 thresholds, 3, 4, 13–14 tying: basic concern, 148 Commission Guidance, 154–5, 158–9, 161–4 efficiencies, 165–7 consumer welfare, 83–4, 157–8, 159 economic analysis, 150–7 defensive leveraging, 152–4, 161 entry deterrence, 154–5, 161

Index 201 impact on innovation, 156–7 Microsoft and, 158–61 offensive leveraging, 151–2 economics-based approach, 149–50 elements, 38 formalistic approach, 70 forms, 147–8 illegality, 147 leverage theory, 148, 150–1 recent EU case law, 148–9, 158–61 recent EU developments, 157–67 United States: antitrust law objectives, 126–9, 143–5 consumer welfare, 127–8, 143 essential facilities, 89 excessive pricing, 170–1 loyalty rebates, 8 margin squeeze, 43n103

predatory pricing, recoupment test, 4, 16, 17 rule of reason, 148 single firm conduct, 29 special responsibility and, 123 tying, 147, 148 vertical restraints, Guidance, 160 Vickers, John, 23 voidness see nullity Waldman, Michael, 152–3, 161 Wanadoo see France Telecom case Whinston, Michael, 151, 152, 154, 156 Whish, Richard, 172n14 Windows, 41–2, 83, 91, 93–5, 97, 101, 114, 147, 153, 160–1 Windows Media Player, 40, 41, 83, 147, 160, 163